[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Status of Sarge Release Issues (Updated for May)

>1 i386 support? (The large thread starting with
>talks about this). Do we have to wait for upstream to fix this? If so that
>might mean a gcc3.4 transition and well, at least half a year before
>freeze (gcc 3.3 is currently in stage three, gcc3.4 will take at least 6
>months to complete the release schedule that
>http://gcc.gnu.org/develop.html describes).

As has been discussed, this may be impossible to fix and is certainly 
extremely difficult.

The fundamental issue affects libstdc++ atomicity.  The implementation is 
exposed as ABI.  Changing this would be a lot of work and might not even make 
it into gcc 3.5 (and would be an ABI change anyway).  

The implementation used for i486+ is the one which all other Linux vendors 
have standardized on, and which was "used" (but didn't work) for 386 in gcc 
3.2.  There is now (gcc 3.3) a version which works for 386, but it's up to 
*50%* slower, which is unacceptable.  Someone may come up with a faster one, 
but the slowdown is still going to be significant, and will still cause ABI 

Accordingly, currently anything using the new libstdc++ is effectively 486+ 
only.  This can't be changed without breaking compatibility with other 
distributions *and* cauing major slowdowns.

An alternate, separate architecture for real i386 would have to accept this 
incompatibility and recompile all libraries and programs linked to libstdc++.

ON another topic...

>2 KDE (3.1?)? Is there a status site? 3.1.1 is in unstable.
KDE 3 is held up by Qt 3, mostly...

which means that the priorities for these two aren't consistent. :-)  Either 
lower KDE3 or raise QT3...

>3 QT transition (Qt2->Qt3)? Mentioned in thread

I think QT is holding up a lot of other things too.


Reply to: