cext: remove PY23()

Since we always run on Python 3 now, we no longer need this
macro to support Python 2. We refactor all users to just use
the 2nd argument.

Differential Revision: https://phab.mercurial-scm.org/D12232

mercurial/cext/base85.c

@@ -1,192 +1,192 b''
 /*
  base85 codec
 
  Copyright 2006 Brendan Cully <[e-mail]>
 
  This software may be used and distributed according to the terms of
  the GNU General Public License, incorporated herein by reference.
 
  Largely based on git's implementation
 */
 
 #define PY_SSIZE_T_CLEAN
 #include <Python.h>
 
 #include "util.h"
 
 static const char b85chars[] =
     "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ"
     "abcdefghijklmnopqrstuvwxyz!#$%&()*+-;<=>?@^_`{|}~";
 static char b85dec[256];
 
 static void b85prep(void)
 {
 	unsigned i;
 
 	memset(b85dec, 0, sizeof(b85dec));
 	for (i = 0; i < sizeof(b85chars); i++) {
 		b85dec[(int)(b85chars[i])] = i + 1;
 	}
 }
 
 static PyObject *b85encode(PyObject *self, PyObject *args)
 {
 	const unsigned char *text;
 	PyObject *out;
 	char *dst;
 	Py_ssize_t len, olen, i;
 	unsigned int acc, val, ch;
 	int pad = 0;
 
-	if (!PyArg_ParseTuple(args, PY23("s#|i", "y#|i"), &text, &len, &pad)) {
+	if (!PyArg_ParseTuple(args, "y#|i", &text, &len, &pad)) {
 		return NULL;
 	}
 
 	if (pad) {
 		olen = ((len + 3) / 4 * 5) - 3;
 	} else {
 		olen = len % 4;
 		if (olen) {
 			olen++;
 		}
 		olen += len / 4 * 5;
 	}
 	if (!(out = PyBytes_FromStringAndSize(NULL, olen + 3))) {
 		return NULL;
 	}
 
 	dst = PyBytes_AsString(out);
 
 	while (len) {
 		acc = 0;
 		for (i = 24; i >= 0; i -= 8) {
 			ch = *text++;
 			acc |= ch << i;
 			if (--len == 0) {
 				break;
 			}
 		}
 		for (i = 4; i >= 0; i--) {
 			val = acc % 85;
 			acc /= 85;
 			dst[i] = b85chars[val];
 		}
 		dst += 5;
 	}
 
 	if (!pad) {
 		_PyBytes_Resize(&out, olen);
 	}
 
 	return out;
 }
 
 static PyObject *b85decode(PyObject *self, PyObject *args)
 {
 	PyObject *out = NULL;
 	const char *text;
 	char *dst;
 	Py_ssize_t len, i, j, olen, cap;
 	int c;
 	unsigned int acc;
 
-	if (!PyArg_ParseTuple(args, PY23("s#", "y#"), &text, &len)) {
+	if (!PyArg_ParseTuple(args, "y#", &text, &len)) {
 		return NULL;
 	}
 
 	olen = len / 5 * 4;
 	i = len % 5;
 	if (i) {
 		olen += i - 1;
 	}
 	if (!(out = PyBytes_FromStringAndSize(NULL, olen))) {
 		return NULL;
 	}
 
 	dst = PyBytes_AsString(out);
 
 	i = 0;
 	while (i < len) {
 		acc = 0;
 		cap = len - i - 1;
 		if (cap > 4) {
 			cap = 4;
 		}
 		for (j = 0; j < cap; i++, j++) {
 			c = b85dec[(int)*text++] - 1;
 			if (c < 0) {
 				PyErr_Format(
 				    PyExc_ValueError,
 				    "bad base85 character at position %d",
 				    (int)i);
 				goto bail;
 			}
 			acc = acc * 85 + c;
 		}
 		if (i++ < len) {
 			c = b85dec[(int)*text++] - 1;
 			if (c < 0) {
 				PyErr_Format(
 				    PyExc_ValueError,
 				    "bad base85 character at position %d",
 				    (int)i);
 				goto bail;
 			}
 			/* overflow detection: 0xffffffff == "|NsC0",
 			 * "|NsC" == 0x03030303 */
 			if (acc > 0x03030303 || (acc *= 85) > 0xffffffff - c) {
 				PyErr_Format(
 				    PyExc_ValueError,
 				    "bad base85 sequence at position %d",
 				    (int)i);
 				goto bail;
 			}
 			acc += c;
 		}
 
 		cap = olen < 4 ? olen : 4;
 		olen -= cap;
 		for (j = 0; j < 4 - cap; j++) {
 			acc *= 85;
 		}
 		if (cap && cap < 4) {
 			acc += 0xffffff >> (cap - 1) * 8;
 		}
 		for (j = 0; j < cap; j++) {
 			acc = (acc << 8) | (acc >> 24);
 			*dst++ = acc;
 		}
 	}
 
 	return out;
 bail:
 	Py_XDECREF(out);
 	return NULL;
 }
 
 static char base85_doc[] = "Base85 Data Encoding";
 
 static PyMethodDef methods[] = {
     {"b85encode", b85encode, METH_VARARGS,
      "Encode text in base85.\n\n"
      "If the second parameter is true, pad the result to a multiple of "
      "five characters.\n"},
     {"b85decode", b85decode, METH_VARARGS, "Decode base85 text.\n"},
     {NULL, NULL},
 };
 
 static const int version = 1;
 
 static struct PyModuleDef base85_module = {
     PyModuleDef_HEAD_INIT, "base85", base85_doc, -1, methods,
 };
 
 PyMODINIT_FUNC PyInit_base85(void)
 {
 	PyObject *m;
 	b85prep();
 
 	m = PyModule_Create(&base85_module);
 	PyModule_AddIntConstant(m, "version", version);
 	return m;
 }

mercurial/cext/bdiff.c

@@ -1,350 +1,348 b''
 /*
  bdiff.c - efficient binary diff extension for Mercurial
 
  Copyright 2005, 2006 Olivia Mackall <[e-mail]>
 
  This software may be used and distributed according to the terms of
  the GNU General Public License, incorporated herein by reference.
 
  Based roughly on Python difflib
 */
 
 #define PY_SSIZE_T_CLEAN
 #include <Python.h>
 #include <limits.h>
 #include <stdlib.h>
 #include <string.h>
 
 #include "bdiff.h"
 #include "bitmanipulation.h"
 #include "thirdparty/xdiff/xdiff.h"
 #include "util.h"
 
 static PyObject *blocks(PyObject *self, PyObject *args)
 {
 	PyObject *sa, *sb, *rl = NULL, *m;
 	struct bdiff_line *a, *b;
 	struct bdiff_hunk l, *h;
 	int an, bn, count, pos = 0;
 
 	l.next = NULL;
 
 	if (!PyArg_ParseTuple(args, "SS:bdiff", &sa, &sb)) {
 		return NULL;
 	}
 
 	an = bdiff_splitlines(PyBytes_AsString(sa), PyBytes_Size(sa), &a);
 	bn = bdiff_splitlines(PyBytes_AsString(sb), PyBytes_Size(sb), &b);
 
 	if (!a || !b) {
 		goto nomem;
 	}
 
 	count = bdiff_diff(a, an, b, bn, &l);
 	if (count < 0) {
 		goto nomem;
 	}
 
 	rl = PyList_New(count);
 	if (!rl) {
 		goto nomem;
 	}
 
 	for (h = l.next; h; h = h->next) {
 		m = Py_BuildValue("iiii", h->a1, h->a2, h->b1, h->b2);
 		PyList_SetItem(rl, pos, m);
 		pos++;
 	}
 
 nomem:
 	free(a);
 	free(b);
 	bdiff_freehunks(l.next);
 	return rl ? rl : PyErr_NoMemory();
 }
 
 static PyObject *bdiff(PyObject *self, PyObject *args)
 {
 	Py_buffer ba, bb;
 	char *rb, *ia, *ib;
 	PyObject *result = NULL;
 	struct bdiff_line *al = NULL, *bl = NULL;
 	struct bdiff_hunk l, *h;
 	int an, bn, count;
 	Py_ssize_t len = 0, la, lb, li = 0, lcommon = 0, lmax;
 	PyThreadState *_save = NULL;
 
 	l.next = NULL;
 
-	if (!PyArg_ParseTuple(args, PY23("s*s*:bdiff", "y*y*:bdiff"), &ba,
-	                      &bb)) {
+	if (!PyArg_ParseTuple(args, "y*y*:bdiff", &ba, &bb)) {
 		return NULL;
 	}
 
 	if (!PyBuffer_IsContiguous(&ba, 'C') || ba.ndim > 1) {
 		PyErr_SetString(PyExc_ValueError, "bdiff input not contiguous");
 		goto cleanup;
 	}
 
 	if (!PyBuffer_IsContiguous(&bb, 'C') || bb.ndim > 1) {
 		PyErr_SetString(PyExc_ValueError, "bdiff input not contiguous");
 		goto cleanup;
 	}
 
 	la = ba.len;
 	lb = bb.len;
 
 	if (la > UINT_MAX || lb > UINT_MAX) {
 		PyErr_SetString(PyExc_ValueError, "bdiff inputs too large");
 		goto cleanup;
 	}
 
 	_save = PyEval_SaveThread();
 
 	lmax = la > lb ? lb : la;
 	for (ia = ba.buf, ib = bb.buf; li < lmax && *ia == *ib;
 	     ++li, ++ia, ++ib) {
 		if (*ia == '\n') {
 			lcommon = li + 1;
 		}
 	}
 	/* we can almost add: if (li == lmax) lcommon = li; */
 
 	an = bdiff_splitlines((char *)ba.buf + lcommon, la - lcommon, &al);
 	bn = bdiff_splitlines((char *)bb.buf + lcommon, lb - lcommon, &bl);
 	if (!al || !bl) {
 		PyErr_NoMemory();
 		goto cleanup;
 	}
 
 	count = bdiff_diff(al, an, bl, bn, &l);
 	if (count < 0) {
 		PyErr_NoMemory();
 		goto cleanup;
 	}
 
 	/* calculate length of output */
 	la = lb = 0;
 	for (h = l.next; h; h = h->next) {
 		if (h->a1 != la || h->b1 != lb) {
 			len += 12 + bl[h->b1].l - bl[lb].l;
 		}
 		la = h->a2;
 		lb = h->b2;
 	}
 	PyEval_RestoreThread(_save);
 	_save = NULL;
 
 	result = PyBytes_FromStringAndSize(NULL, len);
 
 	if (!result) {
 		goto cleanup;
 	}
 
 	/* build binary patch */
 	rb = PyBytes_AsString(result);
 	la = lb = 0;
 
 	for (h = l.next; h; h = h->next) {
 		if (h->a1 != la || h->b1 != lb) {
 			len = bl[h->b1].l - bl[lb].l;
 			putbe32((uint32_t)(al[la].l + lcommon - al->l), rb);
 			putbe32((uint32_t)(al[h->a1].l + lcommon - al->l),
 			        rb + 4);
 			putbe32((uint32_t)len, rb + 8);
 			memcpy(rb + 12, bl[lb].l, len);
 			rb += 12 + len;
 		}
 		la = h->a2;
 		lb = h->b2;
 	}
 
 cleanup:
 	if (_save) {
 		PyEval_RestoreThread(_save);
 	}
 	PyBuffer_Release(&ba);
 	PyBuffer_Release(&bb);
 	free(al);
 	free(bl);
 	bdiff_freehunks(l.next);
 	return result;
 }
 
 /*
  * If allws != 0, remove all whitespace (' ', \t and \r). Otherwise,
  * reduce whitespace sequences to a single space and trim remaining whitespace
  * from end of lines.
  */
 static PyObject *fixws(PyObject *self, PyObject *args)
 {
 	PyObject *s, *result = NULL;
 	char allws, c;
 	const char *r;
 	Py_ssize_t i, rlen, wlen = 0;
 	char *w;
 
 	if (!PyArg_ParseTuple(args, "Sb:fixws", &s, &allws)) {
 		return NULL;
 	}
 	r = PyBytes_AsString(s);
 	rlen = PyBytes_Size(s);
 
 	w = (char *)PyMem_Malloc(rlen ? rlen : 1);
 	if (!w) {
 		goto nomem;
 	}
 
 	for (i = 0; i != rlen; i++) {
 		c = r[i];
 		if (c == ' ' || c == '\t' || c == '\r') {
 			if (!allws && (wlen == 0 || w[wlen - 1] != ' ')) {
 				w[wlen++] = ' ';
 			}
 		} else if (c == '\n' && !allws && wlen > 0 &&
 		           w[wlen - 1] == ' ') {
 			w[wlen - 1] = '\n';
 		} else {
 			w[wlen++] = c;
 		}
 	}
 
 	result = PyBytes_FromStringAndSize(w, wlen);
 
 nomem:
 	PyMem_Free(w);
 	return result ? result : PyErr_NoMemory();
 }
 
 static bool sliceintolist(PyObject *list, Py_ssize_t destidx,
                           const char *source, Py_ssize_t len)
 {
 	PyObject *sliced = PyBytes_FromStringAndSize(source, len);
 	if (sliced == NULL) {
 		return false;
 	}
 	PyList_SET_ITEM(list, destidx, sliced);
 	return true;
 }
 
 static PyObject *splitnewlines(PyObject *self, PyObject *args)
 {
 	const char *text;
 	Py_ssize_t nelts = 0, size, i, start = 0;
 	PyObject *result = NULL;
 
-	if (!PyArg_ParseTuple(args, PY23("s#", "y#"), &text, &size)) {
+	if (!PyArg_ParseTuple(args, "y#", &text, &size)) {
 		goto abort;
 	}
 	if (!size) {
 		return PyList_New(0);
 	}
 	/* This loops to size-1 because if the last byte is a newline,
 	 * we don't want to perform a split there. */
 	for (i = 0; i < size - 1; ++i) {
 		if (text[i] == '\n') {
 			++nelts;
 		}
 	}
 	if ((result = PyList_New(nelts + 1)) == NULL) {
 		goto abort;
 	}
 	nelts = 0;
 	for (i = 0; i < size - 1; ++i) {
 		if (text[i] == '\n') {
 			if (!sliceintolist(result, nelts++, text + start,
 			                   i - start + 1)) {
 				goto abort;
 			}
 			start = i + 1;
 		}
 	}
 	if (!sliceintolist(result, nelts++, text + start, size - start)) {
 		goto abort;
 	}
 	return result;
 abort:
 	Py_XDECREF(result);
 	return NULL;
 }
 
 static int hunk_consumer(int64_t a1, int64_t a2, int64_t b1, int64_t b2,
                          void *priv)
 {
 	PyObject *rl = (PyObject *)priv;
 	PyObject *m = Py_BuildValue("LLLL", a1, a2, b1, b2);
 	int r;
 	if (!m) {
 		return -1;
 	}
 	r = PyList_Append(rl, m);
 	Py_DECREF(m);
 	return r;
 }
 
 static PyObject *xdiffblocks(PyObject *self, PyObject *args)
 {
 	Py_ssize_t la, lb;
 	mmfile_t a, b;
 	PyObject *rl;
 
 	xpparam_t xpp = {
 	    XDF_INDENT_HEURISTIC, /* flags */
 	};
 	xdemitconf_t xecfg = {
 	    XDL_EMIT_BDIFFHUNK, /* flags */
 	    hunk_consumer,      /* hunk_consume_func */
 	};
 	xdemitcb_t ecb = {
 	    NULL, /* priv */
 	};
 
-	if (!PyArg_ParseTuple(args, PY23("s#s#", "y#y#"), &a.ptr, &la, &b.ptr,
-	                      &lb)) {
+	if (!PyArg_ParseTuple(args, "y#y#", &a.ptr, &la, &b.ptr, &lb)) {
 		return NULL;
 	}
 
 	a.size = la;
 	b.size = lb;
 
 	rl = PyList_New(0);
 	if (!rl) {
 		return PyErr_NoMemory();
 	}
 
 	ecb.priv = rl;
 
 	if (xdl_diff(&a, &b, &xpp, &xecfg, &ecb) != 0) {
 		Py_DECREF(rl);
 		return PyErr_NoMemory();
 	}
 
 	return rl;
 }
 
 static char mdiff_doc[] = "Efficient binary diff.";
 
 static PyMethodDef methods[] = {
     {"bdiff", bdiff, METH_VARARGS, "calculate a binary diff\n"},
     {"blocks", blocks, METH_VARARGS, "find a list of matching lines\n"},
     {"fixws", fixws, METH_VARARGS, "normalize diff whitespaces\n"},
     {"splitnewlines", splitnewlines, METH_VARARGS,
      "like str.splitlines, but only split on newlines\n"},
     {"xdiffblocks", xdiffblocks, METH_VARARGS,
      "find a list of matching lines using xdiff algorithm\n"},
     {NULL, NULL},
 };
 
 static const int version = 3;
 
 static struct PyModuleDef bdiff_module = {
     PyModuleDef_HEAD_INIT, "bdiff", mdiff_doc, -1, methods,
 };
 
 PyMODINIT_FUNC PyInit_bdiff(void)
 {
 	PyObject *m;
 	m = PyModule_Create(&bdiff_module);
 	PyModule_AddIntConstant(m, "version", version);
 	return m;
 }

mercurial/cext/charencode.c

@@ -1,405 +1,404 b''
 /*
  charencode.c - miscellaneous character encoding
 
  Copyright 2008 Olivia Mackall <[e-mail]> and others
 
  This software may be used and distributed according to the terms of
  the GNU General Public License, incorporated herein by reference.
 */
 
 #define PY_SSIZE_T_CLEAN
 #include <Python.h>
 #include <assert.h>
 
 #include "charencode.h"
 #include "compat.h"
 #include "util.h"
 
 /* clang-format off */
 static const char lowertable[128] = {
 	'\x00', '\x01', '\x02', '\x03', '\x04', '\x05', '\x06', '\x07',
 	'\x08', '\x09', '\x0a', '\x0b', '\x0c', '\x0d', '\x0e', '\x0f',
 	'\x10', '\x11', '\x12', '\x13', '\x14', '\x15', '\x16', '\x17',
 	'\x18', '\x19', '\x1a', '\x1b', '\x1c', '\x1d', '\x1e', '\x1f',
 	'\x20', '\x21', '\x22', '\x23', '\x24', '\x25', '\x26', '\x27',
 	'\x28', '\x29', '\x2a', '\x2b', '\x2c', '\x2d', '\x2e', '\x2f',
 	'\x30', '\x31', '\x32', '\x33', '\x34', '\x35', '\x36', '\x37',
 	'\x38', '\x39', '\x3a', '\x3b', '\x3c', '\x3d', '\x3e', '\x3f',
 	'\x40',
 	        '\x61', '\x62', '\x63', '\x64', '\x65', '\x66', '\x67', /* A-G */
 	'\x68', '\x69', '\x6a', '\x6b', '\x6c', '\x6d', '\x6e', '\x6f', /* H-O */
 	'\x70', '\x71', '\x72', '\x73', '\x74', '\x75', '\x76', '\x77', /* P-W */
 	'\x78', '\x79', '\x7a',                                         /* X-Z */
 	                        '\x5b', '\x5c', '\x5d', '\x5e', '\x5f',
 	'\x60', '\x61', '\x62', '\x63', '\x64', '\x65', '\x66', '\x67',
 	'\x68', '\x69', '\x6a', '\x6b', '\x6c', '\x6d', '\x6e', '\x6f',
 	'\x70', '\x71', '\x72', '\x73', '\x74', '\x75', '\x76', '\x77',
 	'\x78', '\x79', '\x7a', '\x7b', '\x7c', '\x7d', '\x7e', '\x7f'
 };
 
 static const char uppertable[128] = {
 	'\x00', '\x01', '\x02', '\x03', '\x04', '\x05', '\x06', '\x07',
 	'\x08', '\x09', '\x0a', '\x0b', '\x0c', '\x0d', '\x0e', '\x0f',
 	'\x10', '\x11', '\x12', '\x13', '\x14', '\x15', '\x16', '\x17',
 	'\x18', '\x19', '\x1a', '\x1b', '\x1c', '\x1d', '\x1e', '\x1f',
 	'\x20', '\x21', '\x22', '\x23', '\x24', '\x25', '\x26', '\x27',
 	'\x28', '\x29', '\x2a', '\x2b', '\x2c', '\x2d', '\x2e', '\x2f',
 	'\x30', '\x31', '\x32', '\x33', '\x34', '\x35', '\x36', '\x37',
 	'\x38', '\x39', '\x3a', '\x3b', '\x3c', '\x3d', '\x3e', '\x3f',
 	'\x40', '\x41', '\x42', '\x43', '\x44', '\x45', '\x46', '\x47',
 	'\x48', '\x49', '\x4a', '\x4b', '\x4c', '\x4d', '\x4e', '\x4f',
 	'\x50', '\x51', '\x52', '\x53', '\x54', '\x55', '\x56', '\x57',
 	'\x58', '\x59', '\x5a', '\x5b', '\x5c', '\x5d', '\x5e', '\x5f',
 	'\x60',
 		'\x41', '\x42', '\x43', '\x44', '\x45', '\x46', '\x47', /* a-g */
 	'\x48', '\x49', '\x4a', '\x4b', '\x4c', '\x4d', '\x4e', '\x4f', /* h-o */
 	'\x50', '\x51', '\x52', '\x53', '\x54', '\x55', '\x56', '\x57', /* p-w */
 	'\x58', '\x59', '\x5a', 					/* x-z */
 				'\x7b', '\x7c', '\x7d', '\x7e', '\x7f'
 };
 
 /* 1: no escape, 2: \<c>, 6: \u<x> */
 static const uint8_t jsonlentable[256] = {
 	6, 6, 6, 6, 6, 6, 6, 6, 2, 2, 2, 6, 2, 2, 6, 6, /* b, t, n, f, r */
 	6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6,
 	1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* " */
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, /* \\ */
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 6, /* DEL */
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 };
 
 static const uint8_t jsonparanoidlentable[128] = {
 	6, 6, 6, 6, 6, 6, 6, 6, 2, 2, 2, 6, 2, 2, 6, 6, /* b, t, n, f, r */
 	6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6,
 	1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* " */
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 6, 1, 6, 1, /* <, > */
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, /* \\ */
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 6, /* DEL */
 };
 
 static const char hexchartable[16] = {
 	'0', '1', '2', '3', '4', '5', '6', '7',
 	'8', '9', 'a', 'b', 'c', 'd', 'e', 'f',
 };
 /* clang-format on */
 
 /*
  * Turn a hex-encoded string into binary.
  */
 PyObject *unhexlify(const char *str, Py_ssize_t len)
 {
 	PyObject *ret;
 	char *d;
 	Py_ssize_t i;
 
 	ret = PyBytes_FromStringAndSize(NULL, len / 2);
 
 	if (!ret) {
 		return NULL;
 	}
 
 	d = PyBytes_AsString(ret);
 
 	for (i = 0; i < len;) {
 		int hi = hexdigit(str, i++);
 		int lo = hexdigit(str, i++);
 		*d++ = (hi << 4) | lo;
 	}
 
 	return ret;
 }
 
 PyObject *isasciistr(PyObject *self, PyObject *args)
 {
 	const char *buf;
 	Py_ssize_t i, len;
-	if (!PyArg_ParseTuple(args, PY23("s#:isasciistr", "y#:isasciistr"),
-	                      &buf, &len)) {
+	if (!PyArg_ParseTuple(args, "y#:isasciistr", &buf, &len)) {
 		return NULL;
 	}
 	i = 0;
 	/* char array in PyStringObject should be at least 4-byte aligned */
 	if (((uintptr_t)buf & 3) == 0) {
 		const uint32_t *p = (const uint32_t *)buf;
 		for (; i < len / 4; i++) {
 			if (p[i] & 0x80808080U) {
 				Py_RETURN_FALSE;
 			}
 		}
 		i *= 4;
 	}
 	for (; i < len; i++) {
 		if (buf[i] & 0x80) {
 			Py_RETURN_FALSE;
 		}
 	}
 	Py_RETURN_TRUE;
 }
 
 static inline PyObject *
 _asciitransform(PyObject *str_obj, const char table[128], PyObject *fallback_fn)
 {
 	char *str, *newstr;
 	Py_ssize_t i, len;
 	PyObject *newobj = NULL;
 	PyObject *ret = NULL;
 
 	str = PyBytes_AS_STRING(str_obj);
 	len = PyBytes_GET_SIZE(str_obj);
 
 	newobj = PyBytes_FromStringAndSize(NULL, len);
 	if (!newobj) {
 		goto quit;
 	}
 
 	newstr = PyBytes_AS_STRING(newobj);
 
 	for (i = 0; i < len; i++) {
 		char c = str[i];
 		if (c & 0x80) {
 			if (fallback_fn != NULL) {
 				ret = PyObject_CallFunctionObjArgs(
 				    fallback_fn, str_obj, NULL);
 			} else {
 				PyObject *err = PyUnicodeDecodeError_Create(
 				    "ascii", str, len, i, (i + 1),
 				    "unexpected code byte");
 				PyErr_SetObject(PyExc_UnicodeDecodeError, err);
 				Py_XDECREF(err);
 			}
 			goto quit;
 		}
 		newstr[i] = table[(unsigned char)c];
 	}
 
 	ret = newobj;
 	Py_INCREF(ret);
 quit:
 	Py_XDECREF(newobj);
 	return ret;
 }
 
 PyObject *asciilower(PyObject *self, PyObject *args)
 {
 	PyObject *str_obj;
 	if (!PyArg_ParseTuple(args, "O!:asciilower", &PyBytes_Type, &str_obj)) {
 		return NULL;
 	}
 	return _asciitransform(str_obj, lowertable, NULL);
 }
 
 PyObject *asciiupper(PyObject *self, PyObject *args)
 {
 	PyObject *str_obj;
 	if (!PyArg_ParseTuple(args, "O!:asciiupper", &PyBytes_Type, &str_obj)) {
 		return NULL;
 	}
 	return _asciitransform(str_obj, uppertable, NULL);
 }
 
 PyObject *make_file_foldmap(PyObject *self, PyObject *args)
 {
 	PyObject *dmap, *spec_obj, *normcase_fallback;
 	PyObject *file_foldmap = NULL;
 	enum normcase_spec spec;
 	PyObject *k, *v;
 	dirstateItemObject *tuple;
 	Py_ssize_t pos = 0;
 	const char *table;
 
 	if (!PyArg_ParseTuple(args, "O!O!O!:make_file_foldmap", &PyDict_Type,
 	                      &dmap, &PyLong_Type, &spec_obj, &PyFunction_Type,
 	                      &normcase_fallback)) {
 		goto quit;
 	}
 
 	spec = (int)PyLong_AS_LONG(spec_obj);
 	switch (spec) {
 	case NORMCASE_LOWER:
 		table = lowertable;
 		break;
 	case NORMCASE_UPPER:
 		table = uppertable;
 		break;
 	case NORMCASE_OTHER:
 		table = NULL;
 		break;
 	default:
 		PyErr_SetString(PyExc_TypeError, "invalid normcasespec");
 		goto quit;
 	}
 
 	/* Add some more entries to deal with additions outside this
 	   function. */
 	file_foldmap = _dict_new_presized((PyDict_Size(dmap) / 10) * 11);
 	if (file_foldmap == NULL) {
 		goto quit;
 	}
 
 	while (PyDict_Next(dmap, &pos, &k, &v)) {
 		if (!dirstate_tuple_check(v)) {
 			PyErr_SetString(PyExc_TypeError,
 			                "expected a dirstate tuple");
 			goto quit;
 		}
 
 		tuple = (dirstateItemObject *)v;
 		if (tuple->flags | dirstate_flag_wc_tracked) {
 			PyObject *normed;
 			if (table != NULL) {
 				normed = _asciitransform(k, table,
 				                         normcase_fallback);
 			} else {
 				normed = PyObject_CallFunctionObjArgs(
 				    normcase_fallback, k, NULL);
 			}
 
 			if (normed == NULL) {
 				goto quit;
 			}
 			if (PyDict_SetItem(file_foldmap, normed, k) == -1) {
 				Py_DECREF(normed);
 				goto quit;
 			}
 			Py_DECREF(normed);
 		}
 	}
 	return file_foldmap;
 quit:
 	Py_XDECREF(file_foldmap);
 	return NULL;
 }
 
 /* calculate length of JSON-escaped string; returns -1 if unsupported */
 static Py_ssize_t jsonescapelen(const char *buf, Py_ssize_t len, bool paranoid)
 {
 	Py_ssize_t i, esclen = 0;
 
 	if (paranoid) {
 		/* don't want to process multi-byte escapes in C */
 		for (i = 0; i < len; i++) {
 			char c = buf[i];
 			if (c & 0x80) {
 				PyErr_SetString(PyExc_ValueError,
 				                "cannot process non-ascii str");
 				return -1;
 			}
 			esclen += jsonparanoidlentable[(unsigned char)c];
 			if (esclen < 0) {
 				PyErr_SetString(PyExc_MemoryError,
 				                "overflow in jsonescapelen");
 				return -1;
 			}
 		}
 	} else {
 		for (i = 0; i < len; i++) {
 			char c = buf[i];
 			esclen += jsonlentable[(unsigned char)c];
 			if (esclen < 0) {
 				PyErr_SetString(PyExc_MemoryError,
 				                "overflow in jsonescapelen");
 				return -1;
 			}
 		}
 	}
 
 	return esclen;
 }
 
 /* map '\<c>' escape character */
 static char jsonescapechar2(char c)
 {
 	switch (c) {
 	case '\b':
 		return 'b';
 	case '\t':
 		return 't';
 	case '\n':
 		return 'n';
 	case '\f':
 		return 'f';
 	case '\r':
 		return 'r';
 	case '"':
 		return '"';
 	case '\\':
 		return '\\';
 	}
 	return '\0'; /* should not happen */
 }
 
 /* convert 'origbuf' to JSON-escaped form 'escbuf'; 'origbuf' should only
    include characters mappable by json(paranoid)lentable */
 static void encodejsonescape(char *escbuf, Py_ssize_t esclen,
                              const char *origbuf, Py_ssize_t origlen,
                              bool paranoid)
 {
 	const uint8_t *lentable =
 	    (paranoid) ? jsonparanoidlentable : jsonlentable;
 	Py_ssize_t i, j;
 
 	for (i = 0, j = 0; i < origlen; i++) {
 		char c = origbuf[i];
 		uint8_t l = lentable[(unsigned char)c];
 		assert(j + l <= esclen);
 		switch (l) {
 		case 1:
 			escbuf[j] = c;
 			break;
 		case 2:
 			escbuf[j] = '\\';
 			escbuf[j + 1] = jsonescapechar2(c);
 			break;
 		case 6:
 			memcpy(escbuf + j, "\\u00", 4);
 			escbuf[j + 4] = hexchartable[(unsigned char)c >> 4];
 			escbuf[j + 5] = hexchartable[(unsigned char)c & 0xf];
 			break;
 		}
 		j += l;
 	}
 }
 
 PyObject *jsonescapeu8fast(PyObject *self, PyObject *args)
 {
 	PyObject *origstr, *escstr;
 	const char *origbuf;
 	Py_ssize_t origlen, esclen;
 	int paranoid;
 	if (!PyArg_ParseTuple(args, "O!i:jsonescapeu8fast", &PyBytes_Type,
 	                      &origstr, &paranoid)) {
 		return NULL;
 	}
 
 	origbuf = PyBytes_AS_STRING(origstr);
 	origlen = PyBytes_GET_SIZE(origstr);
 	esclen = jsonescapelen(origbuf, origlen, paranoid);
 	if (esclen < 0) {
 		return NULL; /* unsupported char found or overflow */
 	}
 	if (origlen == esclen) {
 		Py_INCREF(origstr);
 		return origstr;
 	}
 
 	escstr = PyBytes_FromStringAndSize(NULL, esclen);
 	if (!escstr) {
 		return NULL;
 	}
 	encodejsonescape(PyBytes_AS_STRING(escstr), esclen, origbuf, origlen,
 	                 paranoid);
 
 	return escstr;
 }

mercurial/cext/manifest.c

@@ -1,999 +1,999 b''
 /*
  * manifest.c - manifest type that does on-demand parsing.
  *
  * Copyright 2015, Google Inc.
  *
  * This software may be used and distributed according to the terms of
  * the GNU General Public License, incorporated herein by reference.
  */
 #include <Python.h>
 
 #include <assert.h>
 #include <stdlib.h>
 #include <string.h>
 
 #include "charencode.h"
 #include "util.h"
 
 #define DEFAULT_LINES 100000
 
 typedef struct {
 	char *start;
 	Py_ssize_t len; /* length of line including terminal newline */
 	char hash_suffix;
 	bool from_malloc;
 	bool deleted;
 } line;
 
 typedef struct {
 	PyObject_HEAD
 	PyObject *pydata;
 	Py_ssize_t nodelen;
 	line *lines;
 	int numlines; /* number of line entries */
 	int livelines; /* number of non-deleted lines */
 	int maxlines; /* allocated number of lines */
 	bool dirty;
 } lazymanifest;
 
 #define MANIFEST_OOM -1
 #define MANIFEST_NOT_SORTED -2
 #define MANIFEST_MALFORMED -3
 #define MANIFEST_BOGUS_FILENAME -4
 #define MANIFEST_TOO_SHORT_LINE -5
 
 /* get the length of the path for a line */
 static Py_ssize_t pathlen(line *l)
 {
 	const char *end = memchr(l->start, '\0', l->len);
 	return (end) ? (Py_ssize_t)(end - l->start) : l->len;
 }
 
 /* get the node value of a single line */
 static PyObject *nodeof(Py_ssize_t nodelen, line *l, char *flag)
 {
 	char *s = l->start;
 	Py_ssize_t llen = pathlen(l);
 	Py_ssize_t hlen = l->len - llen - 2;
 	PyObject *hash;
 	if (llen + 1 + 40 + 1 > l->len) { /* path '\0' hash '\n' */
 		PyErr_SetString(PyExc_ValueError, "manifest line too short");
 		return NULL;
 	}
 	/* Detect flags after the hash first. */
 	switch (s[llen + hlen]) {
 	case 'l':
 	case 't':
 	case 'x':
 		*flag = s[llen + hlen];
 		--hlen;
 		break;
 	default:
 		*flag = '\0';
 		break;
 	}
 
 	if (hlen != 2 * nodelen) {
 		PyErr_SetString(PyExc_ValueError, "invalid node length in manifest");
 		return NULL;
 	}
 	hash = unhexlify(s + llen + 1, nodelen * 2);
 	if (!hash) {
 		return NULL;
 	}
 	if (l->hash_suffix != '\0') {
 		char newhash[33];
 		memcpy(newhash, PyBytes_AsString(hash), nodelen);
 		Py_DECREF(hash);
 		newhash[nodelen] = l->hash_suffix;
 		hash = PyBytes_FromStringAndSize(newhash, nodelen + 1);
 	}
 	return hash;
 }
 
 /* get the node hash and flags of a line as a tuple */
 static PyObject *hashflags(Py_ssize_t nodelen, line *l)
 {
 	char flag;
 	PyObject *hash = nodeof(nodelen, l, &flag);
 	PyObject *flags;
 	PyObject *tup;
 
 	if (!hash)
 		return NULL;
 	flags = PyBytes_FromStringAndSize(&flag, flag ? 1 : 0);
 	if (!flags) {
 		Py_DECREF(hash);
 		return NULL;
 	}
 	tup = PyTuple_Pack(2, hash, flags);
 	Py_DECREF(flags);
 	Py_DECREF(hash);
 	return tup;
 }
 
 /* if we're about to run out of space in the line index, add more */
 static bool realloc_if_full(lazymanifest *self)
 {
 	if (self->numlines == self->maxlines) {
 		self->maxlines *= 2;
 		self->lines = realloc(self->lines, self->maxlines * sizeof(line));
 	}
 	return !!self->lines;
 }
 
