phases: drop the list with phase of each rev, always comput phase sets

Change the C implementation of phasecache.loadphaserevs to provide only
the sets for draft and secret phase as well as the number of revisions
seen.

Change the pure Python implementation of the same functino to compute
the sets instead of the list of phases for each revision.

Change phasecache.phase to check the phase sets and assume public if the
revision is in neither draft nor secret set. This is computationally
slightly more expensive.

Change phasecache.getrevset for public() based queries to compute the
set of non-matching revisions and return the result as filtered
fullreposet. A shortcut is taken when no draft or secret revision
exists.

Bump the module version for the changed interface contract.

Overall, this saves around 16 Bytes per revision whenever the phasecache
is used, for the test case in issue5691 it is around 3MB. getrevset()
for a large repository is around 13% slower here, that seems an
acceptable trade off. Performance impact for phase() should be similar.

Differential Revision: https://phab.mercurial-scm.org/D1606

mercurial/cext/parsers.c

@@ -1,794 +1,794
 /*
  parsers.c - efficient content parsing
 
  Copyright 2008 Matt 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.
 */
 
 #include <Python.h>
 #include <ctype.h>
 #include <stddef.h>
 #include <string.h>
 
 #include "bitmanipulation.h"
 #include "charencode.h"
 #include "util.h"
 
 #ifdef IS_PY3K
 /* The mapping of Python types is meant to be temporary to get Python
  * 3 to compile. We should remove this once Python 3 support is fully
  * supported and proper types are used in the extensions themselves. */
 #define PyInt_Check PyLong_Check
 #define PyInt_FromLong PyLong_FromLong
 #define PyInt_FromSsize_t PyLong_FromSsize_t
 #define PyInt_AsLong PyLong_AsLong
 #endif
 
 static const char *const versionerrortext = "Python minor version mismatch";
 
 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);
 }
 
 /*
  * This code assumes that a manifest is stitched together with newline
  * ('\n') characters.
  */
 static PyObject *parse_manifest(PyObject *self, PyObject *args)
 {
 	PyObject *mfdict, *fdict;
 	char *str, *start, *end;
 	int len;
 
 	if (!PyArg_ParseTuple(args, "O!O!s#:parse_manifest", &PyDict_Type,
 	                      &mfdict, &PyDict_Type, &fdict, &str, &len))
 		goto quit;
 
 	start = str;
 	end = str + len;
 	while (start < end) {
 		PyObject *file = NULL, *node = NULL;
 		PyObject *flags = NULL;
 		char *zero = NULL, *newline = NULL;
 		ptrdiff_t nlen;
 
 		zero = memchr(start, '\0', end - start);
 		if (!zero) {
 			PyErr_SetString(PyExc_ValueError,
 			                "manifest entry has no separator");
 			goto quit;
 		}
 
 		newline = memchr(zero + 1, '\n', end - (zero + 1));
 		if (!newline) {
 			PyErr_SetString(PyExc_ValueError,
 			                "manifest contains trailing garbage");
 			goto quit;
 		}
 
 		file = PyBytes_FromStringAndSize(start, zero - start);
 
 		if (!file)
 			goto bail;
 
 		nlen = newline - zero - 1;
 
 		node = unhexlify(zero + 1, nlen > 40 ? 40 : (Py_ssize_t)nlen);
 		if (!node)
 			goto bail;
 
 		if (nlen > 40) {
 			flags = PyBytes_FromStringAndSize(zero + 41, nlen - 40);
 			if (!flags)
 				goto bail;
 
 			if (PyDict_SetItem(fdict, file, flags) == -1)
 				goto bail;
 		}
 
 		if (PyDict_SetItem(mfdict, file, node) == -1)
 			goto bail;
 
 		start = newline + 1;
 
 		Py_XDECREF(flags);
 		Py_XDECREF(node);
 		Py_XDECREF(file);
 		continue;
 	bail:
 		Py_XDECREF(flags);
 		Py_XDECREF(node);
 		Py_XDECREF(file);
 		goto quit;
 	}
 
 	Py_INCREF(Py_None);
 	return Py_None;
 quit:
 	return NULL;
 }
 
 static inline dirstateTupleObject *make_dirstate_tuple(char state, int mode,
                                                        int size, int mtime)
 {
 	dirstateTupleObject *t =
 	    PyObject_New(dirstateTupleObject, &dirstateTupleType);
 	if (!t)
 		return NULL;
 	t->state = state;
 	t->mode = mode;
 	t->size = size;
 	t->mtime = mtime;
 	return t;
 }
 
 static PyObject *dirstate_tuple_new(PyTypeObject *subtype, PyObject *args,
                                     PyObject *kwds)
 {
 	/* We do all the initialization here and not a tp_init function because
 	 * dirstate_tuple is immutable. */
 	dirstateTupleObject *t;
 	char state;
 	int size, mode, mtime;
 	if (!PyArg_ParseTuple(args, "ciii", &state, &mode, &size, &mtime))
 		return NULL;
 
 	t = (dirstateTupleObject *)subtype->tp_alloc(subtype, 1);
 	if (!t)
 		return NULL;
 	t->state = state;
 	t->mode = mode;
 	t->size = size;
 	t->mtime = mtime;
 
 	return (PyObject *)t;
 }
 
 static void dirstate_tuple_dealloc(PyObject *o)
 {
 	PyObject_Del(o);
 }
 
 static Py_ssize_t dirstate_tuple_length(PyObject *o)
 {
 	return 4;
 }
 
 static PyObject *dirstate_tuple_item(PyObject *o, Py_ssize_t i)
 {
 	dirstateTupleObject *t = (dirstateTupleObject *)o;
 	switch (i) {
 	case 0:
 		return PyBytes_FromStringAndSize(&t->state, 1);
 	case 1:
 		return PyInt_FromLong(t->mode);
 	case 2:
 		return PyInt_FromLong(t->size);
 	case 3:
 		return PyInt_FromLong(t->mtime);
 	default:
 		PyErr_SetString(PyExc_IndexError, "index out of range");
 		return NULL;
 	}
 }
 
 static PySequenceMethods dirstate_tuple_sq = {
     dirstate_tuple_length, /* sq_length */
     0,                     /* sq_concat */
     0,                     /* sq_repeat */
     dirstate_tuple_item,   /* sq_item */
     0,                     /* sq_ass_item */
     0,                     /* sq_contains */
     0,                     /* sq_inplace_concat */
     0                      /* sq_inplace_repeat */
 };
 
 PyTypeObject dirstateTupleType = {
     PyVarObject_HEAD_INIT(NULL, 0)      /* header */
     "dirstate_tuple",                   /* tp_name */
     sizeof(dirstateTupleObject),        /* tp_basicsize */
     0,                                  /* tp_itemsize */
     (destructor)dirstate_tuple_dealloc, /* tp_dealloc */
     0,                                  /* tp_print */
     0,                                  /* tp_getattr */
     0,                                  /* tp_setattr */
     0,                                  /* tp_compare */
     0,                                  /* tp_repr */
     0,                                  /* tp_as_number */
     &dirstate_tuple_sq,                 /* 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 */
     0,                                  /* 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 */
     0,                                  /* tp_init */
     0,                                  /* tp_alloc */
     dirstate_tuple_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, len, pos = 40;
 	int readlen;
 
 	if (!PyArg_ParseTuple(args, "O!O!s#: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("s#s#", str, 20, str + 20, 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 *)make_dirstate_tuple(state, mode, size, mtime);
 		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;
 }
 
 /*
  * Build a set of non-normal and other parent entries from the dirstate dmap
  */
 static PyObject *nonnormalotherparententries(PyObject *self, PyObject *args)
 {
 	PyObject *dmap, *fname, *v;
 	PyObject *nonnset = NULL, *otherpset = NULL, *result = NULL;
 	Py_ssize_t pos;
 
 	if (!PyArg_ParseTuple(args, "O!:nonnormalentries", &PyDict_Type, &dmap))
 		goto bail;
 
 	nonnset = PySet_New(NULL);
 	if (nonnset == NULL)
 		goto bail;
 
 	otherpset = PySet_New(NULL);
 	if (otherpset == NULL)
 		goto bail;
 
 	pos = 0;
 	while (PyDict_Next(dmap, &pos, &fname, &v)) {
 		dirstateTupleObject *t;
 		if (!dirstate_tuple_check(v)) {
 			PyErr_SetString(PyExc_TypeError,
 			                "expected a dirstate tuple");
 			goto bail;
 		}
 		t = (dirstateTupleObject *)v;
 
 		if (t->state == 'n' && t->size == -2) {
 			if (PySet_Add(otherpset, fname) == -1) {
 				goto bail;
 			}
 		}
 
 		if (t->state == 'n' && t->mtime != -1)
 			continue;
 		if (PySet_Add(nonnset, fname) == -1)
 			goto bail;
 	}
 
 	result = Py_BuildValue("(OO)", nonnset, otherpset);
 	if (result == NULL)
 		goto bail;
 	Py_DECREF(nonnset);
 	Py_DECREF(otherpset);
 	return result;
 bail:
 	Py_XDECREF(nonnset);
 	Py_XDECREF(otherpset);
 	Py_XDECREF(result);
 	return NULL;
 }
 
 /*
  * 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;
 	int now;
 
 	if (!PyArg_ParseTuple(args, "O!O!Oi:pack_dirstate", &PyDict_Type, &map,
 	                      &PyDict_Type, &copymap, &pl, &now))
 		return NULL;
 
 	if (!PySequence_Check(pl) || PySequence_Size(pl) != 2) {
 		PyErr_SetString(PyExc_TypeError, "expected 2-element sequence");
 		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 = PySequence_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 = PySequence_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);) {
 		dirstateTupleObject *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 = (dirstateTupleObject *)v;
 
 		state = tuple->state;
 		mode = tuple->mode;
 		size = tuple->size;
 		mtime = tuple->mtime;
 		if (state == 'n' && mtime == now) {
 			/* See pure/parsers.py:pack_dirstate for why we do
 			 * this. */
 			mtime = -1;
 			mtime_unset = (PyObject *)make_dirstate_tuple(
 			    state, mode, size, mtime);
 			if (!mtime_unset)
 				goto bail;
 			if (PyDict_SetItem(map, k, mtime_unset) == -1)
 				goto bail;
 			Py_DECREF(mtime_unset);
 			mtime_unset = NULL;
 		}
 		*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;
 	int datalen;
 	Py_ssize_t offset, stop;
 	PyObject *markers = NULL;
 
 	if (!PyArg_ParseTuple(args, "s#nn", &data, &datalen, &offset, &stop)) {
 		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);
 
 static PyMethodDef methods[] = {
     {"pack_dirstate", pack_dirstate, METH_VARARGS, "pack a dirstate\n"},
     {"nonnormalotherparententries", nonnormalotherparententries, METH_VARARGS,
      "create a set containing non-normal and other parent entries of given "
      "dirstate\n"},
     {"parse_manifest", parse_manifest, METH_VARARGS, "parse a manifest\n"},
     {"parse_dirstate", parse_dirstate, METH_VARARGS, "parse a dirstate\n"},
     {"parse_index2", parse_index2, METH_VARARGS, "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 = 3;
+static const int version = 4;
 
 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(&dirstateTupleType) < 0)
 		return;
 	Py_INCREF(&dirstateTupleType);
 	PyModule_AddObject(mod, "dirstatetuple",
 	                   (PyObject *)&dirstateTupleType);
 }
 
 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 = PyInt_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;
 }
 
 #ifdef IS_PY3K
 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;
 }
 #else
 PyMODINIT_FUNC initparsers(void)
 {
 	PyObject *mod;
 
 	if (check_python_version() == -1)
 		return;
 	mod = Py_InitModule3("parsers", methods, parsers_doc);
 	module_init(mod);
 }
 #endif

mercurial/cext/revlog.c

@@ -1,2090 +1,2084
 /*
  parsers.c - efficient content parsing
 
  Copyright 2008 Matt 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.
 */
 
 #include <Python.h>
 #include <assert.h>
 #include <ctype.h>
 #include <stddef.h>
 #include <string.h>
 
 #include "bitmanipulation.h"
 #include "charencode.h"
 #include "util.h"
 
 #ifdef IS_PY3K
 /* The mapping of Python types is meant to be temporary to get Python
  * 3 to compile. We should remove this once Python 3 support is fully
  * supported and proper types are used in the extensions themselves. */
 #define PyInt_Check PyLong_Check
 #define PyInt_FromLong PyLong_FromLong
 #define PyInt_FromSsize_t PyLong_FromSsize_t
 #define PyInt_AS_LONG PyLong_AS_LONG
 #define PyInt_AsLong PyLong_AsLong
 #endif
 
 /*
  * 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 + 1)
  * Zero is empty
  */
 typedef struct {
 	int children[16];
 } nodetree;
 
 /*
  * 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. Our last entry is a
  * sentinel, always a nullid.  We have limited support for
  * integer-keyed insert and delete, only at elements right before the
  * sentinel.
  *
  * With string keys, we lazily perform a reverse mapping from node to
  * rev, using a base-16 trie.
  */
 typedef struct {
 	PyObject_HEAD
 	/* Type-specific fields go here. */
 	PyObject *data;        /* raw bytes of index */
 	Py_buffer buf;         /* buffer of data */
 	PyObject **cache;      /* cached tuples */
 	const char **offsets;  /* populated on demand */
 	Py_ssize_t raw_length; /* original number of elements */
 	Py_ssize_t length;     /* current number of elements */
 	PyObject *added;       /* populated on demand */
 	PyObject *headrevs;    /* cache, invalidated on changes */
 	PyObject *filteredrevs;/* filtered revs set */
 	nodetree *nt;          /* base-16 trie */
 	unsigned ntlength;          /* # nodes in use */
 	unsigned ntcapacity;        /* # nodes allocated */
 	int ntdepth;           /* maximum depth of tree */
 	int ntsplits;          /* # splits performed */
 	int ntrev;             /* last rev scanned */
 	int ntlookups;         /* # lookups */
 	int ntmisses;          /* # lookups that miss the cache */
 	int inlined;
 } indexObject;
 
 static Py_ssize_t index_length(const indexObject *self)
 {
 	if (self->added == NULL)
 		return self->length;
 	return self->length + PyList_GET_SIZE(self->added);
 }
 
 static PyObject *nullentry;
 static const char nullid[20];
 
 static Py_ssize_t inline_scan(indexObject *self, const char **offsets);
 
 #if LONG_MAX == 0x7fffffffL
 static char *tuple_format = "Kiiiiiis#";
 #else
 static char *tuple_format = "kiiiiiis#";
 #endif
 
 /* A RevlogNG v1 index entry is 64 bytes long. */
 static const long v1_hdrsize = 64;
 
 /*
  * Return a pointer to the beginning of a RevlogNG record.
  */
 static const char *index_deref(indexObject *self, Py_ssize_t pos)
 {
 	if (self->inlined && pos > 0) {
 		if (self->offsets == NULL) {
 			self->offsets = PyMem_Malloc(self->raw_length *
 					             sizeof(*self->offsets));
 			if (self->offsets == NULL)
 				return (const char *)PyErr_NoMemory();
 			inline_scan(self, self->offsets);
 		}
 		return self->offsets[pos];
 	}
 
 	return (const char *)(self->buf.buf) + pos * v1_hdrsize;
 }
 
 static inline int index_get_parents(indexObject *self, Py_ssize_t rev,
 				    int *ps, int maxrev)
 {
 	if (rev >= self->length - 1) {
 		PyObject *tuple = PyList_GET_ITEM(self->added,
 						  rev - self->length + 1);
 		ps[0] = (int)PyInt_AS_LONG(PyTuple_GET_ITEM(tuple, 5));
 		ps[1] = (int)PyInt_AS_LONG(PyTuple_GET_ITEM(tuple, 6));
 	} else {
 		const char *data = index_deref(self, rev);
 		ps[0] = getbe32(data + 24);
 		ps[1] = getbe32(data + 28);
 	}
 	/* 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] > maxrev || ps[1] > maxrev) {
 		PyErr_SetString(PyExc_ValueError, "parent out of range");
 		return -1;
 	}
 	return 0;
 }
 
 
 /*
  * 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)
  */
 static PyObject *index_get(indexObject *self, Py_ssize_t pos)
 {
 	uint64_t offset_flags;
 	int comp_len, uncomp_len, base_rev, link_rev, parent_1, parent_2;
 	const char *c_node_id;
 	const char *data;
 	Py_ssize_t length = index_length(self);
 	PyObject *entry;
 
 	if (pos < 0)
 		pos += length;
 
 	if (pos < 0 || pos >= length) {
 		PyErr_SetString(PyExc_IndexError, "revlog index out of range");
 		return NULL;
 	}
 
 	if (pos == length - 1) {
 		Py_INCREF(nullentry);
 		return nullentry;
 	}
 
 	if (pos >= self->length - 1) {
 		PyObject *obj;
 		obj = PyList_GET_ITEM(self->added, pos - self->length + 1);
 		Py_INCREF(obj);
 		return obj;
 	}
 
 	if (self->cache) {
 		if (self->cache[pos]) {
 			Py_INCREF(self->cache[pos]);
 			return self->cache[pos];
 		}
 	} else {
 		self->cache = calloc(self->raw_length, sizeof(PyObject *));
 		if (self->cache == NULL)
 			return PyErr_NoMemory();
 	}
 
 	data = index_deref(self, pos);
 	if (data == NULL)
 		return NULL;
 
 	offset_flags = getbe32(data + 4);
 	if (pos == 0) /* mask out version number for the first entry */
 		offset_flags &= 0xFFFF;
 	else {
 		uint32_t offset_high = getbe32(data);
 		offset_flags |= ((uint64_t)offset_high) << 32;
 	}
 
 	comp_len = getbe32(data + 8);
 	uncomp_len = getbe32(data + 12);
 	base_rev = getbe32(data + 16);
 	link_rev = getbe32(data + 20);
 	parent_1 = getbe32(data + 24);
 	parent_2 = getbe32(data + 28);
 	c_node_id = data + 32;
 
 	entry = Py_BuildValue(tuple_format, offset_flags, comp_len,
 			      uncomp_len, base_rev, link_rev,
 			      parent_1, parent_2, c_node_id, 20);
 
 	if (entry) {
 		PyObject_GC_UnTrack(entry);
 		Py_INCREF(entry);
 	}
 
 	self->cache[pos] = entry;
 
 	return entry;
 }
 
 /*
  * Return the 20-byte SHA of the 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;
 
 	if (pos == length - 1 || pos == INT_MAX)
 		return nullid;
 
 	if (pos >= length)
 		return NULL;
 
 	if (pos >= self->length - 1) {
 		PyObject *tuple, *str;
 		tuple = PyList_GET_ITEM(self->added, pos - self->length + 1);
 		str = PyTuple_GetItem(tuple, 7);
 		return str ? PyBytes_AS_STRING(str) : NULL;
 	}
 
 	data = index_deref(self, pos);
 	return data ? data + 32 : NULL;
 }
 
 static int nt_insert(indexObject *self, const char *node, int rev);
 
 static int node_check(PyObject *obj, char **node, Py_ssize_t *nodelen)
 {
 	if (PyBytes_AsStringAndSize(obj, node, nodelen) == -1)
 		return -1;
 	if (*nodelen == 20)
 		return 0;
 	PyErr_SetString(PyExc_ValueError, "20-byte hash required");
 	return -1;
 }
 
 static PyObject *index_insert(indexObject *self, PyObject *args)
 {
 	PyObject *obj;
 	char *node;
 	int index;
 	Py_ssize_t len, nodelen;
 
 	if (!PyArg_ParseTuple(args, "iO", &index, &obj))
 		return NULL;
 
 	if (!PyTuple_Check(obj) || PyTuple_GET_SIZE(obj) != 8) {
 		PyErr_SetString(PyExc_TypeError, "8-tuple required");
 		return NULL;
 	}
 
 	if (node_check(PyTuple_GET_ITEM(obj, 7), &node, &nodelen) == -1)
 		return NULL;
 
 	len = index_length(self);
 
 	if (index < 0)
 		index += len;
 
 	if (index != len - 1) {
 		PyErr_SetString(PyExc_IndexError,
 				"insert only supported at index -1");
 		return NULL;
 	}
 
 	if (self->added == NULL) {
 		self->added = PyList_New(0);
 		if (self->added == NULL)
 			return NULL;
 	}
 
 	if (PyList_Append(self->added, obj) == -1)
 		return NULL;
 
 	if (self->nt)
 		nt_insert(self, node, index);
 
