dirstate-item: introduce a `dm_nonnormal` property

See inline documentation for details.

Differential Revision: https://phab.mercurial-scm.org/D11123

mercurial/cext/parsers.c

@@ -1,968 +1,979
 /*
  parsers.c - efficient content parsing
 
  Copyright 2008 Olivia Mackall <[e-mail]> and others
 
  This software may be used and distributed according to the terms of
  the GNU General Public License, incorporated herein by reference.
 */
 
 #define PY_SSIZE_T_CLEAN
 #include <Python.h>
 #include <ctype.h>
 #include <stddef.h>
 #include <string.h>
 
 #include "bitmanipulation.h"
 #include "charencode.h"
 #include "util.h"
 
 #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 const int dirstate_v1_from_p2 = -2;
 static const int dirstate_v1_nonnormal = -1;
 static const int ambiguous_time = -1;
 
 static PyObject *dict_new_presized(PyObject *self, PyObject *args)
 {
 	Py_ssize_t expected_size;
 
 	if (!PyArg_ParseTuple(args, "n:make_presized_dict", &expected_size)) {
 		return NULL;
 	}
 
 	return _dict_new_presized(expected_size);
 }
 
 static inline dirstateItemObject *make_dirstate_item(char state, int mode,
                                                      int size, int mtime)
 {
 	dirstateItemObject *t =
 	    PyObject_New(dirstateItemObject, &dirstateItemType);
 	if (!t) {
 		return NULL;
 	}
 	t->state = state;
 	t->mode = mode;
 	t->size = size;
 	t->mtime = mtime;
 	return t;
 }
 
 static PyObject *dirstate_item_new(PyTypeObject *subtype, PyObject *args,
                                    PyObject *kwds)
 {
 	/* We do all the initialization here and not a tp_init function because
 	 * dirstate_item is immutable. */
 	dirstateItemObject *t;
 	char state;
 	int size, mode, mtime;
 	if (!PyArg_ParseTuple(args, "ciii", &state, &mode, &size, &mtime)) {
 		return NULL;
 	}
 
 	t = (dirstateItemObject *)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_item_dealloc(PyObject *o)
 {
 	PyObject_Del(o);
 }
 
 static Py_ssize_t dirstate_item_length(PyObject *o)
 {
 	return 4;
 }
 
 static PyObject *dirstate_item_item(PyObject *o, Py_ssize_t i)
 {
 	dirstateItemObject *t = (dirstateItemObject *)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_item_sq = {
     dirstate_item_length, /* sq_length */
     0,                    /* sq_concat */
     0,                    /* sq_repeat */
     dirstate_item_item,   /* sq_item */
     0,                    /* sq_ass_item */
     0,                    /* sq_contains */
     0,                    /* sq_inplace_concat */
     0                     /* sq_inplace_repeat */
 };
 
 static PyObject *dirstate_item_v1_state(dirstateItemObject *self)
 {
 	return PyBytes_FromStringAndSize(&self->state, 1);
 };
 
 static PyObject *dirstate_item_v1_mode(dirstateItemObject *self)
 {
 	return PyInt_FromLong(self->mode);
 };
 
 static PyObject *dirstate_item_v1_size(dirstateItemObject *self)
 {
 	return PyInt_FromLong(self->size);
 };
 
 static PyObject *dirstate_item_v1_mtime(dirstateItemObject *self)
 {
 	return PyInt_FromLong(self->mtime);
 };
 
+static PyObject *dm_nonnormal(dirstateItemObject *self)
+{
+	if (self->state != 'n' || self->mtime == ambiguous_time) {
+		Py_RETURN_TRUE;
+	} else {
+		Py_RETURN_FALSE;
+	}
+};
+
 static PyObject *dirstate_item_need_delay(dirstateItemObject *self,
                                           PyObject *value)
 {
 	long now;
 	if (!pylong_to_long(value, &now)) {
 		return NULL;
 	}
 	if (self->state == 'n' && self->mtime == now) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 /* This will never change since it's bound to V1, unlike `make_dirstate_item`
  */
 static inline dirstateItemObject *
 dirstate_item_from_v1_data(char state, int mode, int size, int mtime)
 {
 	dirstateItemObject *t =
 	    PyObject_New(dirstateItemObject, &dirstateItemType);
 	if (!t) {
 		return NULL;
 	}
 	t->state = state;
 	t->mode = mode;
 	t->size = size;
 	t->mtime = mtime;
 	return t;
 }
 
 /* This will never change since it's bound to V1, unlike `dirstate_item_new` */
 static PyObject *dirstate_item_from_v1_meth(PyTypeObject *subtype,
                                             PyObject *args)
 {
 	/* We do all the initialization here and not a tp_init function because
 	 * dirstate_item is immutable. */
 	dirstateItemObject *t;
 	char state;
 	int size, mode, mtime;
 	if (!PyArg_ParseTuple(args, "ciii", &state, &mode, &size, &mtime)) {
 		return NULL;
 	}
 
 	t = (dirstateItemObject *)subtype->tp_alloc(subtype, 1);
 	if (!t) {
 		return NULL;
 	}
 	t->state = state;
 	t->mode = mode;
 	t->size = size;
 	t->mtime = mtime;
 
 	return (PyObject *)t;
 };
 
 /* This means the next status call will have to actually check its content
    to make sure it is correct. */
 static PyObject *dirstate_item_set_possibly_dirty(dirstateItemObject *self)
 {
 	self->mtime = ambiguous_time;
 	Py_RETURN_NONE;
 }
 
 static PyMethodDef dirstate_item_methods[] = {
     {"v1_state", (PyCFunction)dirstate_item_v1_state, METH_NOARGS,
      "return a \"state\" suitable for v1 serialization"},
     {"v1_mode", (PyCFunction)dirstate_item_v1_mode, METH_NOARGS,
      "return a \"mode\" suitable for v1 serialization"},
     {"v1_size", (PyCFunction)dirstate_item_v1_size, METH_NOARGS,
      "return a \"size\" suitable for v1 serialization"},
     {"v1_mtime", (PyCFunction)dirstate_item_v1_mtime, METH_NOARGS,
      "return a \"mtime\" suitable for v1 serialization"},
     {"need_delay", (PyCFunction)dirstate_item_need_delay, METH_O,
      "True if the stored mtime would be ambiguous with the current time"},
     {"from_v1_data", (PyCFunction)dirstate_item_from_v1_meth, METH_O,
      "build a new DirstateItem object from V1 data"},
     {"set_possibly_dirty", (PyCFunction)dirstate_item_set_possibly_dirty,
      METH_NOARGS, "mark a file as \"possibly dirty\""},
+    {"dm_nonnormal", (PyCFunction)dm_nonnormal, METH_NOARGS,
+     "True is the entry is non-normal in the dirstatemap sense"},
     {NULL} /* Sentinel */
 };
 
 static PyObject *dirstate_item_get_mode(dirstateItemObject *self)
 {
 	return PyInt_FromLong(self->mode);
 };
 
 static PyObject *dirstate_item_get_size(dirstateItemObject *self)
 {
 	return PyInt_FromLong(self->size);
 };
 
 static PyObject *dirstate_item_get_mtime(dirstateItemObject *self)
 {
 	return PyInt_FromLong(self->mtime);
 };
 
 static PyObject *dirstate_item_get_state(dirstateItemObject *self)
 {
 	return PyBytes_FromStringAndSize(&self->state, 1);
 };
 
