Initial implementation of git_diff_blob

This gets the basic plumbing in place for git_diff_blob.
There is a known issue where additional parameters like
the number of lines of context to display on the diff
are not working correctly (which leads one of the new
unit tests to fail).

include/git2/diff.h

+/*
+ * Copyright (C) 2009-2012 the libgit2 contributors
+ *
+ * This file is part of libgit2, distributed under the GNU GPL v2 with
+ * a Linking Exception. For full terms see the included COPYING file.
+ */
+#ifndef INCLUDE_git_diff_h__
+#define INCLUDE_git_diff_h__
+
+#include "common.h"
+#include "types.h"
+#include "oid.h"
+#include "tree.h"
+#include "refs.h"
+
+/**
+ * @file git2/diff.h
+ * @brief Git tree and file differencing routines.
+ * @ingroup Git
+ * @{
+ */
+GIT_BEGIN_DECL
+
+typedef int (*git_diff_file_fn)(
+	void *cb_data,
+	const git_oid *old,
+	const char *old_path,
+	int old_mode,
+	const git_oid *new, /* hashed object if from work tree */
+	const char *new_path,
+	int new_mode);
+
+typedef int (*git_diff_hunk_fn)(
+	void *cb_data,
+	int old_start,
+	int old_lines,
+	int new_start,
+	int new_lines);
+
+#define GIT_DIFF_LINE_CONTEXT  0
+#define GIT_DIFF_LINE_ADDITION 1
+#define GIT_DIFF_LINE_DELETION 2
+
+typedef int (*git_diff_line_fn)(
+	void *cb_data,
+	int origin, /* GIT_DIFF_LINE value from above */
+	const char *content,
+	size_t content_len);
+
+typedef struct {
+	int context_lines;
+	int interhunk_lines;
+	int ignore_whitespace;
+
+	git_diff_file_fn file_cb;
+	git_diff_hunk_fn hunk_cb;
+	git_diff_line_fn line_cb;
+	void *cb_data;
+} git_diff_opts;
+
+
+GIT_EXTERN(int) git_diff_blobs(
+	git_repository *repo,
+	git_blob *old,
+	git_blob *new,
+	git_diff_opts *options);
+
+GIT_EXTERN(int) git_diff_trees(
+	git_repository *repo,
+	git_tree *old,
+	git_tree *new,
+	git_diff_opts *options);
+
+GIT_EXTERN(int) git_diff_index(
+	git_repository *repo,
+	git_tree *old,
+	git_diff_opts *options);
+
+/* pass NULL for the git_tree to diff workdir against index */
+GIT_EXTERN(int) git_diff_workdir(
+	git_repository *repo,
+	git_tree *old,
+	git_diff_opts *options);
+
+GIT_EXTERN(int) git_diff_workdir_file(
+	git_repository *repo,
+	git_blob *old,
+	const char *path,
+	git_diff_opts *options);
+
+/* pass git_objects to diff against or NULL for index.
+ * can handle: blob->blob, tree->index, tree->tree
+ * it will be an error if object types don't match
+ */
+/* pass git_object to diff WT against or NULL for index
+ * can handle: index->wt, tree->wt, blob->wt with path
+ */
+
+GIT_END_DECL
+
+/** @} */
+
+#endif

