git/list[1] front-page[2] threads[3] people[4] search[5] about
 

[PATCH 4/4] diff: add filter for converting binary to text

From
Jeff King <peff@peff.net>
Date
Oct 5, 2008, 21:43 UTC
Message-ID
<20081005214345.GD21925@coredump.intra.peff.net>
In-Reply-To
<20081005214114.GA21875@coredump.intra.peff.net>

When diffing binary files, it is sometimes nice to see the differences of a canonical text form rather than either a binary patch or simply "binary files differ."

Until now, the only option for doing this was to define an external diff command to perform the diff. This was a lot of work, since the external command needed to take care of doing the diff itself (including mode changes), and lost the benefit of git's colorization and other options.

This patch adds a text conversion option, which converts a file to its canonical format before performing the diff. This is less flexible than an arbitrary external diff, but is much less work to set up. For example:

  $ echo '*.jpg diff=exif' >>.gitattributes
  $ git config diff.exif.textconv exiftool
  $ git config diff.exif.binary false

allows one to see jpg diffs represented by the text output of exiftool.

Signed-off-by: Jeff King <peff@peff.net>
---
 diff.c     |   49 +++++++++++++++++++++++++++++++++++++++++++++++--
 userdiff.c |    2 ++
 userdiff.h |    1 +
 3 files changed, 50 insertions(+), 2 deletions(-)
diff --git a/diff.c b/diff.c
index ba4f6fa..421da29 100644
--- a/diff.c
+++ b/diff.c
@@ -38,6 +38,9 @@ static char diff_colors[][COLOR_MAXLEN] = {
 	"\033[41m",	/* WHITESPACE (red background) */
 };
 
+static void diff_filespec_load_driver(struct diff_filespec *one);
+static char *run_textconv(const char *, struct diff_filespec *, size_t *);
+
 static int parse_diff_color_slot(const char *var, int ofs)
 {
 	if (!strcasecmp(var+ofs, "plain"))
@@ -291,8 +294,19 @@ static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
 	}
 	else if (diff_populate_filespec(one, 0))
 		return -1;
-	mf->ptr = one->data;
-	mf->size = one->size;
+
+	diff_filespec_load_driver(one);
+	if (one->driver->textconv) {
+		size_t size;
+		mf->ptr = run_textconv(one->driver->textconv, one, &size);
+		if (!mf->ptr)
+			return -1;
+		mf->size = size;
+	}
+	else {
+		mf->ptr = one->data;
+		mf->size = one->size;
+	}
 	return 0;
 }
 
@@ -3374,3 +3388,34 @@ void diff_unmerge(struct diff_options *options,
 	fill_filespec(one, sha1, mode);
 	diff_queue(&diff_queued_diff, one, two)->is_unmerged = 1;
 }
+
+static char *run_textconv(const char *pgm, struct diff_filespec *spec,
+		size_t *outsize)
+{
+	struct diff_tempfile temp;
+	const char *argv[3];
+	const char **arg = argv;
+	struct child_process child;
+	struct strbuf buf = STRBUF_INIT;
+
+	prepare_temp_file(spec->path, &temp, spec);
+	*arg++ = pgm;
+	*arg++ = temp.name;
+	*arg = NULL;
+
+	memset(&child, 0, sizeof(child));
+	child.argv = argv;
+	child.out = -1;
+	if (start_command(&child) != 0 ||
+	    strbuf_read(&buf, child.out, 0) < 0 ||
+	    finish_command(&child) != 0) {
+		if (temp.name == temp.tmp_path)
+			unlink(temp.name);
+		error("error running textconv command '%s'", pgm);
+		return NULL;
+	}
+	if (temp.name == temp.tmp_path)
+		unlink(temp.name);
+
+	return strbuf_detach(&buf, outsize);
+}
diff --git a/userdiff.c b/userdiff.c
index ac6d4a1..02f225f 100644
--- a/userdiff.c
+++ b/userdiff.c
@@ -131,6 +131,8 @@ int userdiff_config_porcelain(const char *k, const char *v)
 
 	if ((drv = parse_driver(k, v, "command")))
 		return parse_string(&drv->external, k, v);
+	if ((drv = parse_driver(k, v, "textconv")))
+		return parse_string(&drv->textconv, k, v);
 
 	return 0;
 }
diff --git a/userdiff.h b/userdiff.h
index 1c1eb04..f29c18f 100644
--- a/userdiff.h
+++ b/userdiff.h
@@ -11,6 +11,7 @@ struct userdiff_driver {
 	const char *external;
 	int binary;
 	struct userdiff_funcname funcname;
+	const char *textconv;
 };
 
