[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: svn commit: rev 182 - trunk/subversion/libsvn_fs

From: Branko Čibej <brane_at_xbc.nu>
Date: 2001-10-01 19:10:39 CEST

Greg Stein wrote:

>On Fri, Sep 28, 2001 at 07:33:05PM -0500, brane@tigris.org wrote:
>
>>Author: brane
>>Date: 2001-09-29 00:33 GMT
>>New Revision: 182
>>
>>Modified:
>> trunk/subversion/libsvn_fs/key-gen.c
>> trunk/subversion/libsvn_fs/skel.c
>> trunk/subversion/libsvn_fs/txn-table.c
>> trunk/subversion/libsvn_fs/validate.c
>>Log:
>>Fix signed/unsigned comparison warnings.
>>...
>>--- OLD/trunk/subversion/libsvn_fs/key-gen.c Fri Sep 28 19:33:05 2001
>>+++ NEW/trunk/subversion/libsvn_fs/key-gen.c Fri Sep 28 19:33:05 2001
>>@@ -41,14 +41,14 @@
>>
>> So we do the check for overflow before we multiply value and add
>> in the new digit. */
>>- int max_prefix = max / 10;
>>- int max_digit = max % 10;
>>- int i;
>>+ apr_size_t max_prefix = max / 10;
>>+ apr_size_t max_digit = max % 10;
>>+ apr_size_t i;
>>
>
>This is bogus. Digits are integers. The problem is with the "max" parameter,
>not the internal values. A maximum integer shouldn't be an apr_size_t. Some
>other type ought to be used.
>
First, these aren't maximum or minimum numbers, they're digit counters.
Second, the other ways to avoid the warnings are a) use casts (blech),
or b) change the declaration of the skel's length field, which would
open a whole new can of worms.

I would agree with you in principle; unfortunately, this is C, a totally
unprincipled language, where the results of comparisons between signed
and unsigned quantities are undefined. I do /not/ want to add a horde of
casts to the code (the one I did was two too many ...), nor do I want to
see these warnings.

>apr_size_t is for memory sizes.
>
Except for "i", these are memory sizes.

-- 
Brane �ibej   <brane_at_xbc.nu>            http://www.xbc.nu/brane/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 21 14:36:43 2006

This is an archived mail posted to the Subversion Dev mailing list.