 static PyObject *dirstate_item_get_tracked(dirstateItemObject *self)
 {
 	if (self->state == 'a' || self->state == 'm' || self->state == 'n') {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyObject *dirstate_item_get_added(dirstateItemObject *self)
 {
 	if (self->state == 'a') {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyObject *dirstate_item_get_merged(dirstateItemObject *self)
 {
 	if (self->state == 'm') {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyObject *dirstate_item_get_merged_removed(dirstateItemObject *self)
 {
 	if (self->state == 'r' && self->size == dirstate_v1_nonnormal) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyObject *dirstate_item_get_from_p2(dirstateItemObject *self)
 {
 	if (self->state == 'n' && self->size == dirstate_v1_from_p2) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyObject *dirstate_item_get_from_p2_removed(dirstateItemObject *self)
 {
 	if (self->state == 'r' && self->size == dirstate_v1_from_p2) {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyObject *dirstate_item_get_removed(dirstateItemObject *self)
 {
 	if (self->state == 'r') {
 		Py_RETURN_TRUE;
 	} else {
 		Py_RETURN_FALSE;
 	}
 };
 
 static PyGetSetDef dirstate_item_getset[] = {
     {"mode", (getter)dirstate_item_get_mode, NULL, "mode", NULL},
     {"size", (getter)dirstate_item_get_size, NULL, "size", NULL},
     {"mtime", (getter)dirstate_item_get_mtime, NULL, "mtime", NULL},
     {"state", (getter)dirstate_item_get_state, NULL, "state", NULL},
     {"tracked", (getter)dirstate_item_get_tracked, NULL, "tracked", NULL},
     {"added", (getter)dirstate_item_get_added, NULL, "added", NULL},
     {"merged_removed", (getter)dirstate_item_get_merged_removed, NULL,
      "merged_removed", NULL},
     {"merged", (getter)dirstate_item_get_merged, NULL, "merged", NULL},
     {"from_p2_removed", (getter)dirstate_item_get_from_p2_removed, NULL,
      "from_p2_removed", NULL},
     {"from_p2", (getter)dirstate_item_get_from_p2, NULL, "from_p2", NULL},
     {"removed", (getter)dirstate_item_get_removed, NULL, "removed", NULL},
     {NULL} /* Sentinel */
 };
 
 PyTypeObject dirstateItemType = {
     PyVarObject_HEAD_INIT(NULL, 0)     /* header */
     "dirstate_tuple",                  /* tp_name */
     sizeof(dirstateItemObject),        /* tp_basicsize */
     0,                                 /* tp_itemsize */
     (destructor)dirstate_item_dealloc, /* tp_dealloc */
     0,                                 /* tp_print */
     0,                                 /* tp_getattr */
     0,                                 /* tp_setattr */
     0,                                 /* tp_compare */
     0,                                 /* tp_repr */
     0,                                 /* tp_as_number */
     &dirstate_item_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 */
     dirstate_item_methods,             /* tp_methods */
     0,                                 /* tp_members */
     dirstate_item_getset,              /* tp_getset */
     0,                                 /* tp_base */
     0,                                 /* tp_dict */
     0,                                 /* tp_descr_get */
     0,                                 /* tp_descr_set */
     0,                                 /* tp_dictoffset */
     0,                                 /* tp_init */
     0,                                 /* tp_alloc */
     dirstate_item_new,                 /* tp_new */
 };
 
 static PyObject *parse_dirstate(PyObject *self, PyObject *args)
 {
 	PyObject *dmap, *cmap, *parents = NULL, *ret = NULL;
 	PyObject *fname = NULL, *cname = NULL, *entry = NULL;
 	char state, *cur, *str, *cpos;
 	int mode, size, mtime;
 	unsigned int flen, pos = 40;
 	Py_ssize_t len = 40;
 	Py_ssize_t readlen;
 
 	if (!PyArg_ParseTuple(
 	        args, PY23("O!O!s#:parse_dirstate", "O!O!y#:parse_dirstate"),
 	        &PyDict_Type, &dmap, &PyDict_Type, &cmap, &str, &readlen)) {
 		goto quit;
 	}
 
 	len = readlen;
 
 	/* read parents */
 	if (len < 40) {
 		PyErr_SetString(PyExc_ValueError,
 		                "too little data for parents");
 		goto quit;
 	}
 
 	parents = Py_BuildValue(PY23("s#s#", "y#y#"), str, (Py_ssize_t)20,
 	                        str + 20, (Py_ssize_t)20);
 	if (!parents) {
 		goto quit;
 	}
 
 	/* read filenames */
 	while (pos >= 40 && pos < len) {
 		if (pos + 17 > len) {
 			PyErr_SetString(PyExc_ValueError,
 			                "overflow in dirstate");
 			goto quit;
 		}
 		cur = str + pos;
 		/* unpack header */
 		state = *cur;
 		mode = getbe32(cur + 1);
 		size = getbe32(cur + 5);
 		mtime = getbe32(cur + 9);
 		flen = getbe32(cur + 13);
 		pos += 17;
 		cur += 17;
 		if (flen > len - pos) {
 			PyErr_SetString(PyExc_ValueError,
 			                "overflow in dirstate");
 			goto quit;
 		}
 
 		entry = (PyObject *)dirstate_item_from_v1_data(state, mode,
 		                                               size, mtime);
 		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)) {
 		dirstateItemObject *t;
 		if (!dirstate_tuple_check(v)) {
 			PyErr_SetString(PyExc_TypeError,
 			                "expected a dirstate tuple");
 			goto bail;
 		}
 		t = (dirstateItemObject *)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!O!i:pack_dirstate", &PyDict_Type, &map,
 	                      &PyDict_Type, &copymap, &PyTuple_Type, &pl,
 	                      &now)) {
 		return NULL;
 	}
 
 	if (PyTuple_Size(pl) != 2) {
 		PyErr_SetString(PyExc_TypeError, "expected 2-element tuple");
 		return NULL;
 	}
 
 	/* Figure out how much we need to allocate. */
 	for (nbytes = 40, pos = 0; PyDict_Next(map, &pos, &k, &v);) {
 		PyObject *c;
 		if (!PyBytes_Check(k)) {
 			PyErr_SetString(PyExc_TypeError, "expected string key");
 			goto bail;
 		}
 		nbytes += PyBytes_GET_SIZE(k) + 17;
 		c = PyDict_GetItem(copymap, k);
 		if (c) {
 			if (!PyBytes_Check(c)) {
 				PyErr_SetString(PyExc_TypeError,
 				                "expected string key");
 				goto bail;
 			}
 			nbytes += PyBytes_GET_SIZE(c) + 1;
 		}
 	}
 
 	packobj = PyBytes_FromStringAndSize(NULL, nbytes);
 	if (packobj == NULL) {
 		goto bail;
 	}
 
 	p = PyBytes_AS_STRING(packobj);
 
 	pn = PyTuple_GET_ITEM(pl, 0);
 	if (PyBytes_AsStringAndSize(pn, &s, &l) == -1 || l != 20) {
 		PyErr_SetString(PyExc_TypeError, "expected a 20-byte hash");
 		goto bail;
 	}
 	memcpy(p, s, l);
 	p += 20;
 	pn = PyTuple_GET_ITEM(pl, 1);
 	if (PyBytes_AsStringAndSize(pn, &s, &l) == -1 || l != 20) {
 		PyErr_SetString(PyExc_TypeError, "expected a 20-byte hash");
 		goto bail;
 	}
 	memcpy(p, s, l);
 	p += 20;
 