 	Py_CLEAR(self->headrevs);
 	Py_RETURN_NONE;
 }
 
 static void _index_clearcaches(indexObject *self)
 {
 	if (self->cache) {
 		Py_ssize_t i;
 
 		for (i = 0; i < self->raw_length; i++)
 			Py_CLEAR(self->cache[i]);
 		free(self->cache);
 		self->cache = NULL;
 	}
 	if (self->offsets) {
 		PyMem_Free(self->offsets);
 		self->offsets = NULL;
 	}
 	if (self->nt) {
 		free(self->nt);
 		self->nt = NULL;
 	}
 	Py_CLEAR(self->headrevs);
 }
 
 static PyObject *index_clearcaches(indexObject *self)
 {
 	_index_clearcaches(self);
 	self->ntlength = self->ntcapacity = 0;
 	self->ntdepth = self->ntsplits = 0;
 	self->ntrev = -1;
 	self->ntlookups = self->ntmisses = 0;
 	Py_RETURN_NONE;
 }
 
 static PyObject *index_stats(indexObject *self)
 {
 	PyObject *obj = PyDict_New();
 	PyObject *t = NULL;
 
 	if (obj == NULL)
 		return NULL;
 
 #define istat(__n, __d) \
 	do { \
 		t = PyInt_FromSsize_t(self->__n); \
 		if (!t) \
 			goto bail; \
 		if (PyDict_SetItemString(obj, __d, t) == -1) \
 			goto bail; \
 		Py_DECREF(t); \
 	} while (0)
 
 	if (self->added) {
 		Py_ssize_t len = PyList_GET_SIZE(self->added);
 		t = PyInt_FromSsize_t(len);
 		if (!t)
 			goto bail;
 		if (PyDict_SetItemString(obj, "index entries added", t) == -1)
 			goto bail;
 		Py_DECREF(t);
 	}
 
 	if (self->raw_length != self->length - 1)
 		istat(raw_length, "revs on disk");
 	istat(length, "revs in memory");
 	istat(ntcapacity, "node trie capacity");
 	istat(ntdepth, "node trie depth");
 	istat(ntlength, "node trie count");
 	istat(ntlookups, "node trie lookups");
 	istat(ntmisses, "node trie misses");
 	istat(ntrev, "node trie last rev scanned");
 	istat(ntsplits, "node trie splits");
 
 #undef istat
 
 	return obj;
 
 bail:
 	Py_XDECREF(obj);
 	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 = PyEval_CallObject(filter, arglist);
 		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 Py_ssize_t add_roots_get_min(indexObject *self, PyObject *list,
                                     Py_ssize_t marker, char *phases)
 {
 	PyObject *iter = NULL;
 	PyObject *iter_item = NULL;
 	Py_ssize_t min_idx = index_length(self) + 1;
 	long iter_item_long;
 
 	if (PyList_GET_SIZE(list) != 0) {
 		iter = PyObject_GetIter(list);
 		if (iter == NULL)
 			return -2;
 		while ((iter_item = PyIter_Next(iter))) {
 			iter_item_long = PyInt_AS_LONG(iter_item);
 			Py_DECREF(iter_item);
 			if (iter_item_long < min_idx)
 				min_idx = iter_item_long;
 			phases[iter_item_long] = marker;
 		}
 		Py_DECREF(iter);
 	}
 
 	return min_idx;
 }
 
 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) - 1;
 	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 = PyInt_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 = PyInt_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 = PyInt_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 == -1)
 			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 = PyInt_FromLong(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 PyObject *compute_phases_map_sets(indexObject *self, PyObject *args)
 {
 	PyObject *roots = Py_None;
 	PyObject *ret = NULL;
-	PyObject *phaseslist = NULL;
+	PyObject *phasessize = NULL;
 	PyObject *phaseroots = NULL;
 	PyObject *phaseset = NULL;
 	PyObject *phasessetlist = NULL;
 	PyObject *rev = NULL;
 	Py_ssize_t len = index_length(self) - 1;
 	Py_ssize_t numphase = 0;
 	Py_ssize_t minrevallphases = 0;
 	Py_ssize_t minrevphase = 0;
 	Py_ssize_t i = 0;
 	char *phases = NULL;
 	long phase;
 
 	if (!PyArg_ParseTuple(args, "O", &roots))
 		goto done;
 	if (roots == NULL || !PyList_Check(roots))
 		goto done;
 
 	phases = calloc(len, 1); /* phase per rev: {0: public, 1: draft, 2: secret} */
 	if (phases == NULL) {
 		PyErr_NoMemory();
 		goto done;
 	}
 	/* Put the phase information of all the roots in phases */
 	numphase = PyList_GET_SIZE(roots)+1;
 	minrevallphases = len + 1;
 	phasessetlist = PyList_New(numphase);
 	if (phasessetlist == NULL)
 		goto done;
 
 	PyList_SET_ITEM(phasessetlist, 0, Py_None);
 	Py_INCREF(Py_None);
 
 	for (i = 0; i < numphase-1; i++) {
 		phaseroots = PyList_GET_ITEM(roots, i);
 		phaseset = PySet_New(NULL);
 		if (phaseset == NULL)
 			goto release;
 		PyList_SET_ITEM(phasessetlist, i+1, phaseset);
 		if (!PyList_Check(phaseroots))
 			goto release;
 		minrevphase = add_roots_get_min(self, phaseroots, i+1, phases);
 		if (minrevphase == -2) /* Error from add_roots_get_min */
 			goto release;
 		minrevallphases = MIN(minrevallphases, minrevphase);
 	}
 	/* Propagate the phase information from the roots to the revs */
 	if (minrevallphases != -1) {
 		int parents[2];
 		for (i = minrevallphases; i < len; i++) {
 			if (index_get_parents(self, i, parents,
 					      (int)len - 1) < 0)
 				goto release;
 			set_phase_from_parents(phases, parents[0], parents[1], i);
 		}
 	}
 	/* Transform phase list to a python list */
-	phaseslist = PyList_New(len);
-	if (phaseslist == NULL)
+	phasessize = PyInt_FromLong(len);
+	if (phasessize == NULL)
 		goto release;
 	for (i = 0; i < len; i++) {
-		PyObject *phaseval;
-
 		phase = phases[i];
 		/* We only store the sets of phase for non public phase, the public phase
 		 * is computed as a difference */
 		if (phase != 0) {
 			phaseset = PyList_GET_ITEM(phasessetlist, phase);
 			rev = PyInt_FromLong(i);
 			if (rev == NULL)
 				goto release;
 			PySet_Add(phaseset, rev);
 			Py_XDECREF(rev);
 		}
-		phaseval = PyInt_FromLong(phase);
-		if (phaseval == NULL)
-			goto release;
-		PyList_SET_ITEM(phaseslist, i, phaseval);
 	}
-	ret = PyTuple_Pack(2, phaseslist, phasessetlist);
+	ret = PyTuple_Pack(2, phasessize, phasessetlist);
 
 release:
-	Py_XDECREF(phaseslist);
+	Py_XDECREF(phasessize);
 	Py_XDECREF(phasessetlist);
 done:
 	free(phases);
 	return ret;
 }
 
 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) - 1;
 	heads = PyList_New(0);
 	if (heads == NULL)
 		goto bail;
 	if (len == 0) {
 		PyObject *nullid = PyInt_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 = PyInt_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;
 
 	if (rev >= self->length - 1) {
 		PyObject *tuple = PyList_GET_ITEM(self->added,
 			rev - self->length + 1);
 		return (int)PyInt_AS_LONG(PyTuple_GET_ITEM(tuple, 3));
 	}
 	else {
 		data = index_deref(self, rev);
 		if (data == NULL) {
 			return -2;
 		}
 
 		return getbe32(data + 16);
 	}
 }
 
 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 (PyInt_Check(stoparg)) {
 		stoprev = (int)PyInt_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 - 1) {
 		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 = PyInt_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 - 1) {
 			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 = PyInt_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 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(indexObject *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 (nodelen == 20 && node[0] == '\0' && memcmp(node, nullid, 20) == 0)
 		return -1;
 
 	if (self->nt == NULL)
 		return -2;
 
 	if (hex)
 		maxlevel = nodelen > 40 ? 40 : (int)nodelen;
 	else
 		maxlevel = nodelen > 20 ? 40 : ((int)nodelen * 2);
 
 	for (level = off = 0; level < maxlevel; level++) {
 		int k = getnybble(node, level);
 		nodetree *n = &self->nt[off];
 		int v = n->children[k];
 
 		if (v < 0) {
 			const char *n;
 			Py_ssize_t i;
 
 			v = -(v + 1);
 			n = index_node(self, 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(indexObject *self)
 {
 	if (self->ntlength == self->ntcapacity) {
 		if (self->ntcapacity >= INT_MAX / (sizeof(nodetree) * 2)) {
 			PyErr_SetString(PyExc_MemoryError,
 					"overflow in nt_new");
 			return -1;
 		}
 		self->ntcapacity *= 2;
 		self->nt = realloc(self->nt,
 				   self->ntcapacity * sizeof(nodetree));
 		if (self->nt == NULL) {
 			PyErr_SetString(PyExc_MemoryError, "out of memory");
 			return -1;
 		}
 		memset(&self->nt[self->ntlength], 0,
 		       sizeof(nodetree) * (self->ntcapacity - self->ntlength));
 	}
 	return self->ntlength++;
 }
 
 static int nt_insert(indexObject *self, const char *node, int rev)
 {
 	int level = 0;
 	int off = 0;
 
 	while (level < 40) {
 		int k = nt_level(node, level);
 		nodetree *n;
 		int v;
 
 		n = &self->nt[off];
 		v = n->children[k];
 
 		if (v == 0) {
 			n->children[k] = -rev - 1;
 			return 0;
 		}
 		if (v < 0) {
 			const char *oldnode = index_node(self, -(v + 1));
 			int noff;
 
 			if (!oldnode || !memcmp(oldnode, node, 20)) {
 				n->children[k] = -rev - 1;
 				return 0;
 			}
 			noff = nt_new(self);
 			if (noff == -1)
 				return -1;
 			/* self->nt may have been changed by realloc */
 			self->nt[off].children[k] = noff;
 			off = noff;
 			n = &self->nt[off];
 			n->children[nt_level(oldnode, ++level)] = v;
 			if (level > self->ntdepth)
 				self->ntdepth = level;
 			self->ntsplits += 1;
 		} else {
 			level += 1;
 			off = v;
 		}
 	}
 
 	return -1;
 }
 
 static int nt_init(indexObject *self)
 {
 	if (self->nt == NULL) {
 		if ((size_t)self->raw_length > INT_MAX / sizeof(nodetree)) {
 			PyErr_SetString(PyExc_ValueError, "overflow in nt_init");
 			return -1;
 		}
 		self->ntcapacity = self->raw_length < 4
 			? 4 : (int)self->raw_length / 2;
 
 		self->nt = calloc(self->ntcapacity, sizeof(nodetree));
 		if (self->nt == NULL) {
 			PyErr_NoMemory();
 			return -1;
 		}
 		self->ntlength = 1;
 		self->ntrev = (int)index_length(self) - 1;
 		self->ntlookups = 1;
 		self->ntmisses = 0;
 		if (nt_insert(self, nullid, INT_MAX) == -1)
 			return -1;
 	}
 	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, Py_ssize_t nodelen)
 {
 	int rev;
 
 	self->ntlookups++;
 	rev = nt_find(self, node, nodelen, 0);
 	if (rev >= -1)
 		return rev;
 
 	if (nt_init(self) == -1)
 		return -3;
 
 	/*
 	 * 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(self, rev);
 			if (n == NULL)
 				return -2;
 			if (memcmp(node, n, nodelen > 20 ? 20 : nodelen) == 0) {
 				if (nt_insert(self, n, rev) == -1)
 					return -3;
 				break;
 			}
 		}
 	} else {
 		for (rev = self->ntrev - 1; rev >= 0; rev--) {
 			const char *n = index_node(self, rev);
 			if (n == NULL) {
 				self->ntrev = rev + 1;
 				return -2;
 			}
 			if (nt_insert(self, n, rev) == -1) {
 				self->ntrev = rev + 1;
 				return -3;
 			}
 			if (memcmp(node, n, nodelen > 20 ? 20 : nodelen) == 0) {
 				break;
 			}
 		}
 		self->ntrev = rev;
 	}
 
 	if (rev >= 0)
 		return rev;
 	return -2;
 }
 
 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);
 }
 
 static PyObject *index_getitem(indexObject *self, PyObject *value)
 {
 	char *node;
 	Py_ssize_t nodelen;
 	int rev;
 
 	if (PyInt_Check(value))
 		return index_get(self, PyInt_AS_LONG(value));
 
 	if (node_check(value, &node, &nodelen) == -1)
 		return NULL;
 	rev = index_find_node(self, node, nodelen);
 	if (rev >= -1)
 		return PyInt_FromLong(rev);
 	if (rev == -2)
 		raise_revlog_error();
 	return NULL;
 }
 
 static int nt_partialmatch(indexObject *self, const char *node,
 			   Py_ssize_t nodelen)
 {
 	int rev;
 
 	if (nt_init(self) == -1)
 		return -3;
 
 	if (self->ntrev > 0) {
 		/* ensure that the radix tree is fully populated */
 		for (rev = self->ntrev - 1; rev >= 0; rev--) {
 			const char *n = index_node(self, rev);
 			if (n == NULL)
 				return -2;
 			if (nt_insert(self, n, rev) == -1)
 				return -3;
 		}
 		self->ntrev = rev;
 	}
 
 	return nt_find(self, node, nodelen, 1);
 }
 
 static PyObject *index_partialmatch(indexObject *self, PyObject *args)
 {
 	const char *fullnode;
 	int nodelen;
 	char *node;
 	int rev, i;
 
 	if (!PyArg_ParseTuple(args, "s#", &node, &nodelen))
 		return NULL;
 
 	if (nodelen < 4) {
 		PyErr_SetString(PyExc_ValueError, "key too short");
 		return NULL;
 	}
 
 	if (nodelen > 40) {
 		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;
 	}
 
 	rev = nt_partialmatch(self, node, nodelen);
 
 	switch (rev) {
 	case -4:
 		raise_revlog_error();
 	case -3:
 		return NULL;
 	case -2:
 		Py_RETURN_NONE;
 	case -1:
 		return PyBytes_FromStringAndSize(nullid, 20);
 	}
 
 	fullnode = index_node(self, rev);
 	if (fullnode == NULL) {
 		PyErr_Format(PyExc_IndexError,
 			     "could not access rev %d", rev);
 		return NULL;
 	}
 	return PyBytes_FromStringAndSize(fullnode, 20);
 }
 
 static PyObject *index_m_get(indexObject *self, PyObject *args)
 {
 	Py_ssize_t nodelen;
 	PyObject *val;
 	char *node;
 	int rev;
 
 	if (!PyArg_ParseTuple(args, "O", &val))
 		return NULL;
 	if (node_check(val, &node, &nodelen) == -1)
 		return NULL;
 	rev = index_find_node(self, node, nodelen);
 	if (rev == -3)
 		return NULL;
 	if (rev == -2)
 		Py_RETURN_NONE;
 	return PyInt_FromLong(rev);
 }
 
 static int index_contains(indexObject *self, PyObject *value)
 {
 	char *node;
 	Py_ssize_t nodelen;
 
 	if (PyInt_Check(value)) {
 		long rev = PyInt_AS_LONG(value);
 		return rev >= -1 && rev < index_length(self);
 	}
 
 	if (node_check(value, &node, &nodelen) == -1)
 		return -1;
 
 	switch (index_find_node(self, node, nodelen)) {
 	case -3:
 		return -1;
 	case -2:
 		return 0;
 	default:
 		return 1;
 	}
 }
 
 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 = PyInt_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)PyInt_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), 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)PyInt_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) - 1;
 
 	for (i = 0; i < argcount; i++) {
 		static const int capacity = 24;
 		PyObject *obj = PySequence_GetItem(args, i);
 		bitmask x;
 		long val;
 
 		if (!PyInt_Check(obj)) {
 			PyErr_SetString(PyExc_TypeError,
 					"arguments must all be ints");
 			Py_DECREF(obj);
 			goto bail;
 		}
 		val = PyInt_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 = PyInt_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 nt_invalidate_added(indexObject *self, Py_ssize_t start)
 {
 	Py_ssize_t i, len = PyList_GET_SIZE(self->added);
 
 	for (i = start; i < len; i++) {
 		PyObject *tuple = PyList_GET_ITEM(self->added, i);
 		PyObject *node = PyTuple_GET_ITEM(tuple, 7);
 
 		nt_insert(self, PyBytes_AS_STRING(node), -1);
 	}
 
 	if (start == 0)
 		Py_CLEAR(self->added);
 }
 
 /*
  * Delete a numeric range of revs, which must be at the end of the
  * range, but exclude the sentinel nullid entry.
  */
 static int index_slice_del(indexObject *self, PyObject *item)
 {
 	Py_ssize_t start, stop, step, slicelength;
 	Py_ssize_t length = index_length(self);
 	int ret = 0;
 