 /*
  * Find the line boundaries in the manifest that 'data' points to and store
  * information about each line in 'self'.
  */
 static int find_lines(lazymanifest *self, char *data, Py_ssize_t len)
 {
 	char *prev = NULL;
 	while (len > 0) {
 		line *l;
 		char *next;
 		if (*data == '\0') {
 			/* It's implausible there's no filename, don't
 			 * even bother looking for the newline. */
 			return MANIFEST_BOGUS_FILENAME;
 		}
 		next = memchr(data, '\n', len);
 		if (!next) {
 			return MANIFEST_MALFORMED;
 		}
 		if ((next - data) < 42) {
 			/* We should have at least 42 bytes in a line:
 			   1 byte filename
 			   1 NUL
 			   40 bytes of hash
 			   so we can give up here.
 			*/
 			return MANIFEST_TOO_SHORT_LINE;
 		}
 		next++; /* advance past newline */
 		if (prev && strcmp(prev, data) > -1) {
 			/* This data isn't sorted, so we have to abort. */
 			return MANIFEST_NOT_SORTED;
 		}
 		if (!realloc_if_full(self)) {
 			return MANIFEST_OOM; /* no memory */
 		}
 		l = self->lines + ((self->numlines)++);
 		l->start = data;
 		l->len = next - data;
 		l->hash_suffix = '\0';
 		l->from_malloc = false;
 		l->deleted = false;
 		len = len - l->len;
 		prev = data;
 		data = next;
 	}
 	self->livelines = self->numlines;
 	return 0;
 }
 
 static void lazymanifest_init_early(lazymanifest *self)
 {
 	self->pydata = NULL;
 	self->lines = NULL;
 	self->numlines = 0;
 	self->maxlines = 0;
 }
 
 static int lazymanifest_init(lazymanifest *self, PyObject *args)
 {
 	char *data;
 	Py_ssize_t nodelen, len;
 	int err, ret;
 	PyObject *pydata;
 
 	lazymanifest_init_early(self);
 	if (!PyArg_ParseTuple(args, "nS", &nodelen, &pydata)) {
 		return -1;
 	}
 	if (nodelen != 20 && nodelen != 32) {
 		/* See fixed buffer in nodeof */
 		PyErr_Format(PyExc_ValueError, "Unsupported node length");
 		return -1;
 	}
 	self->nodelen = nodelen;
 	self->dirty = false;
 
 	err = PyBytes_AsStringAndSize(pydata, &data, &len);
 	if (err == -1)
 		return -1;
 	self->pydata = pydata;
 	Py_INCREF(self->pydata);
 	Py_BEGIN_ALLOW_THREADS
 	self->lines = malloc(DEFAULT_LINES * sizeof(line));
 	self->maxlines = DEFAULT_LINES;
 	self->numlines = 0;
 	if (!self->lines)
 		ret = MANIFEST_OOM;
 	else
 		ret = find_lines(self, data, len);
 	Py_END_ALLOW_THREADS
 	switch (ret) {
 	case 0:
 		break;
 	case MANIFEST_OOM:
 		PyErr_NoMemory();
 		break;
 	case MANIFEST_NOT_SORTED:
 		PyErr_Format(PyExc_ValueError,
 			     "Manifest lines not in sorted order.");
 		break;
 	case MANIFEST_MALFORMED:
 		PyErr_Format(PyExc_ValueError,
 			     "Manifest did not end in a newline.");
 		break;
 	case MANIFEST_BOGUS_FILENAME:
 		PyErr_Format(
 			PyExc_ValueError,
 			"Manifest had an entry with a zero-length filename.");
 		break;
 	case MANIFEST_TOO_SHORT_LINE:
 		PyErr_Format(
 			PyExc_ValueError,
 			"Manifest had implausibly-short line.");
 		break;
 	default:
 		PyErr_Format(PyExc_ValueError,
 			     "Unknown problem parsing manifest.");
 	}
 	return ret == 0 ? 0 : -1;
 }
 
 static void lazymanifest_dealloc(lazymanifest *self)
 {
 	/* free any extra lines we had to allocate */
 	int i;
 	for (i = 0; self->lines && (i < self->numlines); i++) {
 		if (self->lines[i].from_malloc) {
 			free(self->lines[i].start);
 		}
 	}
 	free(self->lines);
 	self->lines = NULL;
 	if (self->pydata) {
 		Py_DECREF(self->pydata);
 		self->pydata = NULL;
 	}
 	PyObject_Del(self);
 }
 
 /* iteration support */
 
 typedef struct {
 	PyObject_HEAD lazymanifest *m;
 	Py_ssize_t pos;
 } lmIter;
 
 static void lmiter_dealloc(PyObject *o)
 {
 	lmIter *self = (lmIter *)o;
 	Py_DECREF(self->m);
 	PyObject_Del(self);
 }
 
 static line *lmiter_nextline(lmIter *self)
 {
 	do {
 		self->pos++;
 		if (self->pos >= self->m->numlines) {
 			return NULL;
 		}
 		/* skip over deleted manifest entries */
 	} while (self->m->lines[self->pos].deleted);
 	return self->m->lines + self->pos;
 }
 
 static PyObject *lmiter_iterentriesnext(PyObject *o)
 {
 	lmIter *self = (lmIter *)o;
 	Py_ssize_t pl;
 	line *l;
 	char flag;
 	PyObject *ret = NULL, *path = NULL, *hash = NULL, *flags = NULL;
 	l = lmiter_nextline(self);
 	if (!l) {
 		goto done;
 	}
 	pl = pathlen(l);
 	path = PyBytes_FromStringAndSize(l->start, pl);
 	hash = nodeof(self->m->nodelen, l, &flag);
 	if (!path || !hash) {
 		goto done;
 	}
 	flags = PyBytes_FromStringAndSize(&flag, flag ? 1 : 0);
 	if (!flags) {
 		goto done;
 	}
 	ret = PyTuple_Pack(3, path, hash, flags);
 done:
 	Py_XDECREF(path);
 	Py_XDECREF(hash);
 	Py_XDECREF(flags);
 	return ret;
 }
 
 #define LAZYMANIFESTENTRIESITERATOR_TPFLAGS Py_TPFLAGS_DEFAULT
 
 static PyTypeObject lazymanifestEntriesIterator = {
 	PyVarObject_HEAD_INIT(NULL, 0) /* header */
 	"parsers.lazymanifest.entriesiterator", /*tp_name */
 	sizeof(lmIter),                  /*tp_basicsize */
 	0,                               /*tp_itemsize */
 	lmiter_dealloc,                  /*tp_dealloc */
 	0,                               /*tp_print */
 	0,                               /*tp_getattr */
 	0,                               /*tp_setattr */
 	0,                               /*tp_compare */
 	0,                               /*tp_repr */
 	0,                               /*tp_as_number */
 	0,                               /*tp_as_sequence */
 	0,                               /*tp_as_mapping */
 	0,                               /*tp_hash */
 	0,                               /*tp_call */
 	0,                               /*tp_str */
 	0,                               /*tp_getattro */
 	0,                               /*tp_setattro */
 	0,                               /*tp_as_buffer */
 	LAZYMANIFESTENTRIESITERATOR_TPFLAGS, /* tp_flags */
 	"Iterator for 3-tuples in a lazymanifest.",  /* tp_doc */
 	0,                               /* tp_traverse */
 	0,                               /* tp_clear */
 	0,                               /* tp_richcompare */
 	0,                               /* tp_weaklistoffset */
 	PyObject_SelfIter,               /* tp_iter: __iter__() method */
 	lmiter_iterentriesnext,          /* tp_iternext: next() method */
 };
 
 static PyObject *lmiter_iterkeysnext(PyObject *o)
 {
 	Py_ssize_t pl;
 	line *l = lmiter_nextline((lmIter *)o);
 	if (!l) {
 		return NULL;
 	}
 	pl = pathlen(l);
 	return PyBytes_FromStringAndSize(l->start, pl);
 }
 
 #define LAZYMANIFESTKEYSITERATOR_TPFLAGS Py_TPFLAGS_DEFAULT
 
 static PyTypeObject lazymanifestKeysIterator = {
 	PyVarObject_HEAD_INIT(NULL, 0) /* header */
 	"parsers.lazymanifest.keysiterator", /*tp_name */
 	sizeof(lmIter),                  /*tp_basicsize */
 	0,                               /*tp_itemsize */
 	lmiter_dealloc,                  /*tp_dealloc */
 	0,                               /*tp_print */
 	0,                               /*tp_getattr */
 	0,                               /*tp_setattr */
 	0,                               /*tp_compare */
 	0,                               /*tp_repr */
 	0,                               /*tp_as_number */
 	0,                               /*tp_as_sequence */
 	0,                               /*tp_as_mapping */
 	0,                               /*tp_hash */
 	0,                               /*tp_call */
 	0,                               /*tp_str */
 	0,                               /*tp_getattro */
 	0,                               /*tp_setattro */
 	0,                               /*tp_as_buffer */
 	LAZYMANIFESTKEYSITERATOR_TPFLAGS, /* tp_flags */
 	"Keys iterator for a lazymanifest.",  /* tp_doc */
 	0,                               /* tp_traverse */
 	0,                               /* tp_clear */
 	0,                               /* tp_richcompare */
 	0,                               /* tp_weaklistoffset */
 	PyObject_SelfIter,               /* tp_iter: __iter__() method */
 	lmiter_iterkeysnext,             /* tp_iternext: next() method */
 };
 
 static lazymanifest *lazymanifest_copy(lazymanifest *self);
 
 static PyObject *lazymanifest_getentriesiter(lazymanifest *self)
 {
 	lmIter *i = NULL;
 	lazymanifest *t = lazymanifest_copy(self);
 	if (!t) {
 		PyErr_NoMemory();
 		return NULL;
 	}
 	i = PyObject_New(lmIter, &lazymanifestEntriesIterator);
 	if (i) {
 		i->m = t;
 		i->pos = -1;
 	} else {
 		Py_DECREF(t);
 		PyErr_NoMemory();
 	}
 	return (PyObject *)i;
 }
 
 static PyObject *lazymanifest_getkeysiter(lazymanifest *self)
 {
 	lmIter *i = NULL;
 	lazymanifest *t = lazymanifest_copy(self);
 	if (!t) {
 		PyErr_NoMemory();
 		return NULL;
 	}
 	i = PyObject_New(lmIter, &lazymanifestKeysIterator);
 	if (i) {
 		i->m = t;
 		i->pos = -1;
 	} else {
 		Py_DECREF(t);
 		PyErr_NoMemory();
 	}
 	return (PyObject *)i;
 }
 
 /* __getitem__ and __setitem__ support */
 
 static Py_ssize_t lazymanifest_size(lazymanifest *self)
 {
 	return self->livelines;
 }
 
 static int linecmp(const void *left, const void *right)
 {
 	return strcmp(((const line *)left)->start,
 		      ((const line *)right)->start);
 }
 
 static PyObject *lazymanifest_getitem(lazymanifest *self, PyObject *key)
 {
 	line needle;
 	line *hit;
 	if (!PyBytes_Check(key)) {
 		PyErr_Format(PyExc_TypeError,
 			     "getitem: manifest keys must be a string.");
 		return NULL;
 	}
 	needle.start = PyBytes_AsString(key);
 	hit = bsearch(&needle, self->lines, self->numlines, sizeof(line),
 		      &linecmp);
 	if (!hit || hit->deleted) {
 		PyErr_Format(PyExc_KeyError, "No such manifest entry.");
 		return NULL;
 	}
 	return hashflags(self->nodelen, hit);
 }
 
 static int lazymanifest_delitem(lazymanifest *self, PyObject *key)
 {
 	line needle;
 	line *hit;
 	if (!PyBytes_Check(key)) {
 		PyErr_Format(PyExc_TypeError,
 			     "delitem: manifest keys must be a string.");
 		return -1;
 	}
 	needle.start = PyBytes_AsString(key);
 	hit = bsearch(&needle, self->lines, self->numlines, sizeof(line),
 		      &linecmp);
 	if (!hit || hit->deleted) {
 		PyErr_Format(PyExc_KeyError,
 			     "Tried to delete nonexistent manifest entry.");
 		return -1;
 	}
 	self->dirty = true;
 	hit->deleted = true;
 	self->livelines--;
 	return 0;
 }
 
 /* Do a binary search for the insertion point for new, creating the
  * new entry if needed. */
 static int internalsetitem(lazymanifest *self, line *new)
 {
 	int start = 0, end = self->numlines;
 	while (start < end) {
 		int pos = start + (end - start) / 2;
 		int c = linecmp(new, self->lines + pos);
 		if (c < 0)
 			end = pos;
 		else if (c > 0)
 			start = pos + 1;
 		else {
 			if (self->lines[pos].deleted)
 				self->livelines++;
 			if (self->lines[pos].from_malloc)
 				free(self->lines[pos].start);
 			start = pos;
 			goto finish;
 		}
 	}
 	/* being here means we need to do an insert */
 	if (!realloc_if_full(self)) {
 		PyErr_NoMemory();
 		return -1;
 	}
 	memmove(self->lines + start + 1, self->lines + start,
 		(self->numlines - start) * sizeof(line));
 	self->numlines++;
 	self->livelines++;
 finish:
 	self->lines[start] = *new;
 	self->dirty = true;
 	return 0;
 }
 
 static int lazymanifest_setitem(
 	lazymanifest *self, PyObject *key, PyObject *value)
 {
 	char *path;
 	Py_ssize_t plen;
 	PyObject *pyhash;
 	Py_ssize_t hlen;
 	char *hash;
 	PyObject *pyflags;
 	char *flags;
 	Py_ssize_t flen;
 	Py_ssize_t dlen;
 	char *dest;
 	int i;
 	line new;
 	if (!PyBytes_Check(key)) {
 		PyErr_Format(PyExc_TypeError,
 			     "setitem: manifest keys must be a string.");
 		return -1;
 	}
 	if (!value) {
 		return lazymanifest_delitem(self, key);
 	}
 	if (!PyTuple_Check(value) || PyTuple_Size(value) != 2) {
 		PyErr_Format(PyExc_TypeError,
 			     "Manifest values must be a tuple of (node, flags).");
 		return -1;
 	}
 	if (PyBytes_AsStringAndSize(key, &path, &plen) == -1) {
 		return -1;
 	}
 
 	pyhash = PyTuple_GetItem(value, 0);
 	if (!PyBytes_Check(pyhash)) {
 		PyErr_Format(PyExc_TypeError,
 			     "node must be a %zi bytes string", self->nodelen);
 		return -1;
 	}
 	hlen = PyBytes_Size(pyhash);
 	if (hlen != self->nodelen) {
 		PyErr_Format(PyExc_TypeError,
 			     "node must be a %zi bytes string", self->nodelen);
 		return -1;
 	}
 	hash = PyBytes_AsString(pyhash);
 
 	pyflags = PyTuple_GetItem(value, 1);
 	if (!PyBytes_Check(pyflags) || PyBytes_Size(pyflags) > 1) {
 		PyErr_Format(PyExc_TypeError,
 			     "flags must a 0 or 1 bytes string");
 		return -1;
 	}
 	if (PyBytes_AsStringAndSize(pyflags, &flags, &flen) == -1) {
 		return -1;
 	}
 	if (flen == 1) {
 		switch (*flags) {
 		case 'l':
 		case 't':
 		case 'x':
 			break;
 		default:
 			PyErr_Format(PyExc_TypeError, "invalid manifest flag");
 			return -1;
 		}
 	}
 	/* one null byte and one newline */
 	dlen = plen + hlen * 2 + 1 + flen + 1;
 	dest = malloc(dlen);
 	if (!dest) {
 		PyErr_NoMemory();
 		return -1;
 	}
 	memcpy(dest, path, plen + 1);
 	for (i = 0; i < hlen; i++) {
 		/* Cast to unsigned, so it will not get sign-extended when promoted
 		 * to int (as is done when passing to a variadic function)
 		 */
 		sprintf(dest + plen + 1 + (i * 2), "%02x", (unsigned char)hash[i]);
 	}
 	memcpy(dest + plen + 2 * hlen + 1, flags, flen);
 	dest[plen + 2 * hlen + 1 + flen] = '\n';
 	new.start = dest;
 	new.len = dlen;
 	new.hash_suffix = '\0';
 	if (hlen > 20) {
 		new.hash_suffix = hash[20];
 	}
 	new.from_malloc = true;     /* is `start` a pointer we allocated? */
 	new.deleted = false;        /* is this entry deleted? */
 	if (internalsetitem(self, &new)) {
 		return -1;
 	}
 	return 0;
 }
 
 static PyMappingMethods lazymanifest_mapping_methods = {
 	(lenfunc)lazymanifest_size,             /* mp_length */
 	(binaryfunc)lazymanifest_getitem,       /* mp_subscript */
 	(objobjargproc)lazymanifest_setitem,    /* mp_ass_subscript */
 };
 
 /* sequence methods (important or __contains__ builds an iterator) */
 
 static int lazymanifest_contains(lazymanifest *self, PyObject *key)
 {
 	line needle;
 	line *hit;
 	if (!PyBytes_Check(key)) {
 		/* Our keys are always strings, so if the contains
 		 * check is for a non-string, just return false. */
 		return 0;
 	}
 	needle.start = PyBytes_AsString(key);
 	hit = bsearch(&needle, self->lines, self->numlines, sizeof(line),
 		      &linecmp);
 	if (!hit || hit->deleted) {
 		return 0;
 	}
 	return 1;
 }
 
 static PySequenceMethods lazymanifest_seq_meths = {
 	(lenfunc)lazymanifest_size, /* sq_length */
 	0, /* sq_concat */
 	0, /* sq_repeat */
 	0, /* sq_item */
 	0, /* sq_slice */
 	0, /* sq_ass_item */
 	0, /* sq_ass_slice */
 	(objobjproc)lazymanifest_contains, /* sq_contains */
 	0, /* sq_inplace_concat */
 	0, /* sq_inplace_repeat */
 };
 
 
 /* Other methods (copy, diff, etc) */
 static PyTypeObject lazymanifestType;
 
 /* If the manifest has changes, build the new manifest text and reindex it. */
 static int compact(lazymanifest *self)
 {
 	int i;
 	ssize_t need = 0;
 	char *data;
 	line *src, *dst;
 	PyObject *pydata;
 	if (!self->dirty)
 		return 0;
 	for (i = 0; i < self->numlines; i++) {
 		if (!self->lines[i].deleted) {
 			need += self->lines[i].len;
 		}
 	}
 	pydata = PyBytes_FromStringAndSize(NULL, need);
 	if (!pydata)
 		return -1;
 	data = PyBytes_AsString(pydata);
 	if (!data) {
 		return -1;
 	}
 	src = self->lines;
 	dst = self->lines;
 	for (i = 0; i < self->numlines; i++, src++) {
 		char *tofree = NULL;
 		if (src->from_malloc) {
 			tofree = src->start;
 		}
 		if (!src->deleted) {
 			memcpy(data, src->start, src->len);
 			*dst = *src;
 			dst->start = data;
 			dst->from_malloc = false;
 			data += dst->len;
 			dst++;
 		}
 		free(tofree);
 	}
 	Py_DECREF(self->pydata);
 	self->pydata = pydata;
 	self->numlines = self->livelines;
 	self->dirty = false;
 	return 0;
 }
 
 static PyObject *lazymanifest_text(lazymanifest *self)
 {
 	if (compact(self) != 0) {
 		PyErr_NoMemory();
 		return NULL;
 	}
 	Py_INCREF(self->pydata);
 	return self->pydata;
 }
 
 static lazymanifest *lazymanifest_copy(lazymanifest *self)
 {
 	lazymanifest *copy = NULL;
 	if (compact(self) != 0) {
 		goto nomem;
 	}
 	copy = PyObject_New(lazymanifest, &lazymanifestType);
 	if (!copy) {
 		goto nomem;
 	}
 	lazymanifest_init_early(copy);
 	copy->nodelen = self->nodelen;
 	copy->numlines = self->numlines;
 	copy->livelines = self->livelines;
 	copy->dirty = false;
 	copy->lines = malloc(self->maxlines *sizeof(line));
 	if (!copy->lines) {
 		goto nomem;
 	}
 	memcpy(copy->lines, self->lines, self->numlines * sizeof(line));
 	copy->maxlines = self->maxlines;
 	copy->pydata = self->pydata;
 	Py_INCREF(copy->pydata);
 	return copy;
 nomem:
 	PyErr_NoMemory();
 	Py_XDECREF(copy);
 	return NULL;
 }
 
 static lazymanifest *lazymanifest_filtercopy(
 	lazymanifest *self, PyObject *matchfn)
 {
 	lazymanifest *copy = NULL;
 	int i;
 	if (!PyCallable_Check(matchfn)) {
 		PyErr_SetString(PyExc_TypeError, "matchfn must be callable");
 		return NULL;
 	}
 	/* compact ourselves first to avoid double-frees later when we
 	 * compact tmp so that it doesn't have random pointers to our
 	 * underlying from_malloc-data (self->pydata is safe) */
 	if (compact(self) != 0) {
 		goto nomem;
 	}
 	copy = PyObject_New(lazymanifest, &lazymanifestType);
 	if (!copy) {
 		goto nomem;
 	}
 	lazymanifest_init_early(copy);
 	copy->nodelen = self->nodelen;
 	copy->dirty = true;
 	copy->lines = malloc(self->maxlines * sizeof(line));
 	if (!copy->lines) {
 		goto nomem;
 	}
 	copy->maxlines = self->maxlines;
 	copy->numlines = 0;
 	copy->pydata = self->pydata;
 	Py_INCREF(copy->pydata);
 	for (i = 0; i < self->numlines; i++) {
 		PyObject *arglist = NULL, *result = NULL;
-		arglist = Py_BuildValue(PY23("(s)", "(y)"),
+		arglist = Py_BuildValue("(y)",
 					self->lines[i].start);
 		if (!arglist) {
 			goto bail;
 		}
 		result = PyObject_CallObject(matchfn, arglist);
 		Py_DECREF(arglist);
 		/* if the callback raised an exception, just let it
 		 * through and give up */
 		if (!result) {
 			goto bail;
 		}
 		if (PyObject_IsTrue(result)) {
 			assert(!(self->lines[i].from_malloc));
 			copy->lines[copy->numlines++] = self->lines[i];
 		}
 		Py_DECREF(result);
 	}
 	copy->livelines = copy->numlines;
 	return copy;
 nomem:
 	PyErr_NoMemory();
 bail:
 	Py_XDECREF(copy);
 	return NULL;
 }
 
 static PyObject *lazymanifest_diff(lazymanifest *self, PyObject *args)
 {
 	lazymanifest *other;
 	PyObject *pyclean = NULL;
 	bool listclean;
 	PyObject *emptyTup = NULL, *ret = NULL;
 	PyObject *es;
 	int sneedle = 0, oneedle = 0;
 	if (!PyArg_ParseTuple(args, "O!|O", &lazymanifestType, &other, &pyclean)) {
 		return NULL;
 	}
 	listclean = (!pyclean) ? false : PyObject_IsTrue(pyclean);
 	es = PyBytes_FromString("");
 	if (!es) {
 		goto nomem;
 	}
 	emptyTup = PyTuple_Pack(2, Py_None, es);
 	Py_DECREF(es);
 	if (!emptyTup) {
 		goto nomem;
 	}
 	ret = PyDict_New();
 	if (!ret) {
 		goto nomem;
 	}
 	while (sneedle != self->numlines || oneedle != other->numlines) {
 		line *left = self->lines + sneedle;
 		line *right = other->lines + oneedle;
 		int result;
 		PyObject *key;
 		PyObject *outer;
 		/* If we're looking at a deleted entry and it's not
 		 * the end of the manifest, just skip it. */
 		if (sneedle < self->numlines && left->deleted) {
 			sneedle++;
 			continue;
 		}
 		if (oneedle < other->numlines && right->deleted) {
 			oneedle++;
 			continue;
 		}
 		/* if we're at the end of either manifest, then we
 		 * know the remaining items are adds so we can skip
 		 * the strcmp. */
 		if (sneedle == self->numlines) {
 			result = 1;
 		} else if (oneedle == other->numlines) {
 			result = -1;
 		} else {
 			result = linecmp(left, right);
 		}
 		key = result <= 0 ?
 			PyBytes_FromString(left->start) :
 			PyBytes_FromString(right->start);
 		if (!key)
 			goto nomem;
 		if (result < 0) {
 			PyObject *l = hashflags(self->nodelen, left);
 			if (!l) {
 				goto nomem;
 			}
 			outer = PyTuple_Pack(2, l, emptyTup);
 			Py_DECREF(l);
 			if (!outer) {
 				goto nomem;
 			}
 			PyDict_SetItem(ret, key, outer);
 			Py_DECREF(outer);
 			sneedle++;
 		} else if (result > 0) {
 			PyObject *r = hashflags(self->nodelen, right);
 			if (!r) {
 				goto nomem;
 			}
 			outer = PyTuple_Pack(2, emptyTup, r);
 			Py_DECREF(r);
 			if (!outer) {
 				goto nomem;
 			}
 			PyDict_SetItem(ret, key, outer);
 			Py_DECREF(outer);
 			oneedle++;
 		} else {
 			/* file exists in both manifests */
 			if (left->len != right->len
 			    || memcmp(left->start, right->start, left->len)
 			    || left->hash_suffix != right->hash_suffix) {
 				PyObject *l = hashflags(self->nodelen, left);
 				PyObject *r;
 				if (!l) {
 					goto nomem;
 				}
 				r = hashflags(self->nodelen, right);
 				if (!r) {
 					Py_DECREF(l);
 					goto nomem;
 				}
 				outer = PyTuple_Pack(2, l, r);
 				Py_DECREF(l);
 				Py_DECREF(r);
 				if (!outer) {
 					goto nomem;
 				}
 				PyDict_SetItem(ret, key, outer);
 				Py_DECREF(outer);
 			} else if (listclean) {
 				PyDict_SetItem(ret, key, Py_None);
 			}
 			sneedle++;
 			oneedle++;
 		}
 		Py_DECREF(key);
 	}
 	Py_DECREF(emptyTup);
 	return ret;
 nomem:
 	PyErr_NoMemory();
 	Py_XDECREF(ret);
 	Py_XDECREF(emptyTup);
 	return NULL;
 }
 
 static PyMethodDef lazymanifest_methods[] = {
 	{"iterkeys", (PyCFunction)lazymanifest_getkeysiter, METH_NOARGS,
 	 "Iterate over file names in this lazymanifest."},
 	{"iterentries", (PyCFunction)lazymanifest_getentriesiter, METH_NOARGS,
 	 "Iterate over (path, nodeid, flags) tuples in this lazymanifest."},
 	{"copy", (PyCFunction)lazymanifest_copy, METH_NOARGS,
 	 "Make a copy of this lazymanifest."},
 	{"filtercopy", (PyCFunction)lazymanifest_filtercopy, METH_O,
 	 "Make a copy of this manifest filtered by matchfn."},
 	{"diff", (PyCFunction)lazymanifest_diff, METH_VARARGS,
 	 "Compare this lazymanifest to another one."},
 	{"text", (PyCFunction)lazymanifest_text, METH_NOARGS,
 	 "Encode this manifest to text."},
 	{NULL},
 };
 
 #define LAZYMANIFEST_TPFLAGS Py_TPFLAGS_DEFAULT
 
 static PyTypeObject lazymanifestType = {
 	PyVarObject_HEAD_INIT(NULL, 0) /* header */
 	"parsers.lazymanifest",                           /* tp_name */
 	sizeof(lazymanifest),                             /* tp_basicsize */
 	0,                                                /* tp_itemsize */
 	(destructor)lazymanifest_dealloc,                 /* tp_dealloc */
 	0,                                                /* tp_print */
 	0,                                                /* tp_getattr */
 	0,                                                /* tp_setattr */
 	0,                                                /* tp_compare */
 	0,                                                /* tp_repr */
 	0,                                                /* tp_as_number */
 	&lazymanifest_seq_meths,                          /* tp_as_sequence */
 	&lazymanifest_mapping_methods,                    /* tp_as_mapping */
 	0,                                                /* tp_hash */
 	0,                                                /* tp_call */
 	0,                                                /* tp_str */
 	0,                                                /* tp_getattro */
 	0,                                                /* tp_setattro */
 	0,                                                /* tp_as_buffer */
 	LAZYMANIFEST_TPFLAGS,                             /* tp_flags */
 	"TODO(augie)",                                    /* tp_doc */
 	0,                                                /* tp_traverse */
 	0,                                                /* tp_clear */
 	0,                                                /* tp_richcompare */
 	0,                                             /* tp_weaklistoffset */
 	(getiterfunc)lazymanifest_getkeysiter,                /* tp_iter */
 	0,                                                /* tp_iternext */
 	lazymanifest_methods,                             /* tp_methods */
 	0,                                                /* tp_members */
 	0,                                                /* tp_getset */
 	0,                                                /* tp_base */
 	0,                                                /* tp_dict */
 	0,                                                /* tp_descr_get */
 	0,                                                /* tp_descr_set */
 	0,                                                /* tp_dictoffset */
 	(initproc)lazymanifest_init,                      /* tp_init */
 	0,                                                /* tp_alloc */
 };
 
 void manifest_module_init(PyObject * mod)
 {
 	lazymanifestType.tp_new = PyType_GenericNew;
 	if (PyType_Ready(&lazymanifestType) < 0)
 		return;
 	Py_INCREF(&lazymanifestType);
 
 	PyModule_AddObject(mod, "lazymanifest",
 			   (PyObject *)&lazymanifestType);
 }

mercurial/cext/mpatch.c

@@ -1,204 +1,203 b''
 /*
  mpatch.c - efficient binary patching for Mercurial
 
  This implements a patch algorithm that's O(m + nlog n) where m is the
  size of the output and n is the number of patches.
 
  Given a list of binary patches, it unpacks each into a hunk list,
  then combines the hunk lists with a treewise recursion to form a
  single hunk list. This hunk list is then applied to the original
  text.
 
  The text (or binary) fragments are copied directly from their source
  Python objects into a preallocated output string to avoid the
  allocation of intermediate Python objects. Working memory is about 2x
  the total number of hunks.
 
  Copyright 2005, 2006 Olivia Mackall <[e-mail]>
 
  This software may be used and distributed according to the terms
  of the GNU General Public License, incorporated herein by reference.
 */
 
 #define PY_SSIZE_T_CLEAN
 #include <Python.h>
 #include <stdlib.h>
 #include <string.h>
 
 #include "bitmanipulation.h"
 #include "compat.h"
 #include "mpatch.h"
 #include "util.h"
 
 static char mpatch_doc[] = "Efficient binary patching.";
 static PyObject *mpatch_Error;
 
 static void setpyerr(int r)
 {
 	switch (r) {
 	case MPATCH_ERR_NO_MEM:
 		PyErr_NoMemory();
 		break;
 	case MPATCH_ERR_CANNOT_BE_DECODED:
 		PyErr_SetString(mpatch_Error, "patch cannot be decoded");
 		break;
 	case MPATCH_ERR_INVALID_PATCH:
 		PyErr_SetString(mpatch_Error, "invalid patch");
 		break;
 	}
 }
 
 struct mpatch_flist *cpygetitem(void *bins, ssize_t pos)
 {
 	Py_buffer buffer;
 	struct mpatch_flist *res = NULL;
 	int r;
 
 	PyObject *tmp = PyList_GetItem((PyObject *)bins, pos);
 	if (!tmp) {
 		return NULL;
 	}
 	if (PyObject_GetBuffer(tmp, &buffer, PyBUF_CONTIG_RO)) {
 		return NULL;
 	}
 	if ((r = mpatch_decode(buffer.buf, buffer.len, &res)) < 0) {
 		if (!PyErr_Occurred()) {
 			setpyerr(r);
 		}
 		res = NULL;
 	}
 
 	PyBuffer_Release(&buffer);
 	return res;
 }
 
 static PyObject *patches(PyObject *self, PyObject *args)
 {
 	PyObject *text, *bins, *result;
 	struct mpatch_flist *patch;
 	Py_buffer buffer;
 	int r = 0;
 	char *out;
 	Py_ssize_t len, outlen;
 
 	if (!PyArg_ParseTuple(args, "OO:mpatch", &text, &bins)) {
 		return NULL;
 	}
 
 	len = PyList_Size(bins);
 	if (!len) {
 		/* nothing to do */
 		Py_INCREF(text);
 		return text;
 	}
 
 	if (PyObject_GetBuffer(text, &buffer, PyBUF_CONTIG_RO)) {
 		return NULL;
 	}
 
 	patch = mpatch_fold(bins, cpygetitem, 0, len);
 	if (!patch) { /* error already set or memory error */
 		if (!PyErr_Occurred()) {
 			PyErr_NoMemory();
 		}
 		result = NULL;
 		goto cleanup;
 	}
 
 	outlen = mpatch_calcsize(buffer.len, patch);
 	if (outlen < 0) {
 		r = (int)outlen;
 		result = NULL;
 		goto cleanup;
 	}
 	result = PyBytes_FromStringAndSize(NULL, outlen);
 	if (!result) {
 		result = NULL;
 		goto cleanup;
 	}
 	out = PyBytes_AsString(result);
 	/* clang-format off */
 	{
 		Py_BEGIN_ALLOW_THREADS
 		r = mpatch_apply(out, buffer.buf, buffer.len, patch);
 		Py_END_ALLOW_THREADS
 	}
 	/* clang-format on */
 	if (r < 0) {
 		Py_DECREF(result);
 		result = NULL;
 	}
 cleanup:
 	mpatch_lfree(patch);
 	PyBuffer_Release(&buffer);
 	if (!result && !PyErr_Occurred()) {
 		setpyerr(r);
 	}
 	return result;
 }
 
 /* calculate size of a patched file directly */
 static PyObject *patchedsize(PyObject *self, PyObject *args)
 {
 	long orig, start, end, len, outlen = 0, last = 0, pos = 0;
 	Py_ssize_t patchlen;
 	char *bin;
 
-	if (!PyArg_ParseTuple(args, PY23("ls#", "ly#"), &orig, &bin,
-	                      &patchlen)) {
+	if (!PyArg_ParseTuple(args, "ly#", &orig, &bin, &patchlen)) {
 		return NULL;
 	}
 
 	while (pos >= 0 && pos < patchlen) {
 		start = getbe32(bin + pos);
 		end = getbe32(bin + pos + 4);
 		len = getbe32(bin + pos + 8);
 		if (start > end) {
 			break; /* sanity check */
 		}
 		pos += 12 + len;
 		outlen += start - last;
 		last = end;
 		outlen += len;
 	}
 
 	if (pos != patchlen) {
 		if (!PyErr_Occurred()) {
 			PyErr_SetString(mpatch_Error,
 			                "patch cannot be decoded");
 		}
 		return NULL;
 	}
 
 	outlen += orig - last;
 	return Py_BuildValue("l", outlen);
 }
 
 static PyMethodDef methods[] = {
     {"patches", patches, METH_VARARGS, "apply a series of patches\n"},
     {"patchedsize", patchedsize, METH_VARARGS, "calculed patched size\n"},
     {NULL, NULL},
 };
 
 static const int version = 1;
 
 static struct PyModuleDef mpatch_module = {
     PyModuleDef_HEAD_INIT, "mpatch", mpatch_doc, -1, methods,
 };
 
 PyMODINIT_FUNC PyInit_mpatch(void)
 {
 	PyObject *m;
 
 	m = PyModule_Create(&mpatch_module);
 	if (m == NULL)
 		return NULL;
 
 	mpatch_Error =
 	    PyErr_NewException("mercurial.cext.mpatch.mpatchError", NULL, NULL);
 	Py_INCREF(mpatch_Error);
 	PyModule_AddObject(m, "mpatchError", mpatch_Error);
 	PyModule_AddIntConstant(m, "version", version);
 
 	return m;
 }

mercurial/cext/osutil.c

@@ -1,1391 +1,1389 b''
 /*
  osutil.c - native operating system services
 
  Copyright 2007 Olivia Mackall and others
 
  This software may be used and distributed according to the terms of
  the GNU General Public License, incorporated herein by reference.
 */
 
 #define _ATFILE_SOURCE
 #define PY_SSIZE_T_CLEAN
 #include <Python.h>
 #include <errno.h>
 #include <fcntl.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 
 #ifdef _WIN32
 #include <io.h>
 #include <windows.h>
 #else
 #include <dirent.h>
 #include <signal.h>
 #include <sys/socket.h>
 #include <sys/stat.h>
 #include <sys/types.h>
 #include <unistd.h>
 #ifdef HAVE_LINUX_STATFS
 #include <linux/magic.h>
 #include <sys/vfs.h>
 #endif
 #ifdef HAVE_BSD_STATFS
 #include <sys/mount.h>
 #include <sys/param.h>
 #endif
 #endif
 
 #ifdef __APPLE__
 #include <sys/attr.h>
 #include <sys/vnode.h>
 #endif
 
 #include "util.h"
 
 /* some platforms lack the PATH_MAX definition (eg. GNU/Hurd) */
 #ifndef PATH_MAX
 #define PATH_MAX 4096
 #endif
 
 #ifdef _WIN32
 /*
 stat struct compatible with hg expectations
 Mercurial only uses st_mode, st_size and st_mtime
 the rest is kept to minimize changes between implementations
 */
 struct hg_stat {
 	int st_dev;
 	int st_mode;
 	int st_nlink;
 	__int64 st_size;
 	int st_mtime;
 	int st_ctime;
 };
 struct listdir_stat {
 	PyObject_HEAD
 	struct hg_stat st;
 };
 #else
 struct listdir_stat {
 	PyObject_HEAD
 	struct stat st;
 };
 #endif
 