 	for (pos = 0; PyDict_Next(map, &pos, &k, &v);) {
 		dirstateItemObject *tuple;
 		char state;
 		int mode, size, mtime;
 		Py_ssize_t len, l;
 		PyObject *o;
 		char *t;
 
 		if (!dirstate_tuple_check(v)) {
 			PyErr_SetString(PyExc_TypeError,
 			                "expected a dirstate tuple");
 			goto bail;
 		}
 		tuple = (dirstateItemObject *)v;
 
 		state = 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_item(
 			    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;
 	Py_ssize_t datalen, offset, stop;
 	PyObject *markers = NULL;
 
 	if (!PyArg_ParseTuple(args, PY23("s#nn", "y#nn"), &data, &datalen,
 	                      &offset, &stop)) {
 		return NULL;
 	}
 	if (offset < 0) {
 		PyErr_SetString(PyExc_ValueError,
 		                "invalid negative offset in fm1readmarkers");
 		return NULL;
 	}
 	if (stop > datalen) {
 		PyErr_SetString(
 		    PyExc_ValueError,
 		    "stop longer than data length in fm1readmarkers");
 		return NULL;
 	}
 	dataend = data + datalen;
 	data += offset;
 	markers = PyList_New(0);
 	if (!markers) {
 		return NULL;
 	}
 	while (offset < stop) {
 		uint32_t msize;
 		int error;
 		PyObject *record = fm1readmarker(data, dataend, &msize);
 		if (!record) {
 			goto bail;
 		}
 		error = PyList_Append(markers, record);
 		Py_DECREF(record);
 		if (error) {
 			goto bail;
 		}
 		data += msize;
 		offset += msize;
 	}
 	return markers;
 bail:
 	Py_DECREF(markers);
 	return NULL;
 }
 
 static char parsers_doc[] = "Efficient content parsing.";
 
 PyObject *encodedir(PyObject *self, PyObject *args);
 PyObject *pathencode(PyObject *self, PyObject *args);
 PyObject *lowerencode(PyObject *self, PyObject *args);
 PyObject *parse_index2(PyObject *self, PyObject *args, PyObject *kwargs);
 
 static PyMethodDef methods[] = {
     {"pack_dirstate", pack_dirstate, METH_VARARGS, "pack a dirstate\n"},
     {"nonnormalotherparententries", nonnormalotherparententries, METH_VARARGS,
      "create a set containing non-normal and other parent entries of given "
      "dirstate\n"},
     {"parse_dirstate", parse_dirstate, METH_VARARGS, "parse a dirstate\n"},
     {"parse_index2", (PyCFunction)parse_index2, METH_VARARGS | METH_KEYWORDS,
      "parse a revlog index\n"},
     {"isasciistr", isasciistr, METH_VARARGS, "check if an ASCII string\n"},
     {"asciilower", asciilower, METH_VARARGS, "lowercase an ASCII string\n"},
     {"asciiupper", asciiupper, METH_VARARGS, "uppercase an ASCII string\n"},
     {"dict_new_presized", dict_new_presized, METH_VARARGS,
      "construct a dict with an expected size\n"},
     {"make_file_foldmap", make_file_foldmap, METH_VARARGS,
      "make file foldmap\n"},
     {"jsonescapeu8fast", jsonescapeu8fast, METH_VARARGS,
      "escape a UTF-8 byte string to JSON (fast path)\n"},
     {"encodedir", encodedir, METH_VARARGS, "encodedir a path\n"},
     {"pathencode", pathencode, METH_VARARGS, "fncache-encode a path\n"},
     {"lowerencode", lowerencode, METH_VARARGS, "lower-encode a path\n"},
     {"fm1readmarkers", fm1readmarkers, METH_VARARGS,
      "parse v1 obsolete markers\n"},
     {NULL, NULL}};
 
 void dirs_module_init(PyObject *mod);
 void manifest_module_init(PyObject *mod);
 void revlog_module_init(PyObject *mod);
 
 static const int version = 20;
 
 static void module_init(PyObject *mod)
 {
 	PyObject *capsule = NULL;
 	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);
 
 	capsule = PyCapsule_New(
 	    make_dirstate_item,
 	    "mercurial.cext.parsers.make_dirstate_item_CAPI", NULL);
 	if (capsule != NULL)
 		PyModule_AddObject(mod, "make_dirstate_item_CAPI", capsule);
 
 	if (PyType_Ready(&dirstateItemType) < 0) {
 		return;
 	}
 	Py_INCREF(&dirstateItemType);
 	PyModule_AddObject(mod, "DirstateItem", (PyObject *)&dirstateItemType);
 }
 
 static int check_python_version(void)
 {
 	PyObject *sys = PyImport_ImportModule("sys"), *ver;
 	long hexversion;
 	if (!sys) {
 		return -1;
 	}
 	ver = PyObject_GetAttrString(sys, "hexversion");
 	Py_DECREF(sys);
 	if (!ver) {
 		return -1;
 	}
 	hexversion = 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/dirstatemap.py

@@ -1,752 +1,752
 # dirstatemap.py
 #
 # 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
 
 from .i18n import _
 
 from . import (
     error,
     pathutil,
     policy,
     pycompat,
     txnutil,
     util,
 )
 
 from .dirstateutils import (
     docket as docketmod,
 )
 
 parsers = policy.importmod('parsers')
 rustmod = policy.importrust('dirstate')
 
 propertycache = util.propertycache
 
 DirstateItem = parsers.DirstateItem
 
 
 # a special value used internally for `size` if the file come from the other parent
 FROM_P2 = -2
 
 # a special value used internally for `size` if the file is modified/merged/added
 NONNORMAL = -1
 
 # a special value used internally for `time` if the time is ambigeous
 AMBIGUOUS_TIME = -1
 
 rangemask = 0x7FFFFFFF
 
 
 class dirstatemap(object):
     """Map encapsulating the dirstate's contents.
 
     The dirstate contains the following state:
 
     - `identity` is the identity of the dirstate file, which can be used to
       detect when changes have occurred to the dirstate file.
 
     - `parents` is a pair containing the parents of the working copy. The
       parents are updated by calling `setparents`.
 
     - the state map maps filenames to tuples of (state, mode, size, mtime),
       where state is a single character representing 'normal', 'added',
       'removed', or 'merged'. It is read by treating the dirstate as a
       dict.  File state is updated by calling the `addfile`, `removefile` and
       `dropfile` methods.
 
     - `copymap` maps destination filenames to their source filename.
 
     The dirstate also provides the following views onto the state:
 
     - `nonnormalset` is a set of the filenames that have state other
       than 'normal', or are normal but have an mtime of -1 ('normallookup').
 
     - `otherparentset` is a set of the filenames that are marked as coming
       from the second parent when the dirstate is currently being merged.
 
     - `filefoldmap` is a dict mapping normalized filenames to the denormalized
       form that they appear as in the dirstate.
 
