Ingo Weinhold schrieb:
Nope I'm testing on gcc2 *and* gcc4 before I commit changes to the compat layer - since my first commit broke the gcc4 build ;) So I understand your explanation for the cause, but I don't understand how it could break in the first place, meaning how it could have slipped through my eyes. And I cannot sleep until I can reproduce the break here, too. Mainly so that I can prevent such breaks in the future.On 2009-12-22 at 17:30:46 [+0100], Colin Günther <coling@xxxxxx> wrote:coling@xxxxxx schrieb: Also, I'm trying to figure out how and where the build break reported by Ingo (//www.freelists.org/post/haiku-commits/r34716-haikutrunksrclibscompatf reebsd-networkcompatsys) is rooted.Well, as I've explained before, the problem is a clash between differing prototypes for vprintf() and driver_vprintf(), which occurs when <stdio.h> is included after the respective FreeBSD header. That you can't reproduce the issue is probably because you're building with gcc 2, which seems to pull in some standard headers (most notably <string.h> and <stdio.h>) implicitly.CU, Ingo
-ColinPS: For me this is really all about curiousity, getting better in understanding the build system and find a solution we both can live with :)