AngusYoung, please don't touch bugs that are ASSIGNED w/o asking first. There was a reason bug 50926 wasn't closed yet. I spent half of last night at work patching flow-tools to build with as little warnings as possible and to work with gcc 3.4 (the patch is so big it'll have to go on the mirrors). Also, not sure if you tested with USE=mysql, but I tested this morning and that's also going to need some patching to build, which is why I hadn't closed that bug yet.
Ok, I have been already punished on -dev.
Changes: - patched to work with gcc 3.4 - patched configure.in to properly support --with-XXX/--without-XXX without this, running something like --without-mysql will still result in the check for libmysqlclient. - patched to fix the dozens (and dozens) of warnings ; I got almost all of them fixed with the exception of a few - only apply the nodebug patch if ! use debug (debug added to IUSE) - fixed deps - added missing flex/bison - added missing mysql (configure doesnt fail if mysql is missing so even with USE=mysql, it wasn't building the mysql-specific code) - added postgres to IUSE and it's respective dep
Created attachment 53833 [details] hacked ebuild to get .66 to build against gcc-3.4
I am in an odd situation where the new flow-tools .67 is causing my flow-capture to break and go into a loop as per this email: http://www.mail-archive.com/flow-tools%40list.splintered.net/msg00505.html The result is that until I can find a work around I am using flow-tools .66 BUT because of the upgrade I was running .67. Now when I go to downgrade it doesn't work because .66 won't compile with gcc 3.4. As a horrible, but working, hack around this I modified the .66 ebuild as per attached.
I've reverted 0.67's keywords to ~x86 and added the gcc patch to 0.66. Cheers