     - `dirfoldmap` is a dict mapping normalized directory names to the
       denormalized form that they appear as in the dirstate.
     """
 
     def __init__(self, ui, opener, root, nodeconstants, use_dirstate_v2):
         self._ui = ui
         self._opener = opener
         self._root = root
         self._filename = b'dirstate'
         self._nodelen = 20
         self._nodeconstants = nodeconstants
         assert (
             not use_dirstate_v2
         ), "should have detected unsupported requirement"
 
         self._parents = None
         self._dirtyparents = False
 
         # for consistent view between _pl() and _read() invocations
         self._pendingmode = None
 
     @propertycache
     def _map(self):
         self._map = {}
         self.read()
         return self._map
 
     @propertycache
     def copymap(self):
         self.copymap = {}
         self._map
         return self.copymap
 
     def clear(self):
         self._map.clear()
         self.copymap.clear()
         self.setparents(self._nodeconstants.nullid, self._nodeconstants.nullid)
         util.clearcachedproperty(self, b"_dirs")
         util.clearcachedproperty(self, b"_alldirs")
         util.clearcachedproperty(self, b"filefoldmap")
         util.clearcachedproperty(self, b"dirfoldmap")
         util.clearcachedproperty(self, b"nonnormalset")
         util.clearcachedproperty(self, b"otherparentset")
 
     def items(self):
         return pycompat.iteritems(self._map)
 
     # forward for python2,3 compat
     iteritems = items
 
     debug_iter = items
 
     def __len__(self):
         return len(self._map)
 
     def __iter__(self):
         return iter(self._map)
 
     def get(self, key, default=None):
         return self._map.get(key, default)
 
     def __contains__(self, key):
         return key in self._map
 
     def __getitem__(self, key):
         return self._map[key]
 
     def keys(self):
         return self._map.keys()
 
     def preload(self):
         """Loads the underlying data, if it's not already loaded"""
         self._map
 
     def addfile(
         self,
         f,
         mode=0,
         size=None,
         mtime=None,
         added=False,
         merged=False,
         from_p2=False,
         possibly_dirty=False,
     ):
         """Add a tracked file to the dirstate."""
         if added:
             assert not merged
             assert not possibly_dirty
             assert not from_p2
             state = b'a'
             size = NONNORMAL
             mtime = AMBIGUOUS_TIME
         elif merged:
             assert not possibly_dirty
             assert not from_p2
             state = b'm'
             size = FROM_P2
             mtime = AMBIGUOUS_TIME
         elif from_p2:
             assert not possibly_dirty
             state = b'n'
             size = FROM_P2
             mtime = AMBIGUOUS_TIME
         elif possibly_dirty:
             state = b'n'
             size = NONNORMAL
             mtime = AMBIGUOUS_TIME
         else:
             assert size != FROM_P2
             assert size != NONNORMAL
             state = b'n'
             size = size & rangemask
             mtime = mtime & rangemask
         assert state is not None
         assert size is not None
         assert mtime is not None
         old_entry = self.get(f)
         if (
             old_entry is None or old_entry.removed
         ) and "_dirs" in self.__dict__:
             self._dirs.addpath(f)
         if old_entry is None and "_alldirs" in self.__dict__:
             self._alldirs.addpath(f)
-        self._map[f] = DirstateItem(state, mode, size, mtime)
-        if state != b'n' or mtime == AMBIGUOUS_TIME:
+        e = self._map[f] = DirstateItem(state, mode, size, mtime)
+        if e.dm_nonnormal:
             self.nonnormalset.add(f)
         if size == FROM_P2:
             self.otherparentset.add(f)
 
     def removefile(self, f, in_merge=False):
         """
         Mark a file as removed in the dirstate.
 
         The `size` parameter is used to store sentinel values that indicate
         the file's previous state.  In the future, we should refactor this
         to be more explicit about what that state is.
         """
         entry = self.get(f)
         size = 0
         if in_merge:
             # XXX we should not be able to have 'm' state and 'FROM_P2' if not
             # during a merge. So I (marmoute) am not sure we need the
             # conditionnal at all. Adding double checking this with assert
             # would be nice.
             if entry is not None:
                 # backup the previous state
                 if entry.merged:  # merge
                     size = NONNORMAL
                 elif entry.from_p2:
                     size = FROM_P2
                     self.otherparentset.add(f)
         if entry is not None and not (entry.merged or entry.from_p2):
             self.copymap.pop(f, None)
 
         if entry is not None and not entry.removed and "_dirs" in self.__dict__:
             self._dirs.delpath(f)
         if entry is None and "_alldirs" in self.__dict__:
             self._alldirs.addpath(f)
         if "filefoldmap" in self.__dict__:
             normed = util.normcase(f)
             self.filefoldmap.pop(normed, None)
         self._map[f] = DirstateItem(b'r', 0, size, 0)
         self.nonnormalset.add(f)
 
     def dropfile(self, f):
         """
         Remove a file from the dirstate.  Returns True if the file was
         previously recorded.
         """
         old_entry = self._map.pop(f, None)
         exists = False
         oldstate = b'?'
         if old_entry is not None:
             exists = True
             oldstate = old_entry.state
         if exists:
             if oldstate != b"r" and "_dirs" in self.__dict__:
                 self._dirs.delpath(f)
             if "_alldirs" in self.__dict__:
                 self._alldirs.delpath(f)
         if "filefoldmap" in self.__dict__:
             normed = util.normcase(f)
             self.filefoldmap.pop(normed, None)
         self.nonnormalset.discard(f)
         return exists
 
     def clearambiguoustimes(self, files, now):
         for f in files:
             e = self.get(f)
             if e is not None and e.need_delay(now):
                 e.set_possibly_dirty()
                 self.nonnormalset.add(f)
 
     def nonnormalentries(self):
         '''Compute the nonnormal dirstate entries from the dmap'''
         try:
             return parsers.nonnormalotherparententries(self._map)
         except AttributeError:
             nonnorm = set()
             otherparent = set()
             for fname, e in pycompat.iteritems(self._map):
-                if e.state != b'n' or e.mtime == AMBIGUOUS_TIME:
+                if e.dm_nonnormal:
                     nonnorm.add(fname)
                 if e.from_p2:
                     otherparent.add(fname)
             return nonnorm, otherparent
 
     @propertycache
     def filefoldmap(self):
         """Returns a dictionary mapping normalized case paths to their
         non-normalized versions.
         """
         try:
             makefilefoldmap = parsers.make_file_foldmap
         except AttributeError:
             pass
         else:
             return makefilefoldmap(
                 self._map, util.normcasespec, util.normcasefallback
             )
 
         f = {}
         normcase = util.normcase
         for name, s in pycompat.iteritems(self._map):
             if not s.removed:
                 f[normcase(name)] = name
         f[b'.'] = b'.'  # prevents useless util.fspath() invocation
         return f
 
     def hastrackeddir(self, d):
         """
         Returns True if the dirstate contains a tracked (not removed) file
         in this directory.
         """
         return d in self._dirs
 
     def hasdir(self, d):
         """
         Returns True if the dirstate contains a file (tracked or removed)
         in this directory.
         """
         return d in self._alldirs
 
     @propertycache
     def _dirs(self):
         return pathutil.dirs(self._map, b'r')
 
