mercurial/cext/bdiff.c
author Augie Fackler <augie@google.com>
Thu, 24 Jan 2019 10:21:59 -0500
changeset 41336 763b45bc4483
parent 39421 ad76032d27da
child 46819 d4ba4d51f85f
permissions -rw-r--r--
cleanup: use clang-tidy to add missing {} around one-line statements I find this easier to read. Cleanup performed like this: hg files 'set:(**.c or **.cc or **.h) and not "listfile:contrib/clang-format-ignorelist"' | while read f ; do clang-tidy -fix -checks=readability-braces-around-statements $f -- $(python-config --cflags) -Imercurial/cext -Imercurial done make format-c I had to revert chg/chg.c as it's got a construct that seems to confuse clang-tidy, so I'll work on that file later if this change is acceptable. I only tackle files that are under clang-format's authority because otherwise I'd have to do a bunch of manual formatting. A few files didn't get edited because clang-tidy couldn't find some headers. Again, I'll figure that out later assuming this change is accepted. No check-code rule added for now because writing the regex sounds hard. In a perfect world I guess we could write a test that uses clang-tidy on these files, but I think clang-tidy is pretty rarely installed. :/ Differential Revision: https://phab.mercurial-scm.org/D5675
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
     1
/*
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
     2
 bdiff.c - efficient binary diff extension for Mercurial
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
     3
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
     4
 Copyright 2005, 2006 Matt Mackall <mpm@selenic.com>
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
     5
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
     6
 This software may be used and distributed according to the terms of
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
     7
 the GNU General Public License, incorporated herein by reference.
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
     8
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
     9
 Based roughly on Python difflib
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    10
*/
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    11
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    12
#define PY_SSIZE_T_CLEAN
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    13
#include <Python.h>
34438
b90e8da190da cext: reorder #include
Gregory Szorc <gregory.szorc@gmail.com>
parents: 32369
diff changeset
    14
#include <limits.h>
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    15
#include <stdlib.h>
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    16
#include <string.h>
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    17
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    18
#include "bdiff.h"
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    19
#include "bitmanipulation.h"
36675
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
    20
#include "thirdparty/xdiff/xdiff.h"
30170
15635d8b17e0 bdiff: include util.h
Gregory Szorc <gregory.szorc@gmail.com>
parents: 29541
diff changeset
    21
#include "util.h"
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    22
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    23
static PyObject *blocks(PyObject *self, PyObject *args)
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    24
{
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    25
	PyObject *sa, *sb, *rl = NULL, *m;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    26
	struct bdiff_line *a, *b;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    27
	struct bdiff_hunk l, *h;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    28
	int an, bn, count, pos = 0;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    29
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    30
	l.next = NULL;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    31
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
    32
	if (!PyArg_ParseTuple(args, "SS:bdiff", &sa, &sb)) {
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    33
		return NULL;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
    34
	}
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    35
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    36
	an = bdiff_splitlines(PyBytes_AsString(sa), PyBytes_Size(sa), &a);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    37
	bn = bdiff_splitlines(PyBytes_AsString(sb), PyBytes_Size(sb), &b);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    38
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
    39
	if (!a || !b) {
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    40
		goto nomem;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
    41
	}
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    42
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    43
	count = bdiff_diff(a, an, b, bn, &l);
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
    44
	if (count < 0) {
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    45
		goto nomem;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
    46
	}
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    47
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    48
	rl = PyList_New(count);
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
    49
	if (!rl) {
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    50
		goto nomem;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
    51
	}
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    52
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    53
	for (h = l.next; h; h = h->next) {
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    54
		m = Py_BuildValue("iiii", h->a1, h->a2, h->b1, h->b2);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    55
		PyList_SetItem(rl, pos, m);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    56
		pos++;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    57
	}
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    58