 /* Argument changed from PySliceObject* to PyObject* in Python 3. */
 #ifdef IS_PY3K
 	if (PySlice_GetIndicesEx(item, length,
 #else
 	if (PySlice_GetIndicesEx((PySliceObject*)item, length,
 #endif
 				 &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 - 1) {
 		if (self->nt) {
 			Py_ssize_t i;
 
 			for (i = start + 1; i < self->length - 1; i++) {
 				const char *node = index_node(self, i);
 
 				if (node)
 					nt_insert(self, node, -1);
 			}
 			if (self->added)
 				nt_invalidate_added(self, 0);
 			if (self->ntrev > start)
 				self->ntrev = (int)start;
 		}
 		self->length = start + 1;
 		if (start < self->raw_length) {
 			if (self->cache) {
 				Py_ssize_t i;
 				for (i = start; i < self->raw_length; i++)
 					Py_CLEAR(self->cache[i]);
 			}
 			self->raw_length = start;
 		}
 		goto done;
 	}
 
 	if (self->nt) {
 		nt_invalidate_added(self, start - self->length + 1);
 		if (self->ntrev > start)
 			self->ntrev = (int)start;
 	}
 	if (self->added)
 		ret = PyList_SetSlice(self->added, start - self->length + 1,
 				      PyList_GET_SIZE(self->added), NULL);
 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;
 	Py_ssize_t nodelen;
 	long rev;
 
 	if (PySlice_Check(item) && value == NULL)
 		return index_slice_del(self, item);
 
 	if (node_check(item, &node, &nodelen) == -1)
 		return -1;
 
 	if (value == NULL)
 		return self->nt ? nt_insert(self, node, -1) : 0;
 	rev = PyInt_AsLong(value);
 	if (rev > INT_MAX || rev < 0) {
 		if (!PyErr_Occurred())
 			PyErr_SetString(PyExc_ValueError, "rev out of range");
 		return -1;
 	}
 
 	if (nt_init(self) == -1)
 		return -1;
 	return nt_insert(self, 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 = v1_hdrsize;
 	Py_ssize_t len = 0;
 
 	while (pos + v1_hdrsize <= end && pos >= 0) {
 		uint32_t comp_len;
 		/* 3rd element of header is length of compressed inline data */
 		comp_len = getbe32(data + pos + 8);
 		incr = v1_hdrsize + 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 *data_obj, *inlined_obj;
 	Py_ssize_t size;
 
 	/* Initialize before argument-checking to avoid index_dealloc() crash. */
 	self->raw_length = 0;
 	self->added = NULL;
 	self->cache = NULL;
 	self->data = NULL;
 	memset(&self->buf, 0, sizeof(self->buf));
 	self->headrevs = NULL;
 	self->filteredrevs = Py_None;
 	Py_INCREF(Py_None);
 	self->nt = NULL;
 	self->offsets = NULL;
 
 	if (!PyArg_ParseTuple(args, "OO", &data_obj, &inlined_obj))
 		return -1;
 	if (!PyObject_CheckBuffer(data_obj)) {
 		PyErr_SetString(PyExc_TypeError,
 				"data does not support buffer interface");
 		return -1;
 	}
 
 	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->ntlength = self->ntcapacity = 0;
 	self->ntdepth = self->ntsplits = 0;
 	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->raw_length = len;
 		self->length = len + 1;
 	} else {
 		if (size % v1_hdrsize) {
 			PyErr_SetString(PyExc_ValueError, "corrupt index file");
 			goto bail;
 		}
 		self->raw_length = size / v1_hdrsize;
 		self->length = self->raw_length + 1;
 	}
 
 	return 0;
 bail:
 	return -1;
 }
 
 static PyObject *index_nodemap(indexObject *self)
 {
 	Py_INCREF(self);
 	return (PyObject *)self;
 }
 
 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);
 	Py_XDECREF(self->added);
 	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"},
 	{"computephasesmapsets", (PyCFunction)compute_phases_map_sets,
 			METH_VARARGS, "compute phases"},
 	{"reachableroots2", (PyCFunction)reachableroots2, METH_VARARGS,
 		"reachableroots"},
 	{"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 */
 	{"deltachain", (PyCFunction)index_deltachain, METH_VARARGS,
 	 "determine revisions with deltas to reconstruct fulltext"},
 	{"insert", (PyCFunction)index_insert, METH_VARARGS,
 	 "insert an index entry"},
 	{"partialmatch", (PyCFunction)index_partialmatch, METH_VARARGS,
 	 "match a potentially ambiguous node ID"},
 	{"stats", (PyCFunction)index_stats, METH_NOARGS,
 	 "stats for the index"},
 	{NULL} /* Sentinel */
 };
 
 static PyGetSetDef index_getset[] = {
 	{"nodemap", (getter)index_nodemap, NULL, "nodemap", NULL},
 	{NULL} /* Sentinel */
 };
 
 static PyTypeObject indexType = {
 	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 */
 	0,                         /* 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, index, cache) with elements as
  * follows:
  *
  * index: an index object that lazily parses RevlogNG 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 *tuple = NULL, *cache = NULL;
 	indexObject *idx;
 	int ret;
 
 	idx = PyObject_New(indexObject, &indexType);
 	if (idx == NULL)
 		goto bail;
 
 	ret = index_init(idx, args);
 	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);
 	}
 
 	tuple = Py_BuildValue("NN", idx, cache);
 	if (!tuple)
 		goto bail;
 	return tuple;
 
 bail:
 	Py_XDECREF(idx);
 	Py_XDECREF(cache);
 	Py_XDECREF(tuple);
 	return NULL;
 }
 
 void revlog_module_init(PyObject *mod)
 {
 	indexType.tp_new = PyType_GenericNew;
 	if (PyType_Ready(&indexType) < 0)
 		return;
 	Py_INCREF(&indexType);
 	PyModule_AddObject(mod, "index", (PyObject *)&indexType);
 
 	nullentry = Py_BuildValue("iiiiiiis#", 0, 0, 0,
 				  -1, -1, -1, -1, nullid, 20);
 	if (nullentry)
 		PyObject_GC_UnTrack(nullentry);
 }

mercurial/localrepo.py

@@ -1,2274 +1,2274
 # localrepo.py - read/write repository class for mercurial
 #
 # Copyright 2005-2007 Matt Mackall <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import errno
 import hashlib
 import inspect
 import os
 import random
 import time
 import weakref
 
 from .i18n import _
 from .node import (
     hex,
     nullid,
     short,
 )
 from . import (
     bookmarks,
     branchmap,
     bundle2,
     changegroup,
     changelog,
     color,
     context,
     dirstate,
     dirstateguard,
     discovery,
     encoding,
     error,
     exchange,
     extensions,
     filelog,
     hook,
     lock as lockmod,
     manifest,
     match as matchmod,
     merge as mergemod,
     mergeutil,
     namespaces,
     obsolete,
     pathutil,
     peer,
     phases,
     pushkey,
     pycompat,
     repository,
     repoview,
     revset,
     revsetlang,
     scmutil,
     sparse,
     store,
     subrepo,
     tags as tagsmod,
     transaction,
     txnutil,
     util,
     vfs as vfsmod,
 )
 
 release = lockmod.release
 urlerr = util.urlerr
 urlreq = util.urlreq
 
 # set of (path, vfs-location) tuples. vfs-location is:
 # - 'plain for vfs relative paths
 # - '' for svfs relative paths
 _cachedfiles = set()
 
 class _basefilecache(scmutil.filecache):
     """All filecache usage on repo are done for logic that should be unfiltered
     """
     def __get__(self, repo, type=None):
         if repo is None:
             return self
         return super(_basefilecache, self).__get__(repo.unfiltered(), type)
     def __set__(self, repo, value):
         return super(_basefilecache, self).__set__(repo.unfiltered(), value)
     def __delete__(self, repo):
         return super(_basefilecache, self).__delete__(repo.unfiltered())
 
 class repofilecache(_basefilecache):
     """filecache for files in .hg but outside of .hg/store"""
     def __init__(self, *paths):
         super(repofilecache, self).__init__(*paths)
         for path in paths:
             _cachedfiles.add((path, 'plain'))
 
     def join(self, obj, fname):
         return obj.vfs.join(fname)
 
 class storecache(_basefilecache):
     """filecache for files in the store"""
     def __init__(self, *paths):
         super(storecache, self).__init__(*paths)
         for path in paths:
             _cachedfiles.add((path, ''))
 
     def join(self, obj, fname):
         return obj.sjoin(fname)
 
 def isfilecached(repo, name):
     """check if a repo has already cached "name" filecache-ed property
 
     This returns (cachedobj-or-None, iscached) tuple.
     """
     cacheentry = repo.unfiltered()._filecache.get(name, None)
     if not cacheentry:
         return None, False
     return cacheentry.obj, True
 
 class unfilteredpropertycache(util.propertycache):
     """propertycache that apply to unfiltered repo only"""
 
     def __get__(self, repo, type=None):
         unfi = repo.unfiltered()
         if unfi is repo:
             return super(unfilteredpropertycache, self).__get__(unfi)
         return getattr(unfi, self.name)
 
 class filteredpropertycache(util.propertycache):
     """propertycache that must take filtering in account"""
 
     def cachevalue(self, obj, value):
         object.__setattr__(obj, self.name, value)
 
 
 def hasunfilteredcache(repo, name):
     """check if a repo has an unfilteredpropertycache value for <name>"""
     return name in vars(repo.unfiltered())
 
 def unfilteredmethod(orig):
     """decorate method that always need to be run on unfiltered version"""
     def wrapper(repo, *args, **kwargs):
         return orig(repo.unfiltered(), *args, **kwargs)
     return wrapper
 
 moderncaps = {'lookup', 'branchmap', 'pushkey', 'known', 'getbundle',
               'unbundle'}
 legacycaps = moderncaps.union({'changegroupsubset'})
 
 class localpeer(repository.peer):
     '''peer for a local repo; reflects only the most recent API'''
 
     def __init__(self, repo, caps=None):
         super(localpeer, self).__init__()
 
         if caps is None:
             caps = moderncaps.copy()
         self._repo = repo.filtered('served')
         self._ui = repo.ui
         self._caps = repo._restrictcapabilities(caps)
 
     # Begin of _basepeer interface.
 
     @util.propertycache
     def ui(self):
         return self._ui
 
     def url(self):
         return self._repo.url()
 
     def local(self):
         return self._repo
 
     def peer(self):
         return self
 
     def canpush(self):
         return True
 
     def close(self):
         self._repo.close()
 
     # End of _basepeer interface.
 
     # Begin of _basewirecommands interface.
 
     def branchmap(self):
         return self._repo.branchmap()
 
     def capabilities(self):
         return self._caps
 
     def debugwireargs(self, one, two, three=None, four=None, five=None):
         """Used to test argument passing over the wire"""
         return "%s %s %s %s %s" % (one, two, three, four, five)
 
     def getbundle(self, source, heads=None, common=None, bundlecaps=None,
                   **kwargs):
         chunks = exchange.getbundlechunks(self._repo, source, heads=heads,
                                           common=common, bundlecaps=bundlecaps,
                                           **kwargs)
         cb = util.chunkbuffer(chunks)
 
         if exchange.bundle2requested(bundlecaps):
             # When requesting a bundle2, getbundle returns a stream to make the
             # wire level function happier. We need to build a proper object
             # from it in local peer.
             return bundle2.getunbundler(self.ui, cb)
         else:
             return changegroup.getunbundler('01', cb, None)
 
     def heads(self):
         return self._repo.heads()
 
     def known(self, nodes):
         return self._repo.known(nodes)
 
     def listkeys(self, namespace):
         return self._repo.listkeys(namespace)
 
     def lookup(self, key):
         return self._repo.lookup(key)
 
     def pushkey(self, namespace, key, old, new):
         return self._repo.pushkey(namespace, key, old, new)
 
     def stream_out(self):
         raise error.Abort(_('cannot perform stream clone against local '
                             'peer'))
 
     def unbundle(self, cg, heads, url):
         """apply a bundle on a repo
 
         This function handles the repo locking itself."""
         try:
             try:
                 cg = exchange.readbundle(self.ui, cg, None)
                 ret = exchange.unbundle(self._repo, cg, heads, 'push', url)
                 if util.safehasattr(ret, 'getchunks'):
                     # This is a bundle20 object, turn it into an unbundler.
                     # This little dance should be dropped eventually when the
                     # API is finally improved.
                     stream = util.chunkbuffer(ret.getchunks())
                     ret = bundle2.getunbundler(self.ui, stream)
                 return ret
             except Exception as exc:
                 # If the exception contains output salvaged from a bundle2
                 # reply, we need to make sure it is printed before continuing
                 # to fail. So we build a bundle2 with such output and consume
                 # it directly.
                 #
                 # This is not very elegant but allows a "simple" solution for
                 # issue4594
                 output = getattr(exc, '_bundle2salvagedoutput', ())
                 if output:
                     bundler = bundle2.bundle20(self._repo.ui)
                     for out in output:
                         bundler.addpart(out)
                     stream = util.chunkbuffer(bundler.getchunks())
                     b = bundle2.getunbundler(self.ui, stream)
                     bundle2.processbundle(self._repo, b)
                 raise
         except error.PushRaced as exc:
             raise error.ResponseError(_('push failed:'), str(exc))
 
     # End of _basewirecommands interface.
 
     # Begin of peer interface.
 
     def iterbatch(self):
         return peer.localiterbatcher(self)
 
     # End of peer interface.
 
 class locallegacypeer(repository.legacypeer, localpeer):
     '''peer extension which implements legacy methods too; used for tests with
     restricted capabilities'''
 
     def __init__(self, repo):
         super(locallegacypeer, self).__init__(repo, caps=legacycaps)
 
     # Begin of baselegacywirecommands interface.
 
     def between(self, pairs):
         return self._repo.between(pairs)
 
     def branches(self, nodes):
         return self._repo.branches(nodes)
 
     def changegroup(self, basenodes, source):
         outgoing = discovery.outgoing(self._repo, missingroots=basenodes,
                                       missingheads=self._repo.heads())
         return changegroup.makechangegroup(self._repo, outgoing, '01', source)
 
     def changegroupsubset(self, bases, heads, source):
         outgoing = discovery.outgoing(self._repo, missingroots=bases,
                                       missingheads=heads)
         return changegroup.makechangegroup(self._repo, outgoing, '01', source)
 
     # End of baselegacywirecommands interface.
 
 # Increment the sub-version when the revlog v2 format changes to lock out old
 # clients.
 REVLOGV2_REQUIREMENT = 'exp-revlogv2.0'
 
 class localrepository(object):
 
     supportedformats = {
         'revlogv1',
         'generaldelta',
         'treemanifest',
         'manifestv2',
         REVLOGV2_REQUIREMENT,
     }
     _basesupported = supportedformats | {
         'store',
         'fncache',
         'shared',
         'relshared',
         'dotencode',
         'exp-sparse',
     }
     openerreqs = {
         'revlogv1',
         'generaldelta',
         'treemanifest',
         'manifestv2',
     }
 
     # a list of (ui, featureset) functions.
     # only functions defined in module of enabled extensions are invoked
     featuresetupfuncs = set()
 
     # list of prefix for file which can be written without 'wlock'
     # Extensions should extend this list when needed
     _wlockfreeprefix = {
         # We migh consider requiring 'wlock' for the next
         # two, but pretty much all the existing code assume
         # wlock is not needed so we keep them excluded for
         # now.
         'hgrc',
         'requires',
         # XXX cache is a complicatged business someone
         # should investigate this in depth at some point
         'cache/',
         # XXX shouldn't be dirstate covered by the wlock?
         'dirstate',
         # XXX bisect was still a bit too messy at the time
         # this changeset was introduced. Someone should fix
         # the remainig bit and drop this line
         'bisect.state',
     }
 
     def __init__(self, baseui, path, create=False):
         self.requirements = set()
         self.filtername = None
         # wvfs: rooted at the repository root, used to access the working copy
         self.wvfs = vfsmod.vfs(path, expandpath=True, realpath=True)
         # vfs: rooted at .hg, used to access repo files outside of .hg/store
         self.vfs = None
         # svfs: usually rooted at .hg/store, used to access repository history
         # If this is a shared repository, this vfs may point to another
         # repository's .hg/store directory.
         self.svfs = None
         self.root = self.wvfs.base
         self.path = self.wvfs.join(".hg")
         self.origroot = path
         # This is only used by context.workingctx.match in order to
         # detect files in subrepos.
         self.auditor = pathutil.pathauditor(
             self.root, callback=self._checknested)
         # This is only used by context.basectx.match in order to detect
         # files in subrepos.
         self.nofsauditor = pathutil.pathauditor(
             self.root, callback=self._checknested, realfs=False, cached=True)
         self.baseui = baseui
         self.ui = baseui.copy()
         self.ui.copy = baseui.copy # prevent copying repo configuration
         self.vfs = vfsmod.vfs(self.path, cacheaudited=True)
         if (self.ui.configbool('devel', 'all-warnings') or
             self.ui.configbool('devel', 'check-locks')):
             self.vfs.audit = self._getvfsward(self.vfs.audit)
         # A list of callback to shape the phase if no data were found.
         # Callback are in the form: func(repo, roots) --> processed root.
         # This list it to be filled by extension during repo setup
         self._phasedefaults = []
         try:
             self.ui.readconfig(self.vfs.join("hgrc"), self.root)
             self._loadextensions()
         except IOError:
             pass
 
         if self.featuresetupfuncs:
             self.supported = set(self._basesupported) # use private copy
             extmods = set(m.__name__ for n, m
                           in extensions.extensions(self.ui))
             for setupfunc in self.featuresetupfuncs:
                 if setupfunc.__module__ in extmods:
                     setupfunc(self.ui, self.supported)
         else:
             self.supported = self._basesupported
         color.setup(self.ui)
 