 #define listdir_slot(name) \
 	static PyObject *listdir_stat_##name(PyObject *self, void *x) \
 	{ \
 		return PyLong_FromLong(((struct listdir_stat *)self)->st.name); \
 	}
 
 listdir_slot(st_dev)
 listdir_slot(st_mode)
 listdir_slot(st_nlink)
 #ifdef _WIN32
 static PyObject *listdir_stat_st_size(PyObject *self, void *x)
 {
 	return PyLong_FromLongLong(
 		(PY_LONG_LONG)((struct listdir_stat *)self)->st.st_size);
 }
 #else
 listdir_slot(st_size)
 #endif
 listdir_slot(st_mtime)
 listdir_slot(st_ctime)
 
 static struct PyGetSetDef listdir_stat_getsets[] = {
 	{"st_dev", listdir_stat_st_dev, 0, 0, 0},
 	{"st_mode", listdir_stat_st_mode, 0, 0, 0},
 	{"st_nlink", listdir_stat_st_nlink, 0, 0, 0},
 	{"st_size", listdir_stat_st_size, 0, 0, 0},
 	{"st_mtime", listdir_stat_st_mtime, 0, 0, 0},
 	{"st_ctime", listdir_stat_st_ctime, 0, 0, 0},
 	{0, 0, 0, 0, 0}
 };
 
 static PyObject *listdir_stat_new(PyTypeObject *t, PyObject *a, PyObject *k)
 {
 	return t->tp_alloc(t, 0);
 }
 
 static void listdir_stat_dealloc(PyObject *o)
 {
 	Py_TYPE(o)->tp_free(o);
 }
 
 static PyObject *listdir_stat_getitem(PyObject *self, PyObject *key)
 {
 	long index = PyLong_AsLong(key);
 	if (index == -1 && PyErr_Occurred()) {
 		return NULL;
 	}
 	if (index != 8) {
 		PyErr_Format(PyExc_IndexError, "osutil.stat objects only "
 		                               "support stat.ST_MTIME in "
 		                               "__getitem__");
 		return NULL;
 	}
 	return listdir_stat_st_mtime(self, NULL);
 }
 
 static PyMappingMethods listdir_stat_type_mapping_methods = {
 	(lenfunc)NULL,             /* mp_length */
 	(binaryfunc)listdir_stat_getitem,       /* mp_subscript */
 	(objobjargproc)NULL,    /* mp_ass_subscript */
 };
 
 static PyTypeObject listdir_stat_type = {
 	PyVarObject_HEAD_INIT(NULL, 0) /* header */
 	"osutil.stat",             /*tp_name*/
 	sizeof(struct listdir_stat), /*tp_basicsize*/
 	0,                         /*tp_itemsize*/
 	(destructor)listdir_stat_dealloc, /*tp_dealloc*/
 	0,                         /*tp_print*/
 	0,                         /*tp_getattr*/
 	0,                         /*tp_setattr*/
 	0,                         /*tp_compare*/
 	0,                         /*tp_repr*/
 	0,                         /*tp_as_number*/
 	0,                         /*tp_as_sequence*/
 	&listdir_stat_type_mapping_methods, /*tp_as_mapping*/
 	0,                         /*tp_hash */
 	0,                         /*tp_call*/
 	0,                         /*tp_str*/
 	0,                         /*tp_getattro*/
 	0,                         /*tp_setattro*/
 	0,                         /*tp_as_buffer*/
 	Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE, /*tp_flags*/
 	"stat objects",            /* tp_doc */
 	0,                         /* tp_traverse */
 	0,                         /* tp_clear */
 	0,                         /* tp_richcompare */
 	0,                         /* tp_weaklistoffset */
 	0,                         /* tp_iter */
 	0,                         /* tp_iternext */
 	0,                         /* tp_methods */
 	0,                         /* tp_members */
 	listdir_stat_getsets,      /* tp_getset */
 	0,                         /* tp_base */
 	0,                         /* tp_dict */
 	0,                         /* tp_descr_get */
 	0,                         /* tp_descr_set */
 	0,                         /* tp_dictoffset */
 	0,                         /* tp_init */
 	0,                         /* tp_alloc */
 	listdir_stat_new,          /* tp_new */
 };
 
 #ifdef _WIN32
 
 static int to_python_time(const FILETIME *tm)
 {
 	/* number of seconds between epoch and January 1 1601 */
 	const __int64 a0 = (__int64)134774L * (__int64)24L * (__int64)3600L;
 	/* conversion factor from 100ns to 1s */
 	const __int64 a1 = 10000000;
 	/* explicit (int) cast to suspend compiler warnings */
 	return (int)((((__int64)tm->dwHighDateTime << 32)
 			+ tm->dwLowDateTime) / a1 - a0);
 }
 
 static PyObject *make_item(const WIN32_FIND_DATAA *fd, int wantstat)
 {
 	PyObject *py_st;
 	struct hg_stat *stp;
 
 	int kind = (fd->dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY)
 		? _S_IFDIR : _S_IFREG;
 
 	if (!wantstat)
-		return Py_BuildValue(PY23("si", "yi"), fd->cFileName, kind);
+		return Py_BuildValue("yi", fd->cFileName, kind);
 
 	py_st = PyObject_CallObject((PyObject *)&listdir_stat_type, NULL);
 	if (!py_st)
 		return NULL;
 
 	stp = &((struct listdir_stat *)py_st)->st;
 	/*
 	use kind as st_mode
 	rwx bits on Win32 are meaningless
 	and Hg does not use them anyway
 	*/
 	stp->st_mode  = kind;
 	stp->st_mtime = to_python_time(&fd->ftLastWriteTime);
 	stp->st_ctime = to_python_time(&fd->ftCreationTime);
 	if (kind == _S_IFREG)
 		stp->st_size = ((__int64)fd->nFileSizeHigh << 32)
 				+ fd->nFileSizeLow;
-	return Py_BuildValue(PY23("siN", "yiN"), fd->cFileName,
+	return Py_BuildValue("yiN", fd->cFileName,
 		kind, py_st);
 }
 
 static PyObject *_listdir(char *path, Py_ssize_t plen, int wantstat, char *skip)
 {
 	PyObject *rval = NULL; /* initialize - return value */
 	PyObject *list;
 	HANDLE fh;
 	WIN32_FIND_DATAA fd;
 	char *pattern;
 
 	/* build the path + \* pattern string */
 	pattern = PyMem_Malloc(plen + 3); /* path + \* + \0 */
 	if (!pattern) {
 		PyErr_NoMemory();
 		goto error_nomem;
 	}
 	memcpy(pattern, path, plen);
 
 	if (plen > 0) {
 		char c = path[plen-1];
 		if (c != ':' && c != '/' && c != '\\')
 			pattern[plen++] = '\\';
 	}
 	pattern[plen++] = '*';
 	pattern[plen] = '\0';
 
 	fh = FindFirstFileA(pattern, &fd);
 	if (fh == INVALID_HANDLE_VALUE) {
 		PyErr_SetFromWindowsErrWithFilename(GetLastError(), path);
 		goto error_file;
 	}
 
 	list = PyList_New(0);
 	if (!list)
 		goto error_list;
 
 	do {
 		PyObject *item;
 
 		if (fd.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) {
 			if (!strcmp(fd.cFileName, ".")
 			|| !strcmp(fd.cFileName, ".."))
 				continue;
 
 			if (skip && !strcmp(fd.cFileName, skip)) {
 				rval = PyList_New(0);
 				goto error;
 			}
 		}
 
 		item = make_item(&fd, wantstat);
 		if (!item)
 			goto error;
 
 		if (PyList_Append(list, item)) {
 			Py_XDECREF(item);
 			goto error;
 		}
 
 		Py_XDECREF(item);
 	} while (FindNextFileA(fh, &fd));
 
 	if (GetLastError() != ERROR_NO_MORE_FILES) {
 		PyErr_SetFromWindowsErrWithFilename(GetLastError(), path);
 		goto error;
 	}
 
 	rval = list;
 	Py_XINCREF(rval);
 error:
 	Py_XDECREF(list);
 error_list:
 	FindClose(fh);
 error_file:
 	PyMem_Free(pattern);
 error_nomem:
 	return rval;
 }
 
 #else
 
 int entkind(struct dirent *ent)
 {
 #ifdef DT_REG
 	switch (ent->d_type) {
 	case DT_REG: return S_IFREG;
 	case DT_DIR: return S_IFDIR;
 	case DT_LNK: return S_IFLNK;
 	case DT_BLK: return S_IFBLK;
 	case DT_CHR: return S_IFCHR;
 	case DT_FIFO: return S_IFIFO;
 	case DT_SOCK: return S_IFSOCK;
 	}
 #endif
 	return -1;
 }
 
 static PyObject *makestat(const struct stat *st)
 {
 	PyObject *stat;
 
 	stat = PyObject_CallObject((PyObject *)&listdir_stat_type, NULL);
 	if (stat)
 		memcpy(&((struct listdir_stat *)stat)->st, st, sizeof(*st));
 	return stat;
 }
 
 static PyObject *_listdir_stat(char *path, int pathlen, int keepstat,
 			       char *skip)
 {
 	PyObject *list, *elem, *ret = NULL;
 	char fullpath[PATH_MAX + 10];
 	int kind, err;
 	struct stat st;
 	struct dirent *ent;
 	DIR *dir;
 #ifdef AT_SYMLINK_NOFOLLOW
 	int dfd = -1;
 #endif
 
 	if (pathlen >= PATH_MAX) {
 		errno = ENAMETOOLONG;
 		PyErr_SetFromErrnoWithFilename(PyExc_OSError, path);
 		goto error_value;
 	}
 	strncpy(fullpath, path, PATH_MAX);
 	fullpath[pathlen] = '/';
 
 #ifdef AT_SYMLINK_NOFOLLOW
 	dfd = open(path, O_RDONLY);
 	if (dfd == -1) {
 		PyErr_SetFromErrnoWithFilename(PyExc_OSError, path);
 		goto error_value;
 	}
 	dir = fdopendir(dfd);
 #else
 	dir = opendir(path);
 #endif
 	if (!dir) {
 		PyErr_SetFromErrnoWithFilename(PyExc_OSError, path);
 		goto error_dir;
 	}
 
 	list = PyList_New(0);
 	if (!list)
 		goto error_list;
 
 	while ((ent = readdir(dir))) {
 		if (!strcmp(ent->d_name, ".") || !strcmp(ent->d_name, ".."))
 			continue;
 
 		kind = entkind(ent);
 		if (kind == -1 || keepstat) {
 #ifdef AT_SYMLINK_NOFOLLOW
 			err = fstatat(dfd, ent->d_name, &st,
 				      AT_SYMLINK_NOFOLLOW);
 #else
 			strncpy(fullpath + pathlen + 1, ent->d_name,
 				PATH_MAX - pathlen);
 			fullpath[PATH_MAX] = '\0';
 			err = lstat(fullpath, &st);
 #endif
 			if (err == -1) {
 				/* race with file deletion? */
 				if (errno == ENOENT)
 					continue;
 				strncpy(fullpath + pathlen + 1, ent->d_name,
 					PATH_MAX - pathlen);
 				fullpath[PATH_MAX] = 0;
 				PyErr_SetFromErrnoWithFilename(PyExc_OSError,
 							       fullpath);
 				goto error;
 			}
 			kind = st.st_mode & S_IFMT;
 		}
 
 		/* quit early? */
 		if (skip && kind == S_IFDIR && !strcmp(ent->d_name, skip)) {
 			ret = PyList_New(0);
 			goto error;
 		}
 
 		if (keepstat) {
 			PyObject *stat = makestat(&st);
 			if (!stat)
 				goto error;
-			elem = Py_BuildValue(PY23("siN", "yiN"), ent->d_name,
+			elem = Py_BuildValue("yiN", ent->d_name,
 					     kind, stat);
 		} else
-			elem = Py_BuildValue(PY23("si", "yi"), ent->d_name,
+			elem = Py_BuildValue("yi", ent->d_name,
 					     kind);
 		if (!elem)
 			goto error;
 
 		PyList_Append(list, elem);
 		Py_DECREF(elem);
 	}
 
 	ret = list;
 	Py_INCREF(ret);
 
 error:
 	Py_DECREF(list);
 error_list:
 	closedir(dir);
 	/* closedir also closes its dirfd */
 	goto error_value;
 error_dir:
 #ifdef AT_SYMLINK_NOFOLLOW
 	close(dfd);
 #endif
 error_value:
 	return ret;
 }
 
 #ifdef __APPLE__
 
 typedef struct {
 	u_int32_t length;
 	attrreference_t name;
 	fsobj_type_t obj_type;
 	struct timespec mtime;
 #if __LITTLE_ENDIAN__
 	mode_t access_mask;
 	uint16_t padding;
 #else
 	uint16_t padding;
 	mode_t access_mask;
 #endif
 	off_t size;
 } __attribute__((packed)) attrbuf_entry;
 
 int attrkind(attrbuf_entry *entry)
 {
 	switch (entry->obj_type) {
 	case VREG: return S_IFREG;
 	case VDIR: return S_IFDIR;
 	case VLNK: return S_IFLNK;
 	case VBLK: return S_IFBLK;
 	case VCHR: return S_IFCHR;
 	case VFIFO: return S_IFIFO;
 	case VSOCK: return S_IFSOCK;
 	}
 	return -1;
 }
 
 /* get these many entries at a time */
 #define LISTDIR_BATCH_SIZE 50
 
 static PyObject *_listdir_batch(char *path, int pathlen, int keepstat,
 				char *skip, bool *fallback)
 {
 	PyObject *list, *elem, *ret = NULL;
 	int kind, err;
 	unsigned long index;
 	unsigned int count, old_state, new_state;
 	bool state_seen = false;
 	attrbuf_entry *entry;
 	/* from the getattrlist(2) man page: a path can be no longer than
 	   (NAME_MAX * 3 + 1) bytes. Also, "The getattrlist() function will
 	   silently truncate attribute data if attrBufSize is too small." So
 	   pass in a buffer big enough for the worst case. */
 	char attrbuf[LISTDIR_BATCH_SIZE * (sizeof(attrbuf_entry) + NAME_MAX * 3 + 1)];
 	unsigned int basep_unused;
 
 	struct stat st;
 	int dfd = -1;
 
 	/* these must match the attrbuf_entry struct, otherwise you'll end up
 	   with garbage */
 	struct attrlist requested_attr = {0};
 	requested_attr.bitmapcount = ATTR_BIT_MAP_COUNT;
 	requested_attr.commonattr = (ATTR_CMN_NAME | ATTR_CMN_OBJTYPE |
 				     ATTR_CMN_MODTIME | ATTR_CMN_ACCESSMASK);
 	requested_attr.fileattr = ATTR_FILE_DATALENGTH;
 
 	*fallback = false;
 
 	if (pathlen >= PATH_MAX) {
 		errno = ENAMETOOLONG;
 		PyErr_SetFromErrnoWithFilename(PyExc_OSError, path);
 		goto error_value;
 	}
 
 	dfd = open(path, O_RDONLY);
 	if (dfd == -1) {
 		PyErr_SetFromErrnoWithFilename(PyExc_OSError, path);
 		goto error_value;
 	}
 
 	list = PyList_New(0);
 	if (!list)
 		goto error_dir;
 
 	do {
 		count = LISTDIR_BATCH_SIZE;
 		err = getdirentriesattr(dfd, &requested_attr, &attrbuf,
 					sizeof(attrbuf), &count, &basep_unused,
 					&new_state, 0);
 		if (err < 0) {
 			if (errno == ENOTSUP) {
 				/* We're on a filesystem that doesn't support
 				   getdirentriesattr. Fall back to the
 				   stat-based implementation. */
 				*fallback = true;
 			} else
 				PyErr_SetFromErrnoWithFilename(PyExc_OSError, path);
 			goto error;
 		}
 
 		if (!state_seen) {
 			old_state = new_state;
 			state_seen = true;
 		} else if (old_state != new_state) {
 			/* There's an edge case with getdirentriesattr. Consider
 			   the following initial list of files:
 
 			   a
 			   b
 			   <--
 			   c
 			   d
 
 			   If the iteration is paused at the arrow, and b is
 			   deleted before it is resumed, getdirentriesattr will
 			   not return d at all!  Ordinarily we're expected to
 			   restart the iteration from the beginning. To avoid
 			   getting stuck in a retry loop here, fall back to
 			   stat. */
 			*fallback = true;
 			goto error;
 		}
 
 		entry = (attrbuf_entry *)attrbuf;
 
 		for (index = 0; index < count; index++) {
 			char *filename = ((char *)&entry->name) +
 				entry->name.attr_dataoffset;
 
 			if (!strcmp(filename, ".") || !strcmp(filename, ".."))
 				continue;
 
 			kind = attrkind(entry);
 			if (kind == -1) {
 				PyErr_Format(PyExc_OSError,
 					     "unknown object type %u for file "
 					     "%s%s!",
 					     entry->obj_type, path, filename);
 				goto error;
 			}
 
 			/* quit early? */
 			if (skip && kind == S_IFDIR && !strcmp(filename, skip)) {
 				ret = PyList_New(0);
 				goto error;
 			}
 
 			if (keepstat) {
 				PyObject *stat = NULL;
 				/* from the getattrlist(2) man page: "Only the
 				   permission bits ... are valid". */
 				st.st_mode = (entry->access_mask & ~S_IFMT) | kind;
 				st.st_mtime = entry->mtime.tv_sec;
 				st.st_size = entry->size;
 				stat = makestat(&st);
 				if (!stat)
 					goto error;
-				elem = Py_BuildValue(PY23("siN", "yiN"),
+				elem = Py_BuildValue("yiN",
 						     filename, kind, stat);
 			} else
-				elem = Py_BuildValue(PY23("si", "yi"),
+				elem = Py_BuildValue("yi",
 						     filename, kind);
 			if (!elem)
 				goto error;
 
 			PyList_Append(list, elem);
 			Py_DECREF(elem);
 
 			entry = (attrbuf_entry *)((char *)entry + entry->length);
 		}
 	} while (err == 0);
 
 	ret = list;
 	Py_INCREF(ret);
 
 error:
 	Py_DECREF(list);
 error_dir:
 	close(dfd);
 error_value:
 	return ret;
 }
 
 #endif /* __APPLE__ */
 
 static PyObject *_listdir(char *path, int pathlen, int keepstat, char *skip)
 {
 #ifdef __APPLE__
 	PyObject *ret;
 	bool fallback = false;
 
 	ret = _listdir_batch(path, pathlen, keepstat, skip, &fallback);
 	if (ret != NULL || !fallback)
 		return ret;
 #endif
 	return _listdir_stat(path, pathlen, keepstat, skip);
 }
 
 static PyObject *statfiles(PyObject *self, PyObject *args)
 {
 	PyObject *names, *stats;
 	Py_ssize_t i, count;
 
 	if (!PyArg_ParseTuple(args, "O:statfiles", &names))
 		return NULL;
 
 	count = PySequence_Length(names);
 	if (count == -1) {
 		PyErr_SetString(PyExc_TypeError, "not a sequence");
 		return NULL;
 	}
 
 	stats = PyList_New(count);
 	if (stats == NULL)
 		return NULL;
 
 	for (i = 0; i < count; i++) {
 		PyObject *stat, *pypath;
 		struct stat st;
 		int ret, kind;
 		char *path;
 
 		/* With a large file count or on a slow filesystem,
 		   don't block signals for long (issue4878). */
 		if ((i % 1000) == 999 && PyErr_CheckSignals() == -1)
 			goto bail;
 
 		pypath = PySequence_GetItem(names, i);
 		if (!pypath)
 			goto bail;
 		path = PyBytes_AsString(pypath);
 		if (path == NULL) {
 			Py_DECREF(pypath);
 			PyErr_SetString(PyExc_TypeError, "not a string");
 			goto bail;
 		}
 		ret = lstat(path, &st);
 		Py_DECREF(pypath);
 		kind = st.st_mode & S_IFMT;
 		if (ret != -1 && (kind == S_IFREG || kind == S_IFLNK)) {
 			stat = makestat(&st);
 			if (stat == NULL)
 				goto bail;
 			PyList_SET_ITEM(stats, i, stat);
 		} else {
 			Py_INCREF(Py_None);
 			PyList_SET_ITEM(stats, i, Py_None);
 		}
 	}
 
 	return stats;
 
 bail:
 	Py_DECREF(stats);
 	return NULL;
 }
 
 /*
  * recvfds() simply does not release GIL during blocking io operation because
  * command server is known to be single-threaded.
  *
  * Old systems such as Solaris don't provide CMSG_LEN, msg_control, etc.
  * Currently, recvfds() is not supported on these platforms.
  */
 #ifdef CMSG_LEN
 
 static ssize_t recvfdstobuf(int sockfd, int **rfds, void *cbuf, size_t cbufsize)
 {
 	char dummy[1];
 	struct iovec iov = {dummy, sizeof(dummy)};
 	struct msghdr msgh = {0};
 	struct cmsghdr *cmsg;
 
 	msgh.msg_iov = &iov;
 	msgh.msg_iovlen = 1;
 	msgh.msg_control = cbuf;
 	msgh.msg_controllen = (socklen_t)cbufsize;
 	if (recvmsg(sockfd, &msgh, 0) < 0)
 		return -1;
 
 	for (cmsg = CMSG_FIRSTHDR(&msgh); cmsg;
 	     cmsg = CMSG_NXTHDR(&msgh, cmsg)) {
 		if (cmsg->cmsg_level != SOL_SOCKET ||
 		    cmsg->cmsg_type != SCM_RIGHTS)
 			continue;
 		*rfds = (int *)CMSG_DATA(cmsg);
 		return (cmsg->cmsg_len - CMSG_LEN(0)) / sizeof(int);
 	}
 
 	*rfds = cbuf;
 	return 0;
 }
 
 static PyObject *recvfds(PyObject *self, PyObject *args)
 {
 	int sockfd;
 	int *rfds = NULL;
 	ssize_t rfdscount, i;
 	char cbuf[256];
 	PyObject *rfdslist = NULL;
 
 	if (!PyArg_ParseTuple(args, "i", &sockfd))
 		return NULL;
 
 	rfdscount = recvfdstobuf(sockfd, &rfds, cbuf, sizeof(cbuf));
 	if (rfdscount < 0)
 		return PyErr_SetFromErrno(PyExc_OSError);
 
 	rfdslist = PyList_New(rfdscount);
 	if (!rfdslist)
 		goto bail;
 	for (i = 0; i < rfdscount; i++) {
 		PyObject *obj = PyLong_FromLong(rfds[i]);
 		if (!obj)
 			goto bail;
 		PyList_SET_ITEM(rfdslist, i, obj);
 	}
 	return rfdslist;
 
 bail:
 	Py_XDECREF(rfdslist);
 	return NULL;
 }
 
 #endif /* CMSG_LEN */
 
 /* allow disabling setprocname via compiler flags */
 #ifndef SETPROCNAME_USE_NONE
 #if defined(HAVE_SETPROCTITLE)
 /* setproctitle is the first choice - available in FreeBSD */
 #define SETPROCNAME_USE_SETPROCTITLE
 #elif (defined(__linux__) || defined(__APPLE__)) && PY_MAJOR_VERSION == 2
 /* rewrite the argv buffer in place - works in Linux and OS X. Py_GetArgcArgv
  * in Python 3 returns the copied wchar_t **argv, thus unsupported. */
 #define SETPROCNAME_USE_ARGVREWRITE
 #else
 #define SETPROCNAME_USE_NONE
 #endif
 #endif /* ndef SETPROCNAME_USE_NONE */
 
 #ifndef SETPROCNAME_USE_NONE
 static PyObject *setprocname(PyObject *self, PyObject *args)
 {
 	const char *name = NULL;
-	if (!PyArg_ParseTuple(args, PY23("s", "y"), &name))
+	if (!PyArg_ParseTuple(args, "y", &name))
 		return NULL;
 
 #if defined(SETPROCNAME_USE_SETPROCTITLE)
 	setproctitle("%s", name);
 #elif defined(SETPROCNAME_USE_ARGVREWRITE)
 	{
 		static char *argvstart = NULL;
 		static size_t argvsize = 0;
 		if (argvstart == NULL) {
 			int argc = 0, i;
 			char **argv = NULL;
 			char *argvend;
 			extern void Py_GetArgcArgv(int *argc, char ***argv);
 			Py_GetArgcArgv(&argc, &argv);
 			/* Py_GetArgcArgv may not do much if a custom python
 			 * launcher is used that doesn't record the information
 			 * it needs. Let's handle this gracefully instead of
 			 * segfaulting. */
 			if (argv != NULL)
 				argvend = argvstart = argv[0];
 			else
 				argvend = argvstart = NULL;
 
 			/* Check the memory we can use. Typically, argv[i] and
 			 * argv[i + 1] are continuous. */
 			for (i = 0; i < argc; ++i) {
 				size_t len;
 				if (argv[i] > argvend || argv[i] < argvstart)
 					break; /* not continuous */
 				len = strlen(argv[i]);
 				argvend = argv[i] + len + 1 /* '\0' */;
 			}
 			if (argvend > argvstart) /* sanity check */
 				argvsize = argvend - argvstart;
 		}
 
 		if (argvstart && argvsize > 1) {
 			int n = snprintf(argvstart, argvsize, "%s", name);
 			if (n >= 0 && (size_t)n < argvsize)
 				memset(argvstart + n, 0, argvsize - n);
 		}
 	}
 #endif
 
 	Py_RETURN_NONE;
 }
 #endif /* ndef SETPROCNAME_USE_NONE */
 
 #if defined(HAVE_BSD_STATFS)
 static const char *describefstype(const struct statfs *pbuf)
 {
 	/* BSD or OSX provides a f_fstypename field */
 	return pbuf->f_fstypename;
 }
 #elif defined(HAVE_LINUX_STATFS)
 static const char *describefstype(const struct statfs *pbuf)
 {
 	/* Begin of Linux filesystems */
 #ifdef ADFS_SUPER_MAGIC
 	if (pbuf->f_type == ADFS_SUPER_MAGIC)
 		return "adfs";
 #endif
 #ifdef AFFS_SUPER_MAGIC
 	if (pbuf->f_type == AFFS_SUPER_MAGIC)
 		return "affs";
 #endif
 #ifdef AUTOFS_SUPER_MAGIC
 	if (pbuf->f_type == AUTOFS_SUPER_MAGIC)
 		return "autofs";
 #endif
 #ifdef BDEVFS_MAGIC
 	if (pbuf->f_type == BDEVFS_MAGIC)
 		return "bdevfs";
 #endif
 #ifdef BEFS_SUPER_MAGIC
 	if (pbuf->f_type == BEFS_SUPER_MAGIC)
 		return "befs";
 #endif
 #ifdef BFS_MAGIC
 	if (pbuf->f_type == BFS_MAGIC)
 		return "bfs";
 #endif
 #ifdef BINFMTFS_MAGIC
 	if (pbuf->f_type == BINFMTFS_MAGIC)
 		return "binfmtfs";
 #endif
 #ifdef BTRFS_SUPER_MAGIC
 	if (pbuf->f_type == BTRFS_SUPER_MAGIC)
 		return "btrfs";
 #endif
 #ifdef CGROUP_SUPER_MAGIC
 	if (pbuf->f_type == CGROUP_SUPER_MAGIC)
 		return "cgroup";
 #endif
 #ifdef CIFS_MAGIC_NUMBER
 	if (pbuf->f_type == CIFS_MAGIC_NUMBER)
 		return "cifs";
 #endif
 #ifdef CODA_SUPER_MAGIC
 	if (pbuf->f_type == CODA_SUPER_MAGIC)
 		return "coda";
 #endif
 #ifdef COH_SUPER_MAGIC
 	if (pbuf->f_type == COH_SUPER_MAGIC)
 		return "coh";
 #endif
 #ifdef CRAMFS_MAGIC
 	if (pbuf->f_type == CRAMFS_MAGIC)
 		return "cramfs";
 #endif
 #ifdef DEBUGFS_MAGIC
 	if (pbuf->f_type == DEBUGFS_MAGIC)
 		return "debugfs";
 #endif
 #ifdef DEVFS_SUPER_MAGIC
 	if (pbuf->f_type == DEVFS_SUPER_MAGIC)
 		return "devfs";
 #endif
 #ifdef DEVPTS_SUPER_MAGIC
 	if (pbuf->f_type == DEVPTS_SUPER_MAGIC)
 		return "devpts";
 #endif
 #ifdef EFIVARFS_MAGIC
 	if (pbuf->f_type == EFIVARFS_MAGIC)
 		return "efivarfs";
 #endif
 #ifdef EFS_SUPER_MAGIC
 	if (pbuf->f_type == EFS_SUPER_MAGIC)
 		return "efs";
 #endif
 #ifdef EXT_SUPER_MAGIC
 	if (pbuf->f_type == EXT_SUPER_MAGIC)
 		return "ext";
 #endif
 #ifdef EXT2_OLD_SUPER_MAGIC
 	if (pbuf->f_type == EXT2_OLD_SUPER_MAGIC)
 		return "ext2";
 #endif
 #ifdef EXT2_SUPER_MAGIC
 	if (pbuf->f_type == EXT2_SUPER_MAGIC)
 		return "ext2";
 #endif
 #ifdef EXT3_SUPER_MAGIC
 	if (pbuf->f_type == EXT3_SUPER_MAGIC)
 		return "ext3";
 #endif
 #ifdef EXT4_SUPER_MAGIC
 	if (pbuf->f_type == EXT4_SUPER_MAGIC)
 		return "ext4";
 #endif
 #ifdef F2FS_SUPER_MAGIC
 	if (pbuf->f_type == F2FS_SUPER_MAGIC)
 		return "f2fs";
 #endif
 #ifdef FUSE_SUPER_MAGIC
 	if (pbuf->f_type == FUSE_SUPER_MAGIC)
 		return "fuse";
 #endif
 #ifdef FUTEXFS_SUPER_MAGIC
 	if (pbuf->f_type == FUTEXFS_SUPER_MAGIC)
 		return "futexfs";
 #endif
 #ifdef HFS_SUPER_MAGIC
 	if (pbuf->f_type == HFS_SUPER_MAGIC)
 		return "hfs";
 #endif
 #ifdef HOSTFS_SUPER_MAGIC
 	if (pbuf->f_type == HOSTFS_SUPER_MAGIC)
 		return "hostfs";
 #endif
 #ifdef HPFS_SUPER_MAGIC
 	if (pbuf->f_type == HPFS_SUPER_MAGIC)
 		return "hpfs";
 #endif
 #ifdef HUGETLBFS_MAGIC
 	if (pbuf->f_type == HUGETLBFS_MAGIC)
 		return "hugetlbfs";
 #endif
 #ifdef ISOFS_SUPER_MAGIC
 	if (pbuf->f_type == ISOFS_SUPER_MAGIC)
 		return "isofs";
 #endif
 #ifdef JFFS2_SUPER_MAGIC
 	if (pbuf->f_type == JFFS2_SUPER_MAGIC)
 		return "jffs2";
 #endif
 #ifdef JFS_SUPER_MAGIC
 	if (pbuf->f_type == JFS_SUPER_MAGIC)
 		return "jfs";
 #endif
 #ifdef MINIX_SUPER_MAGIC
 	if (pbuf->f_type == MINIX_SUPER_MAGIC)
 		return "minix";
 #endif
 #ifdef MINIX2_SUPER_MAGIC
 	if (pbuf->f_type == MINIX2_SUPER_MAGIC)
 		return "minix2";
 #endif
 #ifdef MINIX3_SUPER_MAGIC
 	if (pbuf->f_type == MINIX3_SUPER_MAGIC)
 		return "minix3";
 #endif
 #ifdef MQUEUE_MAGIC
 	if (pbuf->f_type == MQUEUE_MAGIC)
 		return "mqueue";
 #endif
 #ifdef MSDOS_SUPER_MAGIC
 	if (pbuf->f_type == MSDOS_SUPER_MAGIC)
 		return "msdos";
 #endif
 #ifdef NCP_SUPER_MAGIC
 	if (pbuf->f_type == NCP_SUPER_MAGIC)
 		return "ncp";
 #endif
 #ifdef NFS_SUPER_MAGIC
 	if (pbuf->f_type == NFS_SUPER_MAGIC)
 		return "nfs";
 #endif
 #ifdef NILFS_SUPER_MAGIC
 	if (pbuf->f_type == NILFS_SUPER_MAGIC)
 		return "nilfs";
 #endif
 #ifdef NTFS_SB_MAGIC
 	if (pbuf->f_type == NTFS_SB_MAGIC)
 		return "ntfs-sb";
 #endif
 #ifdef OCFS2_SUPER_MAGIC
 	if (pbuf->f_type == OCFS2_SUPER_MAGIC)
 		return "ocfs2";
 #endif
 #ifdef OPENPROM_SUPER_MAGIC
 	if (pbuf->f_type == OPENPROM_SUPER_MAGIC)
 		return "openprom";
 #endif
 #ifdef OVERLAYFS_SUPER_MAGIC
 	if (pbuf->f_type == OVERLAYFS_SUPER_MAGIC)
 		return "overlay";
 #endif
 #ifdef PIPEFS_MAGIC
 	if (pbuf->f_type == PIPEFS_MAGIC)
 		return "pipefs";
 #endif
 #ifdef PROC_SUPER_MAGIC
 	if (pbuf->f_type == PROC_SUPER_MAGIC)
 		return "proc";
 #endif
 #ifdef PSTOREFS_MAGIC
 	if (pbuf->f_type == PSTOREFS_MAGIC)
 		return "pstorefs";
 #endif
 #ifdef QNX4_SUPER_MAGIC
 	if (pbuf->f_type == QNX4_SUPER_MAGIC)
 		return "qnx4";
 #endif
 #ifdef QNX6_SUPER_MAGIC
 	if (pbuf->f_type == QNX6_SUPER_MAGIC)
 		return "qnx6";
 #endif
 #ifdef RAMFS_MAGIC
 	if (pbuf->f_type == RAMFS_MAGIC)
 		return "ramfs";
 #endif
 #ifdef REISERFS_SUPER_MAGIC
 	if (pbuf->f_type == REISERFS_SUPER_MAGIC)
 		return "reiserfs";
 #endif
 #ifdef ROMFS_MAGIC
 	if (pbuf->f_type == ROMFS_MAGIC)
 		return "romfs";
 #endif
 #ifdef SECURITYFS_MAGIC
 	if (pbuf->f_type == SECURITYFS_MAGIC)
 		return "securityfs";
 #endif
 #ifdef SELINUX_MAGIC
 	if (pbuf->f_type == SELINUX_MAGIC)
 		return "selinux";
 #endif
 #ifdef SMACK_MAGIC
 	if (pbuf->f_type == SMACK_MAGIC)
 		return "smack";
 #endif
 #ifdef SMB_SUPER_MAGIC
 	if (pbuf->f_type == SMB_SUPER_MAGIC)
 		return "smb";
 #endif
 #ifdef SOCKFS_MAGIC
 	if (pbuf->f_type == SOCKFS_MAGIC)
 		return "sockfs";
 #endif
 #ifdef SQUASHFS_MAGIC
 	if (pbuf->f_type == SQUASHFS_MAGIC)
 		return "squashfs";
 #endif
 #ifdef SYSFS_MAGIC
 	if (pbuf->f_type == SYSFS_MAGIC)
 		return "sysfs";
 #endif
 #ifdef SYSV2_SUPER_MAGIC
 	if (pbuf->f_type == SYSV2_SUPER_MAGIC)
 		return "sysv2";
 #endif
 #ifdef SYSV4_SUPER_MAGIC
 	if (pbuf->f_type == SYSV4_SUPER_MAGIC)
 		return "sysv4";
 #endif
 #ifdef TMPFS_MAGIC
 	if (pbuf->f_type == TMPFS_MAGIC)
 		return "tmpfs";
 #endif
 #ifdef UDF_SUPER_MAGIC
 	if (pbuf->f_type == UDF_SUPER_MAGIC)
 		return "udf";
 #endif
 #ifdef UFS_MAGIC
 	if (pbuf->f_type == UFS_MAGIC)
 		return "ufs";
 #endif
 #ifdef USBDEVICE_SUPER_MAGIC
 	if (pbuf->f_type == USBDEVICE_SUPER_MAGIC)
 		return "usbdevice";
 #endif
 #ifdef V9FS_MAGIC
 	if (pbuf->f_type == V9FS_MAGIC)
 		return "v9fs";
 #endif
 #ifdef VXFS_SUPER_MAGIC
 	if (pbuf->f_type == VXFS_SUPER_MAGIC)
 		return "vxfs";
 #endif
 #ifdef XENFS_SUPER_MAGIC
 	if (pbuf->f_type == XENFS_SUPER_MAGIC)
 		return "xenfs";
 #endif
 #ifdef XENIX_SUPER_MAGIC
 	if (pbuf->f_type == XENIX_SUPER_MAGIC)
 		return "xenix";
 #endif
 #ifdef XFS_SUPER_MAGIC
 	if (pbuf->f_type == XFS_SUPER_MAGIC)
 		return "xfs";
 #endif
 	/* End of Linux filesystems */
 	return NULL;
 }
 #endif /* def HAVE_LINUX_STATFS */
 
 #if defined(HAVE_BSD_STATFS) || defined(HAVE_LINUX_STATFS)
 /* given a directory path, return filesystem type name (best-effort) */
 static PyObject *getfstype(PyObject *self, PyObject *args)
 {
 	const char *path = NULL;
 	struct statfs buf;
 	int r;
-	if (!PyArg_ParseTuple(args, PY23("s", "y"), &path))
+	if (!PyArg_ParseTuple(args, "y", &path))
 		return NULL;
 
 	memset(&buf, 0, sizeof(buf));
 	r = statfs(path, &buf);
 	if (r != 0)
 		return PyErr_SetFromErrno(PyExc_OSError);
-	return Py_BuildValue(PY23("s", "y"), describefstype(&buf));
+	return Py_BuildValue("y", describefstype(&buf));
 }
 #endif /* defined(HAVE_LINUX_STATFS) || defined(HAVE_BSD_STATFS) */
 
 #if defined(HAVE_BSD_STATFS)
 /* given a directory path, return filesystem mount point (best-effort) */
 static PyObject *getfsmountpoint(PyObject *self, PyObject *args)
 {
 	const char *path = NULL;
 	struct statfs buf;
 	int r;
-	if (!PyArg_ParseTuple(args, PY23("s", "y"), &path))
+	if (!PyArg_ParseTuple(args, "y", &path))
 		return NULL;
 
 	memset(&buf, 0, sizeof(buf));
 	r = statfs(path, &buf);
 	if (r != 0)
 		return PyErr_SetFromErrno(PyExc_OSError);
-	return Py_BuildValue(PY23("s", "y"), buf.f_mntonname);
+	return Py_BuildValue("y", buf.f_mntonname);
 }
 #endif /* defined(HAVE_BSD_STATFS) */
 
 static PyObject *unblocksignal(PyObject *self, PyObject *args)
 {
 	int sig = 0;
 	sigset_t set;
 	int r;
 	if (!PyArg_ParseTuple(args, "i", &sig))
 		return NULL;
 	r = sigemptyset(&set);
 	if (r != 0)
 		return PyErr_SetFromErrno(PyExc_OSError);
 	r = sigaddset(&set, sig);
 	if (r != 0)
 		return PyErr_SetFromErrno(PyExc_OSError);
 	r = sigprocmask(SIG_UNBLOCK, &set, NULL);
 	if (r != 0)
 		return PyErr_SetFromErrno(PyExc_OSError);
 	Py_RETURN_NONE;
 }
 