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    59
nomem:
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    60
	free(a);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    61
	free(b);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    62
	bdiff_freehunks(l.next);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    63
	return rl ? rl : PyErr_NoMemory();
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    64
}
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    65
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    66
static PyObject *bdiff(PyObject *self, PyObject *args)
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    67
{
36655
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    68
	Py_buffer ba, bb;
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    69
	char *rb, *ia, *ib;
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    70
	PyObject *result = NULL;
36654
b864f4536ca8 cext: refactor cleanup code in bdiff()
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36620
diff changeset
    71
	struct bdiff_line *al = NULL, *bl = NULL;
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    72
	struct bdiff_hunk l, *h;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    73
	int an, bn, count;
30561
7c0c722d568d bdiff: early pruning of common prefix before doing expensive computations
Mads Kiilerich <madski@unity3d.com>
parents: 30170
diff changeset
    74
	Py_ssize_t len = 0, la, lb, li = 0, lcommon = 0, lmax;
36654
b864f4536ca8 cext: refactor cleanup code in bdiff()
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36620
diff changeset
    75
	PyThreadState *_save = NULL;
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    76
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    77
	l.next = NULL;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    78
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
    79
	if (!PyArg_ParseTuple(args, PY23("s*s*:bdiff", "y*y*:bdiff"), &ba,
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
    80
	                      &bb)) {
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    81
		return NULL;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
    82
	}
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    83
36655
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    84
	if (!PyBuffer_IsContiguous(&ba, 'C') || ba.ndim > 1) {
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    85
		PyErr_SetString(PyExc_ValueError, "bdiff input not contiguous");
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    86
		goto cleanup;
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    87
	}
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    88
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    89
	if (!PyBuffer_IsContiguous(&bb, 'C') || bb.ndim > 1) {
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    90
		PyErr_SetString(PyExc_ValueError, "bdiff input not contiguous");
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    91
		goto cleanup;
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    92
	}
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    93
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    94
	la = ba.len;
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    95
	lb = bb.len;
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    96
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    97
	if (la > UINT_MAX || lb > UINT_MAX) {
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
    98
		PyErr_SetString(PyExc_ValueError, "bdiff inputs too large");
36655
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
    99
		goto cleanup;
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   100
	}
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   101
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   102
	_save = PyEval_SaveThread();
30561
7c0c722d568d bdiff: early pruning of common prefix before doing expensive computations
Mads Kiilerich <madski@unity3d.com>
parents: 30170
diff changeset
   103
7c0c722d568d bdiff: early pruning of common prefix before doing expensive computations
Mads Kiilerich <madski@unity3d.com>
parents: 30170
diff changeset
   104
	lmax = la > lb ? lb : la;
36655
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
   105
	for (ia = ba.buf, ib = bb.buf; li < lmax && *ia == *ib;
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
   106
	     ++li, ++ia, ++ib) {
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   107
		if (*ia == '\n') {
30561
7c0c722d568d bdiff: early pruning of common prefix before doing expensive computations
Mads Kiilerich <madski@unity3d.com>
parents: 30170
diff changeset
   108
			lcommon = li + 1;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   109
		}
36655
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
   110
	}
30561
7c0c722d568d bdiff: early pruning of common prefix before doing expensive computations
Mads Kiilerich <madski@unity3d.com>
parents: 30170
diff changeset
   111
	/* we can almost add: if (li == lmax) lcommon = li; */
7c0c722d568d bdiff: early pruning of common prefix before doing expensive computations
Mads Kiilerich <madski@unity3d.com>
parents: 30170
diff changeset
   112
36681
340e4b711df7 bdiff: avoid pointer arithmetic on void*
Matt Harbison <matt_harbison@yahoo.com>
parents: 36675
diff changeset
   113
	an = bdiff_splitlines((char *)ba.buf + lcommon, la - lcommon, &al);
340e4b711df7 bdiff: avoid pointer arithmetic on void*
Matt Harbison <matt_harbison@yahoo.com>
parents: 36675
diff changeset
   114
	bn = bdiff_splitlines((char *)bb.buf + lcommon, lb - lcommon, &bl);
36654
b864f4536ca8 cext: refactor cleanup code in bdiff()
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36620
diff changeset
   115
	if (!al || !bl) {
b864f4536ca8 cext: refactor cleanup code in bdiff()
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36620
diff changeset
   116
		PyErr_NoMemory();
b864f4536ca8 cext: refactor cleanup code in bdiff()