         # Add compression engines.
         for name in util.compengines:
             engine = util.compengines[name]
             if engine.revlogheader():
                 self.supported.add('exp-compression-%s' % name)
 
         if not self.vfs.isdir():
             if create:
                 self.requirements = newreporequirements(self)
 
                 if not self.wvfs.exists():
                     self.wvfs.makedirs()
                 self.vfs.makedir(notindexed=True)
 
                 if 'store' in self.requirements:
                     self.vfs.mkdir("store")
 
                     # create an invalid changelog
                     self.vfs.append(
                         "00changelog.i",
                         '\0\0\0\2' # represents revlogv2
                         ' dummy changelog to prevent using the old repo layout'
                     )
             else:
                 raise error.RepoError(_("repository %s not found") % path)
         elif create:
             raise error.RepoError(_("repository %s already exists") % path)
         else:
             try:
                 self.requirements = scmutil.readrequires(
                         self.vfs, self.supported)
             except IOError as inst:
                 if inst.errno != errno.ENOENT:
                     raise
 
         cachepath = self.vfs.join('cache')
         self.sharedpath = self.path
         try:
             sharedpath = self.vfs.read("sharedpath").rstrip('\n')
             if 'relshared' in self.requirements:
                 sharedpath = self.vfs.join(sharedpath)
             vfs = vfsmod.vfs(sharedpath, realpath=True)
             cachepath = vfs.join('cache')
             s = vfs.base
             if not vfs.exists():
                 raise error.RepoError(
                     _('.hg/sharedpath points to nonexistent directory %s') % s)
             self.sharedpath = s
         except IOError as inst:
             if inst.errno != errno.ENOENT:
                 raise
 
         if 'exp-sparse' in self.requirements and not sparse.enabled:
             raise error.RepoError(_('repository is using sparse feature but '
                                     'sparse is not enabled; enable the '
                                     '"sparse" extensions to access'))
 
         self.store = store.store(
             self.requirements, self.sharedpath,
             lambda base: vfsmod.vfs(base, cacheaudited=True))
         self.spath = self.store.path
         self.svfs = self.store.vfs
         self.sjoin = self.store.join
         self.vfs.createmode = self.store.createmode
         self.cachevfs = vfsmod.vfs(cachepath, cacheaudited=True)
         self.cachevfs.createmode = self.store.createmode
         if (self.ui.configbool('devel', 'all-warnings') or
             self.ui.configbool('devel', 'check-locks')):
             if util.safehasattr(self.svfs, 'vfs'): # this is filtervfs
                 self.svfs.vfs.audit = self._getsvfsward(self.svfs.vfs.audit)
             else: # standard vfs
                 self.svfs.audit = self._getsvfsward(self.svfs.audit)
         self._applyopenerreqs()
         if create:
             self._writerequirements()
 
         self._dirstatevalidatewarned = False
 
         self._branchcaches = {}
         self._revbranchcache = None
         self.filterpats = {}
         self._datafilters = {}
         self._transref = self._lockref = self._wlockref = None
 
         # A cache for various files under .hg/ that tracks file changes,
         # (used by the filecache decorator)
         #
         # Maps a property name to its util.filecacheentry
         self._filecache = {}
 
         # hold sets of revision to be filtered
         # should be cleared when something might have changed the filter value:
         # - new changesets,
         # - phase change,
         # - new obsolescence marker,
         # - working directory parent change,
         # - bookmark changes
         self.filteredrevcache = {}
 
         # post-dirstate-status hooks
         self._postdsstatus = []
 
         # generic mapping between names and nodes
         self.names = namespaces.namespaces()
 
         # Key to signature value.
         self._sparsesignaturecache = {}
         # Signature to cached matcher instance.
         self._sparsematchercache = {}
 
     def _getvfsward(self, origfunc):
         """build a ward for self.vfs"""
         rref = weakref.ref(self)
         def checkvfs(path, mode=None):
             ret = origfunc(path, mode=mode)
             repo = rref()
             if (repo is None
                 or not util.safehasattr(repo, '_wlockref')
                 or not util.safehasattr(repo, '_lockref')):
                 return
             if mode in (None, 'r', 'rb'):
                 return
             if path.startswith(repo.path):
                 # truncate name relative to the repository (.hg)
                 path = path[len(repo.path) + 1:]
             if path.startswith('cache/'):
                 msg = 'accessing cache with vfs instead of cachevfs: "%s"'
                 repo.ui.develwarn(msg % path, stacklevel=2, config="cache-vfs")
             if path.startswith('journal.'):
                 # journal is covered by 'lock'
                 if repo._currentlock(repo._lockref) is None:
                     repo.ui.develwarn('write with no lock: "%s"' % path,
                                       stacklevel=2, config='check-locks')
             elif repo._currentlock(repo._wlockref) is None:
                 # rest of vfs files are covered by 'wlock'
                 #
                 # exclude special files
                 for prefix in self._wlockfreeprefix:
                     if path.startswith(prefix):
                         return
                 repo.ui.develwarn('write with no wlock: "%s"' % path,
                                   stacklevel=2, config='check-locks')
             return ret
         return checkvfs
 
     def _getsvfsward(self, origfunc):
         """build a ward for self.svfs"""
         rref = weakref.ref(self)
         def checksvfs(path, mode=None):
             ret = origfunc(path, mode=mode)
             repo = rref()
             if repo is None or not util.safehasattr(repo, '_lockref'):
                 return
             if mode in (None, 'r', 'rb'):
                 return
             if path.startswith(repo.sharedpath):
                 # truncate name relative to the repository (.hg)
                 path = path[len(repo.sharedpath) + 1:]
             if repo._currentlock(repo._lockref) is None:
                 repo.ui.develwarn('write with no lock: "%s"' % path,
                                   stacklevel=3)
             return ret
         return checksvfs
 
     def close(self):
         self._writecaches()
 
     def _loadextensions(self):
         extensions.loadall(self.ui)
 
     def _writecaches(self):
         if self._revbranchcache:
             self._revbranchcache.write()
 
     def _restrictcapabilities(self, caps):
         if self.ui.configbool('experimental', 'bundle2-advertise'):
             caps = set(caps)
             capsblob = bundle2.encodecaps(bundle2.getrepocaps(self))
             caps.add('bundle2=' + urlreq.quote(capsblob))
         return caps
 
     def _applyopenerreqs(self):
         self.svfs.options = dict((r, 1) for r in self.requirements
                                            if r in self.openerreqs)
         # experimental config: format.chunkcachesize
         chunkcachesize = self.ui.configint('format', 'chunkcachesize')
         if chunkcachesize is not None:
             self.svfs.options['chunkcachesize'] = chunkcachesize
         # experimental config: format.maxchainlen
         maxchainlen = self.ui.configint('format', 'maxchainlen')
         if maxchainlen is not None:
             self.svfs.options['maxchainlen'] = maxchainlen
         # experimental config: format.manifestcachesize
         manifestcachesize = self.ui.configint('format', 'manifestcachesize')
         if manifestcachesize is not None:
             self.svfs.options['manifestcachesize'] = manifestcachesize
         # experimental config: format.aggressivemergedeltas
         aggressivemergedeltas = self.ui.configbool('format',
                                                    'aggressivemergedeltas')
         self.svfs.options['aggressivemergedeltas'] = aggressivemergedeltas
         self.svfs.options['lazydeltabase'] = not scmutil.gddeltaconfig(self.ui)
         chainspan = self.ui.configbytes('experimental', 'maxdeltachainspan')
         if 0 <= chainspan:
             self.svfs.options['maxdeltachainspan'] = chainspan
         mmapindexthreshold = self.ui.configbytes('experimental',
                                                  'mmapindexthreshold')
         if mmapindexthreshold is not None:
             self.svfs.options['mmapindexthreshold'] = mmapindexthreshold
         withsparseread = self.ui.configbool('experimental', 'sparse-read')
         srdensitythres = float(self.ui.config('experimental',
                                               'sparse-read.density-threshold'))
         srmingapsize = self.ui.configbytes('experimental',
                                            'sparse-read.min-gap-size')
         self.svfs.options['with-sparse-read'] = withsparseread
         self.svfs.options['sparse-read-density-threshold'] = srdensitythres
         self.svfs.options['sparse-read-min-gap-size'] = srmingapsize
 
         for r in self.requirements:
             if r.startswith('exp-compression-'):
                 self.svfs.options['compengine'] = r[len('exp-compression-'):]
 
         # TODO move "revlogv2" to openerreqs once finalized.
         if REVLOGV2_REQUIREMENT in self.requirements:
             self.svfs.options['revlogv2'] = True
 
     def _writerequirements(self):
         scmutil.writerequires(self.vfs, self.requirements)
 
     def _checknested(self, path):
         """Determine if path is a legal nested repository."""
         if not path.startswith(self.root):
             return False
         subpath = path[len(self.root) + 1:]
         normsubpath = util.pconvert(subpath)
 
         # XXX: Checking against the current working copy is wrong in
         # the sense that it can reject things like
         #
         #   $ hg cat -r 10 sub/x.txt
         #
         # if sub/ is no longer a subrepository in the working copy
         # parent revision.
         #
         # However, it can of course also allow things that would have
         # been rejected before, such as the above cat command if sub/
         # is a subrepository now, but was a normal directory before.
         # The old path auditor would have rejected by mistake since it
         # panics when it sees sub/.hg/.
         #
         # All in all, checking against the working copy seems sensible
         # since we want to prevent access to nested repositories on
         # the filesystem *now*.
         ctx = self[None]
         parts = util.splitpath(subpath)
         while parts:
             prefix = '/'.join(parts)
             if prefix in ctx.substate:
                 if prefix == normsubpath:
                     return True
                 else:
                     sub = ctx.sub(prefix)
                     return sub.checknested(subpath[len(prefix) + 1:])
             else:
                 parts.pop()
         return False
 
     def peer(self):
         return localpeer(self) # not cached to avoid reference cycle
 
     def unfiltered(self):
         """Return unfiltered version of the repository
 
         Intended to be overwritten by filtered repo."""
         return self
 
     def filtered(self, name):
         """Return a filtered version of a repository"""
         cls = repoview.newtype(self.unfiltered().__class__)
         return cls(self, name)
 
     @repofilecache('bookmarks', 'bookmarks.current')
     def _bookmarks(self):
         return bookmarks.bmstore(self)
 
     @property
     def _activebookmark(self):
         return self._bookmarks.active
 
-    # _phaserevs and _phasesets depend on changelog. what we need is to
-    # call _phasecache.invalidate() if '00changelog.i' was changed, but it
+    # _phasesets depend on changelog. what we need is to call
+    # _phasecache.invalidate() if '00changelog.i' was changed, but it
     # can't be easily expressed in filecache mechanism.
     @storecache('phaseroots', '00changelog.i')
     def _phasecache(self):
         return phases.phasecache(self, self._phasedefaults)
 
     @storecache('obsstore')
     def obsstore(self):
         return obsolete.makestore(self.ui, self)
 
     @storecache('00changelog.i')
     def changelog(self):
         return changelog.changelog(self.svfs,
                                    trypending=txnutil.mayhavepending(self.root))
 
     def _constructmanifest(self):
         # This is a temporary function while we migrate from manifest to
         # manifestlog. It allows bundlerepo and unionrepo to intercept the
         # manifest creation.
         return manifest.manifestrevlog(self.svfs)
 
     @storecache('00manifest.i')
     def manifestlog(self):
         return manifest.manifestlog(self.svfs, self)
 
     @repofilecache('dirstate')
     def dirstate(self):
         sparsematchfn = lambda: sparse.matcher(self)
 
         return dirstate.dirstate(self.vfs, self.ui, self.root,
                                  self._dirstatevalidate, sparsematchfn)
 
     def _dirstatevalidate(self, node):
         try:
             self.changelog.rev(node)
             return node
         except error.LookupError:
             if not self._dirstatevalidatewarned:
                 self._dirstatevalidatewarned = True
                 self.ui.warn(_("warning: ignoring unknown"
                                " working parent %s!\n") % short(node))
             return nullid
 
     def __getitem__(self, changeid):
         if changeid is None:
             return context.workingctx(self)
         if isinstance(changeid, slice):
             # wdirrev isn't contiguous so the slice shouldn't include it
             return [context.changectx(self, i)
                     for i in xrange(*changeid.indices(len(self)))
                     if i not in self.changelog.filteredrevs]
         try:
             return context.changectx(self, changeid)
         except error.WdirUnsupported:
             return context.workingctx(self)
 
     def __contains__(self, changeid):
         """True if the given changeid exists
 
         error.LookupError is raised if an ambiguous node specified.
         """
         try:
             self[changeid]
             return True
         except error.RepoLookupError:
             return False
 
     def __nonzero__(self):
         return True
 
     __bool__ = __nonzero__
 
     def __len__(self):
         return len(self.changelog)
 
     def __iter__(self):
         return iter(self.changelog)
 
     def revs(self, expr, *args):
         '''Find revisions matching a revset.
 
         The revset is specified as a string ``expr`` that may contain
         %-formatting to escape certain types. See ``revsetlang.formatspec``.
 
         Revset aliases from the configuration are not expanded. To expand
         user aliases, consider calling ``scmutil.revrange()`` or
         ``repo.anyrevs([expr], user=True)``.
 
         Returns a revset.abstractsmartset, which is a list-like interface
         that contains integer revisions.
         '''
         expr = revsetlang.formatspec(expr, *args)
         m = revset.match(None, expr)
         return m(self)
 
     def set(self, expr, *args):
         '''Find revisions matching a revset and emit changectx instances.
 
         This is a convenience wrapper around ``revs()`` that iterates the
         result and is a generator of changectx instances.
 
         Revset aliases from the configuration are not expanded. To expand
         user aliases, consider calling ``scmutil.revrange()``.
         '''
         for r in self.revs(expr, *args):
             yield self[r]
 
     def anyrevs(self, specs, user=False, localalias=None):
         '''Find revisions matching one of the given revsets.
 
         Revset aliases from the configuration are not expanded by default. To
         expand user aliases, specify ``user=True``. To provide some local
         definitions overriding user aliases, set ``localalias`` to
         ``{name: definitionstring}``.
         '''
         if user:
             m = revset.matchany(self.ui, specs, repo=self,
                                 localalias=localalias)
         else:
             m = revset.matchany(None, specs, localalias=localalias)
         return m(self)
 
     def url(self):
         return 'file:' + self.root
 
     def hook(self, name, throw=False, **args):
         """Call a hook, passing this repo instance.
 
         This a convenience method to aid invoking hooks. Extensions likely
         won't call this unless they have registered a custom hook or are
         replacing code that is expected to call a hook.
         """
         return hook.hook(self.ui, self, name, throw, **args)
 
     @filteredpropertycache
     def _tagscache(self):
         '''Returns a tagscache object that contains various tags related
         caches.'''
 
         # This simplifies its cache management by having one decorated
         # function (this one) and the rest simply fetch things from it.
         class tagscache(object):
             def __init__(self):
                 # These two define the set of tags for this repository. tags
                 # maps tag name to node; tagtypes maps tag name to 'global' or
                 # 'local'. (Global tags are defined by .hgtags across all
                 # heads, and local tags are defined in .hg/localtags.)
                 # They constitute the in-memory cache of tags.
                 self.tags = self.tagtypes = None
 
                 self.nodetagscache = self.tagslist = None
 
         cache = tagscache()
         cache.tags, cache.tagtypes = self._findtags()
 
         return cache
 
     def tags(self):
         '''return a mapping of tag to node'''
         t = {}
         if self.changelog.filteredrevs:
             tags, tt = self._findtags()
         else:
             tags = self._tagscache.tags
         for k, v in tags.iteritems():
             try:
                 # ignore tags to unknown nodes
                 self.changelog.rev(v)
                 t[k] = v
             except (error.LookupError, ValueError):
                 pass
         return t
 
     def _findtags(self):
         '''Do the hard work of finding tags.  Return a pair of dicts
         (tags, tagtypes) where tags maps tag name to node, and tagtypes
         maps tag name to a string like \'global\' or \'local\'.
         Subclasses or extensions are free to add their own tags, but
         should be aware that the returned dicts will be retained for the
         duration of the localrepo object.'''
 
         # XXX what tagtype should subclasses/extensions use?  Currently
         # mq and bookmarks add tags, but do not set the tagtype at all.
         # Should each extension invent its own tag type?  Should there
         # be one tagtype for all such "virtual" tags?  Or is the status
         # quo fine?
 
 
         # map tag name to (node, hist)
         alltags = tagsmod.findglobaltags(self.ui, self)
         # map tag name to tag type
         tagtypes = dict((tag, 'global') for tag in alltags)
 
         tagsmod.readlocaltags(self.ui, self, alltags, tagtypes)
 
         # Build the return dicts.  Have to re-encode tag names because
         # the tags module always uses UTF-8 (in order not to lose info
         # writing to the cache), but the rest of Mercurial wants them in
         # local encoding.
         tags = {}
         for (name, (node, hist)) in alltags.iteritems():
             if node != nullid:
                 tags[encoding.tolocal(name)] = node
         tags['tip'] = self.changelog.tip()
         tagtypes = dict([(encoding.tolocal(name), value)
                          for (name, value) in tagtypes.iteritems()])
         return (tags, tagtypes)
 
     def tagtype(self, tagname):
         '''
         return the type of the given tag. result can be:
 
         'local'  : a local tag
         'global' : a global tag
         None     : tag does not exist
         '''
 
         return self._tagscache.tagtypes.get(tagname)
 
     def tagslist(self):
         '''return a list of tags ordered by revision'''
         if not self._tagscache.tagslist:
             l = []
             for t, n in self.tags().iteritems():
                 l.append((self.changelog.rev(n), t, n))
             self._tagscache.tagslist = [(t, n) for r, t, n in sorted(l)]
 
         return self._tagscache.tagslist
 
     def nodetags(self, node):
         '''return the tags associated with a node'''
         if not self._tagscache.nodetagscache:
             nodetagscache = {}
             for t, n in self._tagscache.tags.iteritems():
                 nodetagscache.setdefault(n, []).append(t)
             for tags in nodetagscache.itervalues():
                 tags.sort()
             self._tagscache.nodetagscache = nodetagscache
         return self._tagscache.nodetagscache.get(node, [])
 
     def nodebookmarks(self, node):
         """return the list of bookmarks pointing to the specified node"""
         marks = []
         for bookmark, n in self._bookmarks.iteritems():
             if n == node:
                 marks.append(bookmark)
         return sorted(marks)
 
     def branchmap(self):
         '''returns a dictionary {branch: [branchheads]} with branchheads
         ordered by increasing revision number'''
         branchmap.updatecache(self)
         return self._branchcaches[self.filtername]
 
     @unfilteredmethod
     def revbranchcache(self):
         if not self._revbranchcache:
             self._revbranchcache = branchmap.revbranchcache(self.unfiltered())
         return self._revbranchcache
 
     def branchtip(self, branch, ignoremissing=False):
         '''return the tip node for a given branch
 
         If ignoremissing is True, then this method will not raise an error.
         This is helpful for callers that only expect None for a missing branch
         (e.g. namespace).
 
