Thu Sep 11 21:41:47 PDT 2008 Tim Chevalier <[EMAIL PROTECTED]>
* ext-core library: Add dead code eliminator for Core
Added code for dead code elimination to the ext-core library. This can be
used in concert with Language.Core.Merge to produce a single self-contained
module without unneces
Thu Sep 11 20:45:15 PDT 2008 Tim Chevalier <[EMAIL PROTECTED]>
* ext-core library: expose some more modules
M ./utils/ext-core/extcore.cabal -2 +2
View patch online:
http://darcs.haskell.org/ghc/_darcs/patches/20080912034515-d61e2-c2b05cf6329ffacc42b450403dab1e87dec37431.gz
__
Thu Sep 11 20:33:47 PDT 2008 Tim Chevalier <[EMAIL PROTECTED]>
* ext-core library: Change syntax for primitive coercions
Changed the ext-core syntax to include primitive coercions (left, right, sym,
trans, etc.) as syntax rather than referring them to their names as in GHC. (I
hope I upda
Thu Sep 11 20:22:19 PDT 2008 Tim Chevalier <[EMAIL PROTECTED]>
* ext-core library: Export a lot more things from Prims
See comments for details.
M ./utils/ext-core/Language/Core/Overrides.hs -2 +5
M ./utils/ext-core/Language/Core/Prims.hs -13 +78
View patch online:
http://darcs.ha
Thu Sep 11 20:14:52 PDT 2008 Tim Chevalier <[EMAIL PROTECTED]>
* ext-core library: Extend Core preprocessor
See comments for details.
M ./utils/ext-core/Language/Core/Prep.hs -5 +70
View patch online:
http://darcs.haskell.org/ghc/_darcs/patches/20080912031452-d61e2-fcb99278d4aee0d5902
Thu Sep 11 19:56:15 PDT 2008 Tim Chevalier <[EMAIL PROTECTED]>
* ext-core library: Export a bunch more stuff from the parser
M ./utils/ext-core/Language/Core/ParsecParser.hs -1 +3
View patch online:
http://darcs.haskell.org/ghc/_darcs/patches/20080912025615-d61e2-666d607080fb80246d65fa6ae4
Thu Sep 11 19:53:14 PDT 2008 Tim Chevalier <[EMAIL PROTECTED]>
* ext-core library: Fix performance bug
isUtupleTy was implemented inefficiently (and is called a lot by the
typechecker). Replaced with uglier but faster code.
M ./utils/ext-core/Language/Core/Core.hs -2 +11
View patch o
Thu Sep 11 19:38:42 PDT 2008 Tim Chevalier <[EMAIL PROTECTED]>
* ext-core library: Remove some cruft
M ./utils/ext-core/Language/Core/Merge.hs -4 +2
View patch online:
http://darcs.haskell.org/ghc/_darcs/patches/20080912023842-d61e2-1d67e5e5e2e0c998f6df553c02ed9756b7d19d3a.gz
On Wed, Sep 10, 2008 at 4:09 AM, Ian Lynagh <[EMAIL PROTECTED]> wrote:
> As far as I can tell from the error, git failed to create a file in
> /home/darcs/git/ghc.git/refs/remotes/github/ on monk, but the
> permissions look fine so I have no idea why that would happen.
>
> It would be interesting t
Thu Sep 11 19:15:35 PDT 2008 Tim Chevalier <[EMAIL PROTECTED]>
* ext-core library: Add code for merging multiple Core modules into a single
module
I added a new module, Merge, to the ext-core library that combines a list of
ext-core modules into a new, uniquely renamed module.
See co
I think it would probably be a good idea to move the External Core
library out of the GHC tree and onto Hackage (or something), since it
isn't needed to build GHC, and it was presumably only there in the
first place because Hackage didn't exist at the time. (And
admittedly, mostly because I don't
Thanks, I've forward this to Paul.
catamorphism:
> Hi, Don/others --
>
> The bandwidth problems with darcs.haskell.org seem to have returned,
> at least for today. I'm currently getting between 10-40K/s downloading
> a HEAD tarball via wget. By comparison, I get upwards of 1 M/s
> downloading a
Hi, Don/others --
The bandwidth problems with darcs.haskell.org seem to have returned,
at least for today. I'm currently getting between 10-40K/s downloading
a HEAD tarball via wget. By comparison, I get upwards of 1 M/s
downloading a large file via wget from cs.pdx.edu, and if you guys
can't do b
On Thu, Sep 11, 2008 at 09:25:46AM +0100, Simon Marlow wrote:
> Simon Peyton-Jones wrote:
>
> >=> bytestring001(optc)
> >Segmentation fault
>
> This is a segfault from GHC - perhaps you have old interface files in
> your bytestring package?
Nope, it's reproducible:
$ ls bytestring001.
On Thu, Sep 11, 2008 at 09:24:08PM +0100, Simon Marlow wrote:
> Sigh, I did wonder if this might happen. We rely on some undocumented
> behaviour in windres to get it to not mangle the arguments it gets
> passed (I had to read the source), and I guess they changed it. Can
> someone make a tick
Sigh, I did wonder if this might happen. We rely on some undocumented
behaviour in windres to get it to not mangle the arguments it gets
passed (I had to read the source), and I guess they changed it. Can
someone make a ticket please? We should get this fixed before the release.
Cheers,
On Wed, Sep 10, 2008 at 11:14:02PM +0200, Matthias Kilian wrote:
> Is there a preferred way to *not* install some of the tools internally
> used by GHC and built by Cabal? There's at least utils/pwd/pwd
> that really shouldn't be installed anywhere outside the build tree.