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36620
diff changeset
   117
		goto cleanup;
b864f4536ca8 cext: refactor cleanup code in bdiff()
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36620
diff changeset
   118
	}
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   119
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   120
	count = bdiff_diff(al, an, bl, bn, &l);
36654
b864f4536ca8 cext: refactor cleanup code in bdiff()
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36620
diff changeset
   121
	if (count < 0) {
b864f4536ca8 cext: refactor cleanup code in bdiff()
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36620
diff changeset
   122
		PyErr_NoMemory();
b864f4536ca8 cext: refactor cleanup code in bdiff()
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36620
diff changeset
   123
		goto cleanup;
b864f4536ca8 cext: refactor cleanup code in bdiff()
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36620
diff changeset
   124
	}
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   125
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   126
	/* calculate length of output */
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   127
	la = lb = 0;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   128
	for (h = l.next; h; h = h->next) {
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   129
		if (h->a1 != la || h->b1 != lb) {
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   130
			len += 12 + bl[h->b1].l - bl[lb].l;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   131
		}
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   132
		la = h->a2;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   133
		lb = h->b2;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   134
	}
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   135
	PyEval_RestoreThread(_save);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   136
	_save = NULL;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   137
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   138
	result = PyBytes_FromStringAndSize(NULL, len);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   139
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   140
	if (!result) {
36654
b864f4536ca8 cext: refactor cleanup code in bdiff()
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36620
diff changeset
   141
		goto cleanup;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   142
	}
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   143
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   144
	/* build binary patch */
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   145
	rb = PyBytes_AsString(result);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   146
	la = lb = 0;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   147
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   148
	for (h = l.next; h; h = h->next) {
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   149
		if (h->a1 != la || h->b1 != lb) {
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   150
			len = bl[h->b1].l - bl[lb].l;
30561
7c0c722d568d bdiff: early pruning of common prefix before doing expensive computations
Mads Kiilerich <madski@unity3d.com>
parents: 30170
diff changeset
   151
			putbe32((uint32_t)(al[la].l + lcommon - al->l), rb);
36055
b4fdc6177b29 bdiff: add to clang-format oversight
Augie Fackler <augie@google.com>
parents: 34438
diff changeset
   152
			putbe32((uint32_t)(al[h->a1].l + lcommon - al->l),
b4fdc6177b29 bdiff: add to clang-format oversight
Augie Fackler <augie@google.com>
parents: 34438
diff changeset
   153
			        rb + 4);
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   154
			putbe32((uint32_t)len, rb + 8);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   155
			memcpy(rb + 12, bl[lb].l, len);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   156
			rb += 12 + len;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   157
		}
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   158
		la = h->a2;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   159
		lb = h->b2;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   160
	}
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   161
36654
b864f4536ca8 cext: refactor cleanup code in bdiff()
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36620
diff changeset
   162
cleanup:
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   163
	if (_save) {
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   164
		PyEval_RestoreThread(_save);
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   165
	}
36655
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
   166
	PyBuffer_Release(&ba);
68026dd7c4f9 cext: accept arguments as Py_buffer
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36654
diff changeset
   167
	PyBuffer_Release(&bb);
38301
d9e87566f879 cext: stop worrying and love the free(NULL)
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 37980
diff changeset
   168
	free(al);
d9e87566f879 cext: stop worrying and love the free(NULL)
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 37980
diff changeset
   169
	free(bl);
38309
93b812d5b818 bdiff: one more safe call of bdiff_freehunks(NULL)
Yuya Nishihara <yuya@tcha.org>
parents: 38301
diff changeset
   170
	bdiff_freehunks(l.next);
36654
b864f4536ca8 cext: refactor cleanup code in bdiff()
Gregory Szorc <gregory.szorc@gmail.com>
parents: 36620
diff changeset
   171
	return result;
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   172
}
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   173
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   174
/*
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   175
 * If allws != 0, remove all whitespace (' ', \t and \r). Otherwise,
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   176
 * reduce whitespace sequences to a single space and trim remaining whitespace
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   177
 * from end of lines.