         '''
         try:
             return self.branchmap().branchtip(branch)
         except KeyError:
             if not ignoremissing:
                 raise error.RepoLookupError(_("unknown branch '%s'") % branch)
             else:
                 pass
 
     def lookup(self, key):
         return self[key].node()
 
     def lookupbranch(self, key, remote=None):
         repo = remote or self
         if key in repo.branchmap():
             return key
 
         repo = (remote and remote.local()) and remote or self
         return repo[key].branch()
 
     def known(self, nodes):
         cl = self.changelog
         nm = cl.nodemap
         filtered = cl.filteredrevs
         result = []
         for n in nodes:
             r = nm.get(n)
             resp = not (r is None or r in filtered)
             result.append(resp)
         return result
 
     def local(self):
         return self
 
     def publishing(self):
         # it's safe (and desirable) to trust the publish flag unconditionally
         # so that we don't finalize changes shared between users via ssh or nfs
         return self.ui.configbool('phases', 'publish', untrusted=True)
 
     def cancopy(self):
         # so statichttprepo's override of local() works
         if not self.local():
             return False
         if not self.publishing():
             return True
         # if publishing we can't copy if there is filtered content
         return not self.filtered('visible').changelog.filteredrevs
 
     def shared(self):
         '''the type of shared repository (None if not shared)'''
         if self.sharedpath != self.path:
             return 'store'
         return None
 
     def wjoin(self, f, *insidef):
         return self.vfs.reljoin(self.root, f, *insidef)
 
     def file(self, f):
         if f[0] == '/':
             f = f[1:]
         return filelog.filelog(self.svfs, f)
 
     def changectx(self, changeid):
         return self[changeid]
 
     def setparents(self, p1, p2=nullid):
         with self.dirstate.parentchange():
             copies = self.dirstate.setparents(p1, p2)
             pctx = self[p1]
             if copies:
                 # Adjust copy records, the dirstate cannot do it, it
                 # requires access to parents manifests. Preserve them
                 # only for entries added to first parent.
                 for f in copies:
                     if f not in pctx and copies[f] in pctx:
                         self.dirstate.copy(copies[f], f)
             if p2 == nullid:
                 for f, s in sorted(self.dirstate.copies().items()):
                     if f not in pctx and s not in pctx:
                         self.dirstate.copy(None, f)
 
     def filectx(self, path, changeid=None, fileid=None):
         """changeid can be a changeset revision, node, or tag.
            fileid can be a file revision or node."""
         return context.filectx(self, path, changeid, fileid)
 
     def getcwd(self):
         return self.dirstate.getcwd()
 
     def pathto(self, f, cwd=None):
         return self.dirstate.pathto(f, cwd)
 
     def _loadfilter(self, filter):
         if filter not in self.filterpats:
             l = []
             for pat, cmd in self.ui.configitems(filter):
                 if cmd == '!':
                     continue
                 mf = matchmod.match(self.root, '', [pat])
                 fn = None
                 params = cmd
                 for name, filterfn in self._datafilters.iteritems():
                     if cmd.startswith(name):
                         fn = filterfn
                         params = cmd[len(name):].lstrip()
                         break
                 if not fn:
                     fn = lambda s, c, **kwargs: util.filter(s, c)
                 # Wrap old filters not supporting keyword arguments
                 if not inspect.getargspec(fn)[2]:
                     oldfn = fn
                     fn = lambda s, c, **kwargs: oldfn(s, c)
                 l.append((mf, fn, params))
             self.filterpats[filter] = l
         return self.filterpats[filter]
 
     def _filter(self, filterpats, filename, data):
         for mf, fn, cmd in filterpats:
             if mf(filename):
                 self.ui.debug("filtering %s through %s\n" % (filename, cmd))
                 data = fn(data, cmd, ui=self.ui, repo=self, filename=filename)
                 break
 
         return data
 
     @unfilteredpropertycache
     def _encodefilterpats(self):
         return self._loadfilter('encode')
 
     @unfilteredpropertycache
     def _decodefilterpats(self):
         return self._loadfilter('decode')
 
     def adddatafilter(self, name, filter):
         self._datafilters[name] = filter
 
     def wread(self, filename):
         if self.wvfs.islink(filename):
             data = self.wvfs.readlink(filename)
         else:
             data = self.wvfs.read(filename)
         return self._filter(self._encodefilterpats, filename, data)
 
     def wwrite(self, filename, data, flags, backgroundclose=False):
         """write ``data`` into ``filename`` in the working directory
 
         This returns length of written (maybe decoded) data.
         """
         data = self._filter(self._decodefilterpats, filename, data)
         if 'l' in flags:
             self.wvfs.symlink(data, filename)
         else:
             self.wvfs.write(filename, data, backgroundclose=backgroundclose)
             if 'x' in flags:
                 self.wvfs.setflags(filename, False, True)
         return len(data)
 
     def wwritedata(self, filename, data):
         return self._filter(self._decodefilterpats, filename, data)
 
     def currenttransaction(self):
         """return the current transaction or None if non exists"""
         if self._transref:
             tr = self._transref()
         else:
             tr = None
 
         if tr and tr.running():
             return tr
         return None
 
     def transaction(self, desc, report=None):
         if (self.ui.configbool('devel', 'all-warnings')
                 or self.ui.configbool('devel', 'check-locks')):
             if self._currentlock(self._lockref) is None:
                 raise error.ProgrammingError('transaction requires locking')
         tr = self.currenttransaction()
         if tr is not None:
             scmutil.registersummarycallback(self, tr, desc)
             return tr.nest()
 
         # abort here if the journal already exists
         if self.svfs.exists("journal"):
             raise error.RepoError(
                 _("abandoned transaction found"),
                 hint=_("run 'hg recover' to clean up transaction"))
 
         idbase = "%.40f#%f" % (random.random(), time.time())
         ha = hex(hashlib.sha1(idbase).digest())
         txnid = 'TXN:' + ha
         self.hook('pretxnopen', throw=True, txnname=desc, txnid=txnid)
 
         self._writejournal(desc)
         renames = [(vfs, x, undoname(x)) for vfs, x in self._journalfiles()]
         if report:
             rp = report
         else:
             rp = self.ui.warn
         vfsmap = {'plain': self.vfs} # root of .hg/
         # we must avoid cyclic reference between repo and transaction.
         reporef = weakref.ref(self)
         # Code to track tag movement
         #
         # Since tags are all handled as file content, it is actually quite hard
         # to track these movement from a code perspective. So we fallback to a
         # tracking at the repository level. One could envision to track changes
         # to the '.hgtags' file through changegroup apply but that fails to
         # cope with case where transaction expose new heads without changegroup
         # being involved (eg: phase movement).
         #
         # For now, We gate the feature behind a flag since this likely comes
         # with performance impacts. The current code run more often than needed
         # and do not use caches as much as it could.  The current focus is on
         # the behavior of the feature so we disable it by default. The flag
         # will be removed when we are happy with the performance impact.
         #
         # Once this feature is no longer experimental move the following
         # documentation to the appropriate help section:
         #
         # The ``HG_TAG_MOVED`` variable will be set if the transaction touched
         # tags (new or changed or deleted tags). In addition the details of
         # these changes are made available in a file at:
         #     ``REPOROOT/.hg/changes/tags.changes``.
         # Make sure you check for HG_TAG_MOVED before reading that file as it
         # might exist from a previous transaction even if no tag were touched
         # in this one. Changes are recorded in a line base format::
         #
         #     <action> <hex-node> <tag-name>\n
         #
         # Actions are defined as follow:
         #   "-R": tag is removed,
         #   "+A": tag is added,
         #   "-M": tag is moved (old value),
         #   "+M": tag is moved (new value),
         tracktags = lambda x: None
         # experimental config: experimental.hook-track-tags
         shouldtracktags = self.ui.configbool('experimental', 'hook-track-tags')
         if desc != 'strip' and shouldtracktags:
             oldheads = self.changelog.headrevs()
             def tracktags(tr2):
                 repo = reporef()
                 oldfnodes = tagsmod.fnoderevs(repo.ui, repo, oldheads)
                 newheads = repo.changelog.headrevs()
                 newfnodes = tagsmod.fnoderevs(repo.ui, repo, newheads)
                 # notes: we compare lists here.
                 # As we do it only once buiding set would not be cheaper
                 changes = tagsmod.difftags(repo.ui, repo, oldfnodes, newfnodes)
                 if changes:
                     tr2.hookargs['tag_moved'] = '1'
                     with repo.vfs('changes/tags.changes', 'w',
                                   atomictemp=True) as changesfile:
                         # note: we do not register the file to the transaction
                         # because we needs it to still exist on the transaction
                         # is close (for txnclose hooks)
                         tagsmod.writediff(changesfile, changes)
         def validate(tr2):
             """will run pre-closing hooks"""
             # XXX the transaction API is a bit lacking here so we take a hacky
             # path for now
             #
             # We cannot add this as a "pending" hooks since the 'tr.hookargs'
             # dict is copied before these run. In addition we needs the data
             # available to in memory hooks too.
             #
             # Moreover, we also need to make sure this runs before txnclose
             # hooks and there is no "pending" mechanism that would execute
             # logic only if hooks are about to run.
             #
             # Fixing this limitation of the transaction is also needed to track
             # other families of changes (bookmarks, phases, obsolescence).
             #
             # This will have to be fixed before we remove the experimental
             # gating.
             tracktags(tr2)
             repo = reporef()
             if repo.ui.configbool('experimental', 'single-head-per-branch'):
                 scmutil.enforcesinglehead(repo, tr2, desc)
             if hook.hashook(repo.ui, 'pretxnclose-bookmark'):
                 for name, (old, new) in sorted(tr.changes['bookmarks'].items()):
                     args = tr.hookargs.copy()
                     args.update(bookmarks.preparehookargs(name, old, new))
                     repo.hook('pretxnclose-bookmark', throw=True,
                               txnname=desc,
                               **pycompat.strkwargs(args))
             if hook.hashook(repo.ui, 'pretxnclose-phase'):
                 cl = repo.unfiltered().changelog
                 for rev, (old, new) in tr.changes['phases'].items():
                     args = tr.hookargs.copy()
                     node = hex(cl.node(rev))
                     args.update(phases.preparehookargs(node, old, new))
                     repo.hook('pretxnclose-phase', throw=True, txnname=desc,
                               **pycompat.strkwargs(args))
 
             repo.hook('pretxnclose', throw=True,
                       txnname=desc, **pycompat.strkwargs(tr.hookargs))
         def releasefn(tr, success):
             repo = reporef()
             if success:
                 # this should be explicitly invoked here, because
                 # in-memory changes aren't written out at closing
                 # transaction, if tr.addfilegenerator (via
                 # dirstate.write or so) isn't invoked while
                 # transaction running
                 repo.dirstate.write(None)
             else:
                 # discard all changes (including ones already written
                 # out) in this transaction
                 repo.dirstate.restorebackup(None, 'journal.dirstate')
 
                 repo.invalidate(clearfilecache=True)
 
         tr = transaction.transaction(rp, self.svfs, vfsmap,
                                      "journal",
                                      "undo",
                                      aftertrans(renames),
                                      self.store.createmode,
                                      validator=validate,
                                      releasefn=releasefn,
                                      checkambigfiles=_cachedfiles)
         tr.changes['revs'] = xrange(0, 0)
         tr.changes['obsmarkers'] = set()
         tr.changes['phases'] = {}
         tr.changes['bookmarks'] = {}
 
         tr.hookargs['txnid'] = txnid
         # note: writing the fncache only during finalize mean that the file is
         # outdated when running hooks. As fncache is used for streaming clone,
         # this is not expected to break anything that happen during the hooks.
         tr.addfinalize('flush-fncache', self.store.write)
         def txnclosehook(tr2):
             """To be run if transaction is successful, will schedule a hook run
             """
             # Don't reference tr2 in hook() so we don't hold a reference.
             # This reduces memory consumption when there are multiple
             # transactions per lock. This can likely go away if issue5045
             # fixes the function accumulation.
             hookargs = tr2.hookargs
 
             def hookfunc():
                 repo = reporef()
                 if hook.hashook(repo.ui, 'txnclose-bookmark'):
                     bmchanges = sorted(tr.changes['bookmarks'].items())
                     for name, (old, new) in bmchanges:
                         args = tr.hookargs.copy()
                         args.update(bookmarks.preparehookargs(name, old, new))
                         repo.hook('txnclose-bookmark', throw=False,
                                   txnname=desc, **pycompat.strkwargs(args))
 
                 if hook.hashook(repo.ui, 'txnclose-phase'):
                     cl = repo.unfiltered().changelog
                     phasemv = sorted(tr.changes['phases'].items())
                     for rev, (old, new) in phasemv:
                         args = tr.hookargs.copy()
                         node = hex(cl.node(rev))
                         args.update(phases.preparehookargs(node, old, new))
                         repo.hook('txnclose-phase', throw=False, txnname=desc,
                                   **pycompat.strkwargs(args))
 
                 repo.hook('txnclose', throw=False, txnname=desc,
                           **pycompat.strkwargs(hookargs))
             reporef()._afterlock(hookfunc)
         tr.addfinalize('txnclose-hook', txnclosehook)
         tr.addpostclose('warms-cache', self._buildcacheupdater(tr))
         def txnaborthook(tr2):
             """To be run if transaction is aborted
             """
             reporef().hook('txnabort', throw=False, txnname=desc,
                            **tr2.hookargs)
         tr.addabort('txnabort-hook', txnaborthook)
         # avoid eager cache invalidation. in-memory data should be identical
         # to stored data if transaction has no error.
         tr.addpostclose('refresh-filecachestats', self._refreshfilecachestats)
         self._transref = weakref.ref(tr)
         scmutil.registersummarycallback(self, tr, desc)
         return tr
 
     def _journalfiles(self):
         return ((self.svfs, 'journal'),
                 (self.vfs, 'journal.dirstate'),
                 (self.vfs, 'journal.branch'),
                 (self.vfs, 'journal.desc'),
                 (self.vfs, 'journal.bookmarks'),
                 (self.svfs, 'journal.phaseroots'))
 
     def undofiles(self):
         return [(vfs, undoname(x)) for vfs, x in self._journalfiles()]
 
     @unfilteredmethod
     def _writejournal(self, desc):
         self.dirstate.savebackup(None, 'journal.dirstate')
         self.vfs.write("journal.branch",
                           encoding.fromlocal(self.dirstate.branch()))
         self.vfs.write("journal.desc",
                           "%d\n%s\n" % (len(self), desc))
         self.vfs.write("journal.bookmarks",
                           self.vfs.tryread("bookmarks"))
         self.svfs.write("journal.phaseroots",
                            self.svfs.tryread("phaseroots"))
 
     def recover(self):
         with self.lock():
             if self.svfs.exists("journal"):
                 self.ui.status(_("rolling back interrupted transaction\n"))
                 vfsmap = {'': self.svfs,
                           'plain': self.vfs,}
                 transaction.rollback(self.svfs, vfsmap, "journal",
                                      self.ui.warn,
                                      checkambigfiles=_cachedfiles)
                 self.invalidate()
                 return True
             else:
                 self.ui.warn(_("no interrupted transaction available\n"))
                 return False
 
     def rollback(self, dryrun=False, force=False):
         wlock = lock = dsguard = None
         try:
             wlock = self.wlock()
             lock = self.lock()
             if self.svfs.exists("undo"):
                 dsguard = dirstateguard.dirstateguard(self, 'rollback')
 
                 return self._rollback(dryrun, force, dsguard)
             else:
                 self.ui.warn(_("no rollback information available\n"))
                 return 1
         finally:
             release(dsguard, lock, wlock)
 
     @unfilteredmethod # Until we get smarter cache management
     def _rollback(self, dryrun, force, dsguard):
         ui = self.ui
         try:
             args = self.vfs.read('undo.desc').splitlines()
             (oldlen, desc, detail) = (int(args[0]), args[1], None)
             if len(args) >= 3:
                 detail = args[2]
             oldtip = oldlen - 1
 
             if detail and ui.verbose:
                 msg = (_('repository tip rolled back to revision %d'
                          ' (undo %s: %s)\n')
                        % (oldtip, desc, detail))
             else:
                 msg = (_('repository tip rolled back to revision %d'
                          ' (undo %s)\n')
                        % (oldtip, desc))
         except IOError:
             msg = _('rolling back unknown transaction\n')
             desc = None
 
         if not force and self['.'] != self['tip'] and desc == 'commit':
             raise error.Abort(
                 _('rollback of last commit while not checked out '
                   'may lose data'), hint=_('use -f to force'))
 
         ui.status(msg)
         if dryrun:
             return 0
 
         parents = self.dirstate.parents()
         self.destroying()
         vfsmap = {'plain': self.vfs, '': self.svfs}
         transaction.rollback(self.svfs, vfsmap, 'undo', ui.warn,
                              checkambigfiles=_cachedfiles)
         if self.vfs.exists('undo.bookmarks'):
             self.vfs.rename('undo.bookmarks', 'bookmarks', checkambig=True)
         if self.svfs.exists('undo.phaseroots'):
             self.svfs.rename('undo.phaseroots', 'phaseroots', checkambig=True)
         self.invalidate()
 
         parentgone = (parents[0] not in self.changelog.nodemap or
                       parents[1] not in self.changelog.nodemap)
         if parentgone:
             # prevent dirstateguard from overwriting already restored one
             dsguard.close()
 
             self.dirstate.restorebackup(None, 'undo.dirstate')
             try:
                 branch = self.vfs.read('undo.branch')
                 self.dirstate.setbranch(encoding.tolocal(branch))
             except IOError:
                 ui.warn(_('named branch could not be reset: '
                           'current branch is still \'%s\'\n')
                         % self.dirstate.branch())
 
             parents = tuple([p.rev() for p in self[None].parents()])
             if len(parents) > 1:
                 ui.status(_('working directory now based on '
                             'revisions %d and %d\n') % parents)
             else:
                 ui.status(_('working directory now based on '
                             'revision %d\n') % parents)
             mergemod.mergestate.clean(self, self['.'].node())
 
         # TODO: if we know which new heads may result from this rollback, pass
         # them to destroy(), which will prevent the branchhead cache from being
         # invalidated.
         self.destroyed()
         return 0
 
     def _buildcacheupdater(self, newtransaction):
         """called during transaction to build the callback updating cache
 
         Lives on the repository to help extension who might want to augment
         this logic. For this purpose, the created transaction is passed to the
         method.
         """
         # we must avoid cyclic reference between repo and transaction.
         reporef = weakref.ref(self)
         def updater(tr):
             repo = reporef()
             repo.updatecaches(tr)
         return updater
 
     @unfilteredmethod
     def updatecaches(self, tr=None):
         """warm appropriate caches
 
         If this function is called after a transaction closed. The transaction
         will be available in the 'tr' argument. This can be used to selectively
         update caches relevant to the changes in that transaction.
         """
         if tr is not None and tr.hookargs.get('source') == 'strip':
             # During strip, many caches are invalid but
             # later call to `destroyed` will refresh them.
             return
 
         if tr is None or tr.changes['revs']:
             # updating the unfiltered branchmap should refresh all the others,
             self.ui.debug('updating the branch cache\n')
             branchmap.updatecache(self.filtered('served'))
 
     def invalidatecaches(self):
 
         if '_tagscache' in vars(self):
             # can't use delattr on proxy
             del self.__dict__['_tagscache']
 
         self.unfiltered()._branchcaches.clear()
         self.invalidatevolatilesets()
         self._sparsesignaturecache.clear()
 
     def invalidatevolatilesets(self):
         self.filteredrevcache.clear()
         obsolete.clearobscaches(self)
 
     def invalidatedirstate(self):
         '''Invalidates the dirstate, causing the next call to dirstate
         to check if it was modified since the last time it was read,
         rereading it if it has.
 
