Handle binary files in diffs

This teaches all diff-related operations (i.e. ui-log, ui-diff and ui-patch)
how to handle binary files.

Signed-off-by: Lars Hjemli <hjemli@gmail.com>

5 files changed, 52 insertions, 7 deletions

@@ -15,6 +15,7 @@
 #include <revision.h>
 #include <log-tree.h>
 #include <archive.h>
+#include <xdiff-interface.h>
 #include <xdiff/xdiff.h>
 #include <utf8.h>
 
@@ -221,7 +222,8 @@ extern void *cgit_free_commitinfo(struct commitinfo *info);
 
 extern int cgit_diff_files(const unsigned char *old_sha1,
                            const unsigned char *new_sha1,
-                           linediff_fn fn);
+                           unsigned long *old_size, unsigned long *new_size,
+                           int *binary, linediff_fn fn);
 
 extern void cgit_diff_tree(const unsigned char *old_sha1,
                            const unsigned char *new_sha1,
@@ -257,8 +257,8 @@ int filediff_cb(void *priv, mmbuffer_t *mb, int nbuf)
 }
 
 int cgit_diff_files(const unsigned char *old_sha1,
-                     const unsigned char *new_sha1,
-                     linediff_fn fn)
+                    const unsigned char *new_sha1, unsigned long *old_size,
+                    unsigned long *new_size, int *binary, linediff_fn fn)
 {
         mmfile_t file1, file2;
         xpparam_t diff_params;
@@ -268,6 +268,15 @@ int cgit_diff_files(const unsigned char *old_sha1,
         if (!load_mmfile(&file1, old_sha1) || !load_mmfile(&file2, new_sha1))
                 return 1;
 
+        *old_size = file1.size;
+        *new_size = file2.size;
+
+        if (buffer_is_binary(file1.ptr, file1.size) ||
+            buffer_is_binary(file2.ptr, file2.size)) {
+                *binary = 1;
+                return 0;
+        }
+
         memset(&diff_params, 0, sizeof(diff_params));
         memset(&emit_params, 0, sizeof(emit_params));
         memset(&emit_cb, 0, sizeof(emit_cb));
@@ -27,6 +27,9 @@ static struct fileinfo {
         char *new_path;
         unsigned int added;
         unsigned int removed;
+        unsigned long old_size;
+        unsigned long new_size;
+        int binary:1;
 } *items;
 
 
@@ -86,6 +89,11 @@ static void print_fileinfo(struct fileinfo *info)
                       info->status == DIFF_STATUS_COPIED ? "copied" : "renamed",
                       info->old_path);
         html("</td><td class='right'>");
+        if (info->binary) {
+                htmlf("bin</td><td class='graph'>%d -> %d bytes",
+                      info->old_size, info->new_size);
+                return;
+        }
         htmlf("%d", info->added + info->removed);
         html("</td><td class='graph'>");
         htmlf("<table summary='file diffstat' width='%d%%'><tr>", (max_changes > 100 ? 100 : max_changes));
@@ -110,10 +118,14 @@ static void count_diff_lines(char *line, int len)
 
 static void inspect_filepair(struct diff_filepair *pair)
 {
+        int binary = 0;
+        unsigned long old_size = 0;
+        unsigned long new_size = 0;
         files++;
         lines_added = 0;
         lines_removed = 0;
-        cgit_diff_files(pair->one->sha1, pair->two->sha1, count_diff_lines);
+        cgit_diff_files(pair->one->sha1, pair->two->sha1, &old_size, &new_size,
+                        &binary, count_diff_lines);
         if (files >= slots) {
                 if (slots == 0)
                         slots = 4;
@@ -130,6 +142,9 @@ static void inspect_filepair(struct diff_filepair *pair)
         items[files-1].new_path = xstrdup(pair->two->path);
         items[files-1].added = lines_added;
         items[files-1].removed = lines_removed;
+        items[files-1].old_size = old_size;
+        items[files-1].new_size = new_size;
+        items[files-1].binary = binary;
         if (lines_added + lines_removed > max_changes)
                 max_changes = lines_added + lines_removed;
         total_adds += lines_added;
@@ -233,6 +248,10 @@ static void header(unsigned char *sha1, char *path1, int mode1,
 
 static void filepair_cb(struct diff_filepair *pair)
 {
+        unsigned long old_size = 0;
+        unsigned long new_size = 0;
+        int binary = 0;
+
         header(pair->one->sha1, pair->one->path, pair->one->mode,
                pair->two->sha1, pair->two->path, pair->two->mode);
         if (S_ISGITLINK(pair->one->mode) || S_ISGITLINK(pair->two->mode)) {
@@ -242,8 +261,11 @@ static void filepair_cb(struct diff_filepair *pair)
                         print_line(fmt("+Subproject %s", sha1_to_hex(pair->two->sha1)), 52);
                 return;
         }
-        if (cgit_diff_files(pair->one->sha1, pair->two->sha1, print_line))
+        if (cgit_diff_files(pair->one->sha1, pair->two->sha1, &old_size, 
+                            &new_size, &binary, print_line))
                 cgit_print_error("Error running diff");
+        if (binary)
+                html("Binary files differ");
 }
 
 void cgit_print_diff(const char *new_rev, const char *old_rev, const char *prefix)
@@ -26,9 +26,14 @@ void count_lines(char *line, int size)
 
 void inspect_files(struct diff_filepair *pair)
 {
+        unsigned long old_size = 0;
+        unsigned long new_size = 0;
+        int binary = 0;
+
         files++;
         if (ctx.repo->enable_log_linecount)
-                cgit_diff_files(pair->one->sha1, pair->two->sha1, count_lines);
+                cgit_diff_files(pair->one->sha1, pair->two->sha1, &old_size,
+                                &new_size, &binary, count_lines);
 }
 
 void show_commit_decorations(struct commit *commit)
@@ -57,6 +57,10 @@ static void header(unsigned char *sha1, char *path1, int mode1,
 
 static void filepair_cb(struct diff_filepair *pair)
 {
+        unsigned long old_size = 0;
+        unsigned long new_size = 0;
+        int binary = 0;
+
         header(pair->one->sha1, pair->one->path, pair->one->mode,
                pair->two->sha1, pair->two->path, pair->two->mode);
         if (S_ISGITLINK(pair->one->mode) || S_ISGITLINK(pair->two->mode)) {
@@ -66,8 +70,11 @@ static void filepair_cb(struct diff_filepair *pair)
                         print_line(fmt("+Subproject %s", sha1_to_hex(pair->two->sha1)), 52);
                 return;
         }
-        if (cgit_diff_files(pair->one->sha1, pair->two->sha1, print_line))
+        if (cgit_diff_files(pair->one->sha1, pair->two->sha1, &old_size,
+                            &new_size, &binary, print_line))
                 html("Error running diff");
+        if (binary)
+                html("Binary files differ\n");
 }
 
 void cgit_print_patch(char *hex)