bgamari (Ben Gamari)Administrator
Blub... blub... blub...

Today

  • Clear sailing ahead.

Tomorrow

  • Clear sailing ahead.

Tuesday

  • Clear sailing ahead.

User Details

User Since
Jul 20 2014, 6:05 AM (145 w, 11 h)
Roles
Administrator
Availability
Available

I like working on GHC. 'tis all I have to say.

Recent Activity

Today

bgamari committed rGHCaf73f39b3d47: Bump haddock submodule (authored by bgamari).
Bump haddock submodule
Sun, Apr 30, 12:26 PM
bgamari committed rGHCd30ccd45705c: Don't setProgramDynFlags on every :load (authored by simonmar).
Don't setProgramDynFlags on every :load
Sun, Apr 30, 12:26 PM
bgamari committed rGHCf6289e72b266: PPC NCG: Implement callish prim ops (authored by trommler).
PPC NCG: Implement callish prim ops
Sun, Apr 30, 12:26 PM
bgamari committed rGHC70b235c26bae: CSE: Fix cut and paste error (authored by bgamari).
CSE: Fix cut and paste error
Sun, Apr 30, 12:26 PM
bgamari committed rGHCf977b76340aa: Make the tyvars in TH-reified data family instances uniform (authored by RyanGlScott).
Make the tyvars in TH-reified data family instances uniform
Sun, Apr 30, 12:26 PM
bgamari committed rGHC477e7d2860b2: Add regression test for #12104 (authored by RyanGlScott).
Add regression test for #12104
Sun, Apr 30, 12:26 PM
bgamari committed rGHCa97406e978de: Add a missing addDeferredBinding (authored by simonpj).
Add a missing addDeferredBinding
Sun, Apr 30, 12:26 PM
bgamari committed rGHC041463b2ebc2: get-win32-tarballs: Grab perl tarball from haskell.org, not GitHub (authored by bgamari).
get-win32-tarballs: Grab perl tarball from haskell.org, not GitHub
Sun, Apr 30, 12:26 PM
bgamari added a comment to D3441: Compile modules that are needed by template haskell, even with -fno-code..
In D3441#99854, @duog wrote:

I've done some benchmarks of using HscInterpreted vs HscAsm(no optimisation),

Interesting.

and I'm seeing HscAsm being slightly faster. Of course this is on a DEVEL flavour build, so it's probably meaningless. I don't have the horsepower to build an optimized stage 2.

Indeed this does call the result into question. However, I suppose for the time being using the default object target is fine. We can always fine-tune things if we do find there is a significant bit of performance left on the table.

.o files are now unconditionally temporary, .hi files are temporary or not depending on -fwrite-interface.

This sounds like the right solution. Thanks!

However, can we document this? In general it seems that we are missing a Note documenting -fno-code and -fwrite-interface, their purposes, interaction, and implementation. Do you suppose you could provide one, @duog? See Note [Recompilation checking when typechecking only] in GhcMake.hs for an example (and this likely would be a good place to put the new Note as well).

The current version doesn't actually work, and the added test T8025 fails. There is an error trying to load the .dyn_o file. I tried quite a few things without success, then got busy elsewhere. Apologies for having left this unfinished. I'll have another go, but will likely need some help working out what's going on.

Once it is working I would be happy to add the note.

Sun, Apr 30, 9:42 AM
bgamari added a comment to rGHC945c45ad50ed: Prefer #if defined to #ifdef.

I'm on it.