 #endif /* ndef _WIN32 */
 
 static PyObject *listdir(PyObject *self, PyObject *args, PyObject *kwargs)
 {
 	PyObject *statobj = NULL; /* initialize - optional arg */
 	PyObject *skipobj = NULL; /* initialize - optional arg */
 	char *path, *skip = NULL;
 	Py_ssize_t plen;
 	int wantstat;
 
 	static char *kwlist[] = {"path", "stat", "skip", NULL};
 
-	if (!PyArg_ParseTupleAndKeywords(args, kwargs, PY23("s#|OO:listdir",
-							    "y#|OO:listdir"),
+	if (!PyArg_ParseTupleAndKeywords(args, kwargs, "y#|OO:listdir",
 			kwlist, &path, &plen, &statobj, &skipobj))
 		return NULL;
 
 	wantstat = statobj && PyObject_IsTrue(statobj);
 
 	if (skipobj && skipobj != Py_None) {
 		skip = PyBytes_AsString(skipobj);
 		if (!skip)
 			return NULL;
 	}
 
 	return _listdir(path, plen, wantstat, skip);
 }
 
 #ifdef _WIN32
 static PyObject *posixfile(PyObject *self, PyObject *args, PyObject *kwds)
 {
 	static char *kwlist[] = {"name", "mode", "buffering", NULL};
 	PyObject *file_obj = NULL;
 	char *name = NULL;
 	char *mode = "rb";
 	DWORD access = 0;
 	DWORD creation;
 	HANDLE handle;
 	int fd, flags = 0;
 	int bufsize = -1;
 	char m0, m1, m2;
 	char fpmode[4];
 	int fppos = 0;
 	int plus;
 	FILE *fp;
 
-	if (!PyArg_ParseTupleAndKeywords(args, kwds, PY23("et|si:posixfile",
-							  "et|yi:posixfile"),
+	if (!PyArg_ParseTupleAndKeywords(args, kwds, "et|yi:posixfile",
 					 kwlist,
 					 Py_FileSystemDefaultEncoding,
 					 &name, &mode, &bufsize))
 		return NULL;
 
 	m0 = mode[0];
 	m1 = m0 ? mode[1] : '\0';
 	m2 = m1 ? mode[2] : '\0';
 	plus = m1 == '+' || m2 == '+';
 
 	fpmode[fppos++] = m0;
 	if (m1 == 'b' || m2 == 'b') {
 		flags = _O_BINARY;
 		fpmode[fppos++] = 'b';
 	}
 	else
 		flags = _O_TEXT;
 	if (m0 == 'r' && !plus) {
 		flags |= _O_RDONLY;
 		access = GENERIC_READ;
 	} else {
 		/*
 		work around http://support.microsoft.com/kb/899149 and
 		set _O_RDWR for 'w' and 'a', even if mode has no '+'
 		*/
 		flags |= _O_RDWR;
 		access = GENERIC_READ | GENERIC_WRITE;
 		fpmode[fppos++] = '+';
 	}
 	fpmode[fppos++] = '\0';
 
 	switch (m0) {
 	case 'r':
 		creation = OPEN_EXISTING;
 		break;
 	case 'w':
 		creation = CREATE_ALWAYS;
 		break;
 	case 'a':
 		creation = OPEN_ALWAYS;
 		flags |= _O_APPEND;
 		break;
 	default:
 		PyErr_Format(PyExc_ValueError,
 			     "mode string must begin with one of 'r', 'w', "
 			     "or 'a', not '%c'", m0);
 		goto bail;
 	}
 
 	handle = CreateFile(name, access,
 			    FILE_SHARE_READ | FILE_SHARE_WRITE |
 			    FILE_SHARE_DELETE,
 			    NULL,
 			    creation,
 			    FILE_ATTRIBUTE_NORMAL,
 			    0);
 
 	if (handle == INVALID_HANDLE_VALUE) {
 		PyErr_SetFromWindowsErrWithFilename(GetLastError(), name);
 		goto bail;
 	}
 
 	fd = _open_osfhandle((intptr_t)handle, flags);
 
 	if (fd == -1) {
 		CloseHandle(handle);
 		PyErr_SetFromErrnoWithFilename(PyExc_IOError, name);
 		goto bail;
 	}
 	fp = _fdopen(fd, fpmode);
 	if (fp == NULL) {
 		_close(fd);
 		PyErr_SetFromErrnoWithFilename(PyExc_IOError, name);
 		goto bail;
 	}
 
 	file_obj = PyFile_FromFile(fp, name, mode, fclose);
 	if (file_obj == NULL) {
 		fclose(fp);
 		goto bail;
 	}
 
 	PyFile_SetBufSize(file_obj, bufsize);
 bail:
 	PyMem_Free(name);
 	return file_obj;
 }
 #endif
 
 #ifdef __APPLE__
 #include <ApplicationServices/ApplicationServices.h>
 
 static PyObject *isgui(PyObject *self)
 {
 	CFDictionaryRef dict = CGSessionCopyCurrentDictionary();
 
 	if (dict != NULL) {
 		CFRelease(dict);
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 }
 #endif
 
 static char osutil_doc[] = "Native operating system services.";
 
 static PyMethodDef methods[] = {
 	{"listdir", (PyCFunction)listdir, METH_VARARGS | METH_KEYWORDS,
 	 "list a directory\n"},
 #ifdef _WIN32
 	{"posixfile", (PyCFunction)posixfile, METH_VARARGS | METH_KEYWORDS,
 	 "Open a file with POSIX-like semantics.\n"
 "On error, this function may raise either a WindowsError or an IOError."},
 #else
 	{"statfiles", (PyCFunction)statfiles, METH_VARARGS | METH_KEYWORDS,
 	 "stat a series of files or symlinks\n"
 "Returns None for non-existent entries and entries of other types.\n"},
 #ifdef CMSG_LEN
 	{"recvfds", (PyCFunction)recvfds, METH_VARARGS,
 	 "receive list of file descriptors via socket\n"},
 #endif
 #ifndef SETPROCNAME_USE_NONE
 	{"setprocname", (PyCFunction)setprocname, METH_VARARGS,
 	 "set process title (best-effort)\n"},
 #endif
 #if defined(HAVE_BSD_STATFS) || defined(HAVE_LINUX_STATFS)
 	{"getfstype", (PyCFunction)getfstype, METH_VARARGS,
 	 "get filesystem type (best-effort)\n"},
 #endif
 #if defined(HAVE_BSD_STATFS)
 	{"getfsmountpoint", (PyCFunction)getfsmountpoint, METH_VARARGS,
 	 "get filesystem mount point (best-effort)\n"},
 #endif
 	{"unblocksignal", (PyCFunction)unblocksignal, METH_VARARGS,
 	 "change signal mask to unblock a given signal\n"},
 #endif /* ndef _WIN32 */
 #ifdef __APPLE__
 	{
 		"isgui", (PyCFunction)isgui, METH_NOARGS,
 		"Is a CoreGraphics session available?"
 	},
 #endif
 	{NULL, NULL}
 };
 
 static const int version = 4;
 
 static struct PyModuleDef osutil_module = {
 	PyModuleDef_HEAD_INIT,
 	"osutil",
 	osutil_doc,
 	-1,
 	methods
 };
 
 PyMODINIT_FUNC PyInit_osutil(void)
 {
 	PyObject *m;
 	if (PyType_Ready(&listdir_stat_type) < 0)
 		return NULL;
 
 	m = PyModule_Create(&osutil_module);
 	PyModule_AddIntConstant(m, "version", version);
 	return m;
 }

mercurial/cext/parsers.c

@@ -1,1306 +1,1304 b''
 /*
  parsers.c - efficient content parsing
 
  Copyright 2008 Olivia Mackall <[e-mail]> and others
 
  This software may be used and distributed according to the terms of
  the GNU General Public License, incorporated herein by reference.
 */
 
 #define PY_SSIZE_T_CLEAN
 #include <Python.h>
 #include <ctype.h>
 #include <stddef.h>
 #include <string.h>
 
 #include "bitmanipulation.h"
 #include "charencode.h"
 #include "util.h"
 
 static const char *const versionerrortext = "Python minor version mismatch";
 
 static const int dirstate_v1_from_p2 = -2;
 static const int dirstate_v1_nonnormal = -1;
 static const int ambiguous_time = -1;
 
 static PyObject *dict_new_presized(PyObject *self, PyObject *args)
 {
 	Py_ssize_t expected_size;
 
 	if (!PyArg_ParseTuple(args, "n:make_presized_dict", &expected_size)) {
 		return NULL;
 	}
 
 	return _dict_new_presized(expected_size);
 }
 
 static PyObject *dirstate_item_new(PyTypeObject *subtype, PyObject *args,
                                    PyObject *kwds)
 {
 	/* We do all the initialization here and not a tp_init function because
 	 * dirstate_item is immutable. */
 	dirstateItemObject *t;
 	int wc_tracked;
 	int p1_tracked;
 	int p2_info;
 	int has_meaningful_data;
 	int has_meaningful_mtime;
 	int mtime_second_ambiguous;
 	int mode;
 	int size;
 	int mtime_s;
 	int mtime_ns;
 	PyObject *parentfiledata;
 	PyObject *mtime;
 	PyObject *fallback_exec;
 	PyObject *fallback_symlink;
 	static char *keywords_name[] = {
 	    "wc_tracked",          "p1_tracked",           "p2_info",
 	    "has_meaningful_data", "has_meaningful_mtime", "parentfiledata",
 	    "fallback_exec",       "fallback_symlink",     NULL,
 	};
 	wc_tracked = 0;
 	p1_tracked = 0;
 	p2_info = 0;
 	has_meaningful_mtime = 1;
 	has_meaningful_data = 1;
 	mtime_second_ambiguous = 0;
 	parentfiledata = Py_None;
 	fallback_exec = Py_None;
 	fallback_symlink = Py_None;
 	if (!PyArg_ParseTupleAndKeywords(args, kwds, "|iiiiiOOO", keywords_name,
 	                                 &wc_tracked, &p1_tracked, &p2_info,
 	                                 &has_meaningful_data,
 	                                 &has_meaningful_mtime, &parentfiledata,
 	                                 &fallback_exec, &fallback_symlink)) {
 		return NULL;
 	}
 	t = (dirstateItemObject *)subtype->tp_alloc(subtype, 1);
 	if (!t) {
 		return NULL;
 	}
 
 	t->flags = 0;
 	if (wc_tracked) {
 		t->flags |= dirstate_flag_wc_tracked;
 	}
 	if (p1_tracked) {
 		t->flags |= dirstate_flag_p1_tracked;
 	}
 	if (p2_info) {
 		t->flags |= dirstate_flag_p2_info;
 	}
 
 	if (fallback_exec != Py_None) {
 		t->flags |= dirstate_flag_has_fallback_exec;
 		if (PyObject_IsTrue(fallback_exec)) {
 			t->flags |= dirstate_flag_fallback_exec;
 		}
 	}
 	if (fallback_symlink != Py_None) {
 		t->flags |= dirstate_flag_has_fallback_symlink;
 		if (PyObject_IsTrue(fallback_symlink)) {
 			t->flags |= dirstate_flag_fallback_symlink;
 		}
 	}
 
 	if (parentfiledata != Py_None) {
 		if (!PyArg_ParseTuple(parentfiledata, "iiO", &mode, &size,
 		                      &mtime)) {
 			return NULL;
 		}
 		if (mtime != Py_None) {
 			if (!PyArg_ParseTuple(mtime, "iii", &mtime_s, &mtime_ns,
 			                      &mtime_second_ambiguous)) {
 				return NULL;
 			}
 		} else {
 			has_meaningful_mtime = 0;
 		}
 	} else {
 		has_meaningful_data = 0;
 		has_meaningful_mtime = 0;
 	}
 	if (has_meaningful_data) {
 		t->flags |= dirstate_flag_has_meaningful_data;
 		t->mode = mode;
 		t->size = size;
 		if (mtime_second_ambiguous) {
 			t->flags |= dirstate_flag_mtime_second_ambiguous;
 		}
 	} else {
 		t->mode = 0;
 		t->size = 0;
 	}
 	if (has_meaningful_mtime) {
 		t->flags |= dirstate_flag_has_mtime;
 		t->mtime_s = mtime_s;
 		t->mtime_ns = mtime_ns;
 	} else {
 		t->mtime_s = 0;
 		t->mtime_ns = 0;
 	}
 	return (PyObject *)t;
 }
 
 static void dirstate_item_dealloc(PyObject *o)
 {
 	PyObject_Del(o);
 }
 
 static inline bool dirstate_item_c_tracked(dirstateItemObject *self)
 {
 	return (self->flags & dirstate_flag_wc_tracked);
 }
 
 static inline bool dirstate_item_c_any_tracked(dirstateItemObject *self)
 {
 	const int mask = dirstate_flag_wc_tracked | dirstate_flag_p1_tracked |
 	                 dirstate_flag_p2_info;
 	return (self->flags & mask);
 }
 
 static inline bool dirstate_item_c_added(dirstateItemObject *self)
 {
 	const int mask = (dirstate_flag_wc_tracked | dirstate_flag_p1_tracked |
 	                  dirstate_flag_p2_info);
 	const int target = dirstate_flag_wc_tracked;
 	return (self->flags & mask) == target;
 }
 
 static inline bool dirstate_item_c_removed(dirstateItemObject *self)
 {
 	if (self->flags & dirstate_flag_wc_tracked) {
 		return false;
 	}
 	return (self->flags &
 	        (dirstate_flag_p1_tracked | dirstate_flag_p2_info));
 }
 
 static inline bool dirstate_item_c_merged(dirstateItemObject *self)
 {
 	return ((self->flags & dirstate_flag_wc_tracked) &&
 	        (self->flags & dirstate_flag_p1_tracked) &&
 	        (self->flags & dirstate_flag_p2_info));
 }
 
 static inline bool dirstate_item_c_from_p2(dirstateItemObject *self)
 {
 	return ((self->flags & dirstate_flag_wc_tracked) &&
 	        !(self->flags & dirstate_flag_p1_tracked) &&
 	        (self->flags & dirstate_flag_p2_info));
 }
 
 static inline char dirstate_item_c_v1_state(dirstateItemObject *self)
 {
 	if (dirstate_item_c_removed(self)) {
 		return 'r';
 	} else if (dirstate_item_c_merged(self)) {
 		return 'm';
 	} else if (dirstate_item_c_added(self)) {
 		return 'a';
 	} else {
 		return 'n';
 	}
 }
 
 static inline bool dirstate_item_c_has_fallback_exec(dirstateItemObject *self)
 {
 	return (bool)self->flags & dirstate_flag_has_fallback_exec;
 }
 
 static inline bool
 dirstate_item_c_has_fallback_symlink(dirstateItemObject *self)
 {
 	return (bool)self->flags & dirstate_flag_has_fallback_symlink;
 }
 
 static inline int dirstate_item_c_v1_mode(dirstateItemObject *self)
 {
 	if (self->flags & dirstate_flag_has_meaningful_data) {
 		return self->mode;
 	} else {
 		return 0;
 	}
 }
 
 static inline int dirstate_item_c_v1_size(dirstateItemObject *self)
 {
 	if (!(self->flags & dirstate_flag_wc_tracked) &&
 	    (self->flags & dirstate_flag_p2_info)) {
 		if (self->flags & dirstate_flag_p1_tracked) {
 			return dirstate_v1_nonnormal;
 		} else {
 			return dirstate_v1_from_p2;
 		}
 	} else if (dirstate_item_c_removed(self)) {
 		return 0;
 	} else if (self->flags & dirstate_flag_p2_info) {
 		return dirstate_v1_from_p2;
 	} else if (dirstate_item_c_added(self)) {
 		return dirstate_v1_nonnormal;
 	} else if (self->flags & dirstate_flag_has_meaningful_data) {
 		return self->size;
 	} else {
 		return dirstate_v1_nonnormal;
 	}
 }
 
 static inline int dirstate_item_c_v1_mtime(dirstateItemObject *self)
 {
 	if (dirstate_item_c_removed(self)) {
 		return 0;
 	} else if (!(self->flags & dirstate_flag_has_mtime) ||
 	           !(self->flags & dirstate_flag_p1_tracked) ||
 	           !(self->flags & dirstate_flag_wc_tracked) ||
 	           (self->flags & dirstate_flag_p2_info) ||
 	           (self->flags & dirstate_flag_mtime_second_ambiguous)) {
 		return ambiguous_time;
 	} else {
 		return self->mtime_s;
 	}
 }
 
 static PyObject *dirstate_item_v2_data(dirstateItemObject *self)
 {
 	int flags = self->flags;
 	int mode = dirstate_item_c_v1_mode(self);
 #ifdef S_IXUSR
 	/* This is for platforms with an exec bit */
 	if ((mode & S_IXUSR) != 0) {
 		flags |= dirstate_flag_mode_exec_perm;
 	} else {
 		flags &= ~dirstate_flag_mode_exec_perm;
 	}
 #else
 	flags &= ~dirstate_flag_mode_exec_perm;
 #endif
 #ifdef S_ISLNK
 	/* This is for platforms with support for symlinks */
 	if (S_ISLNK(mode)) {
 		flags |= dirstate_flag_mode_is_symlink;
 	} else {
 		flags &= ~dirstate_flag_mode_is_symlink;
 	}
 #else
 	flags &= ~dirstate_flag_mode_is_symlink;
 #endif
 	return Py_BuildValue("iiii", flags, self->size, self->mtime_s,
 	                     self->mtime_ns);
 };
 
 static PyObject *dirstate_item_v1_state(dirstateItemObject *self)
 {
 	char state = dirstate_item_c_v1_state(self);
 	return PyBytes_FromStringAndSize(&state, 1);
 };
 
 static PyObject *dirstate_item_v1_mode(dirstateItemObject *self)
 {
 	return PyLong_FromLong(dirstate_item_c_v1_mode(self));
 };
 
 static PyObject *dirstate_item_v1_size(dirstateItemObject *self)
 {
 	return PyLong_FromLong(dirstate_item_c_v1_size(self));
 };
 
 static PyObject *dirstate_item_v1_mtime(dirstateItemObject *self)
 {
 	return PyLong_FromLong(dirstate_item_c_v1_mtime(self));
 };
 
 static PyObject *dirstate_item_mtime_likely_equal_to(dirstateItemObject *self,
                                                      PyObject *other)
 {
 	int other_s;
 	int other_ns;
 	int other_second_ambiguous;
 	if (!PyArg_ParseTuple(other, "iii", &other_s, &other_ns,
 	                      &other_second_ambiguous)) {
 		return NULL;
 	}
 	if (!(self->flags & dirstate_flag_has_mtime)) {
 		Py_RETURN_FALSE;
 	}
 	if (self->mtime_s != other_s) {
 		Py_RETURN_FALSE;
 	}
 	if (self->mtime_ns == 0 || other_ns == 0) {
 		if (self->flags & dirstate_flag_mtime_second_ambiguous) {
 			Py_RETURN_FALSE;
 		} else {
 			Py_RETURN_TRUE;
 		}
 	}
 	if (self->mtime_ns == other_ns) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 /* This will never change since it's bound to V1
  */
 static inline dirstateItemObject *
 dirstate_item_from_v1_data(char state, int mode, int size, int mtime)
 {
 	dirstateItemObject *t =
 	    PyObject_New(dirstateItemObject, &dirstateItemType);
 	if (!t) {
 		return NULL;
 	}
 	t->flags = 0;
 	t->mode = 0;
 	t->size = 0;
 	t->mtime_s = 0;
 	t->mtime_ns = 0;
 
 	if (state == 'm') {
 		t->flags = (dirstate_flag_wc_tracked |
 		            dirstate_flag_p1_tracked | dirstate_flag_p2_info);
 	} else if (state == 'a') {
 		t->flags = dirstate_flag_wc_tracked;
 	} else if (state == 'r') {
 		if (size == dirstate_v1_nonnormal) {
 			t->flags =
 			    dirstate_flag_p1_tracked | dirstate_flag_p2_info;
 		} else if (size == dirstate_v1_from_p2) {
 			t->flags = dirstate_flag_p2_info;
 		} else {
 			t->flags = dirstate_flag_p1_tracked;
 		}
 	} else if (state == 'n') {
 		if (size == dirstate_v1_from_p2) {
 			t->flags =
 			    dirstate_flag_wc_tracked | dirstate_flag_p2_info;
 		} else if (size == dirstate_v1_nonnormal) {
 			t->flags =
 			    dirstate_flag_wc_tracked | dirstate_flag_p1_tracked;
 		} else if (mtime == ambiguous_time) {
 			t->flags = (dirstate_flag_wc_tracked |
 			            dirstate_flag_p1_tracked |
 			            dirstate_flag_has_meaningful_data);
 			t->mode = mode;
 			t->size = size;
 		} else {
 			t->flags = (dirstate_flag_wc_tracked |
 			            dirstate_flag_p1_tracked |
 			            dirstate_flag_has_meaningful_data |
 			            dirstate_flag_has_mtime);
 			t->mode = mode;
 			t->size = size;
 			t->mtime_s = mtime;
 		}
 	} else {
 		PyErr_Format(PyExc_RuntimeError,
 		             "unknown state: `%c` (%d, %d, %d)", state, mode,
 		             size, mtime, NULL);
 		Py_DECREF(t);
 		return NULL;
 	}
 
 	return t;
 }
 
 /* This will never change since it's bound to V1, unlike `dirstate_item_new` */
 static PyObject *dirstate_item_from_v1_meth(PyTypeObject *subtype,
                                             PyObject *args)
 {
 	/* We do all the initialization here and not a tp_init function because
 	 * dirstate_item is immutable. */
 	char state;
 	int size, mode, mtime;
 	if (!PyArg_ParseTuple(args, "ciii", &state, &mode, &size, &mtime)) {
 		return NULL;
 	}
 	return (PyObject *)dirstate_item_from_v1_data(state, mode, size, mtime);
 };
 
 static PyObject *dirstate_item_from_v2_meth(PyTypeObject *subtype,
                                             PyObject *args)
 {
 	dirstateItemObject *t =
 	    PyObject_New(dirstateItemObject, &dirstateItemType);
 	if (!t) {
 		return NULL;
 	}
 	if (!PyArg_ParseTuple(args, "iiii", &t->flags, &t->size, &t->mtime_s,
 	                      &t->mtime_ns)) {
 		return NULL;
 	}
 	if (t->flags & dirstate_flag_expected_state_is_modified) {
 		t->flags &= ~(dirstate_flag_expected_state_is_modified |
 		              dirstate_flag_has_meaningful_data |
 		              dirstate_flag_has_mtime);
 	}
 	t->mode = 0;
 	if (t->flags & dirstate_flag_has_meaningful_data) {
 		if (t->flags & dirstate_flag_mode_exec_perm) {
 			t->mode = 0755;
 		} else {
 			t->mode = 0644;
 		}
 		if (t->flags & dirstate_flag_mode_is_symlink) {
 			t->mode |= S_IFLNK;
 		} else {
 			t->mode |= S_IFREG;
 		}
 	}
 	return (PyObject *)t;
 };
 
 /* This means the next status call will have to actually check its content
    to make sure it is correct. */
 static PyObject *dirstate_item_set_possibly_dirty(dirstateItemObject *self)
 {
 	self->flags &= ~dirstate_flag_has_mtime;
 	Py_RETURN_NONE;
 }
 
 /* See docstring of the python implementation for details */
 static PyObject *dirstate_item_set_clean(dirstateItemObject *self,
                                          PyObject *args)
 {
 	int size, mode, mtime_s, mtime_ns, mtime_second_ambiguous;
 	PyObject *mtime;
 	mtime_s = 0;
 	mtime_ns = 0;
 	mtime_second_ambiguous = 0;
 	if (!PyArg_ParseTuple(args, "iiO", &mode, &size, &mtime)) {
 		return NULL;
 	}
 	if (mtime != Py_None) {
 		if (!PyArg_ParseTuple(mtime, "iii", &mtime_s, &mtime_ns,
 		                      &mtime_second_ambiguous)) {
 			return NULL;
 		}
 	} else {
 		self->flags &= ~dirstate_flag_has_mtime;
 	}
 	self->flags = dirstate_flag_wc_tracked | dirstate_flag_p1_tracked |
 	              dirstate_flag_has_meaningful_data |
 	              dirstate_flag_has_mtime;
 	if (mtime_second_ambiguous) {
 		self->flags |= dirstate_flag_mtime_second_ambiguous;
 	}
 	self->mode = mode;
 	self->size = size;
 	self->mtime_s = mtime_s;
 	self->mtime_ns = mtime_ns;
 	Py_RETURN_NONE;
 }
 
 static PyObject *dirstate_item_set_tracked(dirstateItemObject *self)
 {
 	self->flags |= dirstate_flag_wc_tracked;
 	self->flags &= ~dirstate_flag_has_mtime;
 	Py_RETURN_NONE;
 }
 
 static PyObject *dirstate_item_set_untracked(dirstateItemObject *self)
 {
 	self->flags &= ~dirstate_flag_wc_tracked;
 	self->mode = 0;
 	self->size = 0;
 	self->mtime_s = 0;
 	self->mtime_ns = 0;
 	Py_RETURN_NONE;
 }
 
 static PyObject *dirstate_item_drop_merge_data(dirstateItemObject *self)
 {
 	if (self->flags & dirstate_flag_p2_info) {
 		self->flags &= ~(dirstate_flag_p2_info |
 		                 dirstate_flag_has_meaningful_data |
 		                 dirstate_flag_has_mtime);
 		self->mode = 0;
 		self->size = 0;
 		self->mtime_s = 0;
 		self->mtime_ns = 0;
 	}
 	Py_RETURN_NONE;
 }
 static PyMethodDef dirstate_item_methods[] = {
     {"v2_data", (PyCFunction)dirstate_item_v2_data, METH_NOARGS,
      "return data suitable for v2 serialization"},
     {"v1_state", (PyCFunction)dirstate_item_v1_state, METH_NOARGS,
      "return a \"state\" suitable for v1 serialization"},
     {"v1_mode", (PyCFunction)dirstate_item_v1_mode, METH_NOARGS,
      "return a \"mode\" suitable for v1 serialization"},
     {"v1_size", (PyCFunction)dirstate_item_v1_size, METH_NOARGS,
      "return a \"size\" suitable for v1 serialization"},
     {"v1_mtime", (PyCFunction)dirstate_item_v1_mtime, METH_NOARGS,
      "return a \"mtime\" suitable for v1 serialization"},
     {"mtime_likely_equal_to", (PyCFunction)dirstate_item_mtime_likely_equal_to,
      METH_O, "True if the stored mtime is likely equal to the given mtime"},
     {"from_v1_data", (PyCFunction)dirstate_item_from_v1_meth,
      METH_VARARGS | METH_CLASS, "build a new DirstateItem object from V1 data"},
     {"from_v2_data", (PyCFunction)dirstate_item_from_v2_meth,
      METH_VARARGS | METH_CLASS, "build a new DirstateItem object from V2 data"},
     {"set_possibly_dirty", (PyCFunction)dirstate_item_set_possibly_dirty,
      METH_NOARGS, "mark a file as \"possibly dirty\""},
     {"set_clean", (PyCFunction)dirstate_item_set_clean, METH_VARARGS,
      "mark a file as \"clean\""},
     {"set_tracked", (PyCFunction)dirstate_item_set_tracked, METH_NOARGS,
      "mark a file as \"tracked\""},
     {"set_untracked", (PyCFunction)dirstate_item_set_untracked, METH_NOARGS,
      "mark a file as \"untracked\""},
     {"drop_merge_data", (PyCFunction)dirstate_item_drop_merge_data, METH_NOARGS,
      "remove all \"merge-only\" from a DirstateItem"},
     {NULL} /* Sentinel */
 };
 
 static PyObject *dirstate_item_get_mode(dirstateItemObject *self)
 {
 	return PyLong_FromLong(dirstate_item_c_v1_mode(self));
 };
 
 static PyObject *dirstate_item_get_size(dirstateItemObject *self)
 {
 	return PyLong_FromLong(dirstate_item_c_v1_size(self));
 };
 
 static PyObject *dirstate_item_get_mtime(dirstateItemObject *self)
 {
 	return PyLong_FromLong(dirstate_item_c_v1_mtime(self));
 };
 
 static PyObject *dirstate_item_get_state(dirstateItemObject *self)
 {
 	char state = dirstate_item_c_v1_state(self);
 	return PyBytes_FromStringAndSize(&state, 1);
 };
 
 static PyObject *dirstate_item_get_has_fallback_exec(dirstateItemObject *self)
 {
 	if (dirstate_item_c_has_fallback_exec(self)) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyObject *dirstate_item_get_fallback_exec(dirstateItemObject *self)
 {
 	if (dirstate_item_c_has_fallback_exec(self)) {
 		if (self->flags & dirstate_flag_fallback_exec) {
 			Py_RETURN_TRUE;
 		} else {
 			Py_RETURN_FALSE;
 		}
 	} else {
 		Py_RETURN_NONE;
 	}
 };
 
 static int dirstate_item_set_fallback_exec(dirstateItemObject *self,
                                            PyObject *value)
 {
 	if ((value == Py_None) || (value == NULL)) {
 		self->flags &= ~dirstate_flag_has_fallback_exec;
 	} else {
 		self->flags |= dirstate_flag_has_fallback_exec;
 		if (PyObject_IsTrue(value)) {
 			self->flags |= dirstate_flag_fallback_exec;
 		} else {
 			self->flags &= ~dirstate_flag_fallback_exec;
 		}
 	}
 	return 0;
 };
 
 static PyObject *
 dirstate_item_get_has_fallback_symlink(dirstateItemObject *self)
 {
 	if (dirstate_item_c_has_fallback_symlink(self)) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyObject *dirstate_item_get_fallback_symlink(dirstateItemObject *self)
 {
 	if (dirstate_item_c_has_fallback_symlink(self)) {
 		if (self->flags & dirstate_flag_fallback_symlink) {
 			Py_RETURN_TRUE;
 		} else {
 			Py_RETURN_FALSE;
 		}
 	} else {
 		Py_RETURN_NONE;
 	}
 };
 
 static int dirstate_item_set_fallback_symlink(dirstateItemObject *self,
                                               PyObject *value)
 {
 	if ((value == Py_None) || (value == NULL)) {
 		self->flags &= ~dirstate_flag_has_fallback_symlink;
 	} else {
 		self->flags |= dirstate_flag_has_fallback_symlink;
 		if (PyObject_IsTrue(value)) {
 			self->flags |= dirstate_flag_fallback_symlink;
 		} else {
 			self->flags &= ~dirstate_flag_fallback_symlink;
 		}
 	}
 	return 0;
 };
 
 static PyObject *dirstate_item_get_tracked(dirstateItemObject *self)
 {
 	if (dirstate_item_c_tracked(self)) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 static PyObject *dirstate_item_get_p1_tracked(dirstateItemObject *self)
 {
 	if (self->flags & dirstate_flag_p1_tracked) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyObject *dirstate_item_get_added(dirstateItemObject *self)
 {
 	if (dirstate_item_c_added(self)) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyObject *dirstate_item_get_p2_info(dirstateItemObject *self)
 {
 	if (self->flags & dirstate_flag_wc_tracked &&
 	    self->flags & dirstate_flag_p2_info) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyObject *dirstate_item_get_merged(dirstateItemObject *self)
 {
 	if (dirstate_item_c_merged(self)) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyObject *dirstate_item_get_from_p2(dirstateItemObject *self)
 {
 	if (dirstate_item_c_from_p2(self)) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyObject *dirstate_item_get_maybe_clean(dirstateItemObject *self)
 {
 	if (!(self->flags & dirstate_flag_wc_tracked)) {
 		Py_RETURN_FALSE;
 	} else if (!(self->flags & dirstate_flag_p1_tracked)) {
 		Py_RETURN_FALSE;
 	} else if (self->flags & dirstate_flag_p2_info) {
 		Py_RETURN_FALSE;
 	} else {
 		Py_RETURN_TRUE;
 	}
 };
 
 static PyObject *dirstate_item_get_any_tracked(dirstateItemObject *self)
 {
 	if (dirstate_item_c_any_tracked(self)) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyObject *dirstate_item_get_removed(dirstateItemObject *self)
 {
 	if (dirstate_item_c_removed(self)) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyGetSetDef dirstate_item_getset[] = {
     {"mode", (getter)dirstate_item_get_mode, NULL, "mode", NULL},
     {"size", (getter)dirstate_item_get_size, NULL, "size", NULL},
     {"mtime", (getter)dirstate_item_get_mtime, NULL, "mtime", NULL},
     {"state", (getter)dirstate_item_get_state, NULL, "state", NULL},
     {"has_fallback_exec", (getter)dirstate_item_get_has_fallback_exec, NULL,
      "has_fallback_exec", NULL},
     {"fallback_exec", (getter)dirstate_item_get_fallback_exec,
      (setter)dirstate_item_set_fallback_exec, "fallback_exec", NULL},
     {"has_fallback_symlink", (getter)dirstate_item_get_has_fallback_symlink,
      NULL, "has_fallback_symlink", NULL},
     {"fallback_symlink", (getter)dirstate_item_get_fallback_symlink,
      (setter)dirstate_item_set_fallback_symlink, "fallback_symlink", NULL},
     {"tracked", (getter)dirstate_item_get_tracked, NULL, "tracked", NULL},
     {"p1_tracked", (getter)dirstate_item_get_p1_tracked, NULL, "p1_tracked",
      NULL},
     {"added", (getter)dirstate_item_get_added, NULL, "added", NULL},
     {"p2_info", (getter)dirstate_item_get_p2_info, NULL, "p2_info", NULL},
     {"merged", (getter)dirstate_item_get_merged, NULL, "merged", NULL},
     {"from_p2", (getter)dirstate_item_get_from_p2, NULL, "from_p2", NULL},
     {"maybe_clean", (getter)dirstate_item_get_maybe_clean, NULL, "maybe_clean",
      NULL},
     {"any_tracked", (getter)dirstate_item_get_any_tracked, NULL, "any_tracked",
      NULL},
     {"removed", (getter)dirstate_item_get_removed, NULL, "removed", NULL},
     {NULL} /* Sentinel */
 };
 
 PyTypeObject dirstateItemType = {
     PyVarObject_HEAD_INIT(NULL, 0)     /* header */
     "dirstate_tuple",                  /* tp_name */
     sizeof(dirstateItemObject),        /* tp_basicsize */
     0,                                 /* tp_itemsize */
     (destructor)dirstate_item_dealloc, /* tp_dealloc */
     0,                                 /* tp_print */
     0,                                 /* tp_getattr */
     0,                                 /* tp_setattr */
     0,                                 /* tp_compare */
     0,                                 /* tp_repr */
     0,                                 /* tp_as_number */
     0,                                 /* tp_as_sequence */
     0,                                 /* tp_as_mapping */
     0,                                 /* tp_hash  */
     0,                                 /* tp_call */
     0,                                 /* tp_str */
     0,                                 /* tp_getattro */
     0,                                 /* tp_setattro */
     0,                                 /* tp_as_buffer */
     Py_TPFLAGS_DEFAULT,                /* tp_flags */
     "dirstate tuple",                  /* tp_doc */
     0,                                 /* tp_traverse */
     0,                                 /* tp_clear */
     0,                                 /* tp_richcompare */
     0,                                 /* tp_weaklistoffset */
     0,                                 /* tp_iter */
     0,                                 /* tp_iternext */
     dirstate_item_methods,             /* tp_methods */
     0,                                 /* tp_members */
     dirstate_item_getset,              /* tp_getset */
     0,                                 /* tp_base */
     0,                                 /* tp_dict */
     0,                                 /* tp_descr_get */
     0,                                 /* tp_descr_set */
     0,                                 /* tp_dictoffset */
     0,                                 /* tp_init */
     0,                                 /* tp_alloc */
     dirstate_item_new,                 /* tp_new */
 };
 
 static PyObject *parse_dirstate(PyObject *self, PyObject *args)
 {
 	PyObject *dmap, *cmap, *parents = NULL, *ret = NULL;
 	PyObject *fname = NULL, *cname = NULL, *entry = NULL;
 	char state, *cur, *str, *cpos;
 	int mode, size, mtime;
 	unsigned int flen, pos = 40;
 	Py_ssize_t len = 40;
 	Py_ssize_t readlen;
 
-	if (!PyArg_ParseTuple(
-	        args, PY23("O!O!s#:parse_dirstate", "O!O!y#:parse_dirstate"),
-	        &PyDict_Type, &dmap, &PyDict_Type, &cmap, &str, &readlen)) {
+	if (!PyArg_ParseTuple(args, "O!O!y#:parse_dirstate", &PyDict_Type,
+	                      &dmap, &PyDict_Type, &cmap, &str, &readlen)) {
 		goto quit;
 	}
 
 	len = readlen;
 
 	/* read parents */
 	if (len < 40) {
 		PyErr_SetString(PyExc_ValueError,
 		                "too little data for parents");
 		goto quit;
 	}
 
-	parents = Py_BuildValue(PY23("s#s#", "y#y#"), str, (Py_ssize_t)20,
-	                        str + 20, (Py_ssize_t)20);
+	parents = Py_BuildValue("y#y#", str, (Py_ssize_t)20, str + 20,
+	                        (Py_ssize_t)20);
 	if (!parents) {
 		goto quit;
 	}
 