     @propertycache
     def _alldirs(self):
         return pathutil.dirs(self._map)
 
     def _opendirstatefile(self):
         fp, mode = txnutil.trypending(self._root, self._opener, self._filename)
         if self._pendingmode is not None and self._pendingmode != mode:
             fp.close()
             raise error.Abort(
                 _(b'working directory state may be changed parallelly')
             )
         self._pendingmode = mode
         return fp
 
     def parents(self):
         if not self._parents:
             try:
                 fp = self._opendirstatefile()
                 st = fp.read(2 * self._nodelen)
                 fp.close()
             except IOError as err:
                 if err.errno != errno.ENOENT:
                     raise
                 # File doesn't exist, so the current state is empty
                 st = b''
 
             l = len(st)
             if l == self._nodelen * 2:
                 self._parents = (
                     st[: self._nodelen],
                     st[self._nodelen : 2 * self._nodelen],
                 )
             elif l == 0:
                 self._parents = (
                     self._nodeconstants.nullid,
                     self._nodeconstants.nullid,
                 )
             else:
                 raise error.Abort(
                     _(b'working directory state appears damaged!')
                 )
 
         return self._parents
 
     def setparents(self, p1, p2):
         self._parents = (p1, p2)
         self._dirtyparents = True
 
     def read(self):
         # ignore HG_PENDING because identity is used only for writing
         self.identity = util.filestat.frompath(
             self._opener.join(self._filename)
         )
 
         try:
             fp = self._opendirstatefile()
             try:
                 st = fp.read()
             finally:
                 fp.close()
         except IOError as err:
             if err.errno != errno.ENOENT:
                 raise
             return
         if not st:
             return
 
         if util.safehasattr(parsers, b'dict_new_presized'):
             # Make an estimate of the number of files in the dirstate based on
             # its size. This trades wasting some memory for avoiding costly
             # resizes. Each entry have a prefix of 17 bytes followed by one or
             # two path names. Studies on various large-scale real-world repositories
             # found 54 bytes a reasonable upper limit for the average path names.
             # Copy entries are ignored for the sake of this estimate.
             self._map = parsers.dict_new_presized(len(st) // 71)
 
         # Python's garbage collector triggers a GC each time a certain number
         # of container objects (the number being defined by
         # gc.get_threshold()) are allocated. parse_dirstate creates a tuple
         # for each file in the dirstate. The C version then immediately marks
         # them as not to be tracked by the collector. However, this has no
         # effect on when GCs are triggered, only on what objects the GC looks
         # into. This means that O(number of files) GCs are unavoidable.
         # Depending on when in the process's lifetime the dirstate is parsed,
         # this can get very expensive. As a workaround, disable GC while
         # parsing the dirstate.
         #
         # (we cannot decorate the function directly since it is in a C module)
         parse_dirstate = util.nogc(parsers.parse_dirstate)
         p = parse_dirstate(self._map, self.copymap, st)
         if not self._dirtyparents:
             self.setparents(*p)
 
         # Avoid excess attribute lookups by fast pathing certain checks
         self.__contains__ = self._map.__contains__
         self.__getitem__ = self._map.__getitem__
         self.get = self._map.get
 
     def write(self, _tr, st, now):
         st.write(
             parsers.pack_dirstate(self._map, self.copymap, self.parents(), now)
         )
         st.close()
         self._dirtyparents = False
         self.nonnormalset, self.otherparentset = self.nonnormalentries()
 
     @propertycache
     def nonnormalset(self):
         nonnorm, otherparents = self.nonnormalentries()
         self.otherparentset = otherparents
         return nonnorm
 
     @propertycache
     def otherparentset(self):
         nonnorm, otherparents = self.nonnormalentries()
         self.nonnormalset = nonnorm
         return otherparents
 
     def non_normal_or_other_parent_paths(self):
         return self.nonnormalset.union(self.otherparentset)
 
     @propertycache
     def identity(self):
         self._map
         return self.identity
 
     @propertycache
     def dirfoldmap(self):
         f = {}
         normcase = util.normcase
         for name in self._dirs:
             f[normcase(name)] = name
         return f
 
 
 if rustmod is not None:
 
     class dirstatemap(object):
         def __init__(self, ui, opener, root, nodeconstants, use_dirstate_v2):
             self._use_dirstate_v2 = use_dirstate_v2
             self._nodeconstants = nodeconstants
             self._ui = ui
             self._opener = opener
             self._root = root
             self._filename = b'dirstate'
             self._nodelen = 20  # Also update Rust code when changing this!
             self._parents = None
             self._dirtyparents = False
             self._docket = None
 
             # for consistent view between _pl() and _read() invocations
             self._pendingmode = None
 
             self._use_dirstate_tree = self._ui.configbool(
                 b"experimental",
                 b"dirstate-tree.in-memory",
                 False,
             )
 
         def addfile(
             self,
             f,
             mode=0,
             size=None,
             mtime=None,
             added=False,
             merged=False,
             from_p2=False,
             possibly_dirty=False,
         ):
             return self._rustmap.addfile(
                 f,
                 mode,
                 size,
                 mtime,
                 added,
                 merged,
                 from_p2,
                 possibly_dirty,
             )
 
         def removefile(self, *args, **kwargs):
             return self._rustmap.removefile(*args, **kwargs)
 
         def dropfile(self, *args, **kwargs):
             return self._rustmap.dropfile(*args, **kwargs)
 
         def clearambiguoustimes(self, *args, **kwargs):
             return self._rustmap.clearambiguoustimes(*args, **kwargs)
 
         def nonnormalentries(self):
             return self._rustmap.nonnormalentries()
 
         def get(self, *args, **kwargs):
             return self._rustmap.get(*args, **kwargs)
 
         @property
         def copymap(self):
             return self._rustmap.copymap()
 
         def directories(self):
             return self._rustmap.directories()
 
         def debug_iter(self):
             return self._rustmap.debug_iter()
 
         def preload(self):
             self._rustmap
 
         def clear(self):
             self._rustmap.clear()
             self.setparents(
                 self._nodeconstants.nullid, self._nodeconstants.nullid
             )
             util.clearcachedproperty(self, b"_dirs")
             util.clearcachedproperty(self, b"_alldirs")
             util.clearcachedproperty(self, b"dirfoldmap")
 
         def items(self):
             return self._rustmap.items()
 
         def keys(self):
             return iter(self._rustmap)
 
         def __contains__(self, key):
             return key in self._rustmap
 
         def __getitem__(self, item):
             return self._rustmap[item]
 
         def __len__(self):
             return len(self._rustmap)
 
         def __iter__(self):
             return iter(self._rustmap)
 
         # forward for python2,3 compat
         iteritems = items
 
         def _opendirstatefile(self):
             fp, mode = txnutil.trypending(
                 self._root, self._opener, self._filename
             )
             if self._pendingmode is not None and self._pendingmode != mode:
                 fp.close()
                 raise error.Abort(
                     _(b'working directory state may be changed parallelly')
                 )
             self._pendingmode = mode
             return fp
 
         def _readdirstatefile(self, size=-1):
             try:
                 with self._opendirstatefile() as fp:
                     return fp.read(size)
             except IOError as err:
                 if err.errno != errno.ENOENT:
                     raise
                 # File doesn't exist, so the current state is empty
                 return b''
 
         def setparents(self, p1, p2):
             self._parents = (p1, p2)
             self._dirtyparents = True
 
         def parents(self):
             if not self._parents:
                 if self._use_dirstate_v2:
                     self._parents = self.docket.parents
                 else:
                     read_len = self._nodelen * 2
                     st = self._readdirstatefile(read_len)
                     l = len(st)
                     if l == read_len:
                         self._parents = (
                             st[: self._nodelen],
                             st[self._nodelen : 2 * self._nodelen],
                         )
                     elif l == 0:
                         self._parents = (
                             self._nodeconstants.nullid,
                             self._nodeconstants.nullid,
                         )
                     else:
                         raise error.Abort(
                             _(b'working directory state appears damaged!')
                         )
 
             return self._parents
 
         @property
         def docket(self):
             if not self._docket:
                 if not self._use_dirstate_v2:
                     raise error.ProgrammingError(
                         b'dirstate only has a docket in v2 format'
                     )
                 self._docket = docketmod.DirstateDocket.parse(
                     self._readdirstatefile(), self._nodeconstants
                 )
             return self._docket
 