Sun, Apr 30, 7:56 AM

Fri, Apr 28

bgamari committed rGHC945c45ad50ed: Prefer #if defined to #ifdef (authored by bgamari).
Prefer #if defined to #ifdef
Fri, Apr 28, 10:12 PM
bgamari committed rGHCe5b3492f23c2: Enable new warning for fragile/incorrect CPP #if usage (authored by erikd).
Enable new warning for fragile/incorrect CPP #if usage
Fri, Apr 28, 10:12 PM
bgamari committed rGHC9ac22183e405: nativeGen: Use SSE2 SQRT instruction (authored by bgamari).
nativeGen: Use SSE2 SQRT instruction
Fri, Apr 28, 2:01 PM
bgamari closed D3508: nativeGen: Use SSE2 SQRT instruction by committing rGHC9ac22183e405: nativeGen: Use SSE2 SQRT instruction.
Fri, Apr 28, 2:01 PM
bgamari added a comment to D3510: CSE: Fix cut and paste error.

Wow! How long has this bug been here? How did you come across it?

Fri, Apr 28, 2:00 PM
bgamari committed rGHCf9aa658ba829: Be a bit more eager to inline in a strict context (authored by simonpj).
Be a bit more eager to inline in a strict context
Fri, Apr 28, 1:29 PM
bgamari committed rGHC9b58c6360b9c: Cure exponential behaviour in the simplifier (authored by simonpj).
Cure exponential behaviour in the simplifier
Fri, Apr 28, 1:29 PM
bgamari committed rGHC09249f930895: Do Note [Improving seq] always (authored by simonpj).
Do Note [Improving seq] always
Fri, Apr 28, 1:29 PM
bgamari committed rGHC3ccdf40fd141: Print module when dumping rules (authored by mpickering).
Print module when dumping rules
Fri, Apr 28, 1:29 PM
bgamari committed rGHC9f9b90f1fb85: CSE: Fix cut and paste error (authored by bgamari).
CSE: Fix cut and paste error
Fri, Apr 28, 1:29 PM
bgamari closed D3510: CSE: Fix cut and paste error by committing rGHC9f9b90f1fb85: CSE: Fix cut and paste error.
Fri, Apr 28, 1:29 PM
bgamari committed rGHC228d4670e98e: Use memcpy in cloneArray (authored by bgamari).
Use memcpy in cloneArray
Fri, Apr 28, 1:29 PM
bgamari closed D3504: Use memcpy in cloneArray by committing rGHC228d4670e98e: Use memcpy in cloneArray.
Fri, Apr 28, 1:29 PM
bgamari committed rGHCb2c38d6b4003: Make the tyvars in TH-reified data family instances uniform (authored by RyanGlScott).
Make the tyvars in TH-reified data family instances uniform
Fri, Apr 28, 1:29 PM
bgamari closed D3505: Make the tyvars in TH-reified data family instances uniform by committing rGHCb2c38d6b4003: Make the tyvars in TH-reified data family instances uniform (authored by RyanGlScott).
Fri, Apr 28, 1:29 PM
bgamari committed rGHC69b9b853e3e6: Add regression test for #12104 (authored by RyanGlScott).
Add regression test for #12104
Fri, Apr 28, 1:29 PM
bgamari closed D3495: Add regression test for #12104 by committing rGHC69b9b853e3e6: Add regression test for #12104 (authored by RyanGlScott).
Fri, Apr 28, 1:29 PM
bgamari updated the diff for D3508: nativeGen: Use SSE2 SQRT instruction.

Add test

Fri, Apr 28, 1:29 PM
bgamari committed rGHCba597c1dd1da: get-win32-tarballs: Grab perl tarball from haskell.org, not GitHub (authored by bgamari).
get-win32-tarballs: Grab perl tarball from haskell.org, not GitHub
Fri, Apr 28, 12:22 PM
bgamari closed D3509: get-win32-tarballs: Grab perl tarball from haskell.org, not GitHub by committing rGHCba597c1dd1da: get-win32-tarballs: Grab perl tarball from haskell.org, not GitHub.
Fri, Apr 28, 12:21 PM
bgamari updated the diff for D3508: nativeGen: Use SSE2 SQRT instruction.

Try again

Fri, Apr 28, 12:18 PM
bgamari updated the diff for D3508: nativeGen: Use SSE2 SQRT instruction.

