From c495cf02bac16e459f7c9e1740798646f12c00c5 Mon Sep 17 00:00:00 2001
From: Lars Hjemli <hjemli@gmail.com>
Date: Sat, 31 Jan 2009 10:40:40 +0100
Subject: 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>
---
 cgit.h     |  4 +++-
 shared.c   | 13 +++++++++++--
 ui-diff.c  | 26 ++++++++++++++++++++++++--
 ui-log.c   |  7 ++++++-
 ui-patch.c |  9 ++++++++-
 5 files changed, 52 insertions(+), 7 deletions(-)

diff --git a/cgit.h b/cgit.h
index 4fe94c6..8b0defa 100644
--- a/cgit.h
+++ b/cgit.h
@@ -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,
diff --git a/shared.c b/shared.c
index 578a544..dbb84d8 100644
--- a/shared.c
+++ b/shared.c
@@ -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));
diff --git a/ui-diff.c b/ui-diff.c
index 878496d..2196745 100644
--- a/ui-diff.c
+++ b/ui-diff.c
@@ -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)
diff --git a/ui-log.c b/ui-log.c
index 3202848..ba2ab03 100644
--- a/ui-log.c
+++ b/ui-log.c
@@ -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)
diff --git a/ui-patch.c b/ui-patch.c
index 1d77336..5d665d3 100644
--- a/ui-patch.c
+++ b/ui-patch.c
@@ -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)
-- 
cgit v1.2.3


From ae1d4d75b2a2eb3534ff4b3685cc5c0b80007ef7 Mon Sep 17 00:00:00 2001
From: Lars Hjemli <hjemli@gmail.com>
Date: Sat, 31 Jan 2009 17:45:48 +0100
Subject: ui-tree: show hexdump of binary blobs

This teaches ui-tree to detect binary blobs and display them similar
to `hexdump -C` (only wider).

Signed-off-by: Lars Hjemli <hjemli@gmail.com>
---
 cgit.css  | 19 +++++++++++++++++
 ui-tree.c | 71 ++++++++++++++++++++++++++++++++++++++++++---------------------
 2 files changed, 67 insertions(+), 23 deletions(-)

diff --git a/cgit.css b/cgit.css
index 957d5aa..adfc8ae 100644
--- a/cgit.css
+++ b/cgit.css
@@ -263,6 +263,25 @@ table.blob a.no a:hover {
 	color: black;
 }
 
+table.bin-blob {
+	margin-top: 0.5em;
+	border: solid 1px black;
+}
+
+table.bin-blob th {
+	font-family: monospace;
+	white-space: pre;
+	border: solid 1px #777;
+	padding: 0.5em 1em;
+}
+
+table.bin-blob td {
+	font-family: monospace;
+	white-space: pre;
+	border-left: solid 1px #777;
+	padding: 0em 1em;
+}
+
 table.nowrap td {
 	white-space: nowrap;
 }
diff --git a/ui-tree.c b/ui-tree.c
index c26ba4c..a37a4e5 100644
--- a/ui-tree.c
+++ b/ui-tree.c
@@ -6,6 +6,7 @@
  *   (see COPYING for full license text)
  */
 
+#include <ctype.h>
 #include "cgit.h"
 #include "html.h"
 #include "ui-shared.h"
@@ -14,12 +15,53 @@ char *curr_rev;
 char *match_path;
 int header = 0;
 