Thanks for the report, no
Thu Sep 11 06:52:53 PDT 2008 Ian Lynagh <[EMAIL PROTECTED]>
* Add a test for trac #1709
A ./tests/ghc-regress/simplCore/should_compile/Simpl020_A.hs
M ./tests/ghc-regress/simplCore/should_compile/all.T +4
A ./tests/ghc-regress/simplCore/should_compile/simpl020.hs
A ./tests/ghc-r
Thu Sep 11 08:38:27 PDT 2008 Ian Lynagh <[EMAIL PROTECTED]>
* Use OSThreadProcAttr for workerStart
M ./rts/Schedule.c -1 +1
M ./rts/Schedule.h -1 +3
View patch online:
http://darcs.haskell.org/ghc/_darcs/patches/20080911153827-3fd76-a321dfda7af61286b22cdd39a5994163571f42ac.gz
Thu Sep 11 07:05:19 PDT 2008 Ian Lynagh <[EMAIL PROTECTED]>
* Include docs in the bindists
M ./mk/bindist.mk -1 +7
View patch online:
http://darcs.haskell.org/ghc/_darcs/patches/20080911140519-3fd76-1653c979aec0a5a23fc8785d56b94223cf0a9e39.gz
__
Thu Sep 11 05:54:22 PDT 2008 Ian Lynagh <[EMAIL PROTECTED]>
* Don't install pwd
M ./utils/Makefile -1 +5
View patch online:
http://darcs.haskell.org/ghc/_darcs/patches/20080911125422-3fd76-b43ce9dd027531bc90e5dec3d15fb35757603300.gz
___
Cvs-ghc
Thu Sep 11 05:08:06 PDT 2008 Ian Lynagh <[EMAIL PROTECTED]>
* In stgReallocForGMP, we need to copy min(old_size,new_size)
We used to always copy old_size
M ./rts/sm/Storage.c -1 +3
View patch online:
http://darcs.haskell.org/ghc/_darcs/patches/20080911120806-3fd76-b74a183d48183bf238f67fe
Wed Sep 10 17:13:48 PDT 2008 Ian Lynagh <[EMAIL PROTECTED]>
* Tell perl to autoflush output in darcs-all. Suggested by Claus Reinke
M ./darcs-all +2
View patch online:
http://darcs.haskell.org/ghc/_darcs/patches/20080911001348-3fd76-e68686b997bb83a86631a3c81a81a322dd5bf44f.gz
Hi,
That was indeed it, and your suggested fix makes it work.
So I guess GHC needs to either warn about a too recent windres, or fix
itself so it can use the latest one.
Thanks
Neil
> -Original Message-
> From: shelarcy [mailto:[EMAIL PROTECTED]
> Sent: 11 September 2008 1:45 pm
> To
On Thu, Sep 11, 2008 at 11:54:35AM +0100, Claus Reinke wrote:
> ok, no dist/ involved there, actually. Still, I do occasionally get
> similar errors and wonder where they originate,
It's a bug in the bootstrapping compiler, fixed in the HEAD.
Thanks
Ian
_
Hi,
On Thu, 11 Sep 2008 20:00:30 +0900, Mitchell, Neil <[EMAIL PROTECTED]> wrote:
> Using a version of GHC HEAD on Windows XP, built in the last hour, I get
> a compilation failure when building haddock. It's possible this is a
> Cabal issue, so I've cc'd Duncan.
>
> (snip)
>
> [24 of 24] Compilin
Hi,
Using a version of GHC HEAD on Windows XP, built in the last hour, I get
a compilation failure when building haddock. It's possible this is a
Cabal issue, so I've cc'd Duncan.
I'm compiling haddock HEAD using GHC Paths 1.0.5. To compile haddock I
required two patches, which I've attached to t
ok, no dist/ involved there, actually. Still, I do occasionally get
similar errors and wonder where they originate, especially in a
clean build like the builbot's. Also, note the segfault.
Claus
From todays head reports (windows, head fast, stage1):
Preprocessing library Cabal-1.5.4...
Buildin
2008/9/10 Claus Reinke <[EMAIL PROTECTED]>:
> Wouldn't it be possible to (a) switch haddock to non--make mode
> and (b) just re-run every 'ghc ' command arising during 'make'
> as 'haddock '?
Haddock would need to write data to file for each module for that to
work, similar to GHC's .hi files.
I
Almost all generated files are now put in dist* directories, which makes
them easy to clean; and even if you do stop using a dist directory and
it doesn't get cleaned, it's hard to see how files in such a directory
could interfere with the build.
Oh, we'll probably find a way;-)
perhaps we alr
Simon Peyton-Jones wrote:
I ran a full testsuite today. Several bytestring tests are seg-faulting
consistently, when -O is on. See below. Does anyone have a clue about why?
Simon
=> bytestring001(normal)
cd ./lib/Data.ByteString && '/64playpen/simonpj/builds/HEAD-1/ghc/stage2-inplace/gh
Build results:
x86-64 Linux head:fail (failed bindisttest nofib.boot.0
nofib.boot.0_2 nofib.boot.0_3 nofib.boot.0_4 nofib.boot.0_5)
x86 Windows head: fail (failed bindisttest failed slave lost)
x86 Windows head fast:pass fail (failed stage1) lost fail (failed
getsu
Build results:
tnaur x86 Linux stable: pass
x86 Windows stable: fail (failed getsubrepos)
x86 Windows stable fast: pass pass pass fail (failed darcs) fail (failed darcs)
pass
x86-64 Linux stable: fail (failed stage1)
Old unexpected test failures:
TyFamUndec 4 gabor stable
33 matches
Mail list logo