Fix it

Fri, Apr 28, 12:13 PM
bgamari updated the diff for D3504: Use memcpy in cloneArray.

Handle cloneSmallArray as well

Fri, Apr 28, 12:07 PM
bgamari added inline comments to D3496: Ignore ANN pragmas with no TH and no external interpreter..
Fri, Apr 28, 12:07 PM
bgamari created D3510: CSE: Fix cut and paste error.
Fri, Apr 28, 11:51 AM
bgamari created D3509: get-win32-tarballs: Grab perl tarball from haskell.org, not GitHub.
Fri, Apr 28, 11:41 AM
bgamari updated the diff for D3508: nativeGen: Use SSE2 SQRT instruction.

Fix it

Fri, Apr 28, 10:23 AM
bgamari updated the Trac tickets for D3508: nativeGen: Use SSE2 SQRT instruction.
Fri, Apr 28, 10:06 AM
bgamari created D3508: nativeGen: Use SSE2 SQRT instruction.
Fri, Apr 28, 10:02 AM

Thu, Apr 27

bgamari committed rGHCd213ef364321: Bump process to 1.6 (authored by bgamari).
Bump process to 1.6
Thu, Apr 27, 3:11 PM
bgamari committed rGHC933fb440ad4a: Document mkWeak# (authored by bgamari).
Document mkWeak#
Thu, Apr 27, 3:11 PM
bgamari committed rGHC768f02d0cb4c: Don't describe tuple sections as "Python-style" (authored by chris-martin).
Don't describe tuple sections as "Python-style"
Thu, Apr 27, 3:11 PM
bgamari committed rGHC8ef9716f8f08: Revert "Remove special casing of Windows in generic files" (authored by bgamari).
Revert "Remove special casing of Windows in generic files"
Thu, Apr 27, 3:11 PM
bgamari added a reverting commit for rGHC48385cb2fc29: Remove special casing of Windows in generic files: rGHC8ef9716f8f08: Revert "Remove special casing of Windows in generic files".
Thu, Apr 27, 3:11 PM
bgamari committed rGHC1bc7429d921a: Add regression test for #13603 (authored by RyanGlScott).
Add regression test for #13603
Thu, Apr 27, 3:11 PM
bgamari committed rGHC5eb185a3dd09: Update Cabal submodule, with necessary wibbles. (authored by ezyang).
Update Cabal submodule, with necessary wibbles.
Thu, Apr 27, 3:11 PM
bgamari committed rGHC5b0058e13910: testsuite/driver: Fix deletion retry logic on Windows (authored by bgamari).
testsuite/driver: Fix deletion retry logic on Windows
Thu, Apr 27, 3:11 PM
bgamari committed rGHCf38e76a3a8d9: Disable terminfo, if we don’t build it. (authored by angerman).
Disable terminfo, if we don’t build it.
Thu, Apr 27, 3:11 PM
bgamari committed rGHC3b4af97b77a0: Document the kind generalization behavior observed in #13555 (authored by RyanGlScott).
Document the kind generalization behavior observed in #13555
Thu, Apr 27, 3:11 PM
bgamari committed rGHC2201cfbc2301: Export function for use in GHC API (authored by alanz).
Export function for use in GHC API
Thu, Apr 27, 3:11 PM
bgamari committed rGHCc82add780b04: ghc.mk: fix 'make install' for cross-mingw32 (authored by trofi).
ghc.mk: fix 'make install' for cross-mingw32
Thu, Apr 27, 3:10 PM
bgamari committed rGHC0992af5f68bd: ghc: tweak cross-compilation to mingw32 (authored by trofi).
ghc: tweak cross-compilation to mingw32
Thu, Apr 27, 3:10 PM
bgamari committed rGHC1a472f18e2ab: Fix compilation for !HAVE_FLOCK (authored by hvr).
Fix compilation for !HAVE_FLOCK
Thu, Apr 27, 3:10 PM
bgamari committed rGHCfe80c3128bb0: Minor reordering of `#include`s fixing compilation on AIX (authored by hvr).
Minor reordering of `#include`s fixing compilation on AIX
Thu, Apr 27, 3:10 PM
bgamari committed rGHCd8d810e281b9: rts: tweak cross-compilation to mingw32 (authored by trofi).
rts: tweak cross-compilation to mingw32
Thu, Apr 27, 3:10 PM
bgamari committed rGHC9fa02bebfb06: aclocal.m4: treat '*-w64-mingw32' targets as windows (authored by trofi).
aclocal.m4: treat '*-w64-mingw32' targets as windows
Thu, Apr 27, 3:10 PM
bgamari committed rGHCfa0110b5d6d0: configure.ac: print paths to dllwrap and windres (authored by trofi).
configure.ac: print paths to dllwrap and windres
Thu, Apr 27, 3:10 PM
bgamari committed rGHC45912ea62008: configure.ac: print resolved 'ar' and 'ranlib' tools (authored by trofi).
configure.ac: print resolved 'ar' and 'ranlib' tools
Thu, Apr 27, 3:10 PM
bgamari committed rGHC83b7f9f5d2e1: Use non-canocalized triple as cross-compiler prefix (authored by trofi).
Use non-canocalized triple as cross-compiler prefix
Thu, Apr 27, 3:10 PM
bgamari committed rGHC7d73ec674e1a: Update haddock submodule (authored by bgamari).
Update haddock submodule
Thu, Apr 27, 3:10 PM
bgamari committed rGHC2744c94124a0: Bump process to 1.6 (authored by bgamari).
Bump process to 1.6
Thu, Apr 27, 2:05 PM
bgamari accepted D3290: Make LLVM output robust to -dead_strip on mach-o platforms.