src/diff.c

+/*
+ * Copyright (C) 2009-2011 the libgit2 contributors
+ *
+ * This file is part of libgit2, distributed under the GNU GPL v2 with
+ * a Linking Exception. For full terms see the included COPYING file.
+ */
+
+#include "common.h"
+#include "git2/diff.h"
+#include "xdiff/xdiff.h"
+#include "blob.h"
+#include <ctype.h>
+
+static int read_next_int(const char **str, int *value)
+{
+	const char *scan = *str;
+	int v = 0, digits = 0;
+	/* find next digit */
+	for (scan = *str; *scan && !isdigit(*scan); scan++);
+	/* parse next number */
+	for (; isdigit(*scan); scan++, digits++)
+		v = (v * 10) + (*scan - '0');
+	*str = scan;
+	*value = v;
+	return (digits > 0) ? GIT_SUCCESS : GIT_ENOTFOUND;
+}
+
+static int diff_output_cb(void *priv, mmbuffer_t *bufs, int len)
+{
+	int err = GIT_SUCCESS;
+	git_diff_opts *opts = priv;
+
+	if (len == 1) {
+		int ostart = -1, olen = 0, nstart = -1, nlen = 0;
+		/* expect something of the form "@@ -%d[,%d] +%d[,%d] @@" */
+		if (opts->hunk_cb && bufs[0].ptr[0] == '@') {
+			const char *scan = bufs[0].ptr;
+			if (!(err = read_next_int(&scan, &ostart)) && *scan == ',')
+				err = read_next_int(&scan, &olen);
+			if (!err && !(err = read_next_int(&scan, &nstart)) && *scan == ',')
+				err = read_next_int(&scan, &nlen);
+			if (!err && ostart >= 0 && nstart >= 0)
+				err = opts->hunk_cb(
+					opts->cb_data, ostart, olen, nstart, nlen);
+		}
+	}
+	else if (len == 2 || len == 3) {
+		int origin;
+		/* expect " "/"-"/"+", then data, then maybe newline */
+		origin =
+			(*bufs[0].ptr == '+') ? GIT_DIFF_LINE_ADDITION :
+			(*bufs[0].ptr == '-') ? GIT_DIFF_LINE_DELETION :
+			GIT_DIFF_LINE_CONTEXT;
+
+		if (opts->line_cb)
+			err = opts->line_cb(
+				opts->cb_data, origin, bufs[1].ptr, bufs[1].size);
+	}
+
+	return err;
+}
+
+int git_diff_blobs(
+	git_repository *repo,
+	git_blob *old_blob,
+	git_blob *new_blob,
+	git_diff_opts *options)
+{
+	mmfile_t old, new;
+	xpparam_t params;
+	xdemitconf_t cfg;
+	xdemitcb_t callback;
+
+	assert(repo && old_blob && new_blob && options);
+
+	old.ptr  = (char *)git_blob_rawcontent(old_blob);
+	old.size = git_blob_rawsize(old_blob);
+
+	new.ptr  = (char *)git_blob_rawcontent(new_blob);
+	new.size = git_blob_rawsize(new_blob);
+
+	memset(&params, 0, sizeof(params));
+
+	memset(&cfg, 0, sizeof(cfg));
+	cfg.ctxlen = options->context_lines || 3;
+	cfg.interhunkctxlen = options->interhunk_lines || 3;
+	if (options->ignore_whitespace)
+		cfg.flags |= XDF_WHITESPACE_FLAGS;
+
+	memset(&callback, 0, sizeof(callback));
+	callback.outf = diff_output_cb;
+	callback.priv = options;
+
+	if (options->file_cb)
+		options->file_cb(
+			options->cb_data,
+			git_object_id((const git_object *)old_blob), NULL, 010644, 
+			git_object_id((const git_object *)new_blob), NULL, 010644);
+
+	xdl_diff(&old, &new, &params, &cfg, &callback);
+
+	return GIT_SUCCESS;
+}
+

src/xdiff/xinclude.h

@@ -26,10 +26,14 @@
 #include <ctype.h>
 #include <stdio.h>
 #include <stdlib.h>
-#include <unistd.h>
 #include <string.h>
 #include <limits.h>
 
+#ifdef WIN32
+#else
+#include <unistd.h>
+#endif
+
 #include "xmacros.h"
 #include "xdiff.h"
 #include "xtypes.h"

