On Mon, 29 Jan 2007, Stefan Küng wrote:
> Daniel Rall wrote:
> >This patch seems conceptually fine. However, I think the crash
> >handler should be compiled in by default (as it is now); this patch
> >changes that behavior.
> Sure. But you can just patch your build scripts and always set that define.
> My problem right now is that the crash handler part doesn't compile on
> x64, so my nightly build fails. Even if I exclude the crash* files from
> the build, the little part in cmdline.c which (right now) is not
> excluded makes the build fail when linking due to the missing functions
> (which I excluded from the build).
With Lieven's changes from today, the problematic code should be
excluded both from being built and from being attached if you undefine
Is this sufficient, Stefan? Or would it be easier if we were treating
SVN_USE_WIN32_CRASHHANDLER like a boolean?
Received on Tue Jan 30 21:42:00 2007
- application/pgp-signature attachment: stored