 	/* read filenames */
 	while (pos >= 40 && pos < len) {
 		if (pos + 17 > len) {
 			PyErr_SetString(PyExc_ValueError,
 			                "overflow in dirstate");
 			goto quit;
 		}
 		cur = str + pos;
 		/* unpack header */
 		state = *cur;
 		mode = getbe32(cur + 1);
 		size = getbe32(cur + 5);
 		mtime = getbe32(cur + 9);
 		flen = getbe32(cur + 13);
 		pos += 17;
 		cur += 17;
 		if (flen > len - pos) {
 			PyErr_SetString(PyExc_ValueError,
 			                "overflow in dirstate");
 			goto quit;
 		}
 
 		entry = (PyObject *)dirstate_item_from_v1_data(state, mode,
 		                                               size, mtime);
 		if (!entry)
 			goto quit;
 		cpos = memchr(cur, 0, flen);
 		if (cpos) {
 			fname = PyBytes_FromStringAndSize(cur, cpos - cur);
 			cname = PyBytes_FromStringAndSize(
 			    cpos + 1, flen - (cpos - cur) - 1);
 			if (!fname || !cname ||
 			    PyDict_SetItem(cmap, fname, cname) == -1 ||
 			    PyDict_SetItem(dmap, fname, entry) == -1) {
 				goto quit;
 			}
 			Py_DECREF(cname);
 		} else {
 			fname = PyBytes_FromStringAndSize(cur, flen);
 			if (!fname ||
 			    PyDict_SetItem(dmap, fname, entry) == -1) {
 				goto quit;
 			}
 		}
 		Py_DECREF(fname);
 		Py_DECREF(entry);
 		fname = cname = entry = NULL;
 		pos += flen;
 	}
 
 	ret = parents;
 	Py_INCREF(ret);
 quit:
 	Py_XDECREF(fname);
 	Py_XDECREF(cname);
 	Py_XDECREF(entry);
 	Py_XDECREF(parents);
 	return ret;
 }
 
 /*
  * Efficiently pack a dirstate object into its on-disk format.
  */
 static PyObject *pack_dirstate(PyObject *self, PyObject *args)
 {
 	PyObject *packobj = NULL;
 	PyObject *map, *copymap, *pl, *mtime_unset = NULL;
 	Py_ssize_t nbytes, pos, l;
 	PyObject *k, *v = NULL, *pn;
 	char *p, *s;
 
 	if (!PyArg_ParseTuple(args, "O!O!O!:pack_dirstate", &PyDict_Type, &map,
 	                      &PyDict_Type, &copymap, &PyTuple_Type, &pl)) {
 		return NULL;
 	}
 
 	if (PyTuple_Size(pl) != 2) {
 		PyErr_SetString(PyExc_TypeError, "expected 2-element tuple");
 		return NULL;
 	}
 
 	/* Figure out how much we need to allocate. */
 	for (nbytes = 40, pos = 0; PyDict_Next(map, &pos, &k, &v);) {
 		PyObject *c;
 		if (!PyBytes_Check(k)) {
 			PyErr_SetString(PyExc_TypeError, "expected string key");
 			goto bail;
 		}
 		nbytes += PyBytes_GET_SIZE(k) + 17;
 		c = PyDict_GetItem(copymap, k);
 		if (c) {
 			if (!PyBytes_Check(c)) {
 				PyErr_SetString(PyExc_TypeError,
 				                "expected string key");
 				goto bail;
 			}
 			nbytes += PyBytes_GET_SIZE(c) + 1;
 		}
 	}
 
 	packobj = PyBytes_FromStringAndSize(NULL, nbytes);
 	if (packobj == NULL) {
 		goto bail;
 	}
 
 	p = PyBytes_AS_STRING(packobj);
 
 	pn = PyTuple_GET_ITEM(pl, 0);
 	if (PyBytes_AsStringAndSize(pn, &s, &l) == -1 || l != 20) {
 		PyErr_SetString(PyExc_TypeError, "expected a 20-byte hash");
 		goto bail;
 	}
 	memcpy(p, s, l);
 	p += 20;
 	pn = PyTuple_GET_ITEM(pl, 1);
 	if (PyBytes_AsStringAndSize(pn, &s, &l) == -1 || l != 20) {
 		PyErr_SetString(PyExc_TypeError, "expected a 20-byte hash");
 		goto bail;
 	}
 	memcpy(p, s, l);
 	p += 20;
 
 	for (pos = 0; PyDict_Next(map, &pos, &k, &v);) {
 		dirstateItemObject *tuple;
 		char state;
 		int mode, size, mtime;
 		Py_ssize_t len, l;
 		PyObject *o;
 		char *t;
 
 		if (!dirstate_tuple_check(v)) {
 			PyErr_SetString(PyExc_TypeError,
 			                "expected a dirstate tuple");
 			goto bail;
 		}
 		tuple = (dirstateItemObject *)v;
 
 		state = dirstate_item_c_v1_state(tuple);
 		mode = dirstate_item_c_v1_mode(tuple);
 		size = dirstate_item_c_v1_size(tuple);
 		mtime = dirstate_item_c_v1_mtime(tuple);
 		*p++ = state;
 		putbe32((uint32_t)mode, p);
 		putbe32((uint32_t)size, p + 4);
 		putbe32((uint32_t)mtime, p + 8);
 		t = p + 12;
 		p += 16;
 		len = PyBytes_GET_SIZE(k);
 		memcpy(p, PyBytes_AS_STRING(k), len);
 		p += len;
 		o = PyDict_GetItem(copymap, k);
 		if (o) {
 			*p++ = '\0';
 			l = PyBytes_GET_SIZE(o);
 			memcpy(p, PyBytes_AS_STRING(o), l);
 			p += l;
 			len += l + 1;
 		}
 		putbe32((uint32_t)len, t);
 	}
 
 	pos = p - PyBytes_AS_STRING(packobj);
 	if (pos != nbytes) {
 		PyErr_Format(PyExc_SystemError, "bad dirstate size: %ld != %ld",
 		             (long)pos, (long)nbytes);
 		goto bail;
 	}
 
 	return packobj;
 bail:
 	Py_XDECREF(mtime_unset);
 	Py_XDECREF(packobj);
 	Py_XDECREF(v);
 	return NULL;
 }
 
 #define BUMPED_FIX 1
 #define USING_SHA_256 2
 #define FM1_HEADER_SIZE (4 + 8 + 2 + 2 + 1 + 1 + 1)
 
 static PyObject *readshas(const char *source, unsigned char num,
                           Py_ssize_t hashwidth)
 {
 	int i;
 	PyObject *list = PyTuple_New(num);
 	if (list == NULL) {
 		return NULL;
 	}
 	for (i = 0; i < num; i++) {
 		PyObject *hash = PyBytes_FromStringAndSize(source, hashwidth);
 		if (hash == NULL) {
 			Py_DECREF(list);
 			return NULL;
 		}
 		PyTuple_SET_ITEM(list, i, hash);
 		source += hashwidth;
 	}
 	return list;
 }
 
 static PyObject *fm1readmarker(const char *databegin, const char *dataend,
                                uint32_t *msize)
 {
 	const char *data = databegin;
 	const char *meta;
 
 	double mtime;
 	int16_t tz;
 	uint16_t flags;
 	unsigned char nsuccs, nparents, nmetadata;
 	Py_ssize_t hashwidth = 20;
 
 	PyObject *prec = NULL, *parents = NULL, *succs = NULL;
 	PyObject *metadata = NULL, *ret = NULL;
 	int i;
 
 	if (data + FM1_HEADER_SIZE > dataend) {
 		goto overflow;
 	}
 
 	*msize = getbe32(data);
 	data += 4;
 	mtime = getbefloat64(data);
 	data += 8;
 	tz = getbeint16(data);
 	data += 2;
 	flags = getbeuint16(data);
 	data += 2;
 
 	if (flags & USING_SHA_256) {
 		hashwidth = 32;
 	}
 
 	nsuccs = (unsigned char)(*data++);
 	nparents = (unsigned char)(*data++);
 	nmetadata = (unsigned char)(*data++);
 
 	if (databegin + *msize > dataend) {
 		goto overflow;
 	}
 	dataend = databegin + *msize; /* narrow down to marker size */
 
 	if (data + hashwidth > dataend) {
 		goto overflow;
 	}
 	prec = PyBytes_FromStringAndSize(data, hashwidth);
 	data += hashwidth;
 	if (prec == NULL) {
 		goto bail;
 	}
 
 	if (data + nsuccs * hashwidth > dataend) {
 		goto overflow;
 	}
 	succs = readshas(data, nsuccs, hashwidth);
 	if (succs == NULL) {
 		goto bail;
 	}
 	data += nsuccs * hashwidth;
 
 	if (nparents == 1 || nparents == 2) {
 		if (data + nparents * hashwidth > dataend) {
 			goto overflow;
 		}
 		parents = readshas(data, nparents, hashwidth);
 		if (parents == NULL) {
 			goto bail;
 		}
 		data += nparents * hashwidth;
 	} else {
 		parents = Py_None;
 		Py_INCREF(parents);
 	}
 
 	if (data + 2 * nmetadata > dataend) {
 		goto overflow;
 	}
 	meta = data + (2 * nmetadata);
 	metadata = PyTuple_New(nmetadata);
 	if (metadata == NULL) {
 		goto bail;
 	}
 	for (i = 0; i < nmetadata; i++) {
 		PyObject *tmp, *left = NULL, *right = NULL;
 		Py_ssize_t leftsize = (unsigned char)(*data++);
 		Py_ssize_t rightsize = (unsigned char)(*data++);
 		if (meta + leftsize + rightsize > dataend) {
 			goto overflow;
 		}
 		left = PyBytes_FromStringAndSize(meta, leftsize);
 		meta += leftsize;
 		right = PyBytes_FromStringAndSize(meta, rightsize);
 		meta += rightsize;
 		tmp = PyTuple_New(2);
 		if (!left || !right || !tmp) {
 			Py_XDECREF(left);
 			Py_XDECREF(right);
 			Py_XDECREF(tmp);
 			goto bail;
 		}
 		PyTuple_SET_ITEM(tmp, 0, left);
 		PyTuple_SET_ITEM(tmp, 1, right);
 		PyTuple_SET_ITEM(metadata, i, tmp);
 	}
 	ret = Py_BuildValue("(OOHO(di)O)", prec, succs, flags, metadata, mtime,
 	                    (int)tz * 60, parents);
 	goto bail; /* return successfully */
 
 overflow:
 	PyErr_SetString(PyExc_ValueError, "overflow in obsstore");
 bail:
 	Py_XDECREF(prec);
 	Py_XDECREF(succs);
 	Py_XDECREF(metadata);
 	Py_XDECREF(parents);
 	return ret;
 }
 
 static PyObject *fm1readmarkers(PyObject *self, PyObject *args)
 {
 	const char *data, *dataend;
 	Py_ssize_t datalen, offset, stop;
 	PyObject *markers = NULL;
 
-	if (!PyArg_ParseTuple(args, PY23("s#nn", "y#nn"), &data, &datalen,
-	                      &offset, &stop)) {
+	if (!PyArg_ParseTuple(args, "y#nn", &data, &datalen, &offset, &stop)) {
 		return NULL;
 	}
 	if (offset < 0) {
 		PyErr_SetString(PyExc_ValueError,
 		                "invalid negative offset in fm1readmarkers");
 		return NULL;
 	}
 	if (stop > datalen) {
 		PyErr_SetString(
 		    PyExc_ValueError,
 		    "stop longer than data length in fm1readmarkers");
 		return NULL;
 	}
 	dataend = data + datalen;
 	data += offset;
 	markers = PyList_New(0);
 	if (!markers) {
 		return NULL;
 	}
 	while (offset < stop) {
 		uint32_t msize;
 		int error;
 		PyObject *record = fm1readmarker(data, dataend, &msize);
 		if (!record) {
 			goto bail;
 		}
 		error = PyList_Append(markers, record);
 		Py_DECREF(record);
 		if (error) {
 			goto bail;
 		}
 		data += msize;
 		offset += msize;
 	}
 	return markers;
 bail:
 	Py_DECREF(markers);
 	return NULL;
 }
 
 static char parsers_doc[] = "Efficient content parsing.";
 
 PyObject *encodedir(PyObject *self, PyObject *args);
 PyObject *pathencode(PyObject *self, PyObject *args);
 PyObject *lowerencode(PyObject *self, PyObject *args);
 PyObject *parse_index2(PyObject *self, PyObject *args, PyObject *kwargs);
 
 static PyMethodDef methods[] = {
     {"pack_dirstate", pack_dirstate, METH_VARARGS, "pack a dirstate\n"},
     {"parse_dirstate", parse_dirstate, METH_VARARGS, "parse a dirstate\n"},
     {"parse_index2", (PyCFunction)parse_index2, METH_VARARGS | METH_KEYWORDS,
      "parse a revlog index\n"},
     {"isasciistr", isasciistr, METH_VARARGS, "check if an ASCII string\n"},
     {"asciilower", asciilower, METH_VARARGS, "lowercase an ASCII string\n"},
     {"asciiupper", asciiupper, METH_VARARGS, "uppercase an ASCII string\n"},
     {"dict_new_presized", dict_new_presized, METH_VARARGS,
      "construct a dict with an expected size\n"},
     {"make_file_foldmap", make_file_foldmap, METH_VARARGS,
      "make file foldmap\n"},
     {"jsonescapeu8fast", jsonescapeu8fast, METH_VARARGS,
      "escape a UTF-8 byte string to JSON (fast path)\n"},
     {"encodedir", encodedir, METH_VARARGS, "encodedir a path\n"},
     {"pathencode", pathencode, METH_VARARGS, "fncache-encode a path\n"},
     {"lowerencode", lowerencode, METH_VARARGS, "lower-encode a path\n"},
     {"fm1readmarkers", fm1readmarkers, METH_VARARGS,
      "parse v1 obsolete markers\n"},
     {NULL, NULL}};
 
 void dirs_module_init(PyObject *mod);
 void manifest_module_init(PyObject *mod);
 void revlog_module_init(PyObject *mod);
 
 static const int version = 20;
 
 static void module_init(PyObject *mod)
 {
 	PyModule_AddIntConstant(mod, "version", version);
 
 	/* This module constant has two purposes.  First, it lets us unit test
 	 * the ImportError raised without hard-coding any error text.  This
 	 * means we can change the text in the future without breaking tests,
 	 * even across changesets without a recompile.  Second, its presence
 	 * can be used to determine whether the version-checking logic is
 	 * present, which also helps in testing across changesets without a
 	 * recompile.  Note that this means the pure-Python version of parsers
 	 * should not have this module constant. */
 	PyModule_AddStringConstant(mod, "versionerrortext", versionerrortext);
 
 	dirs_module_init(mod);
 	manifest_module_init(mod);
 	revlog_module_init(mod);
 
 	if (PyType_Ready(&dirstateItemType) < 0) {
 		return;
 	}
 	Py_INCREF(&dirstateItemType);
 	PyModule_AddObject(mod, "DirstateItem", (PyObject *)&dirstateItemType);
 }
 
 static int check_python_version(void)
 {
 	PyObject *sys = PyImport_ImportModule("sys"), *ver;
 	long hexversion;
 	if (!sys) {
 		return -1;
 	}
 	ver = PyObject_GetAttrString(sys, "hexversion");
 	Py_DECREF(sys);
 	if (!ver) {
 		return -1;
 	}
 	hexversion = PyLong_AsLong(ver);
 	Py_DECREF(ver);
 	/* sys.hexversion is a 32-bit number by default, so the -1 case
 	 * should only occur in unusual circumstances (e.g. if sys.hexversion
 	 * is manually set to an invalid value). */
 	if ((hexversion == -1) || (hexversion >> 16 != PY_VERSION_HEX >> 16)) {
 		PyErr_Format(PyExc_ImportError,
 		             "%s: The Mercurial extension "
 		             "modules were compiled with Python " PY_VERSION
 		             ", but "
 		             "Mercurial is currently using Python with "
 		             "sys.hexversion=%ld: "
 		             "Python %s\n at: %s",
 		             versionerrortext, hexversion, Py_GetVersion(),
 		             Py_GetProgramFullPath());
 		return -1;
 	}
 	return 0;
 }
 
 static struct PyModuleDef parsers_module = {PyModuleDef_HEAD_INIT, "parsers",
                                             parsers_doc, -1, methods};
 
 PyMODINIT_FUNC PyInit_parsers(void)
 {
 	PyObject *mod;
 
 	if (check_python_version() == -1)
 		return NULL;
 	mod = PyModule_Create(&parsers_module);
 	module_init(mod);
 	return mod;
 }

mercurial/cext/pathencode.c

@@ -1,803 +1,802 b''
 /*
  pathencode.c - efficient path name encoding
 
  Copyright 2012 Facebook
 
  This software may be used and distributed according to the terms of
  the GNU General Public License, incorporated herein by reference.
 */
 
 /*
  * An implementation of the name encoding scheme used by the fncache
  * store.  The common case is of a path < 120 bytes long, which is
  * handled either in a single pass with no allocations or two passes
  * with a single allocation.  For longer paths, multiple passes are
  * required.
  */
 
 #define PY_SSIZE_T_CLEAN
 #include <Python.h>
 #include <assert.h>
 #include <ctype.h>
 #include <stdlib.h>
 #include <string.h>
 #include "pythoncapi_compat.h"
 
 #include "util.h"
 
 /* state machine for the fast path */
 enum path_state {
 	START, /* first byte of a path component */
 	A,     /* "AUX" */
 	AU,
 	THIRD, /* third of a 3-byte sequence, e.g. "AUX", "NUL" */
 	C,     /* "CON" or "COMn" */
 	CO,
 	COMLPT, /* "COM" or "LPT" */
 	COMLPTn,
 	L,
 	LP,
 	N,
 	NU,
 	P, /* "PRN" */
 	PR,
 	LDOT, /* leading '.' */
 	DOT,  /* '.' in a non-leading position */
 	H,    /* ".h" */
 	HGDI, /* ".hg", ".d", or ".i" */
 	SPACE,
 	DEFAULT, /* byte of a path component after the first */
 };
 
 /* state machine for dir-encoding */
 enum dir_state {
 	DDOT,
 	DH,
 	DHGDI,
 	DDEFAULT,
 };
 
 static inline int inset(const uint32_t bitset[], char c)
 {
 	return bitset[((uint8_t)c) >> 5] & (1 << (((uint8_t)c) & 31));
 }
 
 static inline void charcopy(char *dest, Py_ssize_t *destlen, size_t destsize,
                             char c)
 {
 	if (dest) {
 		assert(*destlen < destsize);
 		dest[*destlen] = c;
 	}
 	(*destlen)++;
 }
 
 static inline void memcopy(char *dest, Py_ssize_t *destlen, size_t destsize,
                            const void *src, Py_ssize_t len)
 {
 	if (dest) {
 		assert(*destlen + len < destsize);
 		memcpy((void *)&dest[*destlen], src, len);
 	}
 	*destlen += len;
 }
 
 static inline void hexencode(char *dest, Py_ssize_t *destlen, size_t destsize,
                              uint8_t c)
 {
 	static const char hexdigit[] = "0123456789abcdef";
 
 	charcopy(dest, destlen, destsize, hexdigit[c >> 4]);
 	charcopy(dest, destlen, destsize, hexdigit[c & 15]);
 }
 
 /* 3-byte escape: tilde followed by two hex digits */
 static inline void escape3(char *dest, Py_ssize_t *destlen, size_t destsize,
                            char c)
 {
 	charcopy(dest, destlen, destsize, '~');
 	hexencode(dest, destlen, destsize, c);
 }
 
 static Py_ssize_t _encodedir(char *dest, size_t destsize, const char *src,
                              Py_ssize_t len)
 {
 	enum dir_state state = DDEFAULT;
 	Py_ssize_t i = 0, destlen = 0;
 
 	while (i < len) {
 		switch (state) {
 		case DDOT:
 			switch (src[i]) {
 			case 'd':
 			case 'i':
 				state = DHGDI;
 				charcopy(dest, &destlen, destsize, src[i++]);
 				break;
 			case 'h':
 				state = DH;
 				charcopy(dest, &destlen, destsize, src[i++]);
 				break;
 			default:
 				state = DDEFAULT;
 				break;
 			}
 			break;
 		case DH:
 			if (src[i] == 'g') {
 				state = DHGDI;
 				charcopy(dest, &destlen, destsize, src[i++]);
 			} else {
 				state = DDEFAULT;
 			}
 			break;
 		case DHGDI:
 			if (src[i] == '/') {
 				memcopy(dest, &destlen, destsize, ".hg", 3);
 				charcopy(dest, &destlen, destsize, src[i++]);
 			}
 			state = DDEFAULT;
 			break;
 		case DDEFAULT:
 			if (src[i] == '.') {
 				state = DDOT;
 			}
 			charcopy(dest, &destlen, destsize, src[i++]);
 			break;
 		}
 	}
 
 	return destlen;
 }
 
 PyObject *encodedir(PyObject *self, PyObject *args)
 {
 	Py_ssize_t len, newlen;
 	PyObject *pathobj, *newobj;
 	char *path;
 
 	if (!PyArg_ParseTuple(args, "O:encodedir", &pathobj)) {
 		return NULL;
 	}
 
 	if (PyBytes_AsStringAndSize(pathobj, &path, &len) == -1) {
 		PyErr_SetString(PyExc_TypeError, "expected a string");
 		return NULL;
 	}
 
 	newlen = len ? _encodedir(NULL, 0, path, len + 1) : 1;
 
 	if (newlen == len + 1) {
 		Py_INCREF(pathobj);
 		return pathobj;
 	}
 
 	newobj = PyBytes_FromStringAndSize(NULL, newlen);
 
 	if (newobj) {
 		assert(PyBytes_Check(newobj));
 		Py_SET_SIZE(newobj, Py_SIZE(newobj) - 1);
 		_encodedir(PyBytes_AS_STRING(newobj), newlen, path, len + 1);
 	}
 
 	return newobj;
 }
 
 static Py_ssize_t _encode(const uint32_t twobytes[8], const uint32_t onebyte[8],
                           char *dest, Py_ssize_t destlen, size_t destsize,
                           const char *src, Py_ssize_t len, int encodedir)
 {
 	enum path_state state = START;
 	Py_ssize_t i = 0;
 
 	/*
 	 * Python strings end with a zero byte, which we use as a
 	 * terminal token as they are not valid inside path names.
 	 */
 
 	while (i < len) {
 		switch (state) {
 		case START:
 			switch (src[i]) {
 			case '/':
 				charcopy(dest, &destlen, destsize, src[i++]);
 				break;
 			case '.':
 				state = LDOT;
 				escape3(dest, &destlen, destsize, src[i++]);
 				break;
 			case ' ':
 				state = DEFAULT;
 				escape3(dest, &destlen, destsize, src[i++]);
 				break;
 			case 'a':
 				state = A;
 				charcopy(dest, &destlen, destsize, src[i++]);
 				break;
 			case 'c':
 				state = C;
 				charcopy(dest, &destlen, destsize, src[i++]);
 				break;
 			case 'l':
 				state = L;
 				charcopy(dest, &destlen, destsize, src[i++]);
 				break;
 			case 'n':
 				state = N;
 				charcopy(dest, &destlen, destsize, src[i++]);
 				break;
 			case 'p':
 				state = P;
 				charcopy(dest, &destlen, destsize, src[i++]);
 				break;
 			default:
 				state = DEFAULT;
 				break;
 			}
 			break;
 		case A:
 			if (src[i] == 'u') {
 				state = AU;
 				charcopy(dest, &destlen, destsize, src[i++]);
 			} else {
 				state = DEFAULT;
 			}
 			break;
 		case AU:
 			if (src[i] == 'x') {
 				state = THIRD;
 				i++;
 			} else {
 				state = DEFAULT;
 			}
 			break;
 		case THIRD:
 			state = DEFAULT;
 			switch (src[i]) {
 			case '.':
 			case '/':
 			case '\0':
 				escape3(dest, &destlen, destsize, src[i - 1]);
 				break;
 			default:
 				i--;
 				break;
 			}
 			break;
 		case C:
 			if (src[i] == 'o') {
 				state = CO;
 				charcopy(dest, &destlen, destsize, src[i++]);
 			} else {
 				state = DEFAULT;
 			}
 			break;
 		case CO:
 			if (src[i] == 'm') {
 				state = COMLPT;
 				i++;
 			} else if (src[i] == 'n') {
 				state = THIRD;
 				i++;
 			} else {
 				state = DEFAULT;
 			}
 			break;
 		case COMLPT:
 			switch (src[i]) {
 			case '1':
 			case '2':
 			case '3':
 			case '4':
 			case '5':
 			case '6':
 			case '7':
 			case '8':
 			case '9':
 				state = COMLPTn;
 				i++;
 				break;
 			default:
 				state = DEFAULT;
 				charcopy(dest, &destlen, destsize, src[i - 1]);
 				break;
 			}
 			break;
 		case COMLPTn:
 			state = DEFAULT;
 			switch (src[i]) {
 			case '.':
 			case '/':
 			case '\0':
 				escape3(dest, &destlen, destsize, src[i - 2]);
 				charcopy(dest, &destlen, destsize, src[i - 1]);
 				break;
 			default:
 				memcopy(dest, &destlen, destsize, &src[i - 2],
 				        2);
 				break;
 			}
 			break;
 		case L:
 			if (src[i] == 'p') {
 				state = LP;
 				charcopy(dest, &destlen, destsize, src[i++]);
 			} else {
 				state = DEFAULT;
 			}
 			break;
 		case LP:
 			if (src[i] == 't') {
 				state = COMLPT;
 				i++;
 			} else {
 				state = DEFAULT;
 			}
 			break;
 		case N:
 			if (src[i] == 'u') {
 				state = NU;
 				charcopy(dest, &destlen, destsize, src[i++]);
 			} else {
 				state = DEFAULT;
 			}
 			break;
 		case NU:
 			if (src[i] == 'l') {
 				state = THIRD;
 				i++;
 			} else {
 				state = DEFAULT;
 			}
 			break;
 		case P:
 			if (src[i] == 'r') {
 				state = PR;
 				charcopy(dest, &destlen, destsize, src[i++]);
 			} else {
 				state = DEFAULT;
 			}
 			break;
 		case PR:
 			if (src[i] == 'n') {
 				state = THIRD;
 				i++;
 			} else {
 				state = DEFAULT;
 			}
 			break;
 		case LDOT:
 			switch (src[i]) {
 			case 'd':
 			case 'i':
 				state = HGDI;
 				charcopy(dest, &destlen, destsize, src[i++]);
 				break;
 			case 'h':
 				state = H;
 				charcopy(dest, &destlen, destsize, src[i++]);
 				break;
 			default:
 				state = DEFAULT;
 				break;
 			}
 			break;
 		case DOT:
 			switch (src[i]) {
 			case '/':
 			case '\0':
 				state = START;
 				memcopy(dest, &destlen, destsize, "~2e", 3);
 				charcopy(dest, &destlen, destsize, src[i++]);
 				break;
 			case 'd':
 			case 'i':
 				state = HGDI;
 				charcopy(dest, &destlen, destsize, '.');
 				charcopy(dest, &destlen, destsize, src[i++]);
 				break;
 			case 'h':
 				state = H;
 				memcopy(dest, &destlen, destsize, ".h", 2);
 				i++;
 				break;
 			default:
 				state = DEFAULT;
 				charcopy(dest, &destlen, destsize, '.');
 				break;
 			}
 			break;
 		case H:
 			if (src[i] == 'g') {
 				state = HGDI;
 				charcopy(dest, &destlen, destsize, src[i++]);
 			} else {
 				state = DEFAULT;
 			}
 			break;
 		case HGDI:
 			if (src[i] == '/') {
 				state = START;
 				if (encodedir) {
 					memcopy(dest, &destlen, destsize, ".hg",
 					        3);
 				}
 				charcopy(dest, &destlen, destsize, src[i++]);
 			} else {
 				state = DEFAULT;
 			}
 			break;
 		case SPACE:
 			switch (src[i]) {
 			case '/':
 			case '\0':
 				state = START;
 				memcopy(dest, &destlen, destsize, "~20", 3);
 				charcopy(dest, &destlen, destsize, src[i++]);
 				break;
 			default:
 				state = DEFAULT;
 				charcopy(dest, &destlen, destsize, ' ');
 				break;
 			}
 			break;
 		case DEFAULT:
 			while (inset(onebyte, src[i])) {
 				charcopy(dest, &destlen, destsize, src[i++]);
 				if (i == len) {
 					goto done;
 				}
 			}
 			switch (src[i]) {
 			case '.':
 				state = DOT;
 				i++;
 				break;
 			case ' ':
 				state = SPACE;
 				i++;
 				break;
 			case '/':
 				state = START;
 				charcopy(dest, &destlen, destsize, '/');
 				i++;
 				break;
 			default:
 				if (inset(onebyte, src[i])) {
 					do {
 						charcopy(dest, &destlen,
 						         destsize, src[i++]);
 					} while (i < len &&
 					         inset(onebyte, src[i]));
 				} else if (inset(twobytes, src[i])) {
 					char c = src[i++];
 					charcopy(dest, &destlen, destsize, '_');
 					charcopy(dest, &destlen, destsize,
 					         c == '_' ? '_' : c + 32);
 				} else {
 					escape3(dest, &destlen, destsize,
 					        src[i++]);
 				}
 				break;
 			}
 			break;
 		}
 	}
 done:
 	return destlen;
 }
 
 static Py_ssize_t basicencode(char *dest, size_t destsize, const char *src,
                               Py_ssize_t len)
 {
 	static const uint32_t twobytes[8] = {0, 0, 0x87fffffe};
 
 	static const uint32_t onebyte[8] = {
 	    1,
 	    0x2bff3bfa,
 	    0x68000001,
 	    0x2fffffff,
 	};
 
 	Py_ssize_t destlen = 0;
 
 	return _encode(twobytes, onebyte, dest, destlen, destsize, src, len, 1);
 }
 
 static const Py_ssize_t maxstorepathlen = 120;
 
 static Py_ssize_t _lowerencode(char *dest, size_t destsize, const char *src,
                                Py_ssize_t len)
 {
 	static const uint32_t onebyte[8] = {1, 0x2bfffbfb, 0xe8000001,
 	                                    0x2fffffff};
 
 	static const uint32_t lower[8] = {0, 0, 0x7fffffe};
 
 	Py_ssize_t i, destlen = 0;
 
 	for (i = 0; i < len; i++) {
 		if (inset(onebyte, src[i])) {
 			charcopy(dest, &destlen, destsize, src[i]);
 		} else if (inset(lower, src[i])) {
 			charcopy(dest, &destlen, destsize, src[i] + 32);
 		} else {
 			escape3(dest, &destlen, destsize, src[i]);
 		}
 	}
 
 	return destlen;
 }
 
 PyObject *lowerencode(PyObject *self, PyObject *args)
 {
 	char *path;
 	Py_ssize_t len, newlen;
 	PyObject *ret;
 
-	if (!PyArg_ParseTuple(args, PY23("s#:lowerencode", "y#:lowerencode"),
-	                      &path, &len)) {
+	if (!PyArg_ParseTuple(args, "y#:lowerencode", &path, &len)) {
 		return NULL;
 	}
 
 	newlen = _lowerencode(NULL, 0, path, len);
 	ret = PyBytes_FromStringAndSize(NULL, newlen);
 	if (ret) {
 		_lowerencode(PyBytes_AS_STRING(ret), newlen, path, len);
 	}
 
 	return ret;
 }
 
 /* See store.py:_auxencode for a description. */
 static Py_ssize_t auxencode(char *dest, size_t destsize, const char *src,
                             Py_ssize_t len)
 {
 	static const uint32_t twobytes[8];
 
 	static const uint32_t onebyte[8] = {
 	    ~0U, 0xffff3ffe, ~0U, ~0U, ~0U, ~0U, ~0U, ~0U,
 	};
 
 	return _encode(twobytes, onebyte, dest, 0, destsize, src, len, 0);
 }
 
 static PyObject *hashmangle(const char *src, Py_ssize_t len, const char sha[20])
 {
 	static const Py_ssize_t dirprefixlen = 8;
 	static const Py_ssize_t maxshortdirslen = 68;
 	char *dest;
 	PyObject *ret;
 
 	Py_ssize_t i, d, p, lastslash = len - 1, lastdot = -1;
 	Py_ssize_t destsize, destlen = 0, slop, used;
 
 	while (lastslash >= 0 && src[lastslash] != '/') {
 		if (src[lastslash] == '.' && lastdot == -1) {
 			lastdot = lastslash;
 		}
 		lastslash--;
 	}
 
 #if 0
 	/* All paths should end in a suffix of ".i" or ".d".
            Unfortunately, the file names in test-hybridencode.py
            violate this rule.  */
 	if (lastdot != len - 3) {
 		PyErr_SetString(PyExc_ValueError,
 				"suffix missing or wrong length");
 		return NULL;
 	}
 #endif
 
 	/* If src contains a suffix, we will append it to the end of
 	   the new string, so make room. */
 	destsize = 120;
 	if (lastdot >= 0) {
 		destsize += len - lastdot - 1;
 	}
 
 	ret = PyBytes_FromStringAndSize(NULL, destsize);
 	if (ret == NULL) {
 		return NULL;
 	}
 
 	dest = PyBytes_AS_STRING(ret);
 	memcopy(dest, &destlen, destsize, "dh/", 3);
 
 	/* Copy up to dirprefixlen bytes of each path component, up to
 	   a limit of maxshortdirslen bytes. */
 	for (i = d = p = 0; i < lastslash; i++, p++) {
 		if (src[i] == '/') {
 			char d = dest[destlen - 1];
 			/* After truncation, a directory name may end
 			   in a space or dot, which are unportable. */
 			if (d == '.' || d == ' ') {
 				dest[destlen - 1] = '_';
 				/* The + 3 is to account for "dh/" in the
 				 * beginning */
 			}
 			if (destlen > maxshortdirslen + 3) {
 				break;
 			}
 			charcopy(dest, &destlen, destsize, src[i]);
 			p = -1;
 		} else if (p < dirprefixlen) {
 			charcopy(dest, &destlen, destsize, src[i]);
 		}
 	}
 
 	/* Rewind to just before the last slash copied. */
 	if (destlen > maxshortdirslen + 3) {
 		do {
 			destlen--;
 		} while (destlen > 0 && dest[destlen] != '/');
 	}
 
 	if (destlen > 3) {
 		if (lastslash > 0) {
 			char d = dest[destlen - 1];
 			/* The last directory component may be
 			   truncated, so make it safe. */
 			if (d == '.' || d == ' ') {
 				dest[destlen - 1] = '_';
 			}
 		}
 
 		charcopy(dest, &destlen, destsize, '/');
 	}
 
 	/* Add a prefix of the original file's name. Its length
 	   depends on the number of bytes left after accounting for
 	   hash and suffix. */
 	used = destlen + 40;
 	if (lastdot >= 0) {
 		used += len - lastdot - 1;
 	}
 	slop = maxstorepathlen - used;
 	if (slop > 0) {
 		Py_ssize_t basenamelen =
 		    lastslash >= 0 ? len - lastslash - 2 : len - 1;
 
 		if (basenamelen > slop) {
 			basenamelen = slop;
 		}
 		if (basenamelen > 0) {
 			memcopy(dest, &destlen, destsize, &src[lastslash + 1],
 			        basenamelen);
 		}
 	}
 
 	/* Add hash and suffix. */
 	for (i = 0; i < 20; i++) {
 		hexencode(dest, &destlen, destsize, sha[i]);
 	}
 
 	if (lastdot >= 0) {
 		memcopy(dest, &destlen, destsize, &src[lastdot],
 		        len - lastdot - 1);
 	}
 
 	assert(PyBytes_Check(ret));
 	Py_SET_SIZE(ret, destlen);
 
 	return ret;
 }
 
 /*
  * Avoiding a trip through Python would improve performance by 50%,
  * but we don't encounter enough long names to be worth the code.
  */
 static int sha1hash(char hash[20], const char *str, Py_ssize_t len)
 {
 	static PyObject *shafunc;
 	PyObject *shaobj, *hashobj;
 
 	if (shafunc == NULL) {
 		PyObject *hashlib = PyImport_ImportModule("hashlib");
 		if (hashlib == NULL) {
 			PyErr_SetString(PyExc_ImportError,
 			                "pathencode failed to find hashlib");
 			return -1;
 		}
 		shafunc = PyObject_GetAttrString(hashlib, "sha1");
 		Py_DECREF(hashlib);
 
 		if (shafunc == NULL) {
 			PyErr_SetString(PyExc_AttributeError,
 			                "module 'hashlib' has no "
 			                "attribute 'sha1' in pathencode");
 			return -1;
 		}
 	}
 
-	shaobj = PyObject_CallFunction(shafunc, PY23("s#", "y#"), str, len);
+	shaobj = PyObject_CallFunction(shafunc, "y#", str, len);
 
 	if (shaobj == NULL) {
 		return -1;
 	}
 
 	hashobj = PyObject_CallMethod(shaobj, "digest", "");
 	Py_DECREF(shaobj);
 	if (hashobj == NULL) {
 		return -1;
 	}
 
 	if (!PyBytes_Check(hashobj) || PyBytes_GET_SIZE(hashobj) != 20) {
 		PyErr_SetString(PyExc_TypeError,
 		                "result of digest is not a 20-byte hash");
 		Py_DECREF(hashobj);
 		return -1;
 	}
 
 	memcpy(hash, PyBytes_AS_STRING(hashobj), 20);
 	Py_DECREF(hashobj);
 	return 0;
 }
 
