Mercurial > public > mercurial-scm > hg-stable
view mercurial/diffhelpers.c @ 26117:4dc5b51f38fe
revlog: change generaldelta delta parent heuristic
The old generaldelta heuristic was "if p1 (or p2) was closer than the last full text,
use it, otherwise use prev". This was problematic when a repo contained multiple
branches that were very different. If commits to branch A were pushed, and the
last full text was branch B, it would generate a fulltext. Then if branch B was
pushed, it would generate another fulltext. The problem is that the last
fulltext (and delta'ing against `prev` in general) has no correlation with the
contents of the incoming revision, and therefore will always have degenerate
cases.
According to the blame, that algorithm was chosen to minimize the chain length.
Since there is already code that protects against that (the delta-vs-fulltext
code), and since it has been improved since the original generaldelta algorithm
went in (2011), I believe the chain length criteria will still be preserved.
The new algorithm always diffs against p1 (or p2 if it's closer), unless the
resulting delta will fail the delta-vs-fulltext check, in which case we delta
against prev.
Some before and after stats on manifest.d size.
internal large repo
old heuristic - 2.0 GB
new heuristic - 1.2 GB
mozilla-central
old heuristic - 242 MB
new heuristic - 261 MB
The regression in mozilla central is due to the new heuristic choosing p2r as
the delta when it's closer to the tip. Switching the algorithm to always prefer
p1r brings the size back down (242 MB). This is result of the way in which
mozilla does merges and pushes, and the result could easily swing the other
direction in other repos (depending on if they merge X into Y or Y into X), but
will never be as degenerate as before.
I future patch will address the regression by introducing an optional, even more
aggressive delta heuristic which will knock the mozilla manifest size down
dramatically.
author | Durham Goode <durham@fb.com> |
---|---|
date | Sun, 30 Aug 2015 13:58:11 -0700 |
parents | f1c127df7c4f |
children | 2f51f4c535d2 |
line wrap: on
line source
/* * diffhelpers.c - helper routines for mpatch * * Copyright 2007 Chris Mason <chris.mason@oracle.com> * * This software may be used and distributed according to the terms * of the GNU General Public License v2, incorporated herein by reference. */ #include <Python.h> #include <stdlib.h> #include <string.h> #include "util.h" static char diffhelpers_doc[] = "Efficient diff parsing"; static PyObject *diffhelpers_Error; /* fixup the last lines of a and b when the patch has no newline at eof */ static void _fix_newline(PyObject *hunk, PyObject *a, PyObject *b) { Py_ssize_t hunksz = PyList_Size(hunk); PyObject *s = PyList_GET_ITEM(hunk, hunksz-1); char *l = PyBytes_AsString(s); Py_ssize_t alen = PyList_Size(a); Py_ssize_t blen = PyList_Size(b); char c = l[0]; PyObject *hline; Py_ssize_t sz = PyBytes_GET_SIZE(s); if (sz > 1 && l[sz-2] == '\r') /* tolerate CRLF in last line */ sz -= 1; hline = PyBytes_FromStringAndSize(l, sz-1); if (!hline) { return; } if (c == ' ' || c == '+') { PyObject *rline = PyBytes_FromStringAndSize(l + 1, sz - 2); PyList_SetItem(b, blen-1, rline); } if (c == ' ' || c == '-') { Py_INCREF(hline); PyList_SetItem(a, alen-1, hline); } PyList_SetItem(hunk, hunksz-1, hline); } /* python callable form of _fix_newline */ static PyObject * fix_newline(PyObject *self, PyObject *args) { PyObject *hunk, *a, *b; if (!PyArg_ParseTuple(args, "OOO", &hunk, &a, &b)) return NULL; _fix_newline(hunk, a, b); return Py_BuildValue("l", 0); } #if (PY_VERSION_HEX < 0x02050000) static const char *addlines_format = "OOiiOO"; #else static const char *addlines_format = "OOnnOO"; #endif /* * read lines from fp into the hunk. The hunk is parsed into two arrays * a and b. a gets the old state of the text, b gets the new state * The control char from the hunk is saved when inserting into a, but not b * (for performance while deleting files) */ static PyObject * addlines(PyObject *self, PyObject *args) { PyObject *fp, *hunk, *a, *b, *x; Py_ssize_t i; Py_ssize_t lena, lenb; Py_ssize_t num; Py_ssize_t todoa, todob; char *s, c; PyObject *l; if (!PyArg_ParseTuple(args, addlines_format, &fp, &hunk, &lena, &lenb, &a, &b)) return NULL; while (1) { todoa = lena - PyList_Size(a); todob = lenb - PyList_Size(b); num = todoa > todob ? todoa : todob; if (num == 0) break; for (i = 0; i < num; i++) { x = PyFile_GetLine(fp, 0); s = PyBytes_AsString(x); c = *s; if (strcmp(s, "\\ No newline at end of file\n") == 0) { _fix_newline(hunk, a, b); continue; } if (c == '\n') { /* Some patches may be missing the control char * on empty lines. Supply a leading space. */ Py_DECREF(x); x = PyBytes_FromString(" \n"); } PyList_Append(hunk, x); if (c == '+') { l = PyBytes_FromString(s + 1); PyList_Append(b, l); Py_DECREF(l); } else if (c == '-') { PyList_Append(a, x); } else { l = PyBytes_FromString(s + 1); PyList_Append(b, l); Py_DECREF(l); PyList_Append(a, x); } Py_DECREF(x); } } return Py_BuildValue("l", 0); } /* * compare the lines in a with the lines in b. a is assumed to have * a control char at the start of each line, this char is ignored in the * compare */ static PyObject * testhunk(PyObject *self, PyObject *args) { PyObject *a, *b; long bstart; Py_ssize_t alen, blen; Py_ssize_t i; char *sa, *sb; if (!PyArg_ParseTuple(args, "OOl", &a, &b, &bstart)) return NULL; alen = PyList_Size(a); blen = PyList_Size(b); if (alen > blen - bstart || bstart < 0) { return Py_BuildValue("l", -1); } for (i = 0; i < alen; i++) { sa = PyBytes_AsString(PyList_GET_ITEM(a, i)); sb = PyBytes_AsString(PyList_GET_ITEM(b, i + bstart)); if (strcmp(sa + 1, sb) != 0) return Py_BuildValue("l", -1); } return Py_BuildValue("l", 0); } static PyMethodDef methods[] = { {"addlines", addlines, METH_VARARGS, "add lines to a hunk\n"}, {"fix_newline", fix_newline, METH_VARARGS, "fixup newline counters\n"}, {"testhunk", testhunk, METH_VARARGS, "test lines in a hunk\n"}, {NULL, NULL} }; #ifdef IS_PY3K static struct PyModuleDef diffhelpers_module = { PyModuleDef_HEAD_INIT, "diffhelpers", diffhelpers_doc, -1, methods }; PyMODINIT_FUNC PyInit_diffhelpers(void) { PyObject *m; m = PyModule_Create(&diffhelpers_module); if (m == NULL) return NULL; diffhelpers_Error = PyErr_NewException("diffhelpers.diffhelpersError", NULL, NULL); Py_INCREF(diffhelpers_Error); PyModule_AddObject(m, "diffhelpersError", diffhelpers_Error); return m; } #else PyMODINIT_FUNC initdiffhelpers(void) { Py_InitModule3("diffhelpers", methods, diffhelpers_doc); diffhelpers_Error = PyErr_NewException("diffhelpers.diffhelpersError", NULL, NULL); } #endif