         @propertycache
         def _rustmap(self):
             """
             Fills the Dirstatemap when called.
             """
             # ignore HG_PENDING because identity is used only for writing
             self.identity = util.filestat.frompath(
                 self._opener.join(self._filename)
             )
 
             if self._use_dirstate_v2:
                 if self.docket.uuid:
                     # TODO: use mmap when possible
                     data = self._opener.read(self.docket.data_filename())
                 else:
                     data = b''
                 self._rustmap = rustmod.DirstateMap.new_v2(
                     data, self.docket.data_size, self.docket.tree_metadata
                 )
                 parents = self.docket.parents
             else:
                 self._rustmap, parents = rustmod.DirstateMap.new_v1(
                     self._use_dirstate_tree, self._readdirstatefile()
                 )
 
             if parents and not self._dirtyparents:
                 self.setparents(*parents)
 
             self.__contains__ = self._rustmap.__contains__
             self.__getitem__ = self._rustmap.__getitem__
             self.get = self._rustmap.get
             return self._rustmap
 
         def write(self, tr, st, now):
             if not self._use_dirstate_v2:
                 p1, p2 = self.parents()
                 packed = self._rustmap.write_v1(p1, p2, now)
                 st.write(packed)
                 st.close()
                 self._dirtyparents = False
                 return
 
             # We can only append to an existing data file if there is one
             can_append = self.docket.uuid is not None
             packed, meta, append = self._rustmap.write_v2(now, can_append)
             if append:
                 docket = self.docket
                 data_filename = docket.data_filename()
                 if tr:
                     tr.add(data_filename, docket.data_size)
                 with self._opener(data_filename, b'r+b') as fp:
                     fp.seek(docket.data_size)
                     assert fp.tell() == docket.data_size
                     written = fp.write(packed)
                     if written is not None:  # py2 may return None
                         assert written == len(packed), (written, len(packed))
                 docket.data_size += len(packed)
                 docket.parents = self.parents()
                 docket.tree_metadata = meta
                 st.write(docket.serialize())
                 st.close()
             else:
                 old_docket = self.docket
                 new_docket = docketmod.DirstateDocket.with_new_uuid(
                     self.parents(), len(packed), meta
                 )
                 data_filename = new_docket.data_filename()
                 if tr:
                     tr.add(data_filename, 0)
                 self._opener.write(data_filename, packed)
                 # Write the new docket after the new data file has been
                 # written. Because `st` was opened with `atomictemp=True`,
                 # the actual `.hg/dirstate` file is only affected on close.
                 st.write(new_docket.serialize())
                 st.close()
                 # Remove the old data file after the new docket pointing to
                 # the new data file was written.
                 if old_docket.uuid:
                     data_filename = old_docket.data_filename()
                     unlink = lambda _tr=None: self._opener.unlink(data_filename)
                     if tr:
                         category = b"dirstate-v2-clean-" + old_docket.uuid
                         tr.addpostclose(category, unlink)
                     else:
                         unlink()
                 self._docket = new_docket
             # Reload from the newly-written file
             util.clearcachedproperty(self, b"_rustmap")
             self._dirtyparents = False
 
         @propertycache
         def filefoldmap(self):
             """Returns a dictionary mapping normalized case paths to their
             non-normalized versions.
             """
             return self._rustmap.filefoldmapasdict()
 
         def hastrackeddir(self, d):
             return self._rustmap.hastrackeddir(d)
 
         def hasdir(self, d):
             return self._rustmap.hasdir(d)
 
         @propertycache
         def identity(self):
             self._rustmap
             return self.identity
 
         @property
         def nonnormalset(self):
             nonnorm = self._rustmap.non_normal_entries()
             return nonnorm
 
         @propertycache
         def otherparentset(self):
             otherparents = self._rustmap.other_parent_entries()
             return otherparents
 
         def non_normal_or_other_parent_paths(self):
             return self._rustmap.non_normal_or_other_parent_paths()
 
         @propertycache
         def dirfoldmap(self):
             f = {}
             normcase = util.normcase
             for name in self._rustmap.tracked_dirs():
                 f[normcase(name)] = name
             return f

mercurial/pure/parsers.py

@@ -1,605 +1,613
 # parsers.py - Python implementation of parsers.c
 #
 # Copyright 2009 Olivia Mackall <[e-mail]> and others
 #
 # 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 struct
 import zlib
 
 from ..node import (
     nullrev,
     sha1nodeconstants,
 )
 from ..thirdparty import attr
 from .. import (
     error,
     pycompat,
     revlogutils,
     util,
 )
 
 from ..revlogutils import nodemap as nodemaputil
 from ..revlogutils import constants as revlog_constants
 
 stringio = pycompat.bytesio
 
 
 _pack = struct.pack
 _unpack = struct.unpack
 _compress = zlib.compress
 _decompress = zlib.decompress
 
 
 # a special value used internally for `size` if the file come from the other parent
 FROM_P2 = -2
 
 # a special value used internally for `size` if the file is modified/merged/added
 NONNORMAL = -1
 
 # a special value used internally for `time` if the time is ambigeous
 AMBIGUOUS_TIME = -1
 
 
 @attr.s(slots=True, init=False)
 class DirstateItem(object):
     """represent a dirstate entry
 
     It contains:
 
     - state (one of 'n', 'a', 'r', 'm')
     - mode,
     - size,
     - mtime,
     """
 
     _state = attr.ib()
     _mode = attr.ib()
     _size = attr.ib()
     _mtime = attr.ib()
 
     def __init__(self, state, mode, size, mtime):
         self._state = state
         self._mode = mode
         self._size = size
         self._mtime = mtime
 
     @classmethod
     def from_v1_data(cls, state, mode, size, mtime):
         """Build a new DirstateItem object from V1 data
 
         Since the dirstate-v1 format is frozen, the signature of this function
         is not expected to change, unlike the __init__ one.
         """
         return cls(
             state=state,
             mode=mode,
             size=size,
             mtime=mtime,
         )
 
     def set_possibly_dirty(self):
         """Mark a file as "possibly dirty"
 
         This means the next status call will have to actually check its content
         to make sure it is correct.
         """
         self._mtime = AMBIGUOUS_TIME
 
     def __getitem__(self, idx):
         if idx == 0 or idx == -4:
             msg = b"do not use item[x], use item.state"
             util.nouideprecwarn(msg, b'6.0', stacklevel=2)
             return self._state
         elif idx == 1 or idx == -3:
             msg = b"do not use item[x], use item.mode"
             util.nouideprecwarn(msg, b'6.0', stacklevel=2)
             return self._mode
         elif idx == 2 or idx == -2:
             msg = b"do not use item[x], use item.size"
             util.nouideprecwarn(msg, b'6.0', stacklevel=2)
             return self._size
         elif idx == 3 or idx == -1:
             msg = b"do not use item[x], use item.mtime"
             util.nouideprecwarn(msg, b'6.0', stacklevel=2)
             return self._mtime
         else:
             raise IndexError(idx)
 
     @property
     def mode(self):
         return self._mode
 
     @property
     def size(self):
         return self._size
 
     @property
     def mtime(self):
         return self._mtime
 
     @property
     def state(self):
         """
         States are:
           n  normal
           m  needs merging
           r  marked for removal
           a  marked for addition
 
         XXX This "state" is a bit obscure and mostly a direct expression of the
         dirstatev1 format. It would make sense to ultimately deprecate it in
         favor of the more "semantic" attributes.
         """
         return self._state
 
     @property
     def tracked(self):
         """True is the file is tracked in the working copy"""
         return self._state in b"nma"
 
     @property
     def added(self):
         """True if the file has been added"""
         return self._state == b'a'
 
     @property
     def merged(self):
         """True if the file has been merged
 
         Should only be set if a merge is in progress in the dirstate
         """
         return self._state == b'm'
 
     @property
     def from_p2(self):
         """True if the file have been fetched from p2 during the current merge
 
         This is only True is the file is currently tracked.
 
