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

Re: How to get KDE in: total analysis

Steve Langasek wrote:
>And without forcing them, this: 
>> force kdeaddons/4:3.4.2-2.1
>will not work because a force hint will not cause consideration of packages
>which britney eliminates due to dependencies.
In the end, kdeaddons should most likely be force-hint forced into testing 
along with the rest of KDE, since it only breaks three architectures (ia64, 
sparc, m68k).  Making it uninstallable on those arches sounds to me like a 
reasonable tradeoff, as opposed to removing it from testing on all 
architectures.  But it's your call.

>If you want to help move the KDE transition forward, please
>work on figuring out how to get the toolchain back into shape.
Myself, I'd revert to an older version of binutils.  The porting bugs in the 
current version are quite beyond my abilities, since I don't have access to 
any of those architectures.  And #334626 / #334627 don't have enough 
information in the bug logs to do anything with them at all, frankly.

m68k will almost certainly need to be dropped because there's nobody solving 
the recurrent binutils bugs upstream.  Ben Elliston and Mark Kettenis can't 
keep up with it, I guess.  None of the m68k bugs are in the upstream 
bugzilla, so I don't anticipate any progress any time soon.

The major bugs with glibc are porting bugs (can't do anything about them) and 
a NIS bug (don't use it, can't even test it).  At any rate they look like 
they're being worked on by people who have a clue.

GCC-4.0 itself is actually in relatively good shape.  The two grave bugs are 
clearly upstream bugs which haven't been reported upstream (?!?).  I sent the 
requisite nags to the bug trails.

So, thanks, but I already looked at those, and there's really nothing I can do 
except to suggest that the maintainers upload an older version of binutils 
(which I doubt they'll want to do), and ask people to consider contacting GCC 

gcc-4_0-branch is actually relatively unstable, and it's probably not such a 
great idea to keep pulling CVS versions from it, either.

Nathanael Nerode  <neroden@twcny.rr.com>
[Insert famous quote here]

Reply to: