Re: [PATCH 4/5] Make 'diff_populate_filespec()' use the new 'strbuf_readlink()'

From: Linus Torvalds <torvalds@linux-foundation.org>
Date: Dec 18, 2008, 16:55 UTC
Message-ID: <alpine.LFD.2.00.0812180851120.14014@localhost.localdomain>
In-Reply-To: <20081218121118.3635c53c@crow>

On Thu, 18 Dec 2008, Mark Burton wrote:

> 
> Does it need a cast on some architectures?

Gaah. My bad. It should work fine ("unsigned long" is physically the same type as "size_t" in your case), but on 32-bit x86, size_t is generally "unsigned int" - which is the same physical type there (both int and long are 32-bit) but causes a valid warning.

I think we should just make the "size" member "size_t". I use "unsigned long" out of much too long habit, since we traditionally avoided "size_t" in the kernel due to it just being another unnecessary architecture- specific detail.

So the proper patch is probably just the following. Sorry about that,

		Linus
---
 diffcore.h |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/diffcore.h b/diffcore.h
index 5b63458..16a73e6 100644
--- a/diffcore.h
+++ b/diffcore.h
@@ -30,7 +30,7 @@ struct diff_filespec {
 	void *data;
 	void *cnt_data;
 	const char *funcname_pattern_ident;
-	unsigned long size;
+	size_t size;
 	int count;               /* Reference count */
 	int xfrm_flags;		 /* for use by the xfrm */
 	int rename_used;         /* Count of rename users */