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   178
 */
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   179
static PyObject *fixws(PyObject *self, PyObject *args)
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   180
{
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   181
	PyObject *s, *result = NULL;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   182
	char allws, c;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   183
	const char *r;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   184
	Py_ssize_t i, rlen, wlen = 0;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   185
	char *w;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   186
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   187
	if (!PyArg_ParseTuple(args, "Sb:fixws", &s, &allws)) {
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   188
		return NULL;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   189
	}
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   190
	r = PyBytes_AsString(s);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   191
	rlen = PyBytes_Size(s);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   192
31467
08ecec297521 bdiff: use Python memory allocator in fixws
Gregory Szorc <gregory.szorc@gmail.com>
parents: 30561
diff changeset
   193
	w = (char *)PyMem_Malloc(rlen ? rlen : 1);
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   194
	if (!w) {
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   195
		goto nomem;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   196
	}
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   197
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   198
	for (i = 0; i != rlen; i++) {
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   199
		c = r[i];
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   200
		if (c == ' ' || c == '\t' || c == '\r') {
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   201
			if (!allws && (wlen == 0 || w[wlen - 1] != ' ')) {
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   202
				w[wlen++] = ' ';
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   203
			}
36055
b4fdc6177b29 bdiff: add to clang-format oversight
Augie Fackler <augie@google.com>
parents: 34438
diff changeset
   204
		} else if (c == '\n' && !allws && wlen > 0 &&
b4fdc6177b29 bdiff: add to clang-format oversight
Augie Fackler <augie@google.com>
parents: 34438
diff changeset
   205
		           w[wlen - 1] == ' ') {
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   206
			w[wlen - 1] = '\n';
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   207
		} else {
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   208
			w[wlen++] = c;
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   209
		}
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   210
	}
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   211
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   212
	result = PyBytes_FromStringAndSize(w, wlen);
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   213
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   214
nomem:
31467
08ecec297521 bdiff: use Python memory allocator in fixws
Gregory Szorc <gregory.szorc@gmail.com>
parents: 30561
diff changeset
   215
	PyMem_Free(w);
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   216
	return result ? result : PyErr_NoMemory();
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   217
}
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   218
36146
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   219
static bool sliceintolist(PyObject *list, Py_ssize_t destidx,
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   220
                          const char *source, Py_ssize_t len)
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   221
{
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   222
	PyObject *sliced = PyBytes_FromStringAndSize(source, len);
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   223
	if (sliced == NULL) {
36146
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   224
		return false;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   225
	}
36146
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   226
	PyList_SET_ITEM(list, destidx, sliced);
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   227
	return true;
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   228
}
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   229
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   230
static PyObject *splitnewlines(PyObject *self, PyObject *args)
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   231
{
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   232
	const char *text;
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   233
	Py_ssize_t nelts = 0, size, i, start = 0;
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   234
	PyObject *result = NULL;
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   235
36620
186c6df3a373 py3: bulk-replace 'const char*' format specifier passed to PyArg_ParseTuple*()
Yuya Nishihara <yuya@tcha.org>
parents: 36146
diff changeset
   236
	if (!PyArg_ParseTuple(args, PY23("s#", "y#"), &text, &size)) {
36146
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   237
		goto abort;
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   238
	}
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   239
	if (!size) {
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   240
		return PyList_New(0);
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   241
	}
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   242
