[PATCH] builtin-update-index.c (add_file_to_cache): Don't leak a cache entry.

[PATCH] builtin-update-index.c (add_file_to_cache): Don't leak a cache entry.

[Jim Meyering]

Signed-off-by: Jim Meyering <jim@meyering.net>
---
 builtin-update-index.c |    4 +++-
 1 files changed, 3 insertions(+), 1 deletions(-)

diff --git a/builtin-update-index.c b/builtin-update-index.c
index 1ac613a..5b2278e 100644
--- a/builtin-update-index.c
+++ b/builtin-update-index.c
@@ -121,8 +121,10 @@ static int add_file_to_cache(const char *path)
 			ce->ce_mode = create_ce_mode(S_IFREG | 0666);
 	}

-	if (index_path(ce->sha1, path, &st, !info_only))
+	if (index_path(ce->sha1, path, &st, !info_only)) {
+		free(ce);
 		return -1;
+	}
 	option = allow_add ? ADD_CACHE_OK_TO_ADD : 0;
 	option |= allow_replace ? ADD_CACHE_OK_TO_REPLACE : 0;
 	if (add_cache_entry(ce, option))
--
1.5.0.33.gaf997-dirty

Re: [PATCH] builtin-update-index.c (add_file_to_cache): Don't leak a cache entry.

[Linus Torvalds]

On Fri, 16 Feb 2007, Jim Meyering wrote:
> 
> -	if (index_path(ce->sha1, path, &st, !info_only))
> +	if (index_path(ce->sha1, path, &st, !info_only)) {
> +		free(ce);
>  		return -1;
> +	}

Well, the only user of this does:

	if (add_file_to_cache(p))
		die("Unable to process file %s", path);

so the leak is very shortlived ;)

		Linus

Re: [PATCH] builtin-update-index.c (add_file_to_cache): Don't leak a cache entry.

[Jim Meyering]

Linus Torvalds <torvalds@linux-foundation.org> wrote:
> On Fri, 16 Feb 2007, Jim Meyering wrote:
>>
>> -	if (index_path(ce->sha1, path, &st, !info_only))
>> +	if (index_path(ce->sha1, path, &st, !info_only)) {
>> +		free(ce);
>>  		return -1;
>> +	}
>
> Well, the only user of this does:
>
> 	if (add_file_to_cache(p))
> 		die("Unable to process file %s", path);
>
> so the leak is very shortlived ;)

True... for now, and probably for ever after, but what if?  However,
the goal in plugging a leak like this is not to avoid wasting memory,
but rather to avoid spending time investigating it, again and again.
It's analogous to making your code compile warning-free.  When a new
warning/leak pops up, it's easy to spot (in case it really does matter),
not obscured by lots of older ones.

There are other ways to mark a leak as ignorable, but this method works
across all leak-detecting tools.

If you'd prefer to avoid the cost of a technically-unnecessary free,
give it a different name, so it can be a no-op most of the time, and
"free" only when compiled in leak-checking mode.