         This is different to dirstate.invalidate() that it doesn't always
         rereads the dirstate. Use dirstate.invalidate() if you want to
         explicitly read the dirstate again (i.e. restoring it to a previous
         known good state).'''
         if hasunfilteredcache(self, 'dirstate'):
             for k in self.dirstate._filecache:
                 try:
                     delattr(self.dirstate, k)
                 except AttributeError:
                     pass
             delattr(self.unfiltered(), 'dirstate')
 
     def invalidate(self, clearfilecache=False):
         '''Invalidates both store and non-store parts other than dirstate
 
         If a transaction is running, invalidation of store is omitted,
         because discarding in-memory changes might cause inconsistency
         (e.g. incomplete fncache causes unintentional failure, but
         redundant one doesn't).
         '''
         unfiltered = self.unfiltered() # all file caches are stored unfiltered
         for k in list(self._filecache.keys()):
             # dirstate is invalidated separately in invalidatedirstate()
             if k == 'dirstate':
                 continue
             if (k == 'changelog' and
                 self.currenttransaction() and
                 self.changelog._delayed):
                 # The changelog object may store unwritten revisions. We don't
                 # want to lose them.
                 # TODO: Solve the problem instead of working around it.
                 continue
 
             if clearfilecache:
                 del self._filecache[k]
             try:
                 delattr(unfiltered, k)
             except AttributeError:
                 pass
         self.invalidatecaches()
         if not self.currenttransaction():
             # TODO: Changing contents of store outside transaction
             # causes inconsistency. We should make in-memory store
             # changes detectable, and abort if changed.
             self.store.invalidatecaches()
 
     def invalidateall(self):
         '''Fully invalidates both store and non-store parts, causing the
         subsequent operation to reread any outside changes.'''
         # extension should hook this to invalidate its caches
         self.invalidate()
         self.invalidatedirstate()
 
     @unfilteredmethod
     def _refreshfilecachestats(self, tr):
         """Reload stats of cached files so that they are flagged as valid"""
         for k, ce in self._filecache.items():
             if k == 'dirstate' or k not in self.__dict__:
                 continue
             ce.refresh()
 
     def _lock(self, vfs, lockname, wait, releasefn, acquirefn, desc,
               inheritchecker=None, parentenvvar=None):
         parentlock = None
         # the contents of parentenvvar are used by the underlying lock to
         # determine whether it can be inherited
         if parentenvvar is not None:
             parentlock = encoding.environ.get(parentenvvar)
 
         timeout = 0
         warntimeout = 0
         if wait:
             timeout = self.ui.configint("ui", "timeout")
             warntimeout = self.ui.configint("ui", "timeout.warn")
 
         l = lockmod.trylock(self.ui, vfs, lockname, timeout, warntimeout,
                             releasefn=releasefn,
                             acquirefn=acquirefn, desc=desc,
                             inheritchecker=inheritchecker,
                             parentlock=parentlock)
         return l
 
     def _afterlock(self, callback):
         """add a callback to be run when the repository is fully unlocked
 
         The callback will be executed when the outermost lock is released
         (with wlock being higher level than 'lock')."""
         for ref in (self._wlockref, self._lockref):
             l = ref and ref()
             if l and l.held:
                 l.postrelease.append(callback)
                 break
         else: # no lock have been found.
             callback()
 
     def lock(self, wait=True):
         '''Lock the repository store (.hg/store) and return a weak reference
         to the lock. Use this before modifying the store (e.g. committing or
         stripping). If you are opening a transaction, get a lock as well.)
 
         If both 'lock' and 'wlock' must be acquired, ensure you always acquires
         'wlock' first to avoid a dead-lock hazard.'''
         l = self._currentlock(self._lockref)
         if l is not None:
             l.lock()
             return l
 
         l = self._lock(self.svfs, "lock", wait, None,
                        self.invalidate, _('repository %s') % self.origroot)
         self._lockref = weakref.ref(l)
         return l
 
     def _wlockchecktransaction(self):
         if self.currenttransaction() is not None:
             raise error.LockInheritanceContractViolation(
                 'wlock cannot be inherited in the middle of a transaction')
 
     def wlock(self, wait=True):
         '''Lock the non-store parts of the repository (everything under
         .hg except .hg/store) and return a weak reference to the lock.
 
         Use this before modifying files in .hg.
 
         If both 'lock' and 'wlock' must be acquired, ensure you always acquires
         'wlock' first to avoid a dead-lock hazard.'''
         l = self._wlockref and self._wlockref()
         if l is not None and l.held:
             l.lock()
             return l
 
         # We do not need to check for non-waiting lock acquisition.  Such
         # acquisition would not cause dead-lock as they would just fail.
         if wait and (self.ui.configbool('devel', 'all-warnings')
                      or self.ui.configbool('devel', 'check-locks')):
             if self._currentlock(self._lockref) is not None:
                 self.ui.develwarn('"wlock" acquired after "lock"')
 
         def unlock():
             if self.dirstate.pendingparentchange():
                 self.dirstate.invalidate()
             else:
                 self.dirstate.write(None)
 
             self._filecache['dirstate'].refresh()
 
         l = self._lock(self.vfs, "wlock", wait, unlock,
                        self.invalidatedirstate, _('working directory of %s') %
                        self.origroot,
                        inheritchecker=self._wlockchecktransaction,
                        parentenvvar='HG_WLOCK_LOCKER')
         self._wlockref = weakref.ref(l)
         return l
 
     def _currentlock(self, lockref):
         """Returns the lock if it's held, or None if it's not."""
         if lockref is None:
             return None
         l = lockref()
         if l is None or not l.held:
             return None
         return l
 
     def currentwlock(self):
         """Returns the wlock if it's held, or None if it's not."""
         return self._currentlock(self._wlockref)
 
     def _filecommit(self, fctx, manifest1, manifest2, linkrev, tr, changelist):
         """
         commit an individual file as part of a larger transaction
         """
 
         fname = fctx.path()
         fparent1 = manifest1.get(fname, nullid)
         fparent2 = manifest2.get(fname, nullid)
         if isinstance(fctx, context.filectx):
             node = fctx.filenode()
             if node in [fparent1, fparent2]:
                 self.ui.debug('reusing %s filelog entry\n' % fname)
                 if manifest1.flags(fname) != fctx.flags():
                     changelist.append(fname)
                 return node
 
         flog = self.file(fname)
         meta = {}
         copy = fctx.renamed()
         if copy and copy[0] != fname:
             # Mark the new revision of this file as a copy of another
             # file.  This copy data will effectively act as a parent
             # of this new revision.  If this is a merge, the first
             # parent will be the nullid (meaning "look up the copy data")
             # and the second one will be the other parent.  For example:
             #
             # 0 --- 1 --- 3   rev1 changes file foo
             #   \       /     rev2 renames foo to bar and changes it
             #    \- 2 -/      rev3 should have bar with all changes and
             #                      should record that bar descends from
             #                      bar in rev2 and foo in rev1
             #
             # this allows this merge to succeed:
             #
             # 0 --- 1 --- 3   rev4 reverts the content change from rev2
             #   \       /     merging rev3 and rev4 should use bar@rev2
             #    \- 2 --- 4        as the merge base
             #
 
             cfname = copy[0]
             crev = manifest1.get(cfname)
             newfparent = fparent2
 
             if manifest2: # branch merge
                 if fparent2 == nullid or crev is None: # copied on remote side
                     if cfname in manifest2:
                         crev = manifest2[cfname]
                         newfparent = fparent1
 
             # Here, we used to search backwards through history to try to find
             # where the file copy came from if the source of a copy was not in
             # the parent directory. However, this doesn't actually make sense to
             # do (what does a copy from something not in your working copy even
             # mean?) and it causes bugs (eg, issue4476). Instead, we will warn
             # the user that copy information was dropped, so if they didn't
             # expect this outcome it can be fixed, but this is the correct
             # behavior in this circumstance.
 
             if crev:
                 self.ui.debug(" %s: copy %s:%s\n" % (fname, cfname, hex(crev)))
                 meta["copy"] = cfname
                 meta["copyrev"] = hex(crev)
                 fparent1, fparent2 = nullid, newfparent
             else:
                 self.ui.warn(_("warning: can't find ancestor for '%s' "
                                "copied from '%s'!\n") % (fname, cfname))
 
         elif fparent1 == nullid:
             fparent1, fparent2 = fparent2, nullid
         elif fparent2 != nullid:
             # is one parent an ancestor of the other?
             fparentancestors = flog.commonancestorsheads(fparent1, fparent2)
             if fparent1 in fparentancestors:
                 fparent1, fparent2 = fparent2, nullid
             elif fparent2 in fparentancestors:
                 fparent2 = nullid
 
         # is the file changed?
         text = fctx.data()
         if fparent2 != nullid or flog.cmp(fparent1, text) or meta:
             changelist.append(fname)
             return flog.add(text, meta, tr, linkrev, fparent1, fparent2)
         # are just the flags changed during merge?
         elif fname in manifest1 and manifest1.flags(fname) != fctx.flags():
             changelist.append(fname)
 
         return fparent1
 
     def checkcommitpatterns(self, wctx, vdirs, match, status, fail):
         """check for commit arguments that aren't committable"""
         if match.isexact() or match.prefix():
             matched = set(status.modified + status.added + status.removed)
 
             for f in match.files():
                 f = self.dirstate.normalize(f)
                 if f == '.' or f in matched or f in wctx.substate:
                     continue
                 if f in status.deleted:
                     fail(f, _('file not found!'))
                 if f in vdirs: # visited directory
                     d = f + '/'
                     for mf in matched:
                         if mf.startswith(d):
                             break
                     else:
                         fail(f, _("no match under directory!"))
                 elif f not in self.dirstate:
                     fail(f, _("file not tracked!"))
 
     @unfilteredmethod
     def commit(self, text="", user=None, date=None, match=None, force=False,
                editor=False, extra=None):
         """Add a new revision to current repository.
 
         Revision information is gathered from the working directory,
         match can be used to filter the committed files. If editor is
         supplied, it is called to get a commit message.
         """
         if extra is None:
             extra = {}
 
         def fail(f, msg):
             raise error.Abort('%s: %s' % (f, msg))
 
         if not match:
             match = matchmod.always(self.root, '')
 
         if not force:
             vdirs = []
             match.explicitdir = vdirs.append
             match.bad = fail
 
         wlock = lock = tr = None
         try:
             wlock = self.wlock()
             lock = self.lock() # for recent changelog (see issue4368)
 
             wctx = self[None]
             merge = len(wctx.parents()) > 1
 
             if not force and merge and not match.always():
                 raise error.Abort(_('cannot partially commit a merge '
                                    '(do not specify files or patterns)'))
 
             status = self.status(match=match, clean=force)
             if force:
                 status.modified.extend(status.clean) # mq may commit clean files
 
             # check subrepos
             subs, commitsubs, newstate = subrepo.precommit(
                 self.ui, wctx, status, match, force=force)
 
             # make sure all explicit patterns are matched
             if not force:
                 self.checkcommitpatterns(wctx, vdirs, match, status, fail)
 
             cctx = context.workingcommitctx(self, status,
                                             text, user, date, extra)
 
             # internal config: ui.allowemptycommit
             allowemptycommit = (wctx.branch() != wctx.p1().branch()
                                 or extra.get('close') or merge or cctx.files()
                                 or self.ui.configbool('ui', 'allowemptycommit'))
             if not allowemptycommit:
                 return None
 
             if merge and cctx.deleted():
                 raise error.Abort(_("cannot commit merge with missing files"))
 
             ms = mergemod.mergestate.read(self)
             mergeutil.checkunresolved(ms)
 
             if editor:
                 cctx._text = editor(self, cctx, subs)
             edited = (text != cctx._text)
 
             # Save commit message in case this transaction gets rolled back
             # (e.g. by a pretxncommit hook).  Leave the content alone on
             # the assumption that the user will use the same editor again.
             msgfn = self.savecommitmessage(cctx._text)
 
             # commit subs and write new state
             if subs:
                 for s in sorted(commitsubs):
                     sub = wctx.sub(s)
                     self.ui.status(_('committing subrepository %s\n') %
                         subrepo.subrelpath(sub))
                     sr = sub.commit(cctx._text, user, date)
                     newstate[s] = (newstate[s][0], sr)
                 subrepo.writestate(self, newstate)
 
             p1, p2 = self.dirstate.parents()
             hookp1, hookp2 = hex(p1), (p2 != nullid and hex(p2) or '')
             try:
                 self.hook("precommit", throw=True, parent1=hookp1,
                           parent2=hookp2)
                 tr = self.transaction('commit')
                 ret = self.commitctx(cctx, True)
             except: # re-raises
                 if edited:
                     self.ui.write(
                         _('note: commit message saved in %s\n') % msgfn)
                 raise
             # update bookmarks, dirstate and mergestate
             bookmarks.update(self, [p1, p2], ret)
             cctx.markcommitted(ret)
             ms.reset()
             tr.close()
 
         finally:
             lockmod.release(tr, lock, wlock)
 
         def commithook(node=hex(ret), parent1=hookp1, parent2=hookp2):
             # hack for command that use a temporary commit (eg: histedit)
             # temporary commit got stripped before hook release
             if self.changelog.hasnode(ret):
                 self.hook("commit", node=node, parent1=parent1,
                           parent2=parent2)
         self._afterlock(commithook)
         return ret
 
     @unfilteredmethod
     def commitctx(self, ctx, error=False):
         """Add a new revision to current repository.
         Revision information is passed via the context argument.
         """
 
         tr = None
         p1, p2 = ctx.p1(), ctx.p2()
         user = ctx.user()
 
         lock = self.lock()
         try:
             tr = self.transaction("commit")
             trp = weakref.proxy(tr)
 
             if ctx.manifestnode():
                 # reuse an existing manifest revision
                 mn = ctx.manifestnode()
                 files = ctx.files()
             elif ctx.files():
                 m1ctx = p1.manifestctx()
                 m2ctx = p2.manifestctx()
                 mctx = m1ctx.copy()
 
                 m = mctx.read()
                 m1 = m1ctx.read()
                 m2 = m2ctx.read()
 
                 # check in files
                 added = []
                 changed = []
                 removed = list(ctx.removed())
                 linkrev = len(self)
                 self.ui.note(_("committing files:\n"))
                 for f in sorted(ctx.modified() + ctx.added()):
                     self.ui.note(f + "\n")
                     try:
                         fctx = ctx[f]
                         if fctx is None:
                             removed.append(f)
                         else:
                             added.append(f)
                             m[f] = self._filecommit(fctx, m1, m2, linkrev,
                                                     trp, changed)
                             m.setflag(f, fctx.flags())
                     except OSError as inst:
                         self.ui.warn(_("trouble committing %s!\n") % f)
                         raise
                     except IOError as inst:
                         errcode = getattr(inst, 'errno', errno.ENOENT)
                         if error or errcode and errcode != errno.ENOENT:
                             self.ui.warn(_("trouble committing %s!\n") % f)
                         raise
 
                 # update manifest
                 self.ui.note(_("committing manifest\n"))
                 removed = [f for f in sorted(removed) if f in m1 or f in m2]
                 drop = [f for f in removed if f in m]
                 for f in drop:
                     del m[f]
                 mn = mctx.write(trp, linkrev,
                                 p1.manifestnode(), p2.manifestnode(),
                                 added, drop)
                 files = changed + removed
             else:
                 mn = p1.manifestnode()
                 files = []
 
             # update changelog
             self.ui.note(_("committing changelog\n"))
             self.changelog.delayupdate(tr)
             n = self.changelog.add(mn, files, ctx.description(),
                                    trp, p1.node(), p2.node(),
                                    user, ctx.date(), ctx.extra().copy())
             xp1, xp2 = p1.hex(), p2 and p2.hex() or ''
             self.hook('pretxncommit', throw=True, node=hex(n), parent1=xp1,
                       parent2=xp2)
             # set the new commit is proper phase
             targetphase = subrepo.newcommitphase(self.ui, ctx)
             if targetphase:
                 # retract boundary do not alter parent changeset.
                 # if a parent have higher the resulting phase will
                 # be compliant anyway
                 #
                 # if minimal phase was 0 we don't need to retract anything
                 phases.registernew(self, tr, targetphase, [n])
             tr.close()
             return n
         finally:
             if tr:
                 tr.release()
             lock.release()
 
     @unfilteredmethod
     def destroying(self):
         '''Inform the repository that nodes are about to be destroyed.
         Intended for use by strip and rollback, so there's a common
         place for anything that has to be done before destroying history.
 
