[PATCH] No color diff when redirecting to file

[PATCH] No color diff when redirecting to file

[Sascha Peilicke]

[-- Attachment #1: Type: Text/Plain, Size: 961 bytes --]

Previously, when having color diffs enabled and redirecting 'git diff'
into a file, one ends up with a messed up file containing termcap color
stuff. This change disables color when diff output is redirected into a
file.

Signed-off-by: Sascha Peilicke <saschpe@gmx.de>
---
 diff.c |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)

diff --git a/diff.c b/diff.c
index 5422c43..cd1ecb0 100644
--- a/diff.c
+++ b/diff.c
@@ -2854,6 +2854,11 @@ void diff_setup(struct diff_options *options)
 		DIFF_OPT_SET(options, COLOR_DIFF);
 	options->detect_rename = diff_detect_rename_default;
 
+	struct stat buf;
+	if (fstat(fileno(options->file), &buf) != -1 && S_ISREG(buf.st_mode)) {
+		DIFF_OPT_CLR(options, COLOR_DIFF);
+	}
+
 	if (diff_no_prefix) {
 		options->a_prefix = options->b_prefix = "";
 	} else if (!diff_mnemonic_prefix) {
-- 
1.7.3.4

-- 
Mit freundlichen Grüßen,
Sascha Peilicke
http://saschpe.wordpress.com

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [PATCH] No color diff when redirecting to file

[Johannes Sixt]

The subject line reads as if you want that color markup appears in a file.

On Sonntag, 23. Januar 2011, Sascha Peilicke wrote:
> Previously, when having color diffs enabled and redirecting 'git diff'
> into a file, one ends up with a messed up file containing termcap color
> stuff. This change disables color when diff output is redirected into a
> file.

But this description and the patch read as if you don't want it.

That said, I cannot reproduce. Perhaps your configuration says 
color.ui=always? If so, then this is expected behavior and not a bug, IMO.

-- Hannes

Re: [PATCH] No color diff when redirecting to file

[Matthieu Moy]

Johannes Sixt <j6t@kdbg.org> writes:

> That said, I cannot reproduce. Perhaps your configuration says 
> color.ui=always? If so, then this is expected behavior and not a bug, IMO.

I think you're right.

I tested this (without the patch):

git diff
=> I get color, because I have color.ui = auto

git diff | cat
=> I don't get color

git diff --color=auto | cat
=> no color

git diff --color=always | cat
=> colors, because I've asked

this seems to be just the right behavior.

-- 
Matthieu Moy
http://www-verimag.imag.fr/~moy/

Re: [PATCH] No color diff when redirecting to file

[Junio C Hamano]

Matthieu Moy <Matthieu.Moy@grenoble-inp.fr> writes:

> Johannes Sixt <j6t@kdbg.org> writes:
>
>> That said, I cannot reproduce. Perhaps your configuration says 
>> color.ui=always? If so, then this is expected behavior and not a bug, IMO.
>
> I think you're right.
>
> I tested this (without the patch):
>
> git diff
> => I get color, because I have color.ui = auto
>
> git diff | cat
> => I don't get color
>
> git diff --color=auto | cat
> => no color
>
> git diff --color=always | cat
> => colors, because I've asked
>
> this seems to be just the right behavior.

Overall I agree with the above observation, but the original poster says
"redirecting into a _file_".

Another possibility is that the mysterious platform the original poster
did not mention has a broken istty() that mistakes a file descriptor going
to a file (instead of a pipe to another process) as connected to a tty.

Re: [PATCH] No color diff when redirecting to file

[Sascha Peilicke]

[-- Attachment #1: Type: Text/Plain, Size: 752 bytes --]

On Sunday 23 January 2011 18:29:52 you wrote:
> Johannes Sixt <j6t@kdbg.org> writes:
> > That said, I cannot reproduce. Perhaps your configuration says
> > color.ui=always? If so, then this is expected behavior and not a bug,
> > IMO.
> 
> I think you're right.
> 
> I tested this (without the patch):
> 
> git diff
> => I get color, because I have color.ui = auto
> 
> git diff | cat
> => I don't get color
> 
> git diff --color=auto | cat
> => no color
> 
> git diff --color=always | cat
> => colors, because I've asked
> 
> this seems to be just the right behavior.

try the same with

git diff > foo

works when

git diff --no-color > foo
-- 
Mit freundlichen Grüßen,
Sascha Peilicke
http://saschpe.wordpress.com

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [PATCH] No color diff when redirecting to file

[Matthieu Moy]

Sascha Peilicke <saschpe@gmx.de> writes:

> try the same with
>
> git diff > foo

git diff > foo && cat foo
=> no color.

git diff --no-color > foo && cat foo
=> no color either

git diff --color=always > foo && cat foo
=> colors.

Again, just the expected behavior.

If it helps:

$ uname -a
Linux bauges 2.6.36-686 #1 SMP PREEMPT Fri Nov 5 10:35:38 CET 2010 i686 GNU/Linux
$ git --version
git version 1.7.4.rc1.17.ge7ece

-- 
Matthieu Moy
http://www-verimag.imag.fr/~moy/

Re: [PATCH] No color diff when redirecting to file

[Sascha Peilicke]

[-- Attachment #1: Type: text/plain, Size: 749 bytes --]

On Sunday 23 January 2011 15:47:17 you wrote:
> The subject line reads as if you want that color markup appears in a file.
> 
> On Sonntag, 23. Januar 2011, Sascha Peilicke wrote:
> > Previously, when having color diffs enabled and redirecting 'git diff'
> > into a file, one ends up with a messed up file containing termcap color
> > stuff. This change disables color when diff output is redirected into a
> > file.
> 
> But this description and the patch read as if you don't want it.
> 
> That said, I cannot reproduce. Perhaps your configuration says
> color.ui=always? If so, then this is expected behavior and not a bug, IMO.
Nope, it's color.ui=auto
-- 
Mit freundlichen Grüßen,
Sascha Peilicke
http://saschpe.wordpress.com

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 198 bytes --]