	/* This loops to size-1 because if the last byte is a newline,
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   243
	 * we don't want to perform a split there. */
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   244
	for (i = 0; i < size - 1; ++i) {
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   245
		if (text[i] == '\n') {
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   246
			++nelts;
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   247
		}
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   248
	}
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   249
	if ((result = PyList_New(nelts + 1)) == NULL) {
36146
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   250
		goto abort;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   251
	}
36146
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   252
	nelts = 0;
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   253
	for (i = 0; i < size - 1; ++i) {
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   254
		if (text[i] == '\n') {
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   255
			if (!sliceintolist(result, nelts++, text + start,
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   256
			                   i - start + 1)) {
36146
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   257
				goto abort;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   258
			}
36146
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   259
			start = i + 1;
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   260
		}
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   261
	}
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   262
	if (!sliceintolist(result, nelts++, text + start, size - start)) {
36146
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   263
		goto abort;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   264
	}
36146
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   265
	return result;
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   266
abort:
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   267
	Py_XDECREF(result);
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   268
	return NULL;
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   269
}
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   270
36916
1b9f6440506b bdiff: convert more longs to int64_t
Matt Harbison <matt_harbison@yahoo.com>
parents: 36763
diff changeset
   271
static int hunk_consumer(int64_t a1, int64_t a2, int64_t b1, int64_t b2,
1b9f6440506b bdiff: convert more longs to int64_t
Matt Harbison <matt_harbison@yahoo.com>
parents: 36763
diff changeset
   272
                         void *priv)
36675
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   273
{
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   274
	PyObject *rl = (PyObject *)priv;
37980
273ea09f6550 bdiff: fix yet more fallout from xdiff long/int64 conversion (issue5885)
Julien Cristau <jcristau@debian.org>
parents: 36916
diff changeset
   275
	PyObject *m = Py_BuildValue("LLLL", a1, a2, b1, b2);
39421
ad76032d27da xdiff: fix leak in hunk_consumer()
Yuya Nishihara <yuya@tcha.org>
parents: 38309
diff changeset
   276
	int r;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   277
	if (!m) {
36675
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   278
		return -1;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   279
	}
39421
ad76032d27da xdiff: fix leak in hunk_consumer()
Yuya Nishihara <yuya@tcha.org>
parents: 38309
diff changeset
   280
	r = PyList_Append(rl, m);
ad76032d27da xdiff: fix leak in hunk_consumer()
Yuya Nishihara <yuya@tcha.org>
parents: 38309
diff changeset
   281
	Py_DECREF(m);
ad76032d27da xdiff: fix leak in hunk_consumer()
Yuya Nishihara <yuya@tcha.org>
parents: 38309
diff changeset
   282
	return r;
36675
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   283
}
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   284
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   285
static PyObject *xdiffblocks(PyObject *self, PyObject *args)
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   286
{
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   287
	Py_ssize_t la, lb;
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   288
	mmfile_t a, b;
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   289
	PyObject *rl;
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   290
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   291
	xpparam_t xpp = {
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   292
	    XDF_INDENT_HEURISTIC, /* flags */
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   293
	};
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   294
	xdemitconf_t xecfg = {
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   295
	    XDL_EMIT_BDIFFHUNK, /* flags */
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   296
	    hunk_consumer,      /* hunk_consume_func */
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   297
	};
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   298
	xdemitcb_t ecb = {
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   299
	    NULL, /* priv */
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   300
	};
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   301
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   302
	if (!PyArg_ParseTuple(args, PY23("s#s#", "y#y#"), &a.ptr, &la, &b.ptr,
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   303
	                      &lb)) {
36675
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   304
		return NULL;
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   305
	}
36675