         This is mostly useful for saving state that is in memory and waiting
         to be flushed when the current lock is released. Because a call to
         destroyed is imminent, the repo will be invalidated causing those
         changes to stay in memory (waiting for the next unlock), or vanish
         completely.
         '''
         # When using the same lock to commit and strip, the phasecache is left
         # dirty after committing. Then when we strip, the repo is invalidated,
         # causing those changes to disappear.
         if '_phasecache' in vars(self):
             self._phasecache.write()
 
     @unfilteredmethod
     def destroyed(self):
         '''Inform the repository that nodes have been destroyed.
         Intended for use by strip and rollback, so there's a common
         place for anything that has to be done after destroying history.
         '''
         # When one tries to:
         # 1) destroy nodes thus calling this method (e.g. strip)
         # 2) use phasecache somewhere (e.g. commit)
         #
         # then 2) will fail because the phasecache contains nodes that were
         # removed. We can either remove phasecache from the filecache,
         # causing it to reload next time it is accessed, or simply filter
         # the removed nodes now and write the updated cache.
         self._phasecache.filterunknown(self)
         self._phasecache.write()
 
         # refresh all repository caches
         self.updatecaches()
 
         # Ensure the persistent tag cache is updated.  Doing it now
         # means that the tag cache only has to worry about destroyed
         # heads immediately after a strip/rollback.  That in turn
         # guarantees that "cachetip == currenttip" (comparing both rev
         # and node) always means no nodes have been added or destroyed.
 
         # XXX this is suboptimal when qrefresh'ing: we strip the current
         # head, refresh the tag cache, then immediately add a new head.
         # But I think doing it this way is necessary for the "instant
         # tag cache retrieval" case to work.
         self.invalidate()
 
     def walk(self, match, node=None):
         '''
         walk recursively through the directory tree or a given
         changeset, finding all files matched by the match
         function
         '''
         self.ui.deprecwarn('use repo[node].walk instead of repo.walk', '4.3')
         return self[node].walk(match)
 
     def status(self, node1='.', node2=None, match=None,
                ignored=False, clean=False, unknown=False,
                listsubrepos=False):
         '''a convenience method that calls node1.status(node2)'''
         return self[node1].status(node2, match, ignored, clean, unknown,
                                   listsubrepos)
 
     def addpostdsstatus(self, ps):
         """Add a callback to run within the wlock, at the point at which status
         fixups happen.
 
         On status completion, callback(wctx, status) will be called with the
         wlock held, unless the dirstate has changed from underneath or the wlock
         couldn't be grabbed.
 
         Callbacks should not capture and use a cached copy of the dirstate --
         it might change in the meanwhile. Instead, they should access the
         dirstate via wctx.repo().dirstate.
 
         This list is emptied out after each status run -- extensions should
         make sure it adds to this list each time dirstate.status is called.
         Extensions should also make sure they don't call this for statuses
         that don't involve the dirstate.
         """
 
         # The list is located here for uniqueness reasons -- it is actually
         # managed by the workingctx, but that isn't unique per-repo.
         self._postdsstatus.append(ps)
 
     def postdsstatus(self):
         """Used by workingctx to get the list of post-dirstate-status hooks."""
         return self._postdsstatus
 
     def clearpostdsstatus(self):
         """Used by workingctx to clear post-dirstate-status hooks."""
         del self._postdsstatus[:]
 
     def heads(self, start=None):
         if start is None:
             cl = self.changelog
             headrevs = reversed(cl.headrevs())
             return [cl.node(rev) for rev in headrevs]
 
         heads = self.changelog.heads(start)
         # sort the output in rev descending order
         return sorted(heads, key=self.changelog.rev, reverse=True)
 
     def branchheads(self, branch=None, start=None, closed=False):
         '''return a (possibly filtered) list of heads for the given branch
 
         Heads are returned in topological order, from newest to oldest.
         If branch is None, use the dirstate branch.
         If start is not None, return only heads reachable from start.
         If closed is True, return heads that are marked as closed as well.
         '''
         if branch is None:
             branch = self[None].branch()
         branches = self.branchmap()
         if branch not in branches:
             return []
         # the cache returns heads ordered lowest to highest
         bheads = list(reversed(branches.branchheads(branch, closed=closed)))
         if start is not None:
             # filter out the heads that cannot be reached from startrev
             fbheads = set(self.changelog.nodesbetween([start], bheads)[2])
             bheads = [h for h in bheads if h in fbheads]
         return bheads
 
     def branches(self, nodes):
         if not nodes:
             nodes = [self.changelog.tip()]
         b = []
         for n in nodes:
             t = n
             while True:
                 p = self.changelog.parents(n)
                 if p[1] != nullid or p[0] == nullid:
                     b.append((t, n, p[0], p[1]))
                     break
                 n = p[0]
         return b
 
     def between(self, pairs):
         r = []
 
         for top, bottom in pairs:
             n, l, i = top, [], 0
             f = 1
 
             while n != bottom and n != nullid:
                 p = self.changelog.parents(n)[0]
                 if i == f:
                     l.append(n)
                     f = f * 2
                 n = p
                 i += 1
 
             r.append(l)
 
         return r
 
     def checkpush(self, pushop):
         """Extensions can override this function if additional checks have
         to be performed before pushing, or call it if they override push
         command.
         """
 
     @unfilteredpropertycache
     def prepushoutgoinghooks(self):
         """Return util.hooks consists of a pushop with repo, remote, outgoing
         methods, which are called before pushing changesets.
         """
         return util.hooks()
 
     def pushkey(self, namespace, key, old, new):
         try:
             tr = self.currenttransaction()
             hookargs = {}
             if tr is not None:
                 hookargs.update(tr.hookargs)
             hookargs['namespace'] = namespace
             hookargs['key'] = key
             hookargs['old'] = old
             hookargs['new'] = new
             self.hook('prepushkey', throw=True, **hookargs)
         except error.HookAbort as exc:
             self.ui.write_err(_("pushkey-abort: %s\n") % exc)
             if exc.hint:
                 self.ui.write_err(_("(%s)\n") % exc.hint)
             return False
         self.ui.debug('pushing key for "%s:%s"\n' % (namespace, key))
         ret = pushkey.push(self, namespace, key, old, new)
         def runhook():
             self.hook('pushkey', namespace=namespace, key=key, old=old, new=new,
                       ret=ret)
         self._afterlock(runhook)
         return ret
 
     def listkeys(self, namespace):
         self.hook('prelistkeys', throw=True, namespace=namespace)
         self.ui.debug('listing keys for "%s"\n' % namespace)
         values = pushkey.list(self, namespace)
         self.hook('listkeys', namespace=namespace, values=values)
         return values
 
     def debugwireargs(self, one, two, three=None, four=None, five=None):
         '''used to test argument passing over the wire'''
         return "%s %s %s %s %s" % (one, two, three, four, five)
 
     def savecommitmessage(self, text):
         fp = self.vfs('last-message.txt', 'wb')
         try:
             fp.write(text)
         finally:
             fp.close()
         return self.pathto(fp.name[len(self.root) + 1:])
 
 # used to avoid circular references so destructors work
 def aftertrans(files):
     renamefiles = [tuple(t) for t in files]
     def a():
         for vfs, src, dest in renamefiles:
             # if src and dest refer to a same file, vfs.rename is a no-op,
             # leaving both src and dest on disk. delete dest to make sure
             # the rename couldn't be such a no-op.
             vfs.tryunlink(dest)
             try:
                 vfs.rename(src, dest)
             except OSError: # journal file does not yet exist
                 pass
     return a
 
 def undoname(fn):
     base, name = os.path.split(fn)
     assert name.startswith('journal')
     return os.path.join(base, name.replace('journal', 'undo', 1))
 
 def instance(ui, path, create):
     return localrepository(ui, util.urllocalpath(path), create)
 
 def islocal(path):
     return True
 
 def newreporequirements(repo):
     """Determine the set of requirements for a new local repository.
 
     Extensions can wrap this function to specify custom requirements for
     new repositories.
     """
     ui = repo.ui
     requirements = {'revlogv1'}
     if ui.configbool('format', 'usestore'):
         requirements.add('store')
         if ui.configbool('format', 'usefncache'):
             requirements.add('fncache')
             if ui.configbool('format', 'dotencode'):
                 requirements.add('dotencode')
 
     compengine = ui.config('experimental', 'format.compression')
     if compengine not in util.compengines:
         raise error.Abort(_('compression engine %s defined by '
                             'experimental.format.compression not available') %
                           compengine,
                           hint=_('run "hg debuginstall" to list available '
                                  'compression engines'))
 
     # zlib is the historical default and doesn't need an explicit requirement.
     if compengine != 'zlib':
         requirements.add('exp-compression-%s' % compengine)
 
     if scmutil.gdinitconfig(ui):
         requirements.add('generaldelta')
     if ui.configbool('experimental', 'treemanifest'):
         requirements.add('treemanifest')
     if ui.configbool('experimental', 'manifestv2'):
         requirements.add('manifestv2')
 
     revlogv2 = ui.config('experimental', 'revlogv2')
     if revlogv2 == 'enable-unstable-format-and-corrupt-my-data':
         requirements.remove('revlogv1')
         # generaldelta is implied by revlogv2.
         requirements.discard('generaldelta')
         requirements.add(REVLOGV2_REQUIREMENT)
 
     return requirements

mercurial/phases.py

@@ -1,666 +1,674
 """ Mercurial phases support code
 
     ---
 
     Copyright 2011 Pierre-Yves David <[e-mail]>
                    Logilab SA        <[e-mail]>
                    Augie Fackler     <[e-mail]>
 
     This software may be used and distributed according to the terms
     of the GNU General Public License version 2 or any later version.
 
     ---
 
 This module implements most phase logic in mercurial.
 
 
 Basic Concept
 =============
 
 A 'changeset phase' is an indicator that tells us how a changeset is
 manipulated and communicated. The details of each phase is described
 below, here we describe the properties they have in common.
 
 Like bookmarks, phases are not stored in history and thus are not
 permanent and leave no audit trail.
 
 First, no changeset can be in two phases at once. Phases are ordered,
 so they can be considered from lowest to highest. The default, lowest
 phase is 'public' - this is the normal phase of existing changesets. A
 child changeset can not be in a lower phase than its parents.
 
 These phases share a hierarchy of traits:
 
             immutable shared
     public:     X        X
     draft:               X
     secret:
 
 Local commits are draft by default.
 
 Phase Movement and Exchange
 ===========================
 
 Phase data is exchanged by pushkey on pull and push. Some servers have
 a publish option set, we call such a server a "publishing server".
 Pushing a draft changeset to a publishing server changes the phase to
 public.
 
 A small list of fact/rules define the exchange of phase:
 
 * old client never changes server states
 * pull never changes server states
 * publish and old server changesets are seen as public by client
 * any secret changeset seen in another repository is lowered to at
   least draft
 
 Here is the final table summing up the 49 possible use cases of phase
 exchange:
 
                            server
                   old     publish      non-publish
                  N   X    N   D   P    N   D   P
     old client
     pull
      N           -   X/X  -   X/D X/P  -   X/D X/P
      X           -   X/X  -   X/D X/P  -   X/D X/P
     push
      X           X/X X/X  X/P X/P X/P  X/D X/D X/P
     new client
     pull
      N           -   P/X  -   P/D P/P  -   D/D P/P
      D           -   P/X  -   P/D P/P  -   D/D P/P
      P           -   P/X  -   P/D P/P  -   P/D P/P
     push
      D           P/X P/X  P/P P/P P/P  D/D D/D P/P
      P           P/X P/X  P/P P/P P/P  P/P P/P P/P
 
 Legend:
 
     A/B = final state on client / state on server
 
     * N = new/not present,
     * P = public,
     * D = draft,
     * X = not tracked (i.e., the old client or server has no internal
           way of recording the phase.)
 
     passive = only pushes
 
 
     A cell here can be read like this:
 
     "When a new client pushes a draft changeset (D) to a publishing
     server where it's not present (N), it's marked public on both
     sides (P/P)."
 
 Note: old client behave as a publishing server with draft only content
 - other people see it as public
 - content is pushed as draft
 
 """
 
 from __future__ import absolute_import
 
 import errno
 import struct
 
 from .i18n import _
 from .node import (
     bin,
     hex,
     nullid,
     nullrev,
     short,
 )
 from . import (
     error,
+    pycompat,
     smartset,
     txnutil,
     util,
 )
 
 _fphasesentry = struct.Struct('>i20s')
 
 allphases = public, draft, secret = range(3)
 trackedphases = allphases[1:]
 phasenames = ['public', 'draft', 'secret']
 
 def _readroots(repo, phasedefaults=None):
     """Read phase roots from disk
 
     phasedefaults is a list of fn(repo, roots) callable, which are
     executed if the phase roots file does not exist. When phases are
     being initialized on an existing repository, this could be used to
     set selected changesets phase to something else than public.
 
     Return (roots, dirty) where dirty is true if roots differ from
     what is being stored.
     """
     repo = repo.unfiltered()
     dirty = False
     roots = [set() for i in allphases]
     try:
         f, pending = txnutil.trypending(repo.root, repo.svfs, 'phaseroots')
         try:
             for line in f:
                 phase, nh = line.split()
                 roots[int(phase)].add(bin(nh))
         finally:
             f.close()
     except IOError as inst:
         if inst.errno != errno.ENOENT:
             raise
         if phasedefaults:
             for f in phasedefaults:
                 roots = f(repo, roots)
         dirty = True
     return roots, dirty
 
 def binaryencode(phasemapping):
     """encode a 'phase -> nodes' mapping into a binary stream
 
     Since phases are integer the mapping is actually a python list:
     [[PUBLIC_HEADS], [DRAFTS_HEADS], [SECRET_HEADS]]
     """
     binarydata = []
     for phase, nodes in enumerate(phasemapping):
         for head in nodes:
             binarydata.append(_fphasesentry.pack(phase, head))
     return ''.join(binarydata)
 
 def binarydecode(stream):
     """decode a binary stream into a 'phase -> nodes' mapping
 
     Since phases are integer the mapping is actually a python list."""
     headsbyphase = [[] for i in allphases]
     entrysize = _fphasesentry.size
     while True:
         entry = stream.read(entrysize)
         if len(entry) < entrysize:
             if entry:
                 raise error.Abort(_('bad phase-heads stream'))
             break
         phase, node = _fphasesentry.unpack(entry)
         headsbyphase[phase].append(node)
     return headsbyphase
 
 def _trackphasechange(data, rev, old, new):
     """add a phase move the <data> dictionnary
 
     If data is None, nothing happens.
     """
     if data is None:
         return
     existing = data.get(rev)
     if existing is not None:
         old = existing[0]
     data[rev] = (old, new)
 
 class phasecache(object):
     def __init__(self, repo, phasedefaults, _load=True):
         if _load:
             # Cheap trick to allow shallow-copy without copy module
             self.phaseroots, self.dirty = _readroots(repo, phasedefaults)
-            self._phaserevs = None
+            self._phasemaxrev = nullrev
             self._phasesets = None
             self.filterunknown(repo)
             self.opener = repo.svfs
 
     def getrevset(self, repo, phases):
         """return a smartset for the given phases"""
         self.loadphaserevs(repo) # ensure phase's sets are loaded
-
-        if self._phasesets and all(self._phasesets[p] is not None
-                                   for p in phases):
-            # fast path - use _phasesets
-            revs = self._phasesets[phases[0]]
-            if len(phases) > 1:
-                revs = revs.copy() # only copy when needed
-                for p in phases[1:]:
-                    revs.update(self._phasesets[p])
+        phases = set(phases)
+        if public not in phases:
+            # fast path: _phasesets contains the interesting sets,
+            # might only need a union and post-filtering.
+            if len(phases) == 1:
+                [p] = phases
+                revs = self._phasesets[p]
+            else:
+                revs = set.union(*[self._phasesets[p] for p in phases])
             if repo.changelog.filteredrevs:
                 revs = revs - repo.changelog.filteredrevs
             return smartset.baseset(revs)
         else:
-            # slow path - enumerate all revisions
-            phase = self.phase
-            revs = (r for r in repo if phase(repo, r) in phases)
-            return smartset.generatorset(revs, iterasc=True)
+            phases = set(allphases).difference(phases)
+            if not phases:
+                return smartset.fullreposet(repo)
+            if len(phases) == 1:
+                [p] = phases
+                revs = self._phasesets[p]
+            else:
+                revs = set.union(*[self._phasesets[p] for p in phases])
+            if not revs:
+                return smartset.fullreposet(repo)
+            return smartset.fullreposet(repo).filter(lambda r: r not in revs)
 
     def copy(self):
         # Shallow copy meant to ensure isolation in
         # advance/retractboundary(), nothing more.
         ph = self.__class__(None, None, _load=False)
         ph.phaseroots = self.phaseroots[:]
         ph.dirty = self.dirty
         ph.opener = self.opener
-        ph._phaserevs = self._phaserevs
+        ph._phasemaxrev = self._phasemaxrev
         ph._phasesets = self._phasesets
         return ph
 
     def replace(self, phcache):
         """replace all values in 'self' with content of phcache"""
-        for a in ('phaseroots', 'dirty', 'opener', '_phaserevs', '_phasesets'):
+        for a in ('phaseroots', 'dirty', 'opener', '_phasemaxrev',
+                  '_phasesets'):
             setattr(self, a, getattr(phcache, a))
 
     def _getphaserevsnative(self, repo):
         repo = repo.unfiltered()
         nativeroots = []
         for phase in trackedphases:
             nativeroots.append(map(repo.changelog.rev, self.phaseroots[phase]))
         return repo.changelog.computephases(nativeroots)
 
     def _computephaserevspure(self, repo):
         repo = repo.unfiltered()
-        revs = [public] * len(repo.changelog)
-        self._phaserevs = revs
-        self._populatephaseroots(repo)
-        for phase in trackedphases:
-            roots = list(map(repo.changelog.rev, self.phaseroots[phase]))
-            if roots:
-                for rev in roots:
-                    revs[rev] = phase
-                for rev in repo.changelog.descendants(roots):
-                    revs[rev] = phase
+        cl = repo.changelog
+        self._phasesets = [set() for phase in allphases]
+        roots = pycompat.maplist(cl.rev, self.phaseroots[secret])
+        if roots:
+            ps = set(cl.descendants(roots))
+            for root in roots:
+                ps.add(root)
+            self._phasesets[secret] = ps
+        roots = pycompat.maplist(cl.rev, self.phaseroots[draft])
+        if roots:
+            ps = set(cl.descendants(roots))
+            for root in roots:
+                ps.add(root)
+            ps.difference_update(self._phasesets[secret])
+            self._phasesets[draft] = ps
+        self._phasemaxrev = len(cl)
 
     def loadphaserevs(self, repo):
         """ensure phase information is loaded in the object"""
-        if self._phaserevs is None:
+        if self._phasesets is None:
             try:
                 res = self._getphaserevsnative(repo)
-                self._phaserevs, self._phasesets = res
+                self._phasemaxrev, self._phasesets = res
             except AttributeError:
                 self._computephaserevspure(repo)
 
     def invalidate(self):
-        self._phaserevs = None
+        self._phasemaxrev = nullrev
         self._phasesets = None
 