         Should only be set if a merge is in progress in the dirstate
         """
         return self._state == b'n' and self._size == FROM_P2
 
     @property
     def from_p2_removed(self):
         """True if the file has been removed, but was "from_p2" initially
 
         This property seems like an abstraction leakage and should probably be
         dealt in this class (or maybe the dirstatemap) directly.
         """
         return self._state == b'r' and self._size == FROM_P2
 
     @property
     def removed(self):
         """True if the file has been removed"""
         return self._state == b'r'
 
     @property
     def merged_removed(self):
         """True if the file has been removed, but was "merged" initially
 
         This property seems like an abstraction leakage and should probably be
         dealt in this class (or maybe the dirstatemap)  directly.
         """
         return self._state == b'r' and self._size == NONNORMAL
 
+    @property
+    def dm_nonnormal(self):
+        """True is the entry is non-normal in the dirstatemap sense
+
+        There is no reason for any code, but the dirstatemap one to use this.
+        """
+        return self.state != b'n' or self.mtime == AMBIGUOUS_TIME
+
     def v1_state(self):
         """return a "state" suitable for v1 serialization"""
         return self._state
 
     def v1_mode(self):
         """return a "mode" suitable for v1 serialization"""
         return self._mode
 
     def v1_size(self):
         """return a "size" suitable for v1 serialization"""
         return self._size
 
     def v1_mtime(self):
         """return a "mtime" suitable for v1 serialization"""
         return self._mtime
 
     def need_delay(self, now):
         """True if the stored mtime would be ambiguous with the current time"""
         return self._state == b'n' and self._mtime == now
 
 
 def gettype(q):
     return int(q & 0xFFFF)
 
 
 class BaseIndexObject(object):
     # Can I be passed to an algorithme implemented in Rust ?
     rust_ext_compat = 0
     # Format of an index entry according to Python's `struct` language
     index_format = revlog_constants.INDEX_ENTRY_V1
     # Size of a C unsigned long long int, platform independent
     big_int_size = struct.calcsize(b'>Q')
     # Size of a C long int, platform independent
     int_size = struct.calcsize(b'>i')
     # An empty index entry, used as a default value to be overridden, or nullrev
     null_item = (
         0,
         0,
         0,
         -1,
         -1,
         -1,
         -1,
         sha1nodeconstants.nullid,
         0,
         0,
         revlog_constants.COMP_MODE_INLINE,
         revlog_constants.COMP_MODE_INLINE,
     )
 
     @util.propertycache
     def entry_size(self):
         return self.index_format.size
 
     @property
     def nodemap(self):
         msg = b"index.nodemap is deprecated, use index.[has_node|rev|get_rev]"
         util.nouideprecwarn(msg, b'5.3', stacklevel=2)
         return self._nodemap
 
     @util.propertycache
     def _nodemap(self):
         nodemap = nodemaputil.NodeMap({sha1nodeconstants.nullid: nullrev})
         for r in range(0, len(self)):
             n = self[r][7]
             nodemap[n] = r
         return nodemap
 
     def has_node(self, node):
         """return True if the node exist in the index"""
         return node in self._nodemap
 
     def rev(self, node):
         """return a revision for a node
 
         If the node is unknown, raise a RevlogError"""
         return self._nodemap[node]
 
     def get_rev(self, node):
         """return a revision for a node
 
         If the node is unknown, return None"""
         return self._nodemap.get(node)
 
     def _stripnodes(self, start):
         if '_nodemap' in vars(self):
             for r in range(start, len(self)):
                 n = self[r][7]
                 del self._nodemap[n]
 
     def clearcaches(self):
         self.__dict__.pop('_nodemap', None)
 
     def __len__(self):
         return self._lgt + len(self._extra)
 
     def append(self, tup):
         if '_nodemap' in vars(self):
             self._nodemap[tup[7]] = len(self)
         data = self._pack_entry(len(self), tup)
         self._extra.append(data)
 
     def _pack_entry(self, rev, entry):
         assert entry[8] == 0
         assert entry[9] == 0
         return self.index_format.pack(*entry[:8])
 
     def _check_index(self, i):
         if not isinstance(i, int):
             raise TypeError(b"expecting int indexes")
         if i < 0 or i >= len(self):
             raise IndexError
 
     def __getitem__(self, i):
         if i == -1:
             return self.null_item
         self._check_index(i)
         if i >= self._lgt:
             data = self._extra[i - self._lgt]
         else:
             index = self._calculate_index(i)
             data = self._data[index : index + self.entry_size]
         r = self._unpack_entry(i, data)
         if self._lgt and i == 0:
             offset = revlogutils.offset_type(0, gettype(r[0]))
             r = (offset,) + r[1:]
         return r
 
     def _unpack_entry(self, rev, data):
         r = self.index_format.unpack(data)
         r = r + (
             0,
             0,
             revlog_constants.COMP_MODE_INLINE,
             revlog_constants.COMP_MODE_INLINE,
         )
         return r
 
     def pack_header(self, header):
         """pack header information as binary"""
         v_fmt = revlog_constants.INDEX_HEADER
         return v_fmt.pack(header)
 
     def entry_binary(self, rev):
         """return the raw binary string representing a revision"""
         entry = self[rev]
         p = revlog_constants.INDEX_ENTRY_V1.pack(*entry[:8])
         if rev == 0:
             p = p[revlog_constants.INDEX_HEADER.size :]
         return p
 
 
 class IndexObject(BaseIndexObject):
     def __init__(self, data):
         assert len(data) % self.entry_size == 0, (
             len(data),
             self.entry_size,
             len(data) % self.entry_size,
         )
         self._data = data
         self._lgt = len(data) // self.entry_size
         self._extra = []
 
     def _calculate_index(self, i):
         return i * self.entry_size
 
     def __delitem__(self, i):
         if not isinstance(i, slice) or not i.stop == -1 or i.step is not None:
             raise ValueError(b"deleting slices only supports a:-1 with step 1")
         i = i.start
         self._check_index(i)
         self._stripnodes(i)
         if i < self._lgt:
             self._data = self._data[: i * self.entry_size]
             self._lgt = i
             self._extra = []
         else:
             self._extra = self._extra[: i - self._lgt]
 
 
 class PersistentNodeMapIndexObject(IndexObject):
     """a Debug oriented class to test persistent nodemap
 