 int userdiff_config_basic(const char *k, const char *v);
-- 
1.6.0.2.639.g4d7f.dirty
Previous: Johannes SixtNext: Jakub Narebski
Message 73 of 85 in “Implementation of a "textconv" filter for easy custom diff.”
  1. Matthieu MoySep 28, 2008
  2. Facility to have multiple kinds of drivers for diff.Matthieu Moy, Sep 28, 2008
  3. Implement run_command_to_buf (spawn a process and reads its stdout)Matthieu Moy, Sep 28, 2008
  4. Implement a textconv filter for "git diff"Matthieu Moy, Sep 28, 2008
  5. Document the textconv filter.Matthieu Moy, Sep 28, 2008
  6. Add a basic test for the textconv filter.Matthieu Moy, Sep 28, 2008
  7. Johannes SixtSep 28, 2008
  8. Matthieu MoySep 28, 2008
  9. Jeff KingSep 28, 2008
  10. Matthieu MoySep 28, 2008
  11. Jeff KingSep 28, 2008
  12. Jeff KingSep 28, 2008
  13. Matthieu MoySep 28, 2008
  14. Jeff KingSep 28, 2008
  15. Matthieu MoySep 30, 2008
  16. Jeff KingSep 30, 2008
  17. 0/4 diff text conversion filterJeff King, Oct 5, 2008
  18. 1/4 t4012: use test_cmp instead of cmpJeff King, Oct 5, 2008
  19. 2/4 diff: unify external diff and funcname parsing codeJeff King, Oct 5, 2008
  20. 3/4 diff: introduce diff.<driver>.binaryJeff King, Oct 5, 2008
  21. Johannes SixtOct 7, 2008
  22. Jeff KingOct 7, 2008
  23. Johannes SixtOct 7, 2008
  24. Junio C HamanoOct 12, 2008
  25. Jeff KingOct 13, 2008
  26. Junio C HamanoOct 13, 2008
  27. Jeff KingOct 13, 2008
  28. Johannes SixtOct 13, 2008
  29. Junio C HamanoOct 13, 2008
  30. Matthieu MoyOct 13, 2008
  31. Jeff KingOct 24, 2008
  32. 1/5 diff: add missing static declarationJeff King, Oct 24, 2008
  33. 2/5 add userdiff textconv testsJeff King, Oct 24, 2008
  34. 3/5 refactor userdiff textconv codeJeff King, Oct 24, 2008
  35. Johannes SixtOct 24, 2008
  36. Jeff KingOct 24, 2008
  37. Jeff KingOct 24, 2008
  38. Johannes SixtOct 24, 2008
  39. Jeff KingOct 24, 2008
  40. Junio C HamanoOct 24, 2008
  41. Jeff KingOct 24, 2008
  42. Jeff KingOct 24, 2008
  43. Jeff KingOct 25, 2008
  44. 1/7 diff: add missing static declarationJeff King, Oct 25, 2008
  45. 2/7 add userdiff textconv testsJeff King, Oct 25, 2008
  46. 3/7 textconv: assume text-converted contents are not binaryJeff King, Oct 25, 2008
  47. 4/7 textconv: don't convert for every operationJeff King, Oct 25, 2008
  48. Junio C HamanoOct 25, 2008
  49. Jeff KingOct 25, 2008
  50. Junio C HamanoOct 25, 2008
  51. Jeff KingOct 25, 2008
  52. Junio C HamanoOct 25, 2008
  53. Junio C HamanoOct 25, 2008
  54. Jeff KingOct 26, 2008
  55. Jeff KingOct 26, 2008
  56. 1/8 diff: add missing static declarationJeff King, Oct 26, 2008
  57. 2/8 document the diff driver textconv featureJeff King, Oct 26, 2008
  58. 3/8 add userdiff textconv testsJeff King, Oct 26, 2008
  59. 4/8 refactor userdiff textconv codeJeff King, Oct 26, 2008
  60. 5/8 userdiff: require explicitly allowing textconvJeff King, Oct 26, 2008
  61. 6/8 only textconv regular filesJeff King, Oct 26, 2008
  62. 7/8 wt-status: load diff ui configJeff King, Oct 26, 2008
  63. Junio C HamanoOct 27, 2008
  64. Jeff KingOct 27, 2008
  65. 8/8 enable textconv for diff in verbose status/commitJeff King, Oct 26, 2008
  66. 5/7 userdiff: require explicitly allowing textconvJeff King, Oct 25, 2008
  67. 6/7 document the diff driver textconv featureJeff King, Oct 25, 2008
  68. 7/7 only textconv regular filesJeff King, Oct 25, 2008
  69. 4/5 userdiff: require explicitly allowing textconvJeff King, Oct 24, 2008
  70. Johannes SixtOct 24, 2008
  71. 5/5 document the diff driver textconv featureJeff King, Oct 24, 2008
  72. Johannes SixtOct 24, 2008
  73. 4/4 diff: add filter for converting binary to textJeff King, Oct 5, 2008
  74. Jakub NarebskiOct 5, 2008
  75. Johannes SixtOct 6, 2008
  76. Jeff KingOct 6, 2008
  77. Johannes SixtOct 6, 2008
  78. Matthieu MoyOct 6, 2008
  79. Jeff KingOct 7, 2008
  80. Johannes SixtOct 7, 2008
  81. Jeff KingOct 7, 2008
  82. Matthieu MoyOct 7, 2008
  83. Jeff KingOct 7, 2008
  84. Johannes SixtOct 7, 2008
  85. Jeff KingOct 13, 2008

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.