+static void print_text_buffer(char *buf, unsigned long size)
+{
+	unsigned long lineno, idx;
+	const char *numberfmt =
+		"<a class='no' id='n%1$d' name='n%1$d' href='#n%1$d'>%1$d</a>\n";
+
+	html("<table summary='blob content' class='blob'>\n");
+	html("<tr><td class='linenumbers'><pre>");
+	idx = 0;
+	lineno = 0;
+	htmlf(numberfmt, ++lineno);
+	while(idx < size - 1) { // skip absolute last newline
+		if (buf[idx] == '\n')
+			htmlf(numberfmt, ++lineno);
+		idx++;
+	}
+	html("</pre></td>\n");
+	html("<td class='lines'><pre><code>");
+	html_txt(buf);
+	html("</code></pre></td></tr></table>\n");
+}
+
+static void print_binary_buffer(char *buf, unsigned long size)
+{
+	unsigned long ofs, idx;
+
+	html("<table summary='blob content' class='bin-blob'>\n");
+	html("<tr><th>ofs</th><th>hex dump</th><th>ascii</th></tr>");
+	for (ofs = 0; ofs < size; ofs += 32, buf += 32) {
+		htmlf("<tr><td class='right'>%04x</td><td class='hex'>", ofs);
+		for (idx = 0; idx < 32 && ofs + idx < size; idx++)
+			htmlf("%*s%02x",
+			      idx == 16 ? 4 : 1, "",
+			      buf[idx] & 0xff);
+		html(" </td><td class='hex'>");
+		for (idx = 0; idx < 32 && ofs + idx < size; idx++)
+			htmlf("%c", isgraph(buf[idx]) ? buf[idx] : '.');
+		html("</td></tr>\n");
+	}
+	html("</table>\n");
+}
+
 static void print_object(const unsigned char *sha1, char *path)
 {
 	enum object_type type;
 	char *buf;
-	unsigned long size, lineno, idx;
-	const char *numberfmt = "<a class='no' id='n%1$d' name='n%1$d' href='#n%1$d'>%1$d</a>\n";
+	unsigned long size;
 
 	type = sha1_object_info(sha1, &size);
 	if (type == OBJ_BAD) {
@@ -40,27 +82,10 @@ static void print_object(const unsigned char *sha1, char *path)
 		        curr_rev, path);
 	htmlf(")<br/>blob: %s\n", sha1_to_hex(sha1));
 
-	html("<table summary='blob content' class='blob'>\n");
-	html("<tr>\n");
-
-	html("<td class='linenumbers'><pre>");
-	idx = 0;
-	lineno = 0;
-	htmlf(numberfmt, ++lineno);
-	while(idx < size - 1) { // skip absolute last newline
-		if (buf[idx] == '\n') {
-			htmlf(numberfmt, ++lineno);
-		}
-		idx++;
-	}
-	html("</pre></td>\n");
-
-	html("<td class='lines'><pre><code>");
-	html_txt(buf);
-	html("</code></pre></td>\n");
-
-	html("</tr>\n");
-	html("</table>\n");
+	if (buffer_is_binary(buf, size))
+		print_binary_buffer(buf, size);
+	else
+		print_text_buffer(buf, size);
 }
 
 
-- 
cgit v1.2.3


From 481ce5e298e2dcd7edc1d4a30e523dda2ce58b01 Mon Sep 17 00:00:00 2001
From: Lars Hjemli <hjemli@gmail.com>
Date: Sun, 1 Feb 2009 19:29:24 +0100
Subject: shared.c: avoid SEGFAULT when checking for binary buffers

Before calling buffer_is_binary() we need to verify that the buffer
is valid.

Signed-off-by: Lars Hjemli <hjemli@gmail.com>
---
 shared.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/shared.c b/shared.c
index dbb84d8..cce0af4 100644
--- a/shared.c
+++ b/shared.c
@@ -271,8 +271,8 @@ int cgit_diff_files(const unsigned char *old_sha1,
 	*old_size = file1.size;
 	*new_size = file2.size;
 
-	if (buffer_is_binary(file1.ptr, file1.size) ||
-	    buffer_is_binary(file2.ptr, file2.size)) {
+	if ((file1.ptr && buffer_is_binary(file1.ptr, file1.size)) ||
+	    (file2.ptr && buffer_is_binary(file2.ptr, file2.size))) {
 		*binary = 1;
 		return 0;
 	}
-- 
cgit v1.2.3