     We need a simple python object to test API and higher level behavior. See
     the Rust implementation for  more serious usage. This should be used only
     through the dedicated `devel.persistent-nodemap` config.
     """
 
     def nodemap_data_all(self):
         """Return bytes containing a full serialization of a nodemap
 
         The nodemap should be valid for the full set of revisions in the
         index."""
         return nodemaputil.persistent_data(self)
 
     def nodemap_data_incremental(self):
         """Return bytes containing a incremental update to persistent nodemap
 
         This containst the data for an append-only update of the data provided
         in the last call to `update_nodemap_data`.
         """
         if self._nm_root is None:
             return None
         docket = self._nm_docket
         changed, data = nodemaputil.update_persistent_data(
             self, self._nm_root, self._nm_max_idx, self._nm_docket.tip_rev
         )
 
         self._nm_root = self._nm_max_idx = self._nm_docket = None
         return docket, changed, data
 
     def update_nodemap_data(self, docket, nm_data):
         """provide full block of persisted binary data for a nodemap
 
         The data are expected to come from disk. See `nodemap_data_all` for a
         produceur of such data."""
         if nm_data is not None:
             self._nm_root, self._nm_max_idx = nodemaputil.parse_data(nm_data)
             if self._nm_root:
                 self._nm_docket = docket
             else:
                 self._nm_root = self._nm_max_idx = self._nm_docket = None
 
 
 class InlinedIndexObject(BaseIndexObject):
     def __init__(self, data, inline=0):
         self._data = data
         self._lgt = self._inline_scan(None)
         self._inline_scan(self._lgt)
         self._extra = []
 
     def _inline_scan(self, lgt):
         off = 0
         if lgt is not None:
             self._offsets = [0] * lgt
         count = 0
         while off <= len(self._data) - self.entry_size:
             start = off + self.big_int_size
             (s,) = struct.unpack(
                 b'>i',
                 self._data[start : start + self.int_size],
             )
             if lgt is not None:
                 self._offsets[count] = off
             count += 1
             off += self.entry_size + s
         if off != len(self._data):
             raise ValueError(b"corrupted data")
         return count
 
     def __delitem__(self, i):
         if not isinstance(i, slice) or not i.stop == -1 or i.step is not None:
             raise ValueError(b"deleting slices only supports a:-1 with step 1")
         i = i.start
         self._check_index(i)
         self._stripnodes(i)
         if i < self._lgt:
             self._offsets = self._offsets[:i]
             self._lgt = i
             self._extra = []
         else:
             self._extra = self._extra[: i - self._lgt]
 
     def _calculate_index(self, i):
         return self._offsets[i]
 
 
 def parse_index2(data, inline, revlogv2=False):
     if not inline:
         cls = IndexObject2 if revlogv2 else IndexObject
         return cls(data), None
     cls = InlinedIndexObject
     return cls(data, inline), (0, data)
 
 
 def parse_index_cl_v2(data):
     return IndexChangelogV2(data), None
 
 
 class IndexObject2(IndexObject):
     index_format = revlog_constants.INDEX_ENTRY_V2
 
     def replace_sidedata_info(
         self,
         rev,
         sidedata_offset,
         sidedata_length,
         offset_flags,
         compression_mode,
     ):
         """
         Replace an existing index entry's sidedata offset and length with new
         ones.
         This cannot be used outside of the context of sidedata rewriting,
         inside the transaction that creates the revision `rev`.
         """
         if rev < 0:
             raise KeyError
         self._check_index(rev)
         if rev < self._lgt:
             msg = b"cannot rewrite entries outside of this transaction"
             raise KeyError(msg)
         else:
             entry = list(self[rev])
             entry[0] = offset_flags
             entry[8] = sidedata_offset
             entry[9] = sidedata_length
             entry[11] = compression_mode
             entry = tuple(entry)
             new = self._pack_entry(rev, entry)
             self._extra[rev - self._lgt] = new
 
     def _unpack_entry(self, rev, data):
         data = self.index_format.unpack(data)
         entry = data[:10]
         data_comp = data[10] & 3
         sidedata_comp = (data[10] & (3 << 2)) >> 2
         return entry + (data_comp, sidedata_comp)
 
     def _pack_entry(self, rev, entry):
         data = entry[:10]
         data_comp = entry[10] & 3
         sidedata_comp = (entry[11] & 3) << 2
         data += (data_comp | sidedata_comp,)
 
         return self.index_format.pack(*data)
 
     def entry_binary(self, rev):
         """return the raw binary string representing a revision"""
         entry = self[rev]
         return self._pack_entry(rev, entry)
 
     def pack_header(self, header):
         """pack header information as binary"""
         msg = 'version header should go in the docket, not the index: %d'
         msg %= header
         raise error.ProgrammingError(msg)
 
 
 class IndexChangelogV2(IndexObject2):
     index_format = revlog_constants.INDEX_ENTRY_CL_V2
 
     def _unpack_entry(self, rev, data, r=True):
         items = self.index_format.unpack(data)
         entry = items[:3] + (rev, rev) + items[3:8]
         data_comp = items[8] & 3
         sidedata_comp = (items[8] >> 2) & 3
         return entry + (data_comp, sidedata_comp)
 
     def _pack_entry(self, rev, entry):
         assert entry[3] == rev, entry[3]
         assert entry[4] == rev, entry[4]
         data = entry[:3] + entry[5:10]
         data_comp = entry[10] & 3
         sidedata_comp = (entry[11] & 3) << 2
         data += (data_comp | sidedata_comp,)
         return self.index_format.pack(*data)
 
 
 def parse_index_devel_nodemap(data, inline):
     """like parse_index2, but alway return a PersistentNodeMapIndexObject"""
     return PersistentNodeMapIndexObject(data), None
 
 
 def parse_dirstate(dmap, copymap, st):
     parents = [st[:20], st[20:40]]
     # dereference fields so they will be local in loop
     format = b">cllll"
     e_size = struct.calcsize(format)
     pos1 = 40
     l = len(st)
 
     # the inner loop
     while pos1 < l:
         pos2 = pos1 + e_size
         e = _unpack(b">cllll", st[pos1:pos2])  # a literal here is faster
         pos1 = pos2 + e[4]
         f = st[pos2:pos1]
         if b'\0' in f:
             f, c = f.split(b'\0')
             copymap[f] = c
         dmap[f] = DirstateItem.from_v1_data(*e[:4])
     return parents
 
 
 def pack_dirstate(dmap, copymap, pl, now):
     now = int(now)
     cs = stringio()
     write = cs.write
     write(b"".join(pl))
     for f, e in pycompat.iteritems(dmap):
         if e.need_delay(now):
             # The file was last modified "simultaneously" with the current
             # write to dirstate (i.e. within the same second for file-
             # systems with a granularity of 1 sec). This commonly happens
             # for at least a couple of files on 'update'.
             # The user could change the file without changing its size
             # within the same second. Invalidate the file's mtime in
             # dirstate, forcing future 'status' calls to compare the
             # contents of the file if the size is the same. This prevents
             # mistakenly treating such files as clean.
             e.set_possibly_dirty()
 
         if f in copymap:
             f = b"%s\0%s" % (f, copymap[f])
         e = _pack(
             b">cllll",
             e.v1_state(),
             e.v1_mode(),
             e.v1_size(),
             e.v1_mtime(),
             len(f),
         )
         write(e)
         write(f)
     return cs.getvalue()