Yes, I think so. However, I rather ran into a complicated situation. In that particular place, in one mode, arg was used and in another arg wasn't. According to the present situation, we have two options.
1. The patch I proposed (which is logically correct) 2. Changing arg to ARGIN, and just making a dummy call on it, just to make sure it is used Regards, Senaka On Tue, Apr 1, 2008 at 6:28 AM, chromatic <[EMAIL PROTECTED]> wrote: > On Monday 31 March 2008 14:10:13 Senaka Fernando wrote: > > > There is a build failure when using the configure option --gc=libc. The > > attached patch solves this issue. > > Thanks, applied with a corresponding header tweak as r26678. > > Ultimately we're better off unifying the API so that the function > signatures > are completely the same between the two invocations, but it's okay to make > it > compile now and clean it up shortly. > > -- c >
