[cp-patches] RFC: GConf preference api backend

Mark Wielaard mark at klomp.org
Sun Jun 18 14:14:59 UTC 2006

On Sun, 2006-06-18 at 11:23 +0200, Michael Koch wrote:
> On Sun, Jun 18, 2006 at 11:15:26AM +0200, Mark Wielaard wrote:
> > On Sat, 2006-06-17 at 18:07 +0200, Michael Koch wrote:
> > > As Mario has no CVS access yet I commited this for him.
> > 
> > If you commit something for someone else you MUST make sure that it
> > compiles cleanly, even with --enable-Werror and --enable-regen-headers
> > The following fixes the build:
> Sigh. As Tom approved it and Mario told me its fine I just commited it.
> Another excercise for me not to just do this.

No, thanks for your help. But as one of the senior hackers on the
project you should know to double check anything you commit. But no
worries, the issues weren't big ones and easily fixed.

> Isn't --enable-Werror the default anyway?

It would indeed be nice to have --enable-Werror on by default, if you
don't have that yet please do enable it now. In this case it did catch
two typos (chache -> cache and java/ -> gnu/java).

The only reason it isn't on by default is because we don't want it on in
releases since on some older bsd, darwin and cygwin platforms we don't
get a clean compile and users wouldn't be able to test things without
fiddling with the configure options. Would be nice to figure out some
way to make it default enable for cvs builds, but disabled for snapshot
builds. Maybe we can use the VERSION +cvs thing for that? Volunteers
with good build/make/configure/auto-foo please do take a look at this.

Normally builder would have caught this. But the Mauve script seems
wedged. I am looking into this now. Looks like we need a much smaller
timeout, actually running things under xvfb because the new harness
But I have some problem understanding what it uses atm. It seems some
combination of jamvm, gcjx, ecj and gij are used with some using the
just compiled classpath, but other not using it which gives some strange
errors at times. Please watch classpath-testresults to see if I
untangled it all correctly.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://developer.classpath.org/pipermail/classpath-patches/attachments/20060618/e9a5b03d/attachment.pgp

More information about the Classpath-patches mailing list