tests-clay/diff/blob.c

+#include "clay_libgit2.h"
+#include "fileops.h"
+#include "git2/diff.h"
+
+static git_repository *g_repo = NULL;
+
+void test_diff_blob__initialize(void)
+{
+	cl_fixture_sandbox("attr");
+	cl_git_pass(p_rename("attr/.gitted", "attr/.git"));
+	cl_git_pass(p_rename("attr/gitattributes", "attr/.gitattributes"));
+	cl_git_pass(git_repository_open(&g_repo, "attr/.git"));
+}
+
+void test_diff_blob__cleanup(void)
+{
+	git_repository_free(g_repo);
+	g_repo = NULL;
+	cl_fixture_cleanup("attr");
+}
+
+typedef struct {
+	int files;
+	int hunks;
+	int hunk_new_lines;
+	int hunk_old_lines;
+	int lines;
+	int line_ctxt;
+	int line_adds;
+	int line_dels;
+} diff_expects;
+
+static void log(const char *str, int n)
+{
+	FILE *fp = fopen("/Users/rb/tmp/diff.log", "a");
+	if (n > 0)
+		fprintf(fp, "%.*s", n, str);
+	else
+		fputs(str, fp);
+	fclose(fp);
+}
+
+static int diff_file_fn(
+	void *cb_data,
+	const git_oid *old,
+	const char *old_path,
+	int old_mode,
+	const git_oid *new,
+	const char *new_path,
+	int new_mode)
+{
+	diff_expects *e = cb_data;
+	e->files++;
+	log("-- file --\n", 0);
+	return 0;
+}
+
+static int diff_hunk_fn(
+	void *cb_data,
+	int old_start,
+	int old_lines,
+	int new_start,
+	int new_lines)
+{
+	diff_expects *e = cb_data;
+	e->hunks++;
+	e->hunk_old_lines += old_lines;
+	e->hunk_new_lines += new_lines;
+	log("-- hunk --\n", 0);
+	return 0;
+}
+
+static int diff_line_fn(
+	void *cb_data,
+	int origin,
+	const char *content,
+	size_t content_len)
+{
+	diff_expects *e = cb_data;
+	e->lines++;
+	switch (origin) {
+	case GIT_DIFF_LINE_CONTEXT:
+		log("[ ]", 3);
+		e->line_ctxt++;
+		break;
+	case GIT_DIFF_LINE_ADDITION:
+		log("[+]", 3);
+		e->line_adds++;
+		break;
+	case GIT_DIFF_LINE_DELETION:
+		log("[-]", 3);
+		e->line_dels++;
+		break;
+	default:
+		cl_assert("Unknown diff line origin" == 0);
+	}
+	log(content, content_len);
+	return 0;
+}
+
+void test_diff_blob__0(void)
+{
+	int err;
+	git_blob *a, *b, *c, *d;
+	git_oid a_oid, b_oid, c_oid, d_oid;
+	git_diff_opts opts;
+	diff_expects exp;
+
+	/* tests/resources/attr/root_test1 */
+	cl_git_pass(git_oid_fromstrn(&a_oid, "45141a79", 8));
+	cl_git_pass(git_blob_lookup_prefix(&a, g_repo, &a_oid, 4));
+
+	/* tests/resources/attr/root_test2 */
+	cl_git_pass(git_oid_fromstrn(&b_oid, "4d713dc4", 8));
+	cl_git_pass(git_blob_lookup_prefix(&b, g_repo, &b_oid, 4));
+
+	/* tests/resources/attr/root_test3 */
+	cl_git_pass(git_oid_fromstrn(&c_oid, "c96bbb2c2557a832", 16));
+	cl_git_pass(git_blob_lookup_prefix(&c, g_repo, &c_oid, 8));
+
+	/* tests/resources/attr/root_test4.txt */
+	cl_git_pass(git_oid_fromstrn(&d_oid, "fe773770c5a6", 12));
+	cl_git_pass(git_blob_lookup_prefix(&d, g_repo, &d_oid, 6));
+
+	/* Doing the equivalent of a `diff -U 2` on these files */
+
+	opts.context_lines = 2;
+	opts.interhunk_lines = 0;
+	opts.ignore_whitespace = 0;
+	opts.file_cb = diff_file_fn;
+	opts.hunk_cb = diff_hunk_fn;
+	opts.line_cb = diff_line_fn;
+	opts.cb_data = &exp;
+
+	memset(&exp, 0, sizeof(exp));
+	cl_git_pass(git_diff_blobs(g_repo, a, b, &opts));
+
+	cl_assert(exp.files == 1);
+	cl_assert(exp.hunks == 1);
+	cl_assert(exp.lines == 6);
+	cl_assert(exp.line_ctxt == 1);
+	cl_assert(exp.line_adds == 5);
+	cl_assert(exp.line_dels == 0);
+
+	memset(&exp, 0, sizeof(exp));
+	cl_git_pass(git_diff_blobs(g_repo, b, c, &opts));
+
+	cl_assert(exp.files == 1);
+	cl_assert(exp.hunks == 1);
+	cl_assert(exp.lines == 15);
+	cl_assert(exp.line_ctxt == 3);
+	cl_assert(exp.line_adds == 9);
+	cl_assert(exp.line_dels == 3);
+
+	memset(&exp, 0, sizeof(exp));
+	cl_git_pass(git_diff_blobs(g_repo, a, c, &opts));
+
+	cl_assert(exp.files == 1);
+	cl_assert(exp.hunks == 1);
+	cl_assert(exp.lines == 13);
+	cl_assert(exp.line_ctxt == 0);
+	cl_assert(exp.line_adds == 12);
+	cl_assert(exp.line_dels == 1);
+
+	opts.context_lines = 2;
+
+	memset(&exp, 0, sizeof(exp));
+	cl_git_pass(git_diff_blobs(g_repo, c, d, &opts));
+
+	cl_assert(exp.files == 1);
+	cl_assert(exp.hunks == 2);
+	cl_assert(exp.lines == 16);
+	cl_assert(exp.line_ctxt == 6);
+	cl_assert(exp.line_adds == 6);
+	cl_assert(exp.line_dels == 4);
+
+	git_blob_free(a);
+	git_blob_free(b);
+	git_blob_free(c);
+}
+