 #define MAXENCODE 4096 * 4
 
 static PyObject *hashencode(const char *src, Py_ssize_t len)
 {
 	char dired[MAXENCODE];
 	char lowered[MAXENCODE];
 	char auxed[MAXENCODE];
 	Py_ssize_t dirlen, lowerlen, auxlen, baselen;
 	char sha[20];
 
 	baselen = (len - 5) * 3;
 	if (baselen >= MAXENCODE) {
 		PyErr_SetString(PyExc_ValueError, "string too long");
 		return NULL;
 	}
 
 	dirlen = _encodedir(dired, baselen, src, len);
 	if (sha1hash(sha, dired, dirlen - 1) == -1) {
 		return NULL;
 	}
 	lowerlen = _lowerencode(lowered, baselen, dired + 5, dirlen - 5);
 	auxlen = auxencode(auxed, baselen, lowered, lowerlen);
 	return hashmangle(auxed, auxlen, sha);
 }
 
 PyObject *pathencode(PyObject *self, PyObject *args)
 {
 	Py_ssize_t len, newlen;
 	PyObject *pathobj, *newobj;
 	char *path;
 
 	if (!PyArg_ParseTuple(args, "O:pathencode", &pathobj)) {
 		return NULL;
 	}
 
 	if (PyBytes_AsStringAndSize(pathobj, &path, &len) == -1) {
 		PyErr_SetString(PyExc_TypeError, "expected a string");
 		return NULL;
 	}
 
 	if (len > maxstorepathlen) {
 		newlen = maxstorepathlen + 2;
 	} else {
 		newlen = len ? basicencode(NULL, 0, path, len + 1) : 1;
 	}
 
 	if (newlen <= maxstorepathlen + 1) {
 		if (newlen == len + 1) {
 			Py_INCREF(pathobj);
 			return pathobj;
 		}
 
 		newobj = PyBytes_FromStringAndSize(NULL, newlen);
 
 		if (newobj) {
 			assert(PyBytes_Check(newobj));
 			Py_SET_SIZE(newobj, Py_SIZE(newobj) - 1);
 			basicencode(PyBytes_AS_STRING(newobj), newlen, path,
 			            len + 1);
 		}
 	} else {
 		newobj = hashencode(path, len + 1);
 	}
 
 	return newobj;
 }

mercurial/cext/revlog.c

@@ -1,3278 +1,3275 b''
 /*
  parsers.c - efficient content parsing
 
  Copyright 2008 Olivia Mackall <[e-mail]> and others
 
  This software may be used and distributed according to the terms of
  the GNU General Public License, incorporated herein by reference.
 */
 
 #define PY_SSIZE_T_CLEAN
 #include <Python.h>
 #include <assert.h>
 #include <ctype.h>
 #include <limits.h>
 #include <stddef.h>
 #include <stdlib.h>
 #include <string.h>
 #include <structmember.h>
 
 #include "bitmanipulation.h"
 #include "charencode.h"
 #include "compat.h"
 #include "revlog.h"
 #include "util.h"
 
 typedef struct indexObjectStruct indexObject;
 
 typedef struct {
 	int children[16];
 } nodetreenode;
 
 typedef struct {
 	int abi_version;
 	Py_ssize_t (*index_length)(const indexObject *);
 	const char *(*index_node)(indexObject *, Py_ssize_t);
 	int (*index_parents)(PyObject *, int, int *);
 } Revlog_CAPI;
 
 /*
  * A base-16 trie for fast node->rev mapping.
  *
  * Positive value is index of the next node in the trie
  * Negative value is a leaf: -(rev + 2)
  * Zero is empty
  */
 typedef struct {
 	indexObject *index;
 	nodetreenode *nodes;
 	Py_ssize_t nodelen;
 	size_t length;   /* # nodes in use */
 	size_t capacity; /* # nodes allocated */
 	int depth;       /* maximum depth of tree */
 	int splits;      /* # splits performed */
 } nodetree;
 
 typedef struct {
 	PyObject_HEAD /* ; */
 	    nodetree nt;
 } nodetreeObject;
 
 /*
  * This class has two behaviors.
  *
  * When used in a list-like way (with integer keys), we decode an
  * entry in a RevlogNG index file on demand. We have limited support for
  * integer-keyed insert and delete, only at elements right before the
  * end.
  *
  * With string keys, we lazily perform a reverse mapping from node to
  * rev, using a base-16 trie.
  */
 struct indexObjectStruct {
 	PyObject_HEAD
 	    /* Type-specific fields go here. */
 	    PyObject *data;     /* raw bytes of index */
 	Py_ssize_t nodelen;     /* digest size of the hash, 20 for SHA-1 */
 	PyObject *nullentry;    /* fast path for references to null */
 	Py_buffer buf;          /* buffer of data */
 	const char **offsets;   /* populated on demand */
 	Py_ssize_t length;      /* current on-disk number of elements */
 	unsigned new_length;    /* number of added elements */
 	unsigned added_length;  /* space reserved for added elements */
 	char *added;            /* populated on demand */
 	PyObject *headrevs;     /* cache, invalidated on changes */
 	PyObject *filteredrevs; /* filtered revs set */
 	nodetree nt;            /* base-16 trie */
 	int ntinitialized;      /* 0 or 1 */
 	int ntrev;              /* last rev scanned */
 	int ntlookups;          /* # lookups */
 	int ntmisses;           /* # lookups that miss the cache */
 	int inlined;
 	long entry_size; /* size of index headers. Differs in v1 v.s. v2 format
 	                  */
 	long rust_ext_compat; /* compatibility with being used in rust
 	                         extensions */
 	long format_version;  /* format version selector (format_*) */
 };
 
 static Py_ssize_t index_length(const indexObject *self)
 {
 	return self->length + self->new_length;
 }
 
 static const char nullid[32] = {0};
 static const Py_ssize_t nullrev = -1;
 
 static Py_ssize_t inline_scan(indexObject *self, const char **offsets);
 
 static int index_find_node(indexObject *self, const char *node);
 
 #if LONG_MAX == 0x7fffffffL
-static const char *const tuple_format =
-    PY23("Kiiiiiis#KiBBi", "Kiiiiiiy#KiBBi");
+static const char *const tuple_format = "Kiiiiiiy#KiBBi";
 #else
-static const char *const tuple_format =
-    PY23("kiiiiiis#kiBBi", "kiiiiiiy#kiBBi");
+static const char *const tuple_format = "kiiiiiiy#kiBBi";
 #endif
 
 /* A RevlogNG v1 index entry is 64 bytes long. */
 static const long v1_entry_size = 64;
 
 /* A Revlogv2 index entry is 96 bytes long. */
 static const long v2_entry_size = 96;
 
 /* A Changelogv2 index entry is 96 bytes long. */
 static const long cl2_entry_size = 96;
 
 /* Internal format version.
  * Must match their counterparts in revlogutils/constants.py */
 static const long format_v1 = 1;       /* constants.py: REVLOGV1 */
 static const long format_v2 = 0xDEAD;  /* constants.py: REVLOGV2 */
 static const long format_cl2 = 0xD34D; /* constants.py: CHANGELOGV2 */
 
 static const long entry_v1_offset_high = 0;
 static const long entry_v1_offset_offset_flags = 4;
 static const long entry_v1_offset_comp_len = 8;
 static const long entry_v1_offset_uncomp_len = 12;
 static const long entry_v1_offset_base_rev = 16;
 static const long entry_v1_offset_link_rev = 20;
 static const long entry_v1_offset_parent_1 = 24;
 static const long entry_v1_offset_parent_2 = 28;
 static const long entry_v1_offset_node_id = 32;
 
 static const long entry_v2_offset_high = 0;
 static const long entry_v2_offset_offset_flags = 4;
 static const long entry_v2_offset_comp_len = 8;
 static const long entry_v2_offset_uncomp_len = 12;
 static const long entry_v2_offset_base_rev = 16;
 static const long entry_v2_offset_link_rev = 20;
 static const long entry_v2_offset_parent_1 = 24;
 static const long entry_v2_offset_parent_2 = 28;
 static const long entry_v2_offset_node_id = 32;
 static const long entry_v2_offset_sidedata_offset = 64;
 static const long entry_v2_offset_sidedata_comp_len = 72;
 static const long entry_v2_offset_all_comp_mode = 76;
 /* next free offset: 77 */
 
 static const long entry_cl2_offset_high = 0;
 static const long entry_cl2_offset_offset_flags = 4;
 static const long entry_cl2_offset_comp_len = 8;
 static const long entry_cl2_offset_uncomp_len = 12;
 static const long entry_cl2_offset_parent_1 = 16;
 static const long entry_cl2_offset_parent_2 = 20;
 static const long entry_cl2_offset_node_id = 24;
 static const long entry_cl2_offset_sidedata_offset = 56;
 static const long entry_cl2_offset_sidedata_comp_len = 64;
 static const long entry_cl2_offset_all_comp_mode = 68;
 static const long entry_cl2_offset_rank = 69;
 /* next free offset: 73 */
 
 static const char comp_mode_inline = 2;
 static const char rank_unknown = -1;
 
 static void raise_revlog_error(void)
 {
 	PyObject *mod = NULL, *dict = NULL, *errclass = NULL;
 
 	mod = PyImport_ImportModule("mercurial.error");
 	if (mod == NULL) {
 		goto cleanup;
 	}
 
 	dict = PyModule_GetDict(mod);
 	if (dict == NULL) {
 		goto cleanup;
 	}
 	Py_INCREF(dict);
 
 	errclass = PyDict_GetItemString(dict, "RevlogError");
 	if (errclass == NULL) {
 		PyErr_SetString(PyExc_SystemError,
 		                "could not find RevlogError");
 		goto cleanup;
 	}
 
 	/* value of exception is ignored by callers */
 	PyErr_SetString(errclass, "RevlogError");
 
 cleanup:
 	Py_XDECREF(dict);
 	Py_XDECREF(mod);
 }
 
 /*
  * Return a pointer to the beginning of a RevlogNG record.
  */
 static const char *index_deref(indexObject *self, Py_ssize_t pos)
 {
 	if (pos >= self->length)
 		return self->added + (pos - self->length) * self->entry_size;
 
 	if (self->inlined && pos > 0) {
 		if (self->offsets == NULL) {
 			Py_ssize_t ret;
 			self->offsets =
 			    PyMem_Malloc(self->length * sizeof(*self->offsets));
 			if (self->offsets == NULL)
 				return (const char *)PyErr_NoMemory();
 			ret = inline_scan(self, self->offsets);
 			if (ret == -1) {
 				return NULL;
 			};
 		}
 		return self->offsets[pos];
 	}
 
 	return (const char *)(self->buf.buf) + pos * self->entry_size;
 }
 
 /*
  * Get parents of the given rev.
  *
  * The specified rev must be valid and must not be nullrev. A returned
  * parent revision may be nullrev, but is guaranteed to be in valid range.
  */
 static inline int index_get_parents(indexObject *self, Py_ssize_t rev, int *ps,
                                     int maxrev)
 {
 	const char *data = index_deref(self, rev);
 
 	if (self->format_version == format_v1) {
 		ps[0] = getbe32(data + entry_v1_offset_parent_1);
 		ps[1] = getbe32(data + entry_v1_offset_parent_2);
 	} else if (self->format_version == format_v2) {
 		ps[0] = getbe32(data + entry_v2_offset_parent_1);
 		ps[1] = getbe32(data + entry_v2_offset_parent_2);
 	} else if (self->format_version == format_cl2) {
 		ps[0] = getbe32(data + entry_cl2_offset_parent_1);
 		ps[1] = getbe32(data + entry_cl2_offset_parent_2);
 	} else {
 		raise_revlog_error();
 		return -1;
 	}
 
 	/* If index file is corrupted, ps[] may point to invalid revisions. So
 	 * there is a risk of buffer overflow to trust them unconditionally. */
 	if (ps[0] < -1 || ps[0] > maxrev || ps[1] < -1 || ps[1] > maxrev) {
 		PyErr_SetString(PyExc_ValueError, "parent out of range");
 		return -1;
 	}
 	return 0;
 }
 
 /*
  * Get parents of the given rev.
  *
  * If the specified rev is out of range, IndexError will be raised. If the
  * revlog entry is corrupted, ValueError may be raised.
  *
  * Returns 0 on success or -1 on failure.
  */
 static int HgRevlogIndex_GetParents(PyObject *op, int rev, int *ps)
 {
 	int tiprev;
 	if (!op || !HgRevlogIndex_Check(op) || !ps) {
 		PyErr_BadInternalCall();
 		return -1;
 	}
 	tiprev = (int)index_length((indexObject *)op) - 1;
 	if (rev < -1 || rev > tiprev) {
 		PyErr_Format(PyExc_IndexError, "rev out of range: %d", rev);
 		return -1;
 	} else if (rev == -1) {
 		ps[0] = ps[1] = -1;
 		return 0;
 	} else {
 		return index_get_parents((indexObject *)op, rev, ps, tiprev);
 	}
 }
 
 static inline int64_t index_get_start(indexObject *self, Py_ssize_t rev)
 {
 	const char *data;
 	uint64_t offset;
 
 	if (rev == nullrev)
 		return 0;
 
 	data = index_deref(self, rev);
 
 	if (self->format_version == format_v1) {
 		offset = getbe32(data + entry_v1_offset_offset_flags);
 		if (rev == 0) {
 			/* mask out version number for the first entry */
 			offset &= 0xFFFF;
 		} else {
 			uint32_t offset_high =
 			    getbe32(data + entry_v1_offset_high);
 			offset |= ((uint64_t)offset_high) << 32;
 		}
 	} else if (self->format_version == format_v2) {
 		offset = getbe32(data + entry_v2_offset_offset_flags);
 		if (rev == 0) {
 			/* mask out version number for the first entry */
 			offset &= 0xFFFF;
 		} else {
 			uint32_t offset_high =
 			    getbe32(data + entry_v2_offset_high);
 			offset |= ((uint64_t)offset_high) << 32;
 		}
 	} else if (self->format_version == format_cl2) {
 		uint32_t offset_high = getbe32(data + entry_cl2_offset_high);
 		offset = getbe32(data + entry_cl2_offset_offset_flags);
 		offset |= ((uint64_t)offset_high) << 32;
 	} else {
 		raise_revlog_error();
 		return -1;
 	}
 
 	return (int64_t)(offset >> 16);
 }
 
 static inline int index_get_length(indexObject *self, Py_ssize_t rev)
 {
 	const char *data;
 	int tmp;
 
 	if (rev == nullrev)
 		return 0;
 
 	data = index_deref(self, rev);
 
 	if (self->format_version == format_v1) {
 		tmp = (int)getbe32(data + entry_v1_offset_comp_len);
 	} else if (self->format_version == format_v2) {
 		tmp = (int)getbe32(data + entry_v2_offset_comp_len);
 	} else if (self->format_version == format_cl2) {
 		tmp = (int)getbe32(data + entry_cl2_offset_comp_len);
 	} else {
 		raise_revlog_error();
 		return -1;
 	}
 	if (tmp < 0) {
 		PyErr_Format(PyExc_OverflowError,
 		             "revlog entry size out of bound (%d)", tmp);
 		return -1;
 	}
 	return tmp;
 }
 
 /*
  * RevlogNG format (all in big endian, data may be inlined):
  *    6 bytes: offset
  *    2 bytes: flags
  *    4 bytes: compressed length
  *    4 bytes: uncompressed length
  *    4 bytes: base revision
  *    4 bytes: link revision
  *    4 bytes: parent 1 revision
  *    4 bytes: parent 2 revision
  *   32 bytes: nodeid (only 20 bytes used with SHA-1)
  */
 static PyObject *index_get(indexObject *self, Py_ssize_t pos)
 {
 	uint64_t offset_flags, sidedata_offset;
 	int comp_len, uncomp_len, base_rev, link_rev, parent_1, parent_2,
 	    sidedata_comp_len, rank = rank_unknown;
 	char data_comp_mode, sidedata_comp_mode;
 	const char *c_node_id;
 	const char *data;
 	Py_ssize_t length = index_length(self);
 
 	if (pos == nullrev) {
 		Py_INCREF(self->nullentry);
 		return self->nullentry;
 	}
 
 	if (pos < 0 || pos >= length) {
 		PyErr_SetString(PyExc_IndexError, "revlog index out of range");
 		return NULL;
 	}
 
 	data = index_deref(self, pos);
 	if (data == NULL)
 		return NULL;
 
 	if (self->format_version == format_v1) {
 		offset_flags = getbe32(data + entry_v1_offset_offset_flags);
 		/*
 		 * The first entry on-disk needs the version number masked out,
 		 * but this doesn't apply if entries are added to an empty
 		 * index.
 		 */
 		if (self->length && pos == 0)
 			offset_flags &= 0xFFFF;
 		else {
 			uint32_t offset_high =
 			    getbe32(data + entry_v1_offset_high);
 			offset_flags |= ((uint64_t)offset_high) << 32;
 		}
 
 		comp_len = getbe32(data + entry_v1_offset_comp_len);
 		uncomp_len = getbe32(data + entry_v1_offset_uncomp_len);
 		base_rev = getbe32(data + entry_v1_offset_base_rev);
 		link_rev = getbe32(data + entry_v1_offset_link_rev);
 		parent_1 = getbe32(data + entry_v1_offset_parent_1);
 		parent_2 = getbe32(data + entry_v1_offset_parent_2);
 		c_node_id = data + entry_v1_offset_node_id;
 
 		sidedata_offset = 0;
 		sidedata_comp_len = 0;
 		data_comp_mode = comp_mode_inline;
 		sidedata_comp_mode = comp_mode_inline;
 	} else if (self->format_version == format_v2) {
 		offset_flags = getbe32(data + entry_v2_offset_offset_flags);
 		/*
 		 * The first entry on-disk needs the version number masked out,
 		 * but this doesn't apply if entries are added to an empty
 		 * index.
 		 */
 		if (self->length && pos == 0)
 			offset_flags &= 0xFFFF;
 		else {
 			uint32_t offset_high =
 			    getbe32(data + entry_v2_offset_high);
 			offset_flags |= ((uint64_t)offset_high) << 32;
 		}
 
 		comp_len = getbe32(data + entry_v2_offset_comp_len);
 		uncomp_len = getbe32(data + entry_v2_offset_uncomp_len);
 		base_rev = getbe32(data + entry_v2_offset_base_rev);
 		link_rev = getbe32(data + entry_v2_offset_link_rev);
 		parent_1 = getbe32(data + entry_v2_offset_parent_1);
 		parent_2 = getbe32(data + entry_v2_offset_parent_2);
 		c_node_id = data + entry_v2_offset_node_id;
 
 		sidedata_offset =
 		    getbe64(data + entry_v2_offset_sidedata_offset);
 		sidedata_comp_len =
 		    getbe32(data + entry_v2_offset_sidedata_comp_len);
 		data_comp_mode = data[entry_v2_offset_all_comp_mode] & 3;
 		sidedata_comp_mode =
 		    ((data[entry_v2_offset_all_comp_mode] >> 2) & 3);
 	} else if (self->format_version == format_cl2) {
 		uint32_t offset_high = getbe32(data + entry_cl2_offset_high);
 		offset_flags = getbe32(data + entry_cl2_offset_offset_flags);
 		offset_flags |= ((uint64_t)offset_high) << 32;
 		comp_len = getbe32(data + entry_cl2_offset_comp_len);
 		uncomp_len = getbe32(data + entry_cl2_offset_uncomp_len);
 		/* base_rev and link_rev are not stored in changelogv2, but are
 		 still used by some functions shared with the other revlogs.
 		 They are supposed to contain links to other revisions,
 		 but they always point to themselves in the case of a changelog.
 		*/
 		base_rev = pos;
 		link_rev = pos;
 		parent_1 = getbe32(data + entry_cl2_offset_parent_1);
 		parent_2 = getbe32(data + entry_cl2_offset_parent_2);
 		c_node_id = data + entry_cl2_offset_node_id;
 		sidedata_offset =
 		    getbe64(data + entry_cl2_offset_sidedata_offset);
 		sidedata_comp_len =
 		    getbe32(data + entry_cl2_offset_sidedata_comp_len);
 		data_comp_mode = data[entry_cl2_offset_all_comp_mode] & 3;
 		sidedata_comp_mode =
 		    ((data[entry_cl2_offset_all_comp_mode] >> 2) & 3);
 		rank = getbe32(data + entry_cl2_offset_rank);
 	} else {
 		raise_revlog_error();
 		return NULL;
 	}
 
 	return Py_BuildValue(tuple_format, offset_flags, comp_len, uncomp_len,
 	                     base_rev, link_rev, parent_1, parent_2, c_node_id,
 	                     self->nodelen, sidedata_offset, sidedata_comp_len,
 	                     data_comp_mode, sidedata_comp_mode, rank);
 }
 /*
  * Pack header information in binary
  */
 static PyObject *index_pack_header(indexObject *self, PyObject *args)
 {
 	int header;
 	char out[4];
 	if (!PyArg_ParseTuple(args, "I", &header)) {
 		return NULL;
 	}
 	if (self->format_version != format_v1) {
 		PyErr_Format(PyExc_RuntimeError,
 		             "version header should go in the docket, not the "
 		             "index: %lu",
 		             header);
 		return NULL;
 	}
 	putbe32(header, out);
 	return PyBytes_FromStringAndSize(out, 4);
 }
 /*
  * Return the raw binary string representing a revision
  */
 static PyObject *index_entry_binary(indexObject *self, PyObject *value)
 {
 	long rev;
 	const char *data;
 	Py_ssize_t length = index_length(self);
 
 	if (!pylong_to_long(value, &rev)) {
 		return NULL;
 	}
 	if (rev < 0 || rev >= length) {
 		PyErr_Format(PyExc_ValueError, "revlog index out of range: %ld",
 		             rev);
 		return NULL;
 	};
 
 	data = index_deref(self, rev);
 	if (data == NULL)
 		return NULL;
 	if (rev == 0 && self->format_version == format_v1) {
 		/* the header is eating the start of the first entry */
 		return PyBytes_FromStringAndSize(data + 4,
 		                                 self->entry_size - 4);
 	}
 	return PyBytes_FromStringAndSize(data, self->entry_size);
 }
 
 /*
  * Return the hash of node corresponding to the given rev.
  */
 static const char *index_node(indexObject *self, Py_ssize_t pos)
 {
 	Py_ssize_t length = index_length(self);
 	const char *data;
 	const char *node_id;
 
 	if (pos == nullrev)
 		return nullid;
 
 	if (pos >= length)
 		return NULL;
 
 	data = index_deref(self, pos);
 
 	if (self->format_version == format_v1) {
 		node_id = data + entry_v1_offset_node_id;
 	} else if (self->format_version == format_v2) {
 		node_id = data + entry_v2_offset_node_id;
 	} else if (self->format_version == format_cl2) {
 		node_id = data + entry_cl2_offset_node_id;
 	} else {
 		raise_revlog_error();
 		return NULL;
 	}
 
 	return data ? node_id : NULL;
 }
 
 /*
  * Return the hash of the node corresponding to the given rev. The
  * rev is assumed to be existing. If not, an exception is set.
  */
 static const char *index_node_existing(indexObject *self, Py_ssize_t pos)
 {
 	const char *node = index_node(self, pos);
 	if (node == NULL) {
 		PyErr_Format(PyExc_IndexError, "could not access rev %d",
 		             (int)pos);
 	}
 	return node;
 }
 
 static int nt_insert(nodetree *self, const char *node, int rev);
 
 static int node_check(Py_ssize_t nodelen, PyObject *obj, char **node)
 {
 	Py_ssize_t thisnodelen;
 	if (PyBytes_AsStringAndSize(obj, node, &thisnodelen) == -1)
 		return -1;
 	if (nodelen == thisnodelen)
 		return 0;
 	PyErr_Format(PyExc_ValueError, "node len %zd != expected node len %zd",
 	             thisnodelen, nodelen);
 	return -1;
 }
 
 static PyObject *index_append(indexObject *self, PyObject *obj)
 {
 	uint64_t offset_flags, sidedata_offset;
 	int rev, comp_len, uncomp_len, base_rev, link_rev, parent_1, parent_2,
 	    sidedata_comp_len, rank;
 	char data_comp_mode, sidedata_comp_mode;
 	Py_ssize_t c_node_id_len;
 	const char *c_node_id;
 	char comp_field;
 	char *data;
 
 	if (!PyArg_ParseTuple(obj, tuple_format, &offset_flags, &comp_len,
 	                      &uncomp_len, &base_rev, &link_rev, &parent_1,
 	                      &parent_2, &c_node_id, &c_node_id_len,
 	                      &sidedata_offset, &sidedata_comp_len,
 	                      &data_comp_mode, &sidedata_comp_mode, &rank)) {
 		PyErr_SetString(PyExc_TypeError, "12-tuple required");
 		return NULL;
 	}
 
 	if (c_node_id_len != self->nodelen) {
 		PyErr_SetString(PyExc_TypeError, "invalid node");
 		return NULL;
 	}
 	if (self->format_version == format_v1) {
 
 		if (data_comp_mode != comp_mode_inline) {
 			PyErr_Format(PyExc_ValueError,
 			             "invalid data compression mode: %i",
 			             data_comp_mode);
 			return NULL;
 		}
 		if (sidedata_comp_mode != comp_mode_inline) {
 			PyErr_Format(PyExc_ValueError,
 			             "invalid sidedata compression mode: %i",
 			             sidedata_comp_mode);
 			return NULL;
 		}
 	}
 
 	if (self->new_length == self->added_length) {
 		size_t new_added_length =
 		    self->added_length ? self->added_length * 2 : 4096;
 		void *new_added = PyMem_Realloc(
 		    self->added, new_added_length * self->entry_size);
 		if (!new_added)
 			return PyErr_NoMemory();
 		self->added = new_added;
 		self->added_length = new_added_length;
 	}
 	rev = self->length + self->new_length;
 	data = self->added + self->entry_size * self->new_length++;
 
 	memset(data, 0, self->entry_size);
 
 	if (self->format_version == format_v1) {
 		putbe32(offset_flags >> 32, data + entry_v1_offset_high);
 		putbe32(offset_flags & 0xffffffffU,
 		        data + entry_v1_offset_offset_flags);
 		putbe32(comp_len, data + entry_v1_offset_comp_len);
 		putbe32(uncomp_len, data + entry_v1_offset_uncomp_len);
 		putbe32(base_rev, data + entry_v1_offset_base_rev);
 		putbe32(link_rev, data + entry_v1_offset_link_rev);
 		putbe32(parent_1, data + entry_v1_offset_parent_1);
 		putbe32(parent_2, data + entry_v1_offset_parent_2);
 		memcpy(data + entry_v1_offset_node_id, c_node_id,
 		       c_node_id_len);
 	} else if (self->format_version == format_v2) {
 		putbe32(offset_flags >> 32, data + entry_v2_offset_high);
 		putbe32(offset_flags & 0xffffffffU,
 		        data + entry_v2_offset_offset_flags);
 		putbe32(comp_len, data + entry_v2_offset_comp_len);
 		putbe32(uncomp_len, data + entry_v2_offset_uncomp_len);
 		putbe32(base_rev, data + entry_v2_offset_base_rev);
 		putbe32(link_rev, data + entry_v2_offset_link_rev);
 		putbe32(parent_1, data + entry_v2_offset_parent_1);
 		putbe32(parent_2, data + entry_v2_offset_parent_2);
 		memcpy(data + entry_v2_offset_node_id, c_node_id,
 		       c_node_id_len);
 		putbe64(sidedata_offset,
 		        data + entry_v2_offset_sidedata_offset);
 		putbe32(sidedata_comp_len,
 		        data + entry_v2_offset_sidedata_comp_len);
 		comp_field = data_comp_mode & 3;
 		comp_field = comp_field | (sidedata_comp_mode & 3) << 2;
 		data[entry_v2_offset_all_comp_mode] = comp_field;
 	} else if (self->format_version == format_cl2) {
 		putbe32(offset_flags >> 32, data + entry_cl2_offset_high);
 		putbe32(offset_flags & 0xffffffffU,
 		        data + entry_cl2_offset_offset_flags);
 		putbe32(comp_len, data + entry_cl2_offset_comp_len);
 		putbe32(uncomp_len, data + entry_cl2_offset_uncomp_len);
 		putbe32(parent_1, data + entry_cl2_offset_parent_1);
 		putbe32(parent_2, data + entry_cl2_offset_parent_2);
 		memcpy(data + entry_cl2_offset_node_id, c_node_id,
 		       c_node_id_len);
 		putbe64(sidedata_offset,
 		        data + entry_cl2_offset_sidedata_offset);
 		putbe32(sidedata_comp_len,
 		        data + entry_cl2_offset_sidedata_comp_len);
 		comp_field = data_comp_mode & 3;
 		comp_field = comp_field | (sidedata_comp_mode & 3) << 2;
 		data[entry_cl2_offset_all_comp_mode] = comp_field;
 		putbe32(rank, data + entry_cl2_offset_rank);
 	} else {
 		raise_revlog_error();
 		return NULL;
 	}
 
 	if (self->ntinitialized)
 		nt_insert(&self->nt, c_node_id, rev);
 
 	Py_CLEAR(self->headrevs);
 	Py_RETURN_NONE;
 }
 
 /* Replace an existing index entry's sidedata offset and length with new ones.
    This cannot be used outside of the context of sidedata rewriting,
    inside the transaction that creates the given revision. */
 static PyObject *index_replace_sidedata_info(indexObject *self, PyObject *args)
 {
 	uint64_t offset_flags, sidedata_offset;
 	Py_ssize_t rev;
 	int sidedata_comp_len;
 	char comp_mode;
 	char *data;
 #if LONG_MAX == 0x7fffffffL
-	const char *const sidedata_format = PY23("nKiKB", "nKiKB");
+	const char *const sidedata_format = "nKiKB";
 #else
-	const char *const sidedata_format = PY23("nkikB", "nkikB");
+	const char *const sidedata_format = "nkikB";
 #endif
 
 	if (self->entry_size == v1_entry_size || self->inlined) {
 		/*
 		 There is a bug in the transaction handling when going from an
 	   inline revlog to a separate index and data file. Turn it off until
 	   it's fixed, since v2 revlogs sometimes get rewritten on exchange.
 	   See issue6485.
 	  */
 		raise_revlog_error();
 		return NULL;
 	}
 
 	if (!PyArg_ParseTuple(args, sidedata_format, &rev, &sidedata_offset,
 	                      &sidedata_comp_len, &offset_flags, &comp_mode))
 		return NULL;
 
 	if (rev < 0 || rev >= index_length(self)) {
 		PyErr_SetString(PyExc_IndexError, "revision outside index");
 		return NULL;
 	}
 	if (rev < self->length) {
 		PyErr_SetString(
 		    PyExc_IndexError,
 		    "cannot rewrite entries outside of this transaction");
 		return NULL;
 	}
 
 	/* Find the newly added node, offset from the "already on-disk" length
 	 */
 	data = self->added + self->entry_size * (rev - self->length);
 	if (self->format_version == format_v2) {
 		putbe64(offset_flags, data + entry_v2_offset_high);
 		putbe64(sidedata_offset,
 		        data + entry_v2_offset_sidedata_offset);
 		putbe32(sidedata_comp_len,
 		        data + entry_v2_offset_sidedata_comp_len);
 		data[entry_v2_offset_all_comp_mode] =
 		    (data[entry_v2_offset_all_comp_mode] & ~(3 << 2)) |
 		    ((comp_mode & 3) << 2);
 	} else if (self->format_version == format_cl2) {
 		putbe64(offset_flags, data + entry_cl2_offset_high);
 		putbe64(sidedata_offset,
 		        data + entry_cl2_offset_sidedata_offset);
 		putbe32(sidedata_comp_len,
 		        data + entry_cl2_offset_sidedata_comp_len);
 		data[entry_cl2_offset_all_comp_mode] =
 		    (data[entry_cl2_offset_all_comp_mode] & ~(3 << 2)) |
 		    ((comp_mode & 3) << 2);
 	} else {
 		raise_revlog_error();
 		return NULL;
 	}
 
 	Py_RETURN_NONE;
 }
 
 static PyObject *index_stats(indexObject *self)
 {
 	PyObject *obj = PyDict_New();
 	PyObject *s = NULL;
 	PyObject *t = NULL;
 
 	if (obj == NULL)
 		return NULL;
 
 #define istat(__n, __d)                                                        \
 	do {                                                                   \
 		s = PyBytes_FromString(__d);                                   \
 		t = PyLong_FromSsize_t(self->__n);                             \
 		if (!s || !t)                                                  \
 			goto bail;                                             \
 		if (PyDict_SetItem(obj, s, t) == -1)                           \
 			goto bail;                                             \
 		Py_CLEAR(s);                                                   \
 		Py_CLEAR(t);                                                   \
 	} while (0)
 
 	if (self->added_length)
 		istat(new_length, "index entries added");
 	istat(length, "revs in memory");
 	istat(ntlookups, "node trie lookups");
 	istat(ntmisses, "node trie misses");
 	istat(ntrev, "node trie last rev scanned");
 	if (self->ntinitialized) {
 		istat(nt.capacity, "node trie capacity");
 		istat(nt.depth, "node trie depth");
 		istat(nt.length, "node trie count");
 		istat(nt.splits, "node trie splits");
 	}
 
 #undef istat
 
 	return obj;
 
 bail:
 	Py_XDECREF(obj);
 	Py_XDECREF(s);
 	Py_XDECREF(t);
 	return NULL;
 }
 
 /*
  * When we cache a list, we want to be sure the caller can't mutate
  * the cached copy.
  */
 static PyObject *list_copy(PyObject *list)
 {
 	Py_ssize_t len = PyList_GET_SIZE(list);
 	PyObject *newlist = PyList_New(len);
 	Py_ssize_t i;
 
 	if (newlist == NULL)
 		return NULL;
 
 	for (i = 0; i < len; i++) {
 		PyObject *obj = PyList_GET_ITEM(list, i);
 		Py_INCREF(obj);
 		PyList_SET_ITEM(newlist, i, obj);
 	}
 
 	return newlist;
 }
 
 static int check_filter(PyObject *filter, Py_ssize_t arg)
 {
 	if (filter) {
 		PyObject *arglist, *result;
 		int isfiltered;
 
 		arglist = Py_BuildValue("(n)", arg);
 		if (!arglist) {
 			return -1;
 		}
 
 		result = PyObject_Call(filter, arglist, NULL);
 		Py_DECREF(arglist);
 		if (!result) {
 			return -1;
 		}
 
 		/* PyObject_IsTrue returns 1 if true, 0 if false, -1 if error,
 		 * same as this function, so we can just return it directly.*/
 		isfiltered = PyObject_IsTrue(result);
 		Py_DECREF(result);
 		return isfiltered;
 	} else {
 		return 0;
 	}
 }
 
 static inline void set_phase_from_parents(char *phases, int parent_1,
                                           int parent_2, Py_ssize_t i)
 {
 	if (parent_1 >= 0 && phases[parent_1] > phases[i])
 		phases[i] = phases[parent_1];
 	if (parent_2 >= 0 && phases[parent_2] > phases[i])
 		phases[i] = phases[parent_2];
 }
 
 static PyObject *reachableroots2(indexObject *self, PyObject *args)
 {
 
 	/* Input */
 	long minroot;
 	PyObject *includepatharg = NULL;
 	int includepath = 0;
 	/* heads and roots are lists */
 	PyObject *heads = NULL;
 	PyObject *roots = NULL;
 	PyObject *reachable = NULL;
 
 	PyObject *val;
 	Py_ssize_t len = index_length(self);
 	long revnum;
 	Py_ssize_t k;
 	Py_ssize_t i;
 	Py_ssize_t l;
 	int r;
 	int parents[2];
 
 	/* Internal data structure:
 	 * tovisit: array of length len+1 (all revs + nullrev), filled upto
 	 * lentovisit
 	 *
 	 * revstates: array of length len+1 (all revs + nullrev) */
 	int *tovisit = NULL;
 	long lentovisit = 0;
 	enum { RS_SEEN = 1, RS_ROOT = 2, RS_REACHABLE = 4 };
 	char *revstates = NULL;
 
 	/* Get arguments */
 	if (!PyArg_ParseTuple(args, "lO!O!O!", &minroot, &PyList_Type, &heads,
 	                      &PyList_Type, &roots, &PyBool_Type,
 	                      &includepatharg))
 		goto bail;
 
 	if (includepatharg == Py_True)
 		includepath = 1;
 
 	/* Initialize return set */
 	reachable = PyList_New(0);
 	if (reachable == NULL)
 		goto bail;
 
 	/* Initialize internal datastructures */
 	tovisit = (int *)malloc((len + 1) * sizeof(int));
 	if (tovisit == NULL) {
 		PyErr_NoMemory();
 		goto bail;
 	}
 
 	revstates = (char *)calloc(len + 1, 1);
 	if (revstates == NULL) {
 		PyErr_NoMemory();
 		goto bail;
 	}
 
 	l = PyList_GET_SIZE(roots);
 	for (i = 0; i < l; i++) {
 		revnum = PyLong_AsLong(PyList_GET_ITEM(roots, i));
 		if (revnum == -1 && PyErr_Occurred())
 			goto bail;
 		/* If root is out of range, e.g. wdir(), it must be unreachable
 		 * from heads. So we can just ignore it. */
 		if (revnum + 1 < 0 || revnum + 1 >= len + 1)
 			continue;
 		revstates[revnum + 1] |= RS_ROOT;
 	}
 