-    def _populatephaseroots(self, repo):
-        """Fills the _phaserevs cache with phases for the roots.
-        """
-        cl = repo.changelog
-        phaserevs = self._phaserevs
-        for phase in trackedphases:
-            roots = map(cl.rev, self.phaseroots[phase])
-            for root in roots:
-                phaserevs[root] = phase
-
     def phase(self, repo, rev):
-        # We need a repo argument here to be able to build _phaserevs
+        # We need a repo argument here to be able to build _phasesets
         # if necessary. The repository instance is not stored in
         # phasecache to avoid reference cycles. The changelog instance
         # is not stored because it is a filecache() property and can
         # be replaced without us being notified.
         if rev == nullrev:
             return public
         if rev < nullrev:
             raise ValueError(_('cannot lookup negative revision'))
-        if self._phaserevs is None or rev >= len(self._phaserevs):
+        if rev >= self._phasemaxrev:
             self.invalidate()
             self.loadphaserevs(repo)
-        return self._phaserevs[rev]
+        for phase in trackedphases:
+            if rev in self._phasesets[phase]:
+                return phase
+        return public
 
     def write(self):
         if not self.dirty:
             return
         f = self.opener('phaseroots', 'w', atomictemp=True, checkambig=True)
         try:
             self._write(f)
         finally:
             f.close()
 
     def _write(self, fp):
         for phase, roots in enumerate(self.phaseroots):
             for h in roots:
                 fp.write('%i %s\n' % (phase, hex(h)))
         self.dirty = False
 
     def _updateroots(self, phase, newroots, tr):
         self.phaseroots[phase] = newroots
         self.invalidate()
         self.dirty = True
 
         tr.addfilegenerator('phase', ('phaseroots',), self._write)
         tr.hookargs['phases_moved'] = '1'
 
     def registernew(self, repo, tr, targetphase, nodes):
         repo = repo.unfiltered()
         self._retractboundary(repo, tr, targetphase, nodes)
         if tr is not None and 'phases' in tr.changes:
             phasetracking = tr.changes['phases']
             torev = repo.changelog.rev
             phase = self.phase
             for n in nodes:
                 rev = torev(n)
                 revphase = phase(repo, rev)
                 _trackphasechange(phasetracking, rev, None, revphase)
         repo.invalidatevolatilesets()
 
     def advanceboundary(self, repo, tr, targetphase, nodes):
         """Set all 'nodes' to phase 'targetphase'
 
         Nodes with a phase lower than 'targetphase' are not affected.
         """
         # Be careful to preserve shallow-copied values: do not update
         # phaseroots values, replace them.
         if tr is None:
             phasetracking = None
         else:
             phasetracking = tr.changes.get('phases')
 
         repo = repo.unfiltered()
 
         delroots = [] # set of root deleted by this path
         for phase in xrange(targetphase + 1, len(allphases)):
             # filter nodes that are not in a compatible phase already
             nodes = [n for n in nodes
                      if self.phase(repo, repo[n].rev()) >= phase]
             if not nodes:
                 break # no roots to move anymore
 
             olds = self.phaseroots[phase]
 
             affected = repo.revs('%ln::%ln', olds, nodes)
             for r in affected:
                 _trackphasechange(phasetracking, r, self.phase(repo, r),
                                   targetphase)
 
             roots = set(ctx.node() for ctx in repo.set(
                     'roots((%ln::) - %ld)', olds, affected))
             if olds != roots:
                 self._updateroots(phase, roots, tr)
                 # some roots may need to be declared for lower phases
                 delroots.extend(olds - roots)
         # declare deleted root in the target phase
         if targetphase != 0:
             self._retractboundary(repo, tr, targetphase, delroots)
         repo.invalidatevolatilesets()
 
     def retractboundary(self, repo, tr, targetphase, nodes):
         oldroots = self.phaseroots[:targetphase + 1]
         if tr is None:
             phasetracking = None
         else:
             phasetracking = tr.changes.get('phases')
         repo = repo.unfiltered()
         if (self._retractboundary(repo, tr, targetphase, nodes)
             and phasetracking is not None):
 
             # find the affected revisions
             new = self.phaseroots[targetphase]
             old = oldroots[targetphase]
             affected = set(repo.revs('(%ln::) - (%ln::)', new, old))
 
             # find the phase of the affected revision
             for phase in xrange(targetphase, -1, -1):
                 if phase:
                     roots = oldroots[phase]
                     revs = set(repo.revs('%ln::%ld', roots, affected))
                     affected -= revs
                 else: # public phase
                     revs = affected
                 for r in revs:
                     _trackphasechange(phasetracking, r, phase, targetphase)
         repo.invalidatevolatilesets()
 
     def _retractboundary(self, repo, tr, targetphase, nodes):
         # Be careful to preserve shallow-copied values: do not update
         # phaseroots values, replace them.
 
         repo = repo.unfiltered()
         currentroots = self.phaseroots[targetphase]
         finalroots = oldroots = set(currentroots)
         newroots = [n for n in nodes
                     if self.phase(repo, repo[n].rev()) < targetphase]
         if newroots:
 
             if nullid in newroots:
                 raise error.Abort(_('cannot change null revision phase'))
             currentroots = currentroots.copy()
             currentroots.update(newroots)
 
             # Only compute new roots for revs above the roots that are being
             # retracted.
             minnewroot = min(repo[n].rev() for n in newroots)
             aboveroots = [n for n in currentroots
                           if repo[n].rev() >= minnewroot]
             updatedroots = repo.set('roots(%ln::)', aboveroots)
 
             finalroots = set(n for n in currentroots if repo[n].rev() <
                              minnewroot)
             finalroots.update(ctx.node() for ctx in updatedroots)
         if finalroots != oldroots:
             self._updateroots(targetphase, finalroots, tr)
             return True
         return False
 
     def filterunknown(self, repo):
         """remove unknown nodes from the phase boundary
 
         Nothing is lost as unknown nodes only hold data for their descendants.
         """
         filtered = False
         nodemap = repo.changelog.nodemap # to filter unknown nodes
         for phase, nodes in enumerate(self.phaseroots):
             missing = sorted(node for node in nodes if node not in nodemap)
             if missing:
                 for mnode in missing:
                     repo.ui.debug(
                         'removing unknown node %s from %i-phase boundary\n'
                         % (short(mnode), phase))
                 nodes.symmetric_difference_update(missing)
                 filtered = True
         if filtered:
             self.dirty = True
         # filterunknown is called by repo.destroyed, we may have no changes in
-        # root but phaserevs contents is certainly invalid (or at least we
+        # root but _phasesets contents is certainly invalid (or at least we
         # have not proper way to check that). related to issue 3858.
         #
-        # The other caller is __init__ that have no _phaserevs initialized
+        # The other caller is __init__ that have no _phasesets initialized
         # anyway. If this change we should consider adding a dedicated
         # "destroyed" function to phasecache or a proper cache key mechanism
         # (see branchmap one)
         self.invalidate()
 
 def advanceboundary(repo, tr, targetphase, nodes):
     """Add nodes to a phase changing other nodes phases if necessary.
 
     This function move boundary *forward* this means that all nodes
     are set in the target phase or kept in a *lower* phase.
 
     Simplify boundary to contains phase roots only."""
     phcache = repo._phasecache.copy()
     phcache.advanceboundary(repo, tr, targetphase, nodes)
     repo._phasecache.replace(phcache)
 
 def retractboundary(repo, tr, targetphase, nodes):
     """Set nodes back to a phase changing other nodes phases if
     necessary.
 
     This function move boundary *backward* this means that all nodes
     are set in the target phase or kept in a *higher* phase.
 
     Simplify boundary to contains phase roots only."""
     phcache = repo._phasecache.copy()
     phcache.retractboundary(repo, tr, targetphase, nodes)
     repo._phasecache.replace(phcache)
 
 def registernew(repo, tr, targetphase, nodes):
     """register a new revision and its phase
 
     Code adding revisions to the repository should use this function to
     set new changeset in their target phase (or higher).
     """
     phcache = repo._phasecache.copy()
     phcache.registernew(repo, tr, targetphase, nodes)
     repo._phasecache.replace(phcache)
 
 def listphases(repo):
     """List phases root for serialization over pushkey"""
     # Use ordered dictionary so behavior is deterministic.
     keys = util.sortdict()
     value = '%i' % draft
     cl = repo.unfiltered().changelog
     for root in repo._phasecache.phaseroots[draft]:
         if repo._phasecache.phase(repo, cl.rev(root)) <= draft:
             keys[hex(root)] = value
 
     if repo.publishing():
         # Add an extra data to let remote know we are a publishing
         # repo. Publishing repo can't just pretend they are old repo.
         # When pushing to a publishing repo, the client still need to
         # push phase boundary
         #
         # Push do not only push changeset. It also push phase data.
         # New phase data may apply to common changeset which won't be
         # push (as they are common). Here is a very simple example:
         #
         # 1) repo A push changeset X as draft to repo B
         # 2) repo B make changeset X public
         # 3) repo B push to repo A. X is not pushed but the data that
         #    X as now public should
         #
         # The server can't handle it on it's own as it has no idea of
         # client phase data.
         keys['publishing'] = 'True'
     return keys
 
 def pushphase(repo, nhex, oldphasestr, newphasestr):
     """List phases root for serialization over pushkey"""
     repo = repo.unfiltered()
     with repo.lock():
         currentphase = repo[nhex].phase()
         newphase = abs(int(newphasestr)) # let's avoid negative index surprise
         oldphase = abs(int(oldphasestr)) # let's avoid negative index surprise
         if currentphase == oldphase and newphase < oldphase:
             with repo.transaction('pushkey-phase') as tr:
                 advanceboundary(repo, tr, newphase, [bin(nhex)])
             return True
         elif currentphase == newphase:
             # raced, but got correct result
             return True
         else:
             return False
 
 def subsetphaseheads(repo, subset):
     """Finds the phase heads for a subset of a history
 
     Returns a list indexed by phase number where each item is a list of phase
     head nodes.
     """
     cl = repo.changelog
 
     headsbyphase = [[] for i in allphases]
     # No need to keep track of secret phase; any heads in the subset that
     # are not mentioned are implicitly secret.
     for phase in allphases[:-1]:
         revset = "heads(%%ln & %s())" % phasenames[phase]
         headsbyphase[phase] = [cl.node(r) for r in repo.revs(revset, subset)]
     return headsbyphase
 
 def updatephases(repo, trgetter, headsbyphase):
     """Updates the repo with the given phase heads"""
     # Now advance phase boundaries of all but secret phase
     #
     # run the update (and fetch transaction) only if there are actually things
     # to update. This avoid creating empty transaction during no-op operation.
 
     for phase in allphases[:-1]:
         revset = '%%ln - %s()' % phasenames[phase]
         heads = [c.node() for c in repo.set(revset, headsbyphase[phase])]
         if heads:
             advanceboundary(repo, trgetter(), phase, heads)
 
 def analyzeremotephases(repo, subset, roots):
     """Compute phases heads and root in a subset of node from root dict
 
     * subset is heads of the subset
     * roots is {<nodeid> => phase} mapping. key and value are string.
 
     Accept unknown element input
     """
     repo = repo.unfiltered()
     # build list from dictionary
     draftroots = []
     nodemap = repo.changelog.nodemap # to filter unknown nodes
     for nhex, phase in roots.iteritems():
         if nhex == 'publishing': # ignore data related to publish option
             continue
         node = bin(nhex)
         phase = int(phase)
         if phase == public:
             if node != nullid:
                 repo.ui.warn(_('ignoring inconsistent public root'
                                ' from remote: %s\n') % nhex)
         elif phase == draft:
             if node in nodemap:
                 draftroots.append(node)
         else:
             repo.ui.warn(_('ignoring unexpected root from remote: %i %s\n')
                          % (phase, nhex))
     # compute heads
     publicheads = newheads(repo, subset, draftroots)
     return publicheads, draftroots
 
 class remotephasessummary(object):
     """summarize phase information on the remote side
 
     :publishing: True is the remote is publishing
     :publicheads: list of remote public phase heads (nodes)
     :draftheads: list of remote draft phase heads (nodes)
     :draftroots: list of remote draft phase root (nodes)
     """
 
     def __init__(self, repo, remotesubset, remoteroots):
         unfi = repo.unfiltered()
         self._allremoteroots = remoteroots
 
         self.publishing = remoteroots.get('publishing', False)
 
         ana = analyzeremotephases(repo, remotesubset, remoteroots)
         self.publicheads, self.draftroots = ana
         # Get the list of all "heads" revs draft on remote
         dheads = unfi.set('heads(%ln::%ln)', self.draftroots, remotesubset)
         self.draftheads = [c.node() for c in dheads]
 
 def newheads(repo, heads, roots):
     """compute new head of a subset minus another
 
     * `heads`: define the first subset
     * `roots`: define the second we subtract from the first"""
     repo = repo.unfiltered()
     revset = repo.set('heads((%ln + parents(%ln)) - (%ln::%ln))',
                       heads, roots, roots, heads)
     return [c.node() for c in revset]
 
 
 def newcommitphase(ui):
     """helper to get the target phase of new commit
 
     Handle all possible values for the phases.new-commit options.
 
     """
     v = ui.config('phases', 'new-commit')
     try:
         return phasenames.index(v)
     except ValueError:
         try:
             return int(v)
         except ValueError:
             msg = _("phases.new-commit: not a valid phase name ('%s')")
             raise error.ConfigError(msg % v)
 
 def hassecret(repo):
     """utility function that check if a repo have any secret changeset."""
     return bool(repo._phasecache.phaseroots[2])
 
 def preparehookargs(node, old, new):
     if old is None:
         old = ''
     else:
         old = phasenames[old]
     return {'node': node,
             'oldphase': old,
             'phase': phasenames[new]}

mercurial/policy.py

@@ -1,116 +1,116
 # policy.py - module policy logic for Mercurial.
 #
 # Copyright 2015 Gregory Szorc <[e-mail]>
 #
 # This software may be used and distributed according to the terms of the
 # GNU General Public License version 2 or any later version.
 
 from __future__ import absolute_import
 
 import os
 import sys
 
 # Rules for how modules can be loaded. Values are:
 #
 #    c - require C extensions
 #    allow - allow pure Python implementation when C loading fails
 #    cffi - required cffi versions (implemented within pure module)
 #    cffi-allow - allow pure Python implementation if cffi version is missing
 #    py - only load pure Python modules
 #
 # By default, fall back to the pure modules so the in-place build can
 # run without recompiling the C extensions. This will be overridden by
 # __modulepolicy__ generated by setup.py.
 policy = b'allow'
 _packageprefs = {
     # policy: (versioned package, pure package)
     b'c': (r'cext', None),
     b'allow': (r'cext', r'pure'),
     b'cffi': (r'cffi', None),
     b'cffi-allow': (r'cffi', r'pure'),
     b'py': (None, r'pure'),
 }
 
 try:
     from . import __modulepolicy__
     policy = __modulepolicy__.modulepolicy
 except ImportError:
     pass
 
 # PyPy doesn't load C extensions.
 #
 # The canonical way to do this is to test platform.python_implementation().
 # But we don't import platform and don't bloat for it here.
 if r'__pypy__' in sys.builtin_module_names:
     policy = b'cffi'
 
 # Our C extensions aren't yet compatible with Python 3. So use pure Python
 # on Python 3 for now.
 if sys.version_info[0] >= 3:
     policy = b'py'
 
 # Environment variable can always force settings.
 if sys.version_info[0] >= 3:
     if r'HGMODULEPOLICY' in os.environ:
         policy = os.environ[r'HGMODULEPOLICY'].encode(r'utf-8')
 else:
     policy = os.environ.get(r'HGMODULEPOLICY', policy)
 
 def _importfrom(pkgname, modname):
     # from .<pkgname> import <modname> (where . is looked through this module)
     fakelocals = {}
     pkg = __import__(pkgname, globals(), fakelocals, [modname], level=1)
     try:
         fakelocals[modname] = mod = getattr(pkg, modname)
     except AttributeError:
         raise ImportError(r'cannot import name %s' % modname)
     # force import; fakelocals[modname] may be replaced with the real module
     getattr(mod, r'__doc__', None)
     return fakelocals[modname]
 
 # keep in sync with "version" in C modules
 _cextversions = {
     (r'cext', r'base85'): 1,
     (r'cext', r'bdiff'): 1,
     (r'cext', r'diffhelpers'): 1,
     (r'cext', r'mpatch'): 1,
     (r'cext', r'osutil'): 1,
-    (r'cext', r'parsers'): 3,
+    (r'cext', r'parsers'): 4,
 }
 
 # map import request to other package or module
 _modredirects = {
     (r'cext', r'charencode'): (r'cext', r'parsers'),
     (r'cffi', r'base85'): (r'pure', r'base85'),
     (r'cffi', r'charencode'): (r'pure', r'charencode'),
     (r'cffi', r'diffhelpers'): (r'pure', r'diffhelpers'),
     (r'cffi', r'parsers'): (r'pure', r'parsers'),
 }
 
 def _checkmod(pkgname, modname, mod):
     expected = _cextversions.get((pkgname, modname))
     actual = getattr(mod, r'version', None)
     if actual != expected:
         raise ImportError(r'cannot import module %s.%s '
                           r'(expected version: %d, actual: %r)'
                           % (pkgname, modname, expected, actual))
 
 def importmod(modname):
     """Import module according to policy and check API version"""
     try:
         verpkg, purepkg = _packageprefs[policy]
     except KeyError:
         raise ImportError(r'invalid HGMODULEPOLICY %r' % policy)
     assert verpkg or purepkg
     if verpkg:
         pn, mn = _modredirects.get((verpkg, modname), (verpkg, modname))
         try:
             mod = _importfrom(pn, mn)
             if pn == verpkg:
                 _checkmod(pn, mn, mod)
             return mod
         except ImportError:
             if not purepkg:
                 raise
     pn, mn = _modredirects.get((purepkg, modname), (purepkg, modname))
     return _importfrom(pn, mn)