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   306
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   307
	a.size = la;
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   308
	b.size = lb;
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   309
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   310
	rl = PyList_New(0);
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   311
	if (!rl) {
36675
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   312
		return PyErr_NoMemory();
41336
763b45bc4483 cleanup: use clang-tidy to add missing {} around one-line statements
Augie Fackler <augie@google.com>
parents: 39421
diff changeset
   313
	}
36675
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   314
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   315
	ecb.priv = rl;
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   316
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   317
	if (xdl_diff(&a, &b, &xpp, &xecfg, &ecb) != 0) {
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   318
		Py_DECREF(rl);
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   319
		return PyErr_NoMemory();
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   320
	}
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   321
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   322
	return rl;
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   323
}
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   324
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   325
static char mdiff_doc[] = "Efficient binary diff.";
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   326
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   327
static PyMethodDef methods[] = {
36055
b4fdc6177b29 bdiff: add to clang-format oversight
Augie Fackler <augie@google.com>
parents: 34438
diff changeset
   328
    {"bdiff", bdiff, METH_VARARGS, "calculate a binary diff\n"},
b4fdc6177b29 bdiff: add to clang-format oversight
Augie Fackler <augie@google.com>
parents: 34438
diff changeset
   329
    {"blocks", blocks, METH_VARARGS, "find a list of matching lines\n"},
b4fdc6177b29 bdiff: add to clang-format oversight
Augie Fackler <augie@google.com>
parents: 34438
diff changeset
   330
    {"fixws", fixws, METH_VARARGS, "normalize diff whitespaces\n"},
36146
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   331
    {"splitnewlines", splitnewlines, METH_VARARGS,
29dd37a418aa bdiff: write a native version of splitnewlines
Augie Fackler <augie@google.com>
parents: 36055
diff changeset
   332
     "like str.splitlines, but only split on newlines\n"},
36675
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   333
    {"xdiffblocks", xdiffblocks, METH_VARARGS,
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   334
     "find a list of matching lines using xdiff algorithm\n"},
36055
b4fdc6177b29 bdiff: add to clang-format oversight
Augie Fackler <augie@google.com>
parents: 34438
diff changeset
   335
    {NULL, NULL},
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   336
};
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   337
36675
430fdb717549 bdiff: add a xdiffblocks method
Jun Wu <quark@fb.com>
parents: 36655
diff changeset
   338
static const int version = 3;
32355
4195b84940e9 bdiff: add version to help detect breaking binary changes
Jun Wu <quark@fb.com>
parents: 31467
diff changeset
   339
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   340
#ifdef IS_PY3K
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   341
static struct PyModuleDef bdiff_module = {
36055
b4fdc6177b29 bdiff: add to clang-format oversight
Augie Fackler <augie@google.com>
parents: 34438
diff changeset
   342
    PyModuleDef_HEAD_INIT, "bdiff", mdiff_doc, -1, methods,
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   343
};
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   344
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   345
PyMODINIT_FUNC PyInit_bdiff(void)
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   346
{
32355
4195b84940e9 bdiff: add version to help detect breaking binary changes
Jun Wu <quark@fb.com>
parents: 31467
diff changeset
   347
	PyObject *m;
4195b84940e9 bdiff: add version to help detect breaking binary changes
Jun Wu <quark@fb.com>
parents: 31467
diff changeset
   348
	m = PyModule_Create(&bdiff_module);
4195b84940e9 bdiff: add version to help detect breaking binary changes
Jun Wu <quark@fb.com>
parents: 31467
diff changeset
   349
	PyModule_AddIntConstant(m, "version", version);
4195b84940e9 bdiff: add version to help detect breaking binary changes
Jun Wu <quark@fb.com>
parents: 31467
diff changeset
   350
	return m;
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   351
}
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   352
#else
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   353
PyMODINIT_FUNC initbdiff(void)
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   354
{
32355
4195b84940e9 bdiff: add version to help detect breaking binary changes
Jun Wu <quark@fb.com>
parents: 31467
diff changeset
   355
	PyObject *m;
4195b84940e9 bdiff: add version to help detect breaking binary changes
Jun Wu <quark@fb.com>
parents: 31467
diff changeset
   356
	m = Py_InitModule3("bdiff", methods, mdiff_doc);
4195b84940e9 bdiff: add version to help detect breaking binary changes
Jun Wu <quark@fb.com>
parents: 31467
diff changeset
   357
	PyModule_AddIntConstant(m, "version", version);
29541
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   358
}
9631ff5ebbeb bdiff: split bdiff into cpy-aware and cpy-agnostic part
Maciej Fijalkowski <fijall@gmail.com>
parents:
diff changeset
   359
#endif