 	/* Populate tovisit with all the heads */
 	l = PyList_GET_SIZE(heads);
 	for (i = 0; i < l; i++) {
 		revnum = PyLong_AsLong(PyList_GET_ITEM(heads, i));
 		if (revnum == -1 && PyErr_Occurred())
 			goto bail;
 		if (revnum + 1 < 0 || revnum + 1 >= len + 1) {
 			PyErr_SetString(PyExc_IndexError, "head out of range");
 			goto bail;
 		}
 		if (!(revstates[revnum + 1] & RS_SEEN)) {
 			tovisit[lentovisit++] = (int)revnum;
 			revstates[revnum + 1] |= RS_SEEN;
 		}
 	}
 
 	/* Visit the tovisit list and find the reachable roots */
 	k = 0;
 	while (k < lentovisit) {
 		/* Add the node to reachable if it is a root*/
 		revnum = tovisit[k++];
 		if (revstates[revnum + 1] & RS_ROOT) {
 			revstates[revnum + 1] |= RS_REACHABLE;
 			val = PyLong_FromLong(revnum);
 			if (val == NULL)
 				goto bail;
 			r = PyList_Append(reachable, val);
 			Py_DECREF(val);
 			if (r < 0)
 				goto bail;
 			if (includepath == 0)
 				continue;
 		}
 
 		/* Add its parents to the list of nodes to visit */
 		if (revnum == nullrev)
 			continue;
 		r = index_get_parents(self, revnum, parents, (int)len - 1);
 		if (r < 0)
 			goto bail;
 		for (i = 0; i < 2; i++) {
 			if (!(revstates[parents[i] + 1] & RS_SEEN) &&
 			    parents[i] >= minroot) {
 				tovisit[lentovisit++] = parents[i];
 				revstates[parents[i] + 1] |= RS_SEEN;
 			}
 		}
 	}
 
 	/* Find all the nodes in between the roots we found and the heads
 	 * and add them to the reachable set */
 	if (includepath == 1) {
 		long minidx = minroot;
 		if (minidx < 0)
 			minidx = 0;
 		for (i = minidx; i < len; i++) {
 			if (!(revstates[i + 1] & RS_SEEN))
 				continue;
 			r = index_get_parents(self, i, parents, (int)len - 1);
 			/* Corrupted index file, error is set from
 			 * index_get_parents */
 			if (r < 0)
 				goto bail;
 			if (((revstates[parents[0] + 1] |
 			      revstates[parents[1] + 1]) &
 			     RS_REACHABLE) &&
 			    !(revstates[i + 1] & RS_REACHABLE)) {
 				revstates[i + 1] |= RS_REACHABLE;
 				val = PyLong_FromSsize_t(i);
 				if (val == NULL)
 					goto bail;
 				r = PyList_Append(reachable, val);
 				Py_DECREF(val);
 				if (r < 0)
 					goto bail;
 			}
 		}
 	}
 
 	free(revstates);
 	free(tovisit);
 	return reachable;
 bail:
 	Py_XDECREF(reachable);
 	free(revstates);
 	free(tovisit);
 	return NULL;
 }
 
 static int add_roots_get_min(indexObject *self, PyObject *roots, char *phases,
                              char phase)
 {
 	Py_ssize_t len = index_length(self);
 	PyObject *item;
 	PyObject *iterator;
 	int rev, minrev = -1;
 	char *node;
 
 	if (!PySet_Check(roots)) {
 		PyErr_SetString(PyExc_TypeError,
 		                "roots must be a set of nodes");
 		return -2;
 	}
 	iterator = PyObject_GetIter(roots);
 	if (iterator == NULL)
 		return -2;
 	while ((item = PyIter_Next(iterator))) {
 		if (node_check(self->nodelen, item, &node) == -1)
 			goto failed;
 		rev = index_find_node(self, node);
 		/* null is implicitly public, so negative is invalid */
 		if (rev < 0 || rev >= len)
 			goto failed;
 		phases[rev] = phase;
 		if (minrev == -1 || minrev > rev)
 			minrev = rev;
 		Py_DECREF(item);
 	}
 	Py_DECREF(iterator);
 	return minrev;
 failed:
 	Py_DECREF(iterator);
 	Py_DECREF(item);
 	return -2;
 }
 
 static PyObject *compute_phases_map_sets(indexObject *self, PyObject *args)
 {
 	/* 0: public (untracked), 1: draft, 2: secret, 32: archive,
 	   96: internal */
 	static const char trackedphases[] = {1, 2, 32, 96};
 	PyObject *roots = Py_None;
 	PyObject *phasesetsdict = NULL;
 	PyObject *phasesets[4] = {NULL, NULL, NULL, NULL};
 	Py_ssize_t len = index_length(self);
 	char *phases = NULL;
 	int minphaserev = -1, rev, i;
 	const int numphases = (int)(sizeof(phasesets) / sizeof(phasesets[0]));
 
 	if (!PyArg_ParseTuple(args, "O", &roots))
 		return NULL;
 	if (roots == NULL || !PyDict_Check(roots)) {
 		PyErr_SetString(PyExc_TypeError, "roots must be a dictionary");
 		return NULL;
 	}
 
 	phases = calloc(len, 1);
 	if (phases == NULL) {
 		PyErr_NoMemory();
 		return NULL;
 	}
 
 	for (i = 0; i < numphases; ++i) {
 		PyObject *pyphase = PyLong_FromLong(trackedphases[i]);
 		PyObject *phaseroots = NULL;
 		if (pyphase == NULL)
 			goto release;
 		phaseroots = PyDict_GetItem(roots, pyphase);
 		Py_DECREF(pyphase);
 		if (phaseroots == NULL)
 			continue;
 		rev = add_roots_get_min(self, phaseroots, phases,
 		                        trackedphases[i]);
 		if (rev == -2)
 			goto release;
 		if (rev != -1 && (minphaserev == -1 || rev < minphaserev))
 			minphaserev = rev;
 	}
 
 	for (i = 0; i < numphases; ++i) {
 		phasesets[i] = PySet_New(NULL);
 		if (phasesets[i] == NULL)
 			goto release;
 	}
 
 	if (minphaserev == -1)
 		minphaserev = len;
 	for (rev = minphaserev; rev < len; ++rev) {
 		PyObject *pyphase = NULL;
 		PyObject *pyrev = NULL;
 		int parents[2];
 		/*
 		 * The parent lookup could be skipped for phaseroots, but
 		 * phase --force would historically not recompute them
 		 * correctly, leaving descendents with a lower phase around.
 		 * As such, unconditionally recompute the phase.
 		 */
 		if (index_get_parents(self, rev, parents, (int)len - 1) < 0)
 			goto release;
 		set_phase_from_parents(phases, parents[0], parents[1], rev);
 		switch (phases[rev]) {
 		case 0:
 			continue;
 		case 1:
 			pyphase = phasesets[0];
 			break;
 		case 2:
 			pyphase = phasesets[1];
 			break;
 		case 32:
 			pyphase = phasesets[2];
 			break;
 		case 96:
 			pyphase = phasesets[3];
 			break;
 		default:
 			/* this should never happen since the phase number is
 			 * specified by this function. */
 			PyErr_SetString(PyExc_SystemError,
 			                "bad phase number in internal list");
 			goto release;
 		}
 		pyrev = PyLong_FromLong(rev);
 		if (pyrev == NULL)
 			goto release;
 		if (PySet_Add(pyphase, pyrev) == -1) {
 			Py_DECREF(pyrev);
 			goto release;
 		}
 		Py_DECREF(pyrev);
 	}
 
 	phasesetsdict = _dict_new_presized(numphases);
 	if (phasesetsdict == NULL)
 		goto release;
 	for (i = 0; i < numphases; ++i) {
 		PyObject *pyphase = PyLong_FromLong(trackedphases[i]);
 		if (pyphase == NULL)
 			goto release;
 		if (PyDict_SetItem(phasesetsdict, pyphase, phasesets[i]) ==
 		    -1) {
 			Py_DECREF(pyphase);
 			goto release;
 		}
 		Py_DECREF(phasesets[i]);
 		phasesets[i] = NULL;
 	}
 
 	free(phases);
 	return Py_BuildValue("nN", len, phasesetsdict);
 
 release:
 	for (i = 0; i < numphases; ++i)
 		Py_XDECREF(phasesets[i]);
 	Py_XDECREF(phasesetsdict);
 
 	free(phases);
 	return NULL;
 }
 
 static PyObject *index_headrevs(indexObject *self, PyObject *args)
 {
 	Py_ssize_t i, j, len;
 	char *nothead = NULL;
 	PyObject *heads = NULL;
 	PyObject *filter = NULL;
 	PyObject *filteredrevs = Py_None;
 
 	if (!PyArg_ParseTuple(args, "|O", &filteredrevs)) {
 		return NULL;
 	}
 
 	if (self->headrevs && filteredrevs == self->filteredrevs)
 		return list_copy(self->headrevs);
 
 	Py_DECREF(self->filteredrevs);
 	self->filteredrevs = filteredrevs;
 	Py_INCREF(filteredrevs);
 
 	if (filteredrevs != Py_None) {
 		filter = PyObject_GetAttrString(filteredrevs, "__contains__");
 		if (!filter) {
 			PyErr_SetString(
 			    PyExc_TypeError,
 			    "filteredrevs has no attribute __contains__");
 			goto bail;
 		}
 	}
 
 	len = index_length(self);
 	heads = PyList_New(0);
 	if (heads == NULL)
 		goto bail;
 	if (len == 0) {
 		PyObject *nullid = PyLong_FromLong(-1);
 		if (nullid == NULL || PyList_Append(heads, nullid) == -1) {
 			Py_XDECREF(nullid);
 			goto bail;
 		}
 		goto done;
 	}
 
 	nothead = calloc(len, 1);
 	if (nothead == NULL) {
 		PyErr_NoMemory();
 		goto bail;
 	}
 
 	for (i = len - 1; i >= 0; i--) {
 		int isfiltered;
 		int parents[2];
 
 		/* If nothead[i] == 1, it means we've seen an unfiltered child
 		 * of this node already, and therefore this node is not
 		 * filtered. So we can skip the expensive check_filter step.
 		 */
 		if (nothead[i] != 1) {
 			isfiltered = check_filter(filter, i);
 			if (isfiltered == -1) {
 				PyErr_SetString(PyExc_TypeError,
 				                "unable to check filter");
 				goto bail;
 			}
 
 			if (isfiltered) {
 				nothead[i] = 1;
 				continue;
 			}
 		}
 
 		if (index_get_parents(self, i, parents, (int)len - 1) < 0)
 			goto bail;
 		for (j = 0; j < 2; j++) {
 			if (parents[j] >= 0)
 				nothead[parents[j]] = 1;
 		}
 	}
 
 	for (i = 0; i < len; i++) {
 		PyObject *head;
 
 		if (nothead[i])
 			continue;
 		head = PyLong_FromSsize_t(i);
 		if (head == NULL || PyList_Append(heads, head) == -1) {
 			Py_XDECREF(head);
 			goto bail;
 		}
 	}
 
 done:
 	self->headrevs = heads;
 	Py_XDECREF(filter);
 	free(nothead);
 	return list_copy(self->headrevs);
 bail:
 	Py_XDECREF(filter);
 	Py_XDECREF(heads);
 	free(nothead);
 	return NULL;
 }
 
 /**
  * Obtain the base revision index entry.
  *
  * Callers must ensure that rev >= 0 or illegal memory access may occur.
  */
 static inline int index_baserev(indexObject *self, int rev)
 {
 	const char *data;
 	int result;
 
 	data = index_deref(self, rev);
 	if (data == NULL)
 		return -2;
 
 	if (self->format_version == format_v1) {
 		result = getbe32(data + entry_v1_offset_base_rev);
 	} else if (self->format_version == format_v2) {
 		result = getbe32(data + entry_v2_offset_base_rev);
 	} else if (self->format_version == format_cl2) {
 		return rev;
 	} else {
 		raise_revlog_error();
 		return -1;
 	}
 
 	if (result > rev) {
 		PyErr_Format(
 		    PyExc_ValueError,
 		    "corrupted revlog, revision base above revision: %d, %d",
 		    rev, result);
 		return -2;
 	}
 	if (result < -1) {
 		PyErr_Format(
 		    PyExc_ValueError,
 		    "corrupted revlog, revision base out of range: %d, %d", rev,
 		    result);
 		return -2;
 	}
 	return result;
 }
 
 /**
  * Find if a revision is a snapshot or not
  *
  * Only relevant for sparse-revlog case.
  * Callers must ensure that rev is in a valid range.
  */
 static int index_issnapshotrev(indexObject *self, Py_ssize_t rev)
 {
 	int ps[2];
 	Py_ssize_t base;
 	while (rev >= 0) {
 		base = (Py_ssize_t)index_baserev(self, rev);
 		if (base == rev) {
 			base = -1;
 		}
 		if (base == -2) {
 			assert(PyErr_Occurred());
 			return -1;
 		}
 		if (base == -1) {
 			return 1;
 		}
 		if (index_get_parents(self, rev, ps, (int)rev) < 0) {
 			assert(PyErr_Occurred());
 			return -1;
 		};
 		if (base == ps[0] || base == ps[1]) {
 			return 0;
 		}
 		rev = base;
 	}
 	return rev == -1;
 }
 
 static PyObject *index_issnapshot(indexObject *self, PyObject *value)
 {
 	long rev;
 	int issnap;
 	Py_ssize_t length = index_length(self);
 
 	if (!pylong_to_long(value, &rev)) {
 		return NULL;
 	}
 	if (rev < -1 || rev >= length) {
 		PyErr_Format(PyExc_ValueError, "revlog index out of range: %ld",
 		             rev);
 		return NULL;
 	};
 	issnap = index_issnapshotrev(self, (Py_ssize_t)rev);
 	if (issnap < 0) {
 		return NULL;
 	};
 	return PyBool_FromLong((long)issnap);
 }
 
 static PyObject *index_findsnapshots(indexObject *self, PyObject *args)
 {
 	Py_ssize_t start_rev;
 	PyObject *cache;
 	Py_ssize_t base;
 	Py_ssize_t rev;
 	PyObject *key = NULL;
 	PyObject *value = NULL;
 	const Py_ssize_t length = index_length(self);
 	if (!PyArg_ParseTuple(args, "O!n", &PyDict_Type, &cache, &start_rev)) {
 		return NULL;
 	}
 	for (rev = start_rev; rev < length; rev++) {
 		int issnap;
 		PyObject *allvalues = NULL;
 		issnap = index_issnapshotrev(self, rev);
 		if (issnap < 0) {
 			goto bail;
 		}
 		if (issnap == 0) {
 			continue;
 		}
 		base = (Py_ssize_t)index_baserev(self, rev);
 		if (base == rev) {
 			base = -1;
 		}
 		if (base == -2) {
 			assert(PyErr_Occurred());
 			goto bail;
 		}
 		key = PyLong_FromSsize_t(base);
 		allvalues = PyDict_GetItem(cache, key);
 		if (allvalues == NULL && PyErr_Occurred()) {
 			goto bail;
 		}
 		if (allvalues == NULL) {
 			int r;
 			allvalues = PyList_New(0);
 			if (!allvalues) {
 				goto bail;
 			}
 			r = PyDict_SetItem(cache, key, allvalues);
 			Py_DECREF(allvalues);
 			if (r < 0) {
 				goto bail;
 			}
 		}
 		value = PyLong_FromSsize_t(rev);
 		if (PyList_Append(allvalues, value)) {
 			goto bail;
 		}
 		Py_CLEAR(key);
 		Py_CLEAR(value);
 	}
 	Py_RETURN_NONE;
 bail:
 	Py_XDECREF(key);
 	Py_XDECREF(value);
 	return NULL;
 }
 
 static PyObject *index_deltachain(indexObject *self, PyObject *args)
 {
 	int rev, generaldelta;
 	PyObject *stoparg;
 	int stoprev, iterrev, baserev = -1;
 	int stopped;
 	PyObject *chain = NULL, *result = NULL;
 	const Py_ssize_t length = index_length(self);
 
 	if (!PyArg_ParseTuple(args, "iOi", &rev, &stoparg, &generaldelta)) {
 		return NULL;
 	}
 
 	if (PyLong_Check(stoparg)) {
 		stoprev = (int)PyLong_AsLong(stoparg);
 		if (stoprev == -1 && PyErr_Occurred()) {
 			return NULL;
 		}
 	} else if (stoparg == Py_None) {
 		stoprev = -2;
 	} else {
 		PyErr_SetString(PyExc_ValueError,
 		                "stoprev must be integer or None");
 		return NULL;
 	}
 
 	if (rev < 0 || rev >= length) {
 		PyErr_SetString(PyExc_ValueError, "revlog index out of range");
 		return NULL;
 	}
 
 	chain = PyList_New(0);
 	if (chain == NULL) {
 		return NULL;
 	}
 
 	baserev = index_baserev(self, rev);
 
 	/* This should never happen. */
 	if (baserev <= -2) {
 		/* Error should be set by index_deref() */
 		assert(PyErr_Occurred());
 		goto bail;
 	}
 
 	iterrev = rev;
 
 	while (iterrev != baserev && iterrev != stoprev) {
 		PyObject *value = PyLong_FromLong(iterrev);
 		if (value == NULL) {
 			goto bail;
 		}
 		if (PyList_Append(chain, value)) {
 			Py_DECREF(value);
 			goto bail;
 		}
 		Py_DECREF(value);
 
 		if (generaldelta) {
 			iterrev = baserev;
 		} else {
 			iterrev--;
 		}
 
 		if (iterrev < 0) {
 			break;
 		}
 
 		if (iterrev >= length) {
 			PyErr_SetString(PyExc_IndexError,
 			                "revision outside index");
 			return NULL;
 		}
 
 		baserev = index_baserev(self, iterrev);
 
 		/* This should never happen. */
 		if (baserev <= -2) {
 			/* Error should be set by index_deref() */
 			assert(PyErr_Occurred());
 			goto bail;
 		}
 	}
 
 	if (iterrev == stoprev) {
 		stopped = 1;
 	} else {
 		PyObject *value = PyLong_FromLong(iterrev);
 		if (value == NULL) {
 			goto bail;
 		}
 		if (PyList_Append(chain, value)) {
 			Py_DECREF(value);
 			goto bail;
 		}
 		Py_DECREF(value);
 
 		stopped = 0;
 	}
 
 	if (PyList_Reverse(chain)) {
 		goto bail;
 	}
 
 	result = Py_BuildValue("OO", chain, stopped ? Py_True : Py_False);
 	Py_DECREF(chain);
 	return result;
 
 bail:
 	Py_DECREF(chain);
 	return NULL;
 }
 
 static inline int64_t
 index_segment_span(indexObject *self, Py_ssize_t start_rev, Py_ssize_t end_rev)
 {
 	int64_t start_offset;
 	int64_t end_offset;
 	int end_size;
 	start_offset = index_get_start(self, start_rev);
 	if (start_offset < 0) {
 		return -1;
 	}
 	end_offset = index_get_start(self, end_rev);
 	if (end_offset < 0) {
 		return -1;
 	}
 	end_size = index_get_length(self, end_rev);
 	if (end_size < 0) {
 		return -1;
 	}
 	if (end_offset < start_offset) {
 		PyErr_Format(PyExc_ValueError,
 		             "corrupted revlog index: inconsistent offset "
 		             "between revisions (%zd) and (%zd)",
 		             start_rev, end_rev);
 		return -1;
 	}
 	return (end_offset - start_offset) + (int64_t)end_size;
 }
 
 /* returns endidx so that revs[startidx:endidx] has no empty trailing revs */
 static Py_ssize_t trim_endidx(indexObject *self, const Py_ssize_t *revs,
                               Py_ssize_t startidx, Py_ssize_t endidx)
 {
 	int length;
 	while (endidx > 1 && endidx > startidx) {
 		length = index_get_length(self, revs[endidx - 1]);
 		if (length < 0) {
 			return -1;
 		}
 		if (length != 0) {
 			break;
 		}
 		endidx -= 1;
 	}
 	return endidx;
 }
 
 struct Gap {
 	int64_t size;
 	Py_ssize_t idx;
 };
 
 static int gap_compare(const void *left, const void *right)
 {
 	const struct Gap *l_left = ((const struct Gap *)left);
 	const struct Gap *l_right = ((const struct Gap *)right);
 	if (l_left->size < l_right->size) {
 		return -1;
 	} else if (l_left->size > l_right->size) {
 		return 1;
 	}
 	return 0;
 }
 static int Py_ssize_t_compare(const void *left, const void *right)
 {
 	const Py_ssize_t l_left = *(const Py_ssize_t *)left;
 	const Py_ssize_t l_right = *(const Py_ssize_t *)right;
 	if (l_left < l_right) {
 		return -1;
 	} else if (l_left > l_right) {
 		return 1;
 	}
 	return 0;
 }
 
 static PyObject *index_slicechunktodensity(indexObject *self, PyObject *args)
 {
 	/* method arguments */
 	PyObject *list_revs = NULL; /* revisions in the chain */
 	double targetdensity = 0;   /* min density to achieve */
 	Py_ssize_t mingapsize = 0;  /* threshold to ignore gaps */
 
 	/* other core variables */
 	Py_ssize_t idxlen = index_length(self);
 	Py_ssize_t i;            /* used for various iteration */
 	PyObject *result = NULL; /* the final return of the function */
 
 	/* generic information about the delta chain being slice */
 	Py_ssize_t num_revs = 0;    /* size of the full delta chain */
 	Py_ssize_t *revs = NULL;    /* native array of revision in the chain */
 	int64_t chainpayload = 0;   /* sum of all delta in the chain */
 	int64_t deltachainspan = 0; /* distance from first byte to last byte */
 
 	/* variable used for slicing the delta chain */
 	int64_t readdata = 0; /* amount of data currently planned to be read */
 	double density = 0;   /* ration of payload data compared to read ones */
 	int64_t previous_end;
 	struct Gap *gaps = NULL; /* array of notable gap in the chain */
 	Py_ssize_t num_gaps =
 	    0; /* total number of notable gap recorded so far */
 	Py_ssize_t *selected_indices = NULL; /* indices of gap skipped over */
 	Py_ssize_t num_selected = 0;         /* number of gaps skipped */
 	PyObject *chunk = NULL;              /* individual slice */
 	PyObject *allchunks = NULL;          /* all slices */
 	Py_ssize_t previdx;
 
 	/* parsing argument */
 	if (!PyArg_ParseTuple(args, "O!dn", &PyList_Type, &list_revs,
 	                      &targetdensity, &mingapsize)) {
 		goto bail;
 	}
 
 	/* If the delta chain contains a single element, we do not need slicing
 	 */
 	num_revs = PyList_GET_SIZE(list_revs);
 	if (num_revs <= 1) {
 		result = PyTuple_Pack(1, list_revs);
 		goto done;
 	}
 
 	/* Turn the python list into a native integer array (for efficiency) */
 	revs = (Py_ssize_t *)calloc(num_revs, sizeof(Py_ssize_t));
 	if (revs == NULL) {
 		PyErr_NoMemory();
 		goto bail;
 	}
 	for (i = 0; i < num_revs; i++) {
 		Py_ssize_t revnum =
 		    PyLong_AsLong(PyList_GET_ITEM(list_revs, i));
 		if (revnum == -1 && PyErr_Occurred()) {
 			goto bail;
 		}
 		if (revnum < nullrev || revnum >= idxlen) {
 			PyErr_Format(PyExc_IndexError,
 			             "index out of range: %zd", revnum);
 			goto bail;
 		}
 		revs[i] = revnum;
 	}
 
 	/* Compute and check various property of the unsliced delta chain */
 	deltachainspan = index_segment_span(self, revs[0], revs[num_revs - 1]);
 	if (deltachainspan < 0) {
 		goto bail;
 	}
 
 	if (deltachainspan <= mingapsize) {
 		result = PyTuple_Pack(1, list_revs);
 		goto done;
 	}
 	chainpayload = 0;
 	for (i = 0; i < num_revs; i++) {
 		int tmp = index_get_length(self, revs[i]);
 		if (tmp < 0) {
 			goto bail;
 		}
 		chainpayload += tmp;
 	}
 
 	readdata = deltachainspan;
 	density = 1.0;
 
 	if (0 < deltachainspan) {
 		density = (double)chainpayload / (double)deltachainspan;
 	}
 
 	if (density >= targetdensity) {
 		result = PyTuple_Pack(1, list_revs);
 		goto done;
 	}
 
 	/* if chain is too sparse, look for relevant gaps */
 	gaps = (struct Gap *)calloc(num_revs, sizeof(struct Gap));
 	if (gaps == NULL) {
 		PyErr_NoMemory();
 		goto bail;
 	}
 
 	previous_end = -1;
 	for (i = 0; i < num_revs; i++) {
 		int64_t revstart;
 		int revsize;
 		revstart = index_get_start(self, revs[i]);
 		if (revstart < 0) {
 			goto bail;
 		};
 		revsize = index_get_length(self, revs[i]);
 		if (revsize < 0) {
 			goto bail;
 		};
 		if (revsize == 0) {
 			continue;
 		}
 		if (previous_end >= 0) {
 			int64_t gapsize = revstart - previous_end;
 			if (gapsize > mingapsize) {
 				gaps[num_gaps].size = gapsize;
 				gaps[num_gaps].idx = i;
 				num_gaps += 1;
 			}
 		}
 		previous_end = revstart + revsize;
 	}
 	if (num_gaps == 0) {
 		result = PyTuple_Pack(1, list_revs);
 		goto done;
 	}
 	qsort(gaps, num_gaps, sizeof(struct Gap), &gap_compare);
 
 	/* Slice the largest gap first, they improve the density the most */
 	selected_indices =
 	    (Py_ssize_t *)malloc((num_gaps + 1) * sizeof(Py_ssize_t));
 	if (selected_indices == NULL) {
 		PyErr_NoMemory();
 		goto bail;
 	}
 
 	for (i = num_gaps - 1; i >= 0; i--) {
 		selected_indices[num_selected] = gaps[i].idx;
 		readdata -= gaps[i].size;
 		num_selected += 1;
 		if (readdata <= 0) {
 			density = 1.0;
 		} else {
 			density = (double)chainpayload / (double)readdata;
 		}
 		if (density >= targetdensity) {
 			break;
 		}
 	}
 	qsort(selected_indices, num_selected, sizeof(Py_ssize_t),
 	      &Py_ssize_t_compare);
 
 	/* create the resulting slice */
 	allchunks = PyList_New(0);
 	if (allchunks == NULL) {
 		goto bail;
 	}
 	previdx = 0;
 	selected_indices[num_selected] = num_revs;
 	for (i = 0; i <= num_selected; i++) {
 		Py_ssize_t idx = selected_indices[i];
 		Py_ssize_t endidx = trim_endidx(self, revs, previdx, idx);
 		if (endidx < 0) {
 			goto bail;
 		}
 		if (previdx < endidx) {
 			chunk = PyList_GetSlice(list_revs, previdx, endidx);
 			if (chunk == NULL) {
 				goto bail;
 			}
 			if (PyList_Append(allchunks, chunk) == -1) {
 				goto bail;
 			}
 			Py_DECREF(chunk);
 			chunk = NULL;
 		}
 		previdx = idx;
 	}
 	result = allchunks;
 	goto done;
 
 bail:
 	Py_XDECREF(allchunks);
 	Py_XDECREF(chunk);
 done:
 	free(revs);
 	free(gaps);
 	free(selected_indices);
 	return result;
 }
 
 static inline int nt_level(const char *node, Py_ssize_t level)
 {
 	int v = node[level >> 1];
 	if (!(level & 1))
 		v >>= 4;
 	return v & 0xf;
 }
 
 /*
  * Return values:
  *
  *   -4: match is ambiguous (multiple candidates)
  *   -2: not found
  * rest: valid rev
  */
 static int nt_find(nodetree *self, const char *node, Py_ssize_t nodelen,
                    int hex)
 {
 	int (*getnybble)(const char *, Py_ssize_t) = hex ? hexdigit : nt_level;
 	int level, maxlevel, off;
 
 	/* If the input is binary, do a fast check for the nullid first. */
 	if (!hex && nodelen == self->nodelen && node[0] == '\0' &&
 	    node[1] == '\0' && memcmp(node, nullid, self->nodelen) == 0)
 		return -1;
 
 	if (hex)
 		maxlevel = nodelen;
 	else
 		maxlevel = 2 * nodelen;
 	if (maxlevel > 2 * self->nodelen)
 		maxlevel = 2 * self->nodelen;
 
 	for (level = off = 0; level < maxlevel; level++) {
 		int k = getnybble(node, level);
 		nodetreenode *n = &self->nodes[off];
 		int v = n->children[k];
 
 		if (v < 0) {
 			const char *n;
 			Py_ssize_t i;
 
 			v = -(v + 2);
 			n = index_node(self->index, v);
 			if (n == NULL)
 				return -2;
 			for (i = level; i < maxlevel; i++)
 				if (getnybble(node, i) != nt_level(n, i))
 					return -2;
 			return v;
 		}
 		if (v == 0)
 			return -2;
 		off = v;
 	}
 	/* multiple matches against an ambiguous prefix */
 	return -4;
 }
 
 static int nt_new(nodetree *self)
 {
 	if (self->length == self->capacity) {
 		size_t newcapacity;
 		nodetreenode *newnodes;
 		newcapacity = self->capacity * 2;
 		if (newcapacity >= SIZE_MAX / sizeof(nodetreenode)) {
 			PyErr_SetString(PyExc_MemoryError,
 			                "overflow in nt_new");
 			return -1;
 		}
 		newnodes =
 		    realloc(self->nodes, newcapacity * sizeof(nodetreenode));
 		if (newnodes == NULL) {
 			PyErr_SetString(PyExc_MemoryError, "out of memory");
 			return -1;
 		}
 		self->capacity = newcapacity;
 		self->nodes = newnodes;
 		memset(&self->nodes[self->length], 0,
 		       sizeof(nodetreenode) * (self->capacity - self->length));
 	}
 	return self->length++;
 }
 
 static int nt_insert(nodetree *self, const char *node, int rev)
 {
 	int level = 0;
 	int off = 0;
 
 	while (level < 2 * self->nodelen) {
 		int k = nt_level(node, level);
 		nodetreenode *n;
 		int v;
 
 		n = &self->nodes[off];
 		v = n->children[k];
 
 		if (v == 0) {
 			n->children[k] = -rev - 2;
 			return 0;
 		}
 		if (v < 0) {
 			const char *oldnode =
 			    index_node_existing(self->index, -(v + 2));
 			int noff;
 
 			if (oldnode == NULL)
 				return -1;
 			if (!memcmp(oldnode, node, self->nodelen)) {
 				n->children[k] = -rev - 2;
 				return 0;
 			}
 			noff = nt_new(self);
 			if (noff == -1)
 				return -1;
 			/* self->nodes may have been changed by realloc */
 			self->nodes[off].children[k] = noff;
 			off = noff;
 			n = &self->nodes[off];
 			n->children[nt_level(oldnode, ++level)] = v;
 			if (level > self->depth)
 				self->depth = level;
 			self->splits += 1;
 		} else {
 			level += 1;
 			off = v;
 		}
 	}
 
 	return -1;
 }
 
 static PyObject *ntobj_insert(nodetreeObject *self, PyObject *args)
 {
 	Py_ssize_t rev;
 	const char *node;
 	Py_ssize_t length;
 	if (!PyArg_ParseTuple(args, "n", &rev))
 		return NULL;
 	length = index_length(self->nt.index);
 	if (rev < 0 || rev >= length) {
 		PyErr_SetString(PyExc_ValueError, "revlog index out of range");
 		return NULL;
 	}
 	node = index_node_existing(self->nt.index, rev);
 	if (nt_insert(&self->nt, node, (int)rev) == -1)
 		return NULL;
 	Py_RETURN_NONE;
 }
 
 static int nt_delete_node(nodetree *self, const char *node)
 {
 	/* rev==-2 happens to get encoded as 0, which is interpreted as not set
 	 */
 	return nt_insert(self, node, -2);
 }
 
 static int nt_init(nodetree *self, indexObject *index, unsigned capacity)
 {
 	/* Initialize before overflow-checking to avoid nt_dealloc() crash. */
 	self->nodes = NULL;
 
 	self->index = index;
 	/* The input capacity is in terms of revisions, while the field is in
 	 * terms of nodetree nodes. */
 	self->capacity = (capacity < 4 ? 4 : capacity / 2);
 	self->nodelen = index->nodelen;
 	self->depth = 0;
 	self->splits = 0;
 	if (self->capacity > SIZE_MAX / sizeof(nodetreenode)) {
 		PyErr_SetString(PyExc_ValueError, "overflow in init_nt");
 		return -1;
 	}
 	self->nodes = calloc(self->capacity, sizeof(nodetreenode));
 	if (self->nodes == NULL) {
 		PyErr_NoMemory();
 		return -1;
 	}
 	self->length = 1;
 	return 0;
 }
 
 static int ntobj_init(nodetreeObject *self, PyObject *args)
 {
 	PyObject *index;
 	unsigned capacity;
 	if (!PyArg_ParseTuple(args, "O!I", &HgRevlogIndex_Type, &index,
 	                      &capacity))
 		return -1;
 	Py_INCREF(index);
 	return nt_init(&self->nt, (indexObject *)index, capacity);
 }
 
 static int nt_partialmatch(nodetree *self, const char *node, Py_ssize_t nodelen)
 {
 	return nt_find(self, node, nodelen, 1);
 }
 
 /*
  * Find the length of the shortest unique prefix of node.
  *
  * Return values:
  *
  *   -3: error (exception set)
  *   -2: not found (no exception set)
  * rest: length of shortest prefix
  */
 static int nt_shortest(nodetree *self, const char *node)
 {
 	int level, off;
 
 	for (level = off = 0; level < 2 * self->nodelen; level++) {
 		int k, v;
 		nodetreenode *n = &self->nodes[off];
 		k = nt_level(node, level);
 		v = n->children[k];
 		if (v < 0) {
 			const char *n;
 			v = -(v + 2);
 			n = index_node_existing(self->index, v);
 			if (n == NULL)
 				return -3;
 			if (memcmp(node, n, self->nodelen) != 0)
 				/*
 				 * Found a unique prefix, but it wasn't for the
 				 * requested node (i.e the requested node does
 				 * not exist).
 				 */
 				return -2;
 			return level + 1;
 		}
 		if (v == 0)
 			return -2;
 		off = v;
 	}
 	/*
 	 * The node was still not unique after 40 hex digits, so this won't
 	 * happen. Also, if we get here, then there's a programming error in
 	 * this file that made us insert a node longer than 40 hex digits.
 	 */
 	PyErr_SetString(PyExc_Exception, "broken node tree");
 	return -3;
 }
 
 static PyObject *ntobj_shortest(nodetreeObject *self, PyObject *args)
 {
 	PyObject *val;
 	char *node;
 	int length;
 
 	if (!PyArg_ParseTuple(args, "O", &val))
 		return NULL;
 	if (node_check(self->nt.nodelen, val, &node) == -1)
 		return NULL;
 
 	length = nt_shortest(&self->nt, node);
 	if (length == -3)
 		return NULL;
 	if (length == -2) {
 		raise_revlog_error();
 		return NULL;
 	}
 	return PyLong_FromLong(length);
 }
 
 static void nt_dealloc(nodetree *self)
 {
 	free(self->nodes);
 	self->nodes = NULL;
 }
 
 static void ntobj_dealloc(nodetreeObject *self)
 {
 	Py_XDECREF(self->nt.index);
 	nt_dealloc(&self->nt);
 	PyObject_Del(self);
 }
 
 static PyMethodDef ntobj_methods[] = {
     {"insert", (PyCFunction)ntobj_insert, METH_VARARGS,
      "insert an index entry"},
     {"shortest", (PyCFunction)ntobj_shortest, METH_VARARGS,
      "find length of shortest hex nodeid of a binary ID"},
     {NULL} /* Sentinel */
 };
 
 static PyTypeObject nodetreeType = {
     PyVarObject_HEAD_INIT(NULL, 0) /* header */
     "parsers.nodetree",            /* tp_name */
     sizeof(nodetreeObject),        /* tp_basicsize */
     0,                             /* tp_itemsize */
     (destructor)ntobj_dealloc,     /* tp_dealloc */
     0,                             /* tp_print */
     0,                             /* tp_getattr */
     0,                             /* tp_setattr */
     0,                             /* tp_compare */
     0,                             /* tp_repr */
     0,                             /* tp_as_number */
     0,                             /* tp_as_sequence */
     0,                             /* tp_as_mapping */
     0,                             /* tp_hash */
     0,                             /* tp_call */
     0,                             /* tp_str */
     0,                             /* tp_getattro */
     0,                             /* tp_setattro */
     0,                             /* tp_as_buffer */
     Py_TPFLAGS_DEFAULT,            /* tp_flags */
     "nodetree",                    /* tp_doc */
     0,                             /* tp_traverse */
     0,                             /* tp_clear */
     0,                             /* tp_richcompare */
     0,                             /* tp_weaklistoffset */
     0,                             /* tp_iter */
     0,                             /* tp_iternext */
     ntobj_methods,                 /* tp_methods */
     0,                             /* tp_members */
     0,                             /* tp_getset */
     0,                             /* tp_base */
     0,                             /* tp_dict */
     0,                             /* tp_descr_get */
     0,                             /* tp_descr_set */
     0,                             /* tp_dictoffset */
     (initproc)ntobj_init,          /* tp_init */
     0,                             /* tp_alloc */
 };
 
 static int index_init_nt(indexObject *self)
 {
 	if (!self->ntinitialized) {
 		if (nt_init(&self->nt, self, (int)self->length) == -1) {
 			nt_dealloc(&self->nt);
 			return -1;
 		}
 		if (nt_insert(&self->nt, nullid, -1) == -1) {
 			nt_dealloc(&self->nt);
 			return -1;
 		}
 		self->ntinitialized = 1;
 		self->ntrev = (int)index_length(self);
 		self->ntlookups = 1;
 		self->ntmisses = 0;
 	}
 	return 0;
 }
 
