Matthias Klose <[EMAIL PROTECTED]> wrote: > This package build-depends for some reason on g++-3.4 (most likely, > because it could not be built with a newer g++ version. We would > like to get rid of g++-3.4 for the etch release, although currently > not a hard release goal. > > Please (re)check, if the package can be built by g++, and if not, if it > can be built using gcc-snapshot (or g++-4.1, currently found in > experimental). Please keep the bug report open until you can avoid the > build dependency on g++-3.4. Collect further information in this bug > report (i.e. if it builds with a newer/experimental version).
icu built on all platforms except m68k with g++ 4.0, but it had various test suite failures on all including i386. The nature of the failures varied from platform to platform, if I recall. There have been many upstream changes, some of which are apparently to address gcc 4.0 issues, but the changes are too massive and spread apart to try to isolate them. It is likely that the next upstream version of icu will work with g++ 4.0. When I get time, most likely over the holidays, I'll try building icu and running its test suite on i386. If I have suitable support on developer-accessible build engines, I may also give it a shot on one 64-bit system and one big-endian system (powerpc), but I won't promise that. If there is a version of g++ 4.0 in sid that can build an icu 3.4 on i386 that passes its test suite, I can try an upload to sid and see if it works on other platforms, but if it does not, my inclination would be to revert the change and go back to the 3.4 version. Does this sound reasonable? If not, is there support available for someone to help with the process of trying to sort out the problems? I'll update this bug with my progress when I have time to work on this, which won't be until at least 12/23. Thanks for the report and your continuing efforts on the g++-4.0 issues! (Since I know you wear the Ubuntu hat as well, I notice that Ubuntu's icu presently appears to be synchronized with debian's. I'd be curious to know whether there will effort within the context of dapper on this.) --Jay -- Jay Berkenbilt <[EMAIL PROTECTED]> -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
