Trent W. Buck wrote:
> Um, that bug only affects architectures which do not support the
> threaded runtime (e.g. ARM). You won't be able to reproduce it on
> linux-i386 or linux-amd64 (and perhaps some other architectures).
>
> The fix for it was a patch to darcs.cabal (to add the "threaded" flag,
> which defaults to true), and a patch to debian/rules, which adds
> -f-threaded (disabling threading) when it isn't supported.
>
> The former patch was applied to upstream's HEAD, and I also applied it
> temporarily as a Debian-specific patch until it was released (in 2.4,
> IIRC).
Sorry for the lack of clarity. My attention was drawn to that bug
by rc-alert, which was treating it as an rc bug in lenny. But in
lenny, there _was_ no darcs.cabal; lenny uses 2.0 and darcs.cabal
was not added until 2.2. So the full history looks like this:
2.0.1rc1 - darcs starts using threaded runtime by default.
an autoconf test ensures that platforms not
supporting that cope fine.
2.0.2 - lenny release
2.2.0 - darcs switches build system to Cabal and
starts erroring out on platforms without threaded
runtime.
2.3.0-3 - patch to make use of threaded runtime configurable
with -threaded flag (and patch to debian/rules
to use it when appropriate)
2.4 - patch applied upstream
I used your reproduction recipe from
http://bugs.debian.org/cgi-bin/bugreport.cgi?msg=33;bug=541848
which just moves the threaded runtime out of the way.
Due to my poor bts-fu, I’d write that as:
# abusing “fixed” to mean “bug not present”
fixed 541848 2.0.2-1
found 541848 2.2.0-1
fixed 541848 2.3.0-3
and the only point is to convince the version tracking that this bug
is not present in lenny (and is obviously fixed in squeeze).
Sorry for the noise.
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