tests/resources/attr/.gitted/logs/HEAD

 0000000000000000000000000000000000000000 6bab5c79cd5140d0f800917f550eb2a3dc32b0da Russell Belfer <arrbee@arrbee.com> 1324416995 -0800	commit (initial): initial test data
 6bab5c79cd5140d0f800917f550eb2a3dc32b0da 605812ab7fe421fdd325a935d35cb06a9234a7d7 Russell Belfer <arrbee@arrbee.com> 1325143098 -0800	commit: latest test updates
 605812ab7fe421fdd325a935d35cb06a9234a7d7 a5d76cad53f66f1312bd995909a5bab3c0820770 Russell Belfer <arrbee@arrbee.com> 1325281762 -0800	commit: more macro tests
+a5d76cad53f66f1312bd995909a5bab3c0820770 370fe9ec224ce33e71f9e5ec2bd1142ce9937a6a Russell Belfer <arrbee@arrbee.com> 1327611749 -0800	commit: Updating files so we can do diffs
+370fe9ec224ce33e71f9e5ec2bd1142ce9937a6a f5b0af1fb4f5c0cd7aad880711d368a07333c307 Russell Belfer <arrbee@arrbee.com> 1327621027 -0800	commit: Updating test data

tests/resources/attr/.gitted/logs/refs/heads/master

 0000000000000000000000000000000000000000 6bab5c79cd5140d0f800917f550eb2a3dc32b0da Russell Belfer <arrbee@arrbee.com> 1324416995 -0800	commit (initial): initial test data
 6bab5c79cd5140d0f800917f550eb2a3dc32b0da 605812ab7fe421fdd325a935d35cb06a9234a7d7 Russell Belfer <arrbee@arrbee.com> 1325143098 -0800	commit: latest test updates
 605812ab7fe421fdd325a935d35cb06a9234a7d7 a5d76cad53f66f1312bd995909a5bab3c0820770 Russell Belfer <arrbee@arrbee.com> 1325281762 -0800	commit: more macro tests
+a5d76cad53f66f1312bd995909a5bab3c0820770 370fe9ec224ce33e71f9e5ec2bd1142ce9937a6a Russell Belfer <arrbee@arrbee.com> 1327611749 -0800	commit: Updating files so we can do diffs
+370fe9ec224ce33e71f9e5ec2bd1142ce9937a6a f5b0af1fb4f5c0cd7aad880711d368a07333c307 Russell Belfer <arrbee@arrbee.com> 1327621027 -0800	commit: Updating test data

tests/resources/attr/.gitted/refs/heads/master

-a5d76cad53f66f1312bd995909a5bab3c0820770
+f5b0af1fb4f5c0cd7aad880711d368a07333c307

tests/resources/attr/root_test2

 Hello from the root
+
+Some additional lines
+
+Down here below
+

tests/resources/attr/root_test3

-Hello from the root
+Some additional lines
+
+Down here below the other lines
+
+With even more at the end
+
+And lots of good stuff
+
+Anywhere you want
+
+Don't you think
+

tests/resources/attr/root_test4.txt

-Hello again
+Here is some stuff at the start
+
+This should go in one hunk
+
+Some additional lines
+
+Down here below the other lines
+
+With even more at the end
+
+Followed by a second hunk of stuff
+
+That happens down here
+