 /*
  * Return values:
  *
  *   -3: error (exception set)
  *   -2: not found (no exception set)
  * rest: valid rev
  */
 static int index_find_node(indexObject *self, const char *node)
 {
 	int rev;
 
 	if (index_init_nt(self) == -1)
 		return -3;
 
 	self->ntlookups++;
 	rev = nt_find(&self->nt, node, self->nodelen, 0);
 	if (rev >= -1)
 		return rev;
 
 	/*
 	 * For the first handful of lookups, we scan the entire index,
 	 * and cache only the matching nodes. This optimizes for cases
 	 * like "hg tip", where only a few nodes are accessed.
 	 *
 	 * After that, we cache every node we visit, using a single
 	 * scan amortized over multiple lookups.  This gives the best
 	 * bulk performance, e.g. for "hg log".
 	 */
 	if (self->ntmisses++ < 4) {
 		for (rev = self->ntrev - 1; rev >= 0; rev--) {
 			const char *n = index_node_existing(self, rev);
 			if (n == NULL)
 				return -3;
 			if (memcmp(node, n, self->nodelen) == 0) {
 				if (nt_insert(&self->nt, n, rev) == -1)
 					return -3;
 				break;
 			}
 		}
 	} else {
 		for (rev = self->ntrev - 1; rev >= 0; rev--) {
 			const char *n = index_node_existing(self, rev);
 			if (n == NULL)
 				return -3;
 			if (nt_insert(&self->nt, n, rev) == -1) {
 				self->ntrev = rev + 1;
 				return -3;
 			}
 			if (memcmp(node, n, self->nodelen) == 0) {
 				break;
 			}
 		}
 		self->ntrev = rev;
 	}
 
 	if (rev >= 0)
 		return rev;
 	return -2;
 }
 
 static PyObject *index_getitem(indexObject *self, PyObject *value)
 {
 	char *node;
 	int rev;
 
 	if (PyLong_Check(value)) {
 		long idx;
 		if (!pylong_to_long(value, &idx)) {
 			return NULL;
 		}
 		return index_get(self, idx);
 	}
 
 	if (node_check(self->nodelen, value, &node) == -1)
 		return NULL;
 	rev = index_find_node(self, node);
 	if (rev >= -1)
 		return PyLong_FromLong(rev);
 	if (rev == -2)
 		raise_revlog_error();
 	return NULL;
 }
 
 /*
  * Fully populate the radix tree.
  */
 static int index_populate_nt(indexObject *self)
 {
 	int rev;
 	if (self->ntrev > 0) {
 		for (rev = self->ntrev - 1; rev >= 0; rev--) {
 			const char *n = index_node_existing(self, rev);
 			if (n == NULL)
 				return -1;
 			if (nt_insert(&self->nt, n, rev) == -1)
 				return -1;
 		}
 		self->ntrev = -1;
 	}
 	return 0;
 }
 
 static PyObject *index_partialmatch(indexObject *self, PyObject *args)
 {
 	const char *fullnode;
 	Py_ssize_t nodelen;
 	char *node;
 	int rev, i;
 
-	if (!PyArg_ParseTuple(args, PY23("s#", "y#"), &node, &nodelen))
+	if (!PyArg_ParseTuple(args, "y#", &node, &nodelen))
 		return NULL;
 
 	if (nodelen < 1) {
 		PyErr_SetString(PyExc_ValueError, "key too short");
 		return NULL;
 	}
 
 	if (nodelen > 2 * self->nodelen) {
 		PyErr_SetString(PyExc_ValueError, "key too long");
 		return NULL;
 	}
 
 	for (i = 0; i < nodelen; i++)
 		hexdigit(node, i);
 	if (PyErr_Occurred()) {
 		/* input contains non-hex characters */
 		PyErr_Clear();
 		Py_RETURN_NONE;
 	}
 
 	if (index_init_nt(self) == -1)
 		return NULL;
 	if (index_populate_nt(self) == -1)
 		return NULL;
 	rev = nt_partialmatch(&self->nt, node, nodelen);
 
 	switch (rev) {
 	case -4:
 		raise_revlog_error();
 		return NULL;
 	case -2:
 		Py_RETURN_NONE;
 	case -1:
 		return PyBytes_FromStringAndSize(nullid, self->nodelen);
 	}
 
 	fullnode = index_node_existing(self, rev);
 	if (fullnode == NULL) {
 		return NULL;
 	}
 	return PyBytes_FromStringAndSize(fullnode, self->nodelen);
 }
 
 static PyObject *index_shortest(indexObject *self, PyObject *args)
 {
 	PyObject *val;
 	char *node;
 	int length;
 
 	if (!PyArg_ParseTuple(args, "O", &val))
 		return NULL;
 	if (node_check(self->nodelen, val, &node) == -1)
 		return NULL;
 
 	self->ntlookups++;
 	if (index_init_nt(self) == -1)
 		return NULL;
 	if (index_populate_nt(self) == -1)
 		return NULL;
 	length = nt_shortest(&self->nt, node);
 	if (length == -3)
 		return NULL;
 	if (length == -2) {
 		raise_revlog_error();
 		return NULL;
 	}
 	return PyLong_FromLong(length);
 }
 
 static PyObject *index_m_get(indexObject *self, PyObject *args)
 {
 	PyObject *val;
 	char *node;
 	int rev;
 
 	if (!PyArg_ParseTuple(args, "O", &val))
 		return NULL;
 	if (node_check(self->nodelen, val, &node) == -1)
 		return NULL;
 	rev = index_find_node(self, node);
 	if (rev == -3)
 		return NULL;
 	if (rev == -2)
 		Py_RETURN_NONE;
 	return PyLong_FromLong(rev);
 }
 
 static int index_contains(indexObject *self, PyObject *value)
 {
 	char *node;
 
 	if (PyLong_Check(value)) {
 		long rev;
 		if (!pylong_to_long(value, &rev)) {
 			return -1;
 		}
 		return rev >= -1 && rev < index_length(self);
 	}
 
 	if (node_check(self->nodelen, value, &node) == -1)
 		return -1;
 
 	switch (index_find_node(self, node)) {
 	case -3:
 		return -1;
 	case -2:
 		return 0;
 	default:
 		return 1;
 	}
 }
 
 static PyObject *index_m_has_node(indexObject *self, PyObject *args)
 {
 	int ret = index_contains(self, args);
 	if (ret < 0)
 		return NULL;
 	return PyBool_FromLong((long)ret);
 }
 
 static PyObject *index_m_rev(indexObject *self, PyObject *val)
 {
 	char *node;
 	int rev;
 
 	if (node_check(self->nodelen, val, &node) == -1)
 		return NULL;
 	rev = index_find_node(self, node);
 	if (rev >= -1)
 		return PyLong_FromLong(rev);
 	if (rev == -2)
 		raise_revlog_error();
 	return NULL;
 }
 
 typedef uint64_t bitmask;
 
 /*
  * Given a disjoint set of revs, return all candidates for the
  * greatest common ancestor. In revset notation, this is the set
  * "heads(::a and ::b and ...)"
  */
 static PyObject *find_gca_candidates(indexObject *self, const int *revs,
                                      int revcount)
 {
 	const bitmask allseen = (1ull << revcount) - 1;
 	const bitmask poison = 1ull << revcount;
 	PyObject *gca = PyList_New(0);
 	int i, v, interesting;
 	int maxrev = -1;
 	bitmask sp;
 	bitmask *seen;
 
 	if (gca == NULL)
 		return PyErr_NoMemory();
 
 	for (i = 0; i < revcount; i++) {
 		if (revs[i] > maxrev)
 			maxrev = revs[i];
 	}
 
 	seen = calloc(sizeof(*seen), maxrev + 1);
 	if (seen == NULL) {
 		Py_DECREF(gca);
 		return PyErr_NoMemory();
 	}
 
 	for (i = 0; i < revcount; i++)
 		seen[revs[i]] = 1ull << i;
 
 	interesting = revcount;
 
 	for (v = maxrev; v >= 0 && interesting; v--) {
 		bitmask sv = seen[v];
 		int parents[2];
 
 		if (!sv)
 			continue;
 
 		if (sv < poison) {
 			interesting -= 1;
 			if (sv == allseen) {
 				PyObject *obj = PyLong_FromLong(v);
 				if (obj == NULL)
 					goto bail;
 				if (PyList_Append(gca, obj) == -1) {
 					Py_DECREF(obj);
 					goto bail;
 				}
 				sv |= poison;
 				for (i = 0; i < revcount; i++) {
 					if (revs[i] == v)
 						goto done;
 				}
 			}
 		}
 		if (index_get_parents(self, v, parents, maxrev) < 0)
 			goto bail;
 
 		for (i = 0; i < 2; i++) {
 			int p = parents[i];
 			if (p == -1)
 				continue;
 			sp = seen[p];
 			if (sv < poison) {
 				if (sp == 0) {
 					seen[p] = sv;
 					interesting++;
 				} else if (sp != sv)
 					seen[p] |= sv;
 			} else {
 				if (sp && sp < poison)
 					interesting--;
 				seen[p] = sv;
 			}
 		}
 	}
 
 done:
 	free(seen);
 	return gca;
 bail:
 	free(seen);
 	Py_XDECREF(gca);
 	return NULL;
 }
 
 /*
  * Given a disjoint set of revs, return the subset with the longest
  * path to the root.
  */
 static PyObject *find_deepest(indexObject *self, PyObject *revs)
 {
 	const Py_ssize_t revcount = PyList_GET_SIZE(revs);
 	static const Py_ssize_t capacity = 24;
 	int *depth, *interesting = NULL;
 	int i, j, v, ninteresting;
 	PyObject *dict = NULL, *keys = NULL;
 	long *seen = NULL;
 	int maxrev = -1;
 	long final;
 
 	if (revcount > capacity) {
 		PyErr_Format(PyExc_OverflowError,
 		             "bitset size (%ld) > capacity (%ld)",
 		             (long)revcount, (long)capacity);
 		return NULL;
 	}
 
 	for (i = 0; i < revcount; i++) {
 		int n = (int)PyLong_AsLong(PyList_GET_ITEM(revs, i));
 		if (n > maxrev)
 			maxrev = n;
 	}
 
 	depth = calloc(sizeof(*depth), maxrev + 1);
 	if (depth == NULL)
 		return PyErr_NoMemory();
 
 	seen = calloc(sizeof(*seen), maxrev + 1);
 	if (seen == NULL) {
 		PyErr_NoMemory();
 		goto bail;
 	}
 
 	interesting = calloc(sizeof(*interesting), ((size_t)1) << revcount);
 	if (interesting == NULL) {
 		PyErr_NoMemory();
 		goto bail;
 	}
 
 	if (PyList_Sort(revs) == -1)
 		goto bail;
 
 	for (i = 0; i < revcount; i++) {
 		int n = (int)PyLong_AsLong(PyList_GET_ITEM(revs, i));
 		long b = 1l << i;
 		depth[n] = 1;
 		seen[n] = b;
 		interesting[b] = 1;
 	}
 
 	/* invariant: ninteresting is the number of non-zero entries in
 	 * interesting. */
 	ninteresting = (int)revcount;
 
 	for (v = maxrev; v >= 0 && ninteresting > 1; v--) {
 		int dv = depth[v];
 		int parents[2];
 		long sv;
 
 		if (dv == 0)
 			continue;
 
 		sv = seen[v];
 		if (index_get_parents(self, v, parents, maxrev) < 0)
 			goto bail;
 
 		for (i = 0; i < 2; i++) {
 			int p = parents[i];
 			long sp;
 			int dp;
 
 			if (p == -1)
 				continue;
 
 			dp = depth[p];
 			sp = seen[p];
 			if (dp <= dv) {
 				depth[p] = dv + 1;
 				if (sp != sv) {
 					interesting[sv] += 1;
 					seen[p] = sv;
 					if (sp) {
 						interesting[sp] -= 1;
 						if (interesting[sp] == 0)
 							ninteresting -= 1;
 					}
 				}
 			} else if (dv == dp - 1) {
 				long nsp = sp | sv;
 				if (nsp == sp)
 					continue;
 				seen[p] = nsp;
 				interesting[sp] -= 1;
 				if (interesting[sp] == 0)
 					ninteresting -= 1;
 				if (interesting[nsp] == 0)
 					ninteresting += 1;
 				interesting[nsp] += 1;
 			}
 		}
 		interesting[sv] -= 1;
 		if (interesting[sv] == 0)
 			ninteresting -= 1;
 	}
 
 	final = 0;
 	j = ninteresting;
 	for (i = 0; i < (int)(2 << revcount) && j > 0; i++) {
 		if (interesting[i] == 0)
 			continue;
 		final |= i;
 		j -= 1;
 	}
 	if (final == 0) {
 		keys = PyList_New(0);
 		goto bail;
 	}
 
 	dict = PyDict_New();
 	if (dict == NULL)
 		goto bail;
 
 	for (i = 0; i < revcount; i++) {
 		PyObject *key;
 
 		if ((final & (1 << i)) == 0)
 			continue;
 
 		key = PyList_GET_ITEM(revs, i);
 		Py_INCREF(key);
 		Py_INCREF(Py_None);
 		if (PyDict_SetItem(dict, key, Py_None) == -1) {
 			Py_DECREF(key);
 			Py_DECREF(Py_None);
 			goto bail;
 		}
 	}
 
 	keys = PyDict_Keys(dict);
 
 bail:
 	free(depth);
 	free(seen);
 	free(interesting);
 	Py_XDECREF(dict);
 
 	return keys;
 }
 
 /*
  * Given a (possibly overlapping) set of revs, return all the
  * common ancestors heads: heads(::args[0] and ::a[1] and ...)
  */
 static PyObject *index_commonancestorsheads(indexObject *self, PyObject *args)
 {
 	PyObject *ret = NULL;
 	Py_ssize_t argcount, i, len;
 	bitmask repeat = 0;
 	int revcount = 0;
 	int *revs;
 
 	argcount = PySequence_Length(args);
 	revs = PyMem_Malloc(argcount * sizeof(*revs));
 	if (argcount > 0 && revs == NULL)
 		return PyErr_NoMemory();
 	len = index_length(self);
 
 	for (i = 0; i < argcount; i++) {
 		static const int capacity = 24;
 		PyObject *obj = PySequence_GetItem(args, i);
 		bitmask x;
 		long val;
 
 		if (!PyLong_Check(obj)) {
 			PyErr_SetString(PyExc_TypeError,
 			                "arguments must all be ints");
 			Py_DECREF(obj);
 			goto bail;
 		}
 		val = PyLong_AsLong(obj);
 		Py_DECREF(obj);
 		if (val == -1) {
 			ret = PyList_New(0);
 			goto done;
 		}
 		if (val < 0 || val >= len) {
 			PyErr_SetString(PyExc_IndexError, "index out of range");
 			goto bail;
 		}
 		/* this cheesy bloom filter lets us avoid some more
 		 * expensive duplicate checks in the common set-is-disjoint
 		 * case */
 		x = 1ull << (val & 0x3f);
 		if (repeat & x) {
 			int k;
 			for (k = 0; k < revcount; k++) {
 				if (val == revs[k])
 					goto duplicate;
 			}
 		} else
 			repeat |= x;
 		if (revcount >= capacity) {
 			PyErr_Format(PyExc_OverflowError,
 			             "bitset size (%d) > capacity (%d)",
 			             revcount, capacity);
 			goto bail;
 		}
 		revs[revcount++] = (int)val;
 	duplicate:;
 	}
 
 	if (revcount == 0) {
 		ret = PyList_New(0);
 		goto done;
 	}
 	if (revcount == 1) {
 		PyObject *obj;
 		ret = PyList_New(1);
 		if (ret == NULL)
 			goto bail;
 		obj = PyLong_FromLong(revs[0]);
 		if (obj == NULL)
 			goto bail;
 		PyList_SET_ITEM(ret, 0, obj);
 		goto done;
 	}
 
 	ret = find_gca_candidates(self, revs, revcount);
 	if (ret == NULL)
 		goto bail;
 
 done:
 	PyMem_Free(revs);
 	return ret;
 
 bail:
 	PyMem_Free(revs);
 	Py_XDECREF(ret);
 	return NULL;
 }
 
 /*
  * Given a (possibly overlapping) set of revs, return the greatest
  * common ancestors: those with the longest path to the root.
  */
 static PyObject *index_ancestors(indexObject *self, PyObject *args)
 {
 	PyObject *ret;
 	PyObject *gca = index_commonancestorsheads(self, args);
 	if (gca == NULL)
 		return NULL;
 
 	if (PyList_GET_SIZE(gca) <= 1) {
 		return gca;
 	}
 
 	ret = find_deepest(self, gca);
 	Py_DECREF(gca);
 	return ret;
 }
 
 /*
  * Invalidate any trie entries introduced by added revs.
  */
 static void index_invalidate_added(indexObject *self, Py_ssize_t start)
 {
 	Py_ssize_t i, len;
 
 	len = self->length + self->new_length;
 	i = start - self->length;
 	if (i < 0)
 		return;
 
 	for (i = start; i < len; i++) {
 		const char *node = index_node(self, i);
 		nt_delete_node(&self->nt, node);
 	}
 
 	self->new_length = start - self->length;
 }
 
 /*
  * Delete a numeric range of revs, which must be at the end of the
  * range.
  */
 static int index_slice_del(indexObject *self, PyObject *item)
 {
 	Py_ssize_t start, stop, step, slicelength;
 	Py_ssize_t length = index_length(self) + 1;
 	int ret = 0;
 
 	if (PySlice_GetIndicesEx(item, length, &start, &stop, &step,
 	                         &slicelength) < 0)
 		return -1;
 
 	if (slicelength <= 0)
 		return 0;
 
 	if ((step < 0 && start < stop) || (step > 0 && start > stop))
 		stop = start;
 
 	if (step < 0) {
 		stop = start + 1;
 		start = stop + step * (slicelength - 1) - 1;
 		step = -step;
 	}
 
 	if (step != 1) {
 		PyErr_SetString(PyExc_ValueError,
 		                "revlog index delete requires step size of 1");
 		return -1;
 	}
 
 	if (stop != length - 1) {
 		PyErr_SetString(PyExc_IndexError,
 		                "revlog index deletion indices are invalid");
 		return -1;
 	}
 
 	if (start < self->length) {
 		if (self->ntinitialized) {
 			Py_ssize_t i;
 
 			for (i = start; i < self->length; i++) {
 				const char *node = index_node_existing(self, i);
 				if (node == NULL)
 					return -1;
 
 				nt_delete_node(&self->nt, node);
 			}
 			if (self->new_length)
 				index_invalidate_added(self, self->length);
 			if (self->ntrev > start)
 				self->ntrev = (int)start;
 		} else if (self->new_length) {
 			self->new_length = 0;
 		}
 
 		self->length = start;
 		goto done;
 	}
 
 	if (self->ntinitialized) {
 		index_invalidate_added(self, start);
 		if (self->ntrev > start)
 			self->ntrev = (int)start;
 	} else {
 		self->new_length = start - self->length;
 	}
 done:
 	Py_CLEAR(self->headrevs);
 	return ret;
 }
 
 /*
  * Supported ops:
  *
  * slice deletion
  * string assignment (extend node->rev mapping)
  * string deletion (shrink node->rev mapping)
  */
 static int index_assign_subscript(indexObject *self, PyObject *item,
                                   PyObject *value)
 {
 	char *node;
 	long rev;
 
 	if (PySlice_Check(item) && value == NULL)
 		return index_slice_del(self, item);
 
 	if (node_check(self->nodelen, item, &node) == -1)
 		return -1;
 
 	if (value == NULL)
 		return self->ntinitialized ? nt_delete_node(&self->nt, node)
 		                           : 0;
 	rev = PyLong_AsLong(value);
 	if (rev > INT_MAX || rev < 0) {
 		if (!PyErr_Occurred())
 			PyErr_SetString(PyExc_ValueError, "rev out of range");
 		return -1;
 	}
 
 	if (index_init_nt(self) == -1)
 		return -1;
 	return nt_insert(&self->nt, node, (int)rev);
 }
 
 /*
  * Find all RevlogNG entries in an index that has inline data. Update
  * the optional "offsets" table with those entries.
  */
 static Py_ssize_t inline_scan(indexObject *self, const char **offsets)
 {
 	const char *data = (const char *)self->buf.buf;
 	Py_ssize_t pos = 0;
 	Py_ssize_t end = self->buf.len;
 	long incr = self->entry_size;
 	Py_ssize_t len = 0;
 
 	while (pos + self->entry_size <= end && pos >= 0) {
 		uint32_t comp_len, sidedata_comp_len = 0;
 		/* 3rd element of header is length of compressed inline data */
 		if (self->format_version == format_v1) {
 			comp_len =
 			    getbe32(data + pos + entry_v1_offset_comp_len);
 			sidedata_comp_len = 0;
 		} else if (self->format_version == format_v2) {
 			comp_len =
 			    getbe32(data + pos + entry_v2_offset_comp_len);
 			sidedata_comp_len = getbe32(
 			    data + pos + entry_v2_offset_sidedata_comp_len);
 		} else {
 			raise_revlog_error();
 			return -1;
 		}
 		incr = self->entry_size + comp_len + sidedata_comp_len;
 		if (offsets)
 			offsets[len] = data + pos;
 		len++;
 		pos += incr;
 	}
 
 	if (pos != end) {
 		if (!PyErr_Occurred())
 			PyErr_SetString(PyExc_ValueError, "corrupt index file");
 		return -1;
 	}
 
 	return len;
 }
 
 static int index_init(indexObject *self, PyObject *args, PyObject *kwargs)
 {
 	PyObject *data_obj, *inlined_obj;
 	Py_ssize_t size;
 
 	static char *kwlist[] = {"data", "inlined", "format", NULL};
 
 	/* Initialize before argument-checking to avoid index_dealloc() crash.
 	 */
 	self->added = NULL;
 	self->new_length = 0;
 	self->added_length = 0;
 	self->data = NULL;
 	memset(&self->buf, 0, sizeof(self->buf));
 	self->headrevs = NULL;
 	self->filteredrevs = Py_None;
 	Py_INCREF(Py_None);
 	self->ntinitialized = 0;
 	self->offsets = NULL;
 	self->nodelen = 20;
 	self->nullentry = NULL;
 	self->rust_ext_compat = 1;
 	self->format_version = format_v1;
 
 	if (!PyArg_ParseTupleAndKeywords(args, kwargs, "OO|l", kwlist,
 	                                 &data_obj, &inlined_obj,
 	                                 &(self->format_version)))
 		return -1;
 	if (!PyObject_CheckBuffer(data_obj)) {
 		PyErr_SetString(PyExc_TypeError,
 		                "data does not support buffer interface");
 		return -1;
 	}
 	if (self->nodelen < 20 || self->nodelen > (Py_ssize_t)sizeof(nullid)) {
 		PyErr_SetString(PyExc_RuntimeError, "unsupported node size");
 		return -1;
 	}
 
 	if (self->format_version == format_v1) {
 		self->entry_size = v1_entry_size;
 	} else if (self->format_version == format_v2) {
 		self->entry_size = v2_entry_size;
 	} else if (self->format_version == format_cl2) {
 		self->entry_size = cl2_entry_size;
 	}
 
-	self->nullentry =
-	    Py_BuildValue(PY23("iiiiiiis#iiBBi", "iiiiiiiy#iiBBi"), 0, 0, 0, -1,
-	                  -1, -1, -1, nullid, self->nodelen, 0, 0,
-	                  comp_mode_inline, comp_mode_inline, rank_unknown);
+	self->nullentry = Py_BuildValue(
+	    "iiiiiiiy#iiBBi", 0, 0, 0, -1, -1, -1, -1, nullid, self->nodelen, 0,
+	    0, comp_mode_inline, comp_mode_inline, rank_unknown);
 
 	if (!self->nullentry)
 		return -1;
 	PyObject_GC_UnTrack(self->nullentry);
 
 	if (PyObject_GetBuffer(data_obj, &self->buf, PyBUF_SIMPLE) == -1)
 		return -1;
 	size = self->buf.len;
 
 	self->inlined = inlined_obj && PyObject_IsTrue(inlined_obj);
 	self->data = data_obj;
 
 	self->ntlookups = self->ntmisses = 0;
 	self->ntrev = -1;
 	Py_INCREF(self->data);
 
 	if (self->inlined) {
 		Py_ssize_t len = inline_scan(self, NULL);
 		if (len == -1)
 			goto bail;
 		self->length = len;
 	} else {
 		if (size % self->entry_size) {
 			PyErr_SetString(PyExc_ValueError, "corrupt index file");
 			goto bail;
 		}
 		self->length = size / self->entry_size;
 	}
 
 	return 0;
 bail:
 	return -1;
 }
 
 static PyObject *index_nodemap(indexObject *self)
 {
 	Py_INCREF(self);
 	return (PyObject *)self;
 }
 
 static void _index_clearcaches(indexObject *self)
 {
 	if (self->offsets) {
 		PyMem_Free((void *)self->offsets);
 		self->offsets = NULL;
 	}
 	if (self->ntinitialized) {
 		nt_dealloc(&self->nt);
 	}
 	self->ntinitialized = 0;
 	Py_CLEAR(self->headrevs);
 }
 
 static PyObject *index_clearcaches(indexObject *self)
 {
 	_index_clearcaches(self);
 	self->ntrev = -1;
 	self->ntlookups = self->ntmisses = 0;
 	Py_RETURN_NONE;
 }
 
 static void index_dealloc(indexObject *self)
 {
 	_index_clearcaches(self);
 	Py_XDECREF(self->filteredrevs);
 	if (self->buf.buf) {
 		PyBuffer_Release(&self->buf);
 		memset(&self->buf, 0, sizeof(self->buf));
 	}
 	Py_XDECREF(self->data);
 	PyMem_Free(self->added);
 	Py_XDECREF(self->nullentry);
 	PyObject_Del(self);
 }
 
 static PySequenceMethods index_sequence_methods = {
     (lenfunc)index_length,      /* sq_length */
     0,                          /* sq_concat */
     0,                          /* sq_repeat */
     (ssizeargfunc)index_get,    /* sq_item */
     0,                          /* sq_slice */
     0,                          /* sq_ass_item */
     0,                          /* sq_ass_slice */
     (objobjproc)index_contains, /* sq_contains */
 };
 
 static PyMappingMethods index_mapping_methods = {
     (lenfunc)index_length,                 /* mp_length */
     (binaryfunc)index_getitem,             /* mp_subscript */
     (objobjargproc)index_assign_subscript, /* mp_ass_subscript */
 };
 
 static PyMethodDef index_methods[] = {
     {"ancestors", (PyCFunction)index_ancestors, METH_VARARGS,
      "return the gca set of the given revs"},
     {"commonancestorsheads", (PyCFunction)index_commonancestorsheads,
      METH_VARARGS,
      "return the heads of the common ancestors of the given revs"},
     {"clearcaches", (PyCFunction)index_clearcaches, METH_NOARGS,
      "clear the index caches"},
     {"get", (PyCFunction)index_m_get, METH_VARARGS, "get an index entry"},
     {"get_rev", (PyCFunction)index_m_get, METH_VARARGS,
      "return `rev` associated with a node or None"},
     {"has_node", (PyCFunction)index_m_has_node, METH_O,
      "return True if the node exist in the index"},
     {"rev", (PyCFunction)index_m_rev, METH_O,
      "return `rev` associated with a node or raise RevlogError"},
     {"computephasesmapsets", (PyCFunction)compute_phases_map_sets, METH_VARARGS,
      "compute phases"},
     {"reachableroots2", (PyCFunction)reachableroots2, METH_VARARGS,
      "reachableroots"},
     {"replace_sidedata_info", (PyCFunction)index_replace_sidedata_info,
      METH_VARARGS, "replace an existing index entry with a new value"},
     {"headrevs", (PyCFunction)index_headrevs, METH_VARARGS,
      "get head revisions"}, /* Can do filtering since 3.2 */
     {"headrevsfiltered", (PyCFunction)index_headrevs, METH_VARARGS,
      "get filtered head revisions"}, /* Can always do filtering */
     {"issnapshot", (PyCFunction)index_issnapshot, METH_O,
      "True if the object is a snapshot"},
     {"findsnapshots", (PyCFunction)index_findsnapshots, METH_VARARGS,
      "Gather snapshot data in a cache dict"},
     {"deltachain", (PyCFunction)index_deltachain, METH_VARARGS,
      "determine revisions with deltas to reconstruct fulltext"},
     {"slicechunktodensity", (PyCFunction)index_slicechunktodensity,
      METH_VARARGS, "determine revisions with deltas to reconstruct fulltext"},
     {"append", (PyCFunction)index_append, METH_O, "append an index entry"},
     {"partialmatch", (PyCFunction)index_partialmatch, METH_VARARGS,
      "match a potentially ambiguous node ID"},
     {"shortest", (PyCFunction)index_shortest, METH_VARARGS,
      "find length of shortest hex nodeid of a binary ID"},
     {"stats", (PyCFunction)index_stats, METH_NOARGS, "stats for the index"},
     {"entry_binary", (PyCFunction)index_entry_binary, METH_O,
      "return an entry in binary form"},
     {"pack_header", (PyCFunction)index_pack_header, METH_VARARGS,
      "pack the revlog header information into binary"},
     {NULL} /* Sentinel */
 };
 
 static PyGetSetDef index_getset[] = {
     {"nodemap", (getter)index_nodemap, NULL, "nodemap", NULL},
     {NULL} /* Sentinel */
 };
 
 static PyMemberDef index_members[] = {
     {"entry_size", T_LONG, offsetof(indexObject, entry_size), 0,
      "size of an index entry"},
     {"rust_ext_compat", T_LONG, offsetof(indexObject, rust_ext_compat), 0,
      "size of an index entry"},
     {NULL} /* Sentinel */
 };
 
 PyTypeObject HgRevlogIndex_Type = {
     PyVarObject_HEAD_INIT(NULL, 0) /* header */
     "parsers.index",               /* tp_name */
     sizeof(indexObject),           /* tp_basicsize */
     0,                             /* tp_itemsize */
     (destructor)index_dealloc,     /* tp_dealloc */
     0,                             /* tp_print */
     0,                             /* tp_getattr */
     0,                             /* tp_setattr */
     0,                             /* tp_compare */
     0,                             /* tp_repr */
     0,                             /* tp_as_number */
     &index_sequence_methods,       /* tp_as_sequence */
     &index_mapping_methods,        /* tp_as_mapping */
     0,                             /* tp_hash */
     0,                             /* tp_call */
     0,                             /* tp_str */
     0,                             /* tp_getattro */
     0,                             /* tp_setattro */
     0,                             /* tp_as_buffer */
     Py_TPFLAGS_DEFAULT,            /* tp_flags */
     "revlog index",                /* tp_doc */
     0,                             /* tp_traverse */
     0,                             /* tp_clear */
     0,                             /* tp_richcompare */
     0,                             /* tp_weaklistoffset */
     0,                             /* tp_iter */
     0,                             /* tp_iternext */
     index_methods,                 /* tp_methods */
     index_members,                 /* tp_members */
     index_getset,                  /* tp_getset */
     0,                             /* tp_base */
     0,                             /* tp_dict */
     0,                             /* tp_descr_get */
     0,                             /* tp_descr_set */
     0,                             /* tp_dictoffset */
     (initproc)index_init,          /* tp_init */
     0,                             /* tp_alloc */
 };
 
 /*
  * returns a tuple of the form (index, cache) with elements as
  * follows:
  *
  * index: an index object that lazily parses Revlog (v1 or v2) records
  * cache: if data is inlined, a tuple (0, index_file_content), else None
  *        index_file_content could be a string, or a buffer
  *
  * added complications are for backwards compatibility
  */
 PyObject *parse_index2(PyObject *self, PyObject *args, PyObject *kwargs)
 {
 	PyObject *cache = NULL;
 	indexObject *idx;
 	int ret;
 
 	idx = PyObject_New(indexObject, &HgRevlogIndex_Type);
 	if (idx == NULL)
 		goto bail;
 
 	ret = index_init(idx, args, kwargs);
 	if (ret == -1)
 		goto bail;
 
 	if (idx->inlined) {
 		cache = Py_BuildValue("iO", 0, idx->data);
 		if (cache == NULL)
 			goto bail;
 	} else {
 		cache = Py_None;
 		Py_INCREF(cache);
 	}
 
 	return Py_BuildValue("NN", idx, cache);
 
 bail:
 	Py_XDECREF(idx);
 	Py_XDECREF(cache);
 	return NULL;
 }
 
 static Revlog_CAPI CAPI = {
     /* increment the abi_version field upon each change in the Revlog_CAPI
        struct or in the ABI of the listed functions */
     2,
     index_length,
     index_node,
     HgRevlogIndex_GetParents,
 };
 
 void revlog_module_init(PyObject *mod)
 {
 	PyObject *caps = NULL;
 	HgRevlogIndex_Type.tp_new = PyType_GenericNew;
 	if (PyType_Ready(&HgRevlogIndex_Type) < 0)
 		return;
 	Py_INCREF(&HgRevlogIndex_Type);
 	PyModule_AddObject(mod, "index", (PyObject *)&HgRevlogIndex_Type);
 
 	nodetreeType.tp_new = PyType_GenericNew;
 	if (PyType_Ready(&nodetreeType) < 0)
 		return;
 	Py_INCREF(&nodetreeType);
 	PyModule_AddObject(mod, "nodetree", (PyObject *)&nodetreeType);
 
 	caps = PyCapsule_New(&CAPI, "mercurial.cext.parsers.revlog_CAPI", NULL);
 	if (caps != NULL)
 		PyModule_AddObject(mod, "revlog_CAPI", caps);
 }

mercurial/cext/util.h

@@ -1,92 +1,85 b''
 /*
  util.h - utility functions for interfacing with the various python APIs.
 
  This software may be used and distributed according to the terms of
  the GNU General Public License, incorporated herein by reference.
 */
 
 #ifndef _HG_UTIL_H_
 #define _HG_UTIL_H_
 
 #include "compat.h"
 
 #if PY_MAJOR_VERSION >= 3
 #define IS_PY3K
 #endif
 
-/* helper to switch things like string literal depending on Python version */
-#ifdef IS_PY3K
-#define PY23(py2, py3) py3
-#else
-#define PY23(py2, py3) py2
-#endif
-
 /* clang-format off */
 typedef struct {
 	PyObject_HEAD
 	int flags;
 	int mode;
 	int size;
 	int mtime_s;
 	int mtime_ns;
 } dirstateItemObject;
 /* clang-format on */
 
 static const int dirstate_flag_wc_tracked = 1 << 0;
 static const int dirstate_flag_p1_tracked = 1 << 1;
 static const int dirstate_flag_p2_info = 1 << 2;
 static const int dirstate_flag_mode_exec_perm = 1 << 3;
 static const int dirstate_flag_mode_is_symlink = 1 << 4;
 static const int dirstate_flag_has_fallback_exec = 1 << 5;
 static const int dirstate_flag_fallback_exec = 1 << 6;
 static const int dirstate_flag_has_fallback_symlink = 1 << 7;
 static const int dirstate_flag_fallback_symlink = 1 << 8;
 static const int dirstate_flag_expected_state_is_modified = 1 << 9;
 static const int dirstate_flag_has_meaningful_data = 1 << 10;
 static const int dirstate_flag_has_mtime = 1 << 11;
 static const int dirstate_flag_mtime_second_ambiguous = 1 << 12;
 static const int dirstate_flag_directory = 1 << 13;
 static const int dirstate_flag_all_unknown_recorded = 1 << 14;
 static const int dirstate_flag_all_ignored_recorded = 1 << 15;
 
 extern PyTypeObject dirstateItemType;
 #define dirstate_tuple_check(op) (Py_TYPE(op) == &dirstateItemType)
 
 #ifndef MIN
 #define MIN(a, b) (((a) < (b)) ? (a) : (b))
 #endif
 /* VC9 doesn't include bool and lacks stdbool.h based on my searching */
 #if defined(_MSC_VER) || __STDC_VERSION__ < 199901L
 #define true 1
 #define false 0
 typedef unsigned char bool;
 #else
 #include <stdbool.h>
 #endif
 
 static inline PyObject *_dict_new_presized(Py_ssize_t expected_size)
 {
 	/* _PyDict_NewPresized expects a minused parameter, but it actually
 	   creates a dictionary that's the nearest power of two bigger than the
 	   parameter. For example, with the initial minused = 1000, the
 	   dictionary created has size 1024. Of course in a lot of cases that
 	   can be greater than the maximum load factor Python's dict object
 	   expects (= 2/3), so as soon as we cross the threshold we'll resize
 	   anyway. So create a dictionary that's at least 3/2 the size. */
 	return _PyDict_NewPresized(((1 + expected_size) / 2) * 3);
 }
 
 /* Convert a PyInt or PyLong to a long. Returns false if there is an
    error, in which case an exception will already have been set. */
 static inline bool pylong_to_long(PyObject *pylong, long *out)
 {
 	*out = PyLong_AsLong(pylong);
 	/* Fast path to avoid hitting PyErr_Occurred if the value was obviously
 	 * not an error. */
 	if (*out != -1) {
 		return true;
 	}
 	return PyErr_Occurred() == NULL;
 }
 
 #endif /* _HG_UTIL_H_ */