Control: tag -1 + fixed-upstream
Control: forwarded http://hackage.haskell.org/trac/ghc/ticket/6063
Hi,
upstream has no implemented run-time-detection of the linker in
question, so this bug will eventually be fixed in Debian as well.
Greetings,
Joachim
--
Joachim nomeata Breitner
Debian
reassign 673081 ghc 7.4.1-4
tags 673081 =
quit
Joachim Breitner wrote:
ghc (7.4.1-4) unstable; urgency=low
[ Erik de Castro Lopo ]
* Add debian/patches/fix-PPC-right-shift-bug which fixes upstream GHC bug:
http://hackage.haskell.org/trac/ghc/ticket/6156 (Closes: #677591)
*
(resending because I forgot to cc ghc@packages on the reassignment.
Sorry for the noise.)
Joachim Breitner wrote:
ghc (7.4.1-4) unstable; urgency=low
[ Erik de Castro Lopo ]
* Add debian/patches/fix-PPC-right-shift-bug which fixes upstream GHC bug:
On Thu, May 17, 2012 at 16:47:52 -0500, Jonathan Nieder wrote:
Dear release team,
Joachim Breitner wrote[1]:
I’d rather like to be able to transition the current set of Haskell
packages to testing first and then, if there is time before the freeze,
tackle this bug. For that, the
Julien Cristau wrote:
What's the rationale for this bug being 'serious' in the first place?
That seems rather inflated to me.
ghc is unusable when binutils-gold is installed. There is no conflict
between them declared.
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
severity 673081 wishlist
kthxbye
On Fri, May 18, 2012 at 02:04:01PM -0500, Jonathan Nieder wrote:
Do you mean that GNU gold should be bug-for-bug compatible with
the libbfd-based ld? (Then I would disagree --- for example, gold
does not support linker scripts, for good reasons.)
What I mean
Hi again,
Clint Adams wrote:
Since you haven't responded to anything I said that was actually
relevant,
Sorry, I guess I misunderstood your messages. Which was the relevant
part? I was assuming you've seen the upstream ghc report, but that
might be a bad assumption. Have you?
A little
reassign 673081 binutils-gold 2.22-6
severity 673081 important
affects 673081 + ghc
tags 673081 + upstream moreinfo
quit
Jonathan Nieder wrote:
In the long term, it might make sense for gold to learn --hash-size.
Do you think that gold should learn --hash-size?
I can only guess so, because
On Sat, May 19, 2012 at 01:29:22PM -0500, Jonathan Nieder wrote:
Sorry, I guess I misunderstood your messages. Which was the relevant
part? I was assuming you've seen the upstream ghc report, but that
The relevant part is that it's absurd for ghc to Breaks any linker
at all.
might be a bad
Clint Adams wrote:
The relevant part is that it's absurd for ghc to Breaks any linker
at all.
Based on the English language meaning of Breaks, I agree.
Would Conflicts have been a better suggestion? I marked the bug as RC
because one package is broken when the other is installed but the
On Sat, May 19, 2012 at 01:39:14PM -0500, Jonathan Nieder wrote:
Do you think that gold should learn --hash-size?
I think that if people want gold to become the default linker
and for some reason don't want me to despise them, they should
do whatever it takes to make gold be less broken than
On Sat, May 19, 2012 at 01:47:26PM -0500, Jonathan Nieder wrote:
It all seems academic, though, unless configuring ghc to use ld.bfd
is very difficult for some reason.
That does seem like it could be a reasonable compromise for
the moment.
--
To UNSUBSCRIBE, email to
On Thu, May 17, 2012 at 05:34:15PM -0500, Jonathan Nieder wrote:
Not really. I mean that ghc Depends on !binutils-gold because its
configure script is too zealous in adopting features from the linker
it happens to be built against, but there's not any way to express
that in debian/control.
I
Clint Adams wrote:
ghc seems to work just fine with real ld.
What exactly is the real ld? ghc works just fine with many linkers,
but only if you configure and run with the same one.
I'm not sure what the point of this sort of meta-questioning is,
though. If there's some practical reason for
Hi Jonathan,
Am Mittwoch, den 16.05.2012, 10:01 -0500 schrieb Jonathan Nieder:
Joachim Breitner wrote:
Am Dienstag, den 15.05.2012, 22:22 +0100 schrieb Finn Lawler:
ghc 7.4.1-3 (unstable) appears to be configured against the ordinary
binutils ld. The ld.gold that binutils-gold 2.22-6
# undeclared dependency (policy §3.5)
severity 673081 serious
quit
Hi,
Joachim Breitner wrote:
thanks. I’d like to avoid a ghc upload at this stage of the release
process, so this change will likely not make it into stable. Do you see
this as a major problem?
I think it's a serious bug,
Hi,
Am Donnerstag, den 17.05.2012, 12:58 -0500 schrieb Jonathan Nieder:
Joachim Breitner wrote:
thanks. I’d like to avoid a ghc upload at this stage of the release
process, so this change will likely not make it into stable. Do you see
this as a major problem?
I think it's a serious
On Thu, May 17, 2012 at 12:58:22PM -0500, Jonathan Nieder wrote:
I think it's a serious bug, but it could be fixed after the release if
you want. Would it be problematic to make an upload that either
changes the linker used to ld.bfd or (less pleasant) adds a Breaks
against binutils-gold?
Dear release team,
Joachim Breitner wrote[1]:
I’d rather like to be able to transition the current set of Haskell
packages to testing first and then, if there is time before the freeze,
tackle this bug. For that, the severity needs to be lowered, though, as
otherwise nothing will migrate.
I
Clint Adams wrote:
On Thu, May 17, 2012 at 12:58:22PM -0500, Jonathan Nieder wrote:
I think it's a serious bug, but it could be fixed after the release if
you want. Would it be problematic to make an upload that either
changes the linker used to ld.bfd or (less pleasant) adds a Breaks
Hi,
Am Dienstag, den 15.05.2012, 22:22 +0100 schrieb Finn Lawler:
Package: ghc
Version: 7.4.1-3
Severity: normal
ghc 7.4.1-3 (unstable) appears to be configured against the ordinary binutils
ld. The ld.gold that binutils-gold 2.22-6 (testing) replaces it with does
not accept the
Hi Joachim,
Joachim Breitner wrote:
Am Dienstag, den 15.05.2012, 22:22 +0100 schrieb Finn Lawler:
ghc 7.4.1-3 (unstable) appears to be configured against the ordinary
binutils ld. The ld.gold that binutils-gold 2.22-6 (testing) replaces it
with does not accept the `--hash-size' option,
Package: ghc
Version: 7.4.1-3
Severity: normal
ghc 7.4.1-3 (unstable) appears to be configured against the ordinary binutils
ld. The ld.gold that binutils-gold 2.22-6 (testing) replaces it with does not
accept the `--hash-size' option, resulting in a fatal error at the linking
stage.
See
23 matches
Mail list logo