I'm still not entirely sure why this is better than the mangler approach given that ultimately we will eventually need to move to some more permanent solution anyways, but fair enough.

Thu, Apr 27, 11:32 AM
bgamari accepted D3505: Make the tyvars in TH-reified data family instances uniform.

Lovely.

Thu, Apr 27, 11:22 AM
bgamari created D3506: [WIP] users guide: Some notes on -feager-blackholing.
Thu, Apr 27, 11:00 AM

Wed, Apr 26

bgamari created D3504: Use memcpy in cloneArray.
Wed, Apr 26, 3:36 PM
bgamari committed rGHC579bb7669f40: Update Cabal submodule, with necessary wibbles. (authored by ezyang).
Update Cabal submodule, with necessary wibbles.
Wed, Apr 26, 3:04 PM
bgamari closed D3501: Update Cabal submodule, with necessary wibbles. by committing rGHC579bb7669f40: Update Cabal submodule, with necessary wibbles. (authored by ezyang).
Wed, Apr 26, 3:04 PM
bgamari accepted D3501: Update Cabal submodule, with necessary wibbles..

Thanks @ezyang!

Wed, Apr 26, 2:26 PM
bgamari added a comment to D3503: Cmake file for rts..

Hmm, I'm honestly not sure that this is something that we should include in the tree as it will almost inevitably bitrot.

Wed, Apr 26, 2:25 PM
bgamari committed rGHC350d268aa625: Update hsc2hs submodule to 0.68.2 (authored by bgamari).
Update hsc2hs submodule to 0.68.2
Wed, Apr 26, 1:17 PM

Tue, Apr 25

bgamari committed rGHC47be6444d357: Add instances for Data.Ord.Down (authored by adamse).
Add instances for Data.Ord.Down
Tue, Apr 25, 8:12 PM
bgamari closed D3500: Add instances `Num`, `Functor`, `Applicative`, `Monad`, `Semigroup` and `Monoid` for `Data.Ord.Down` (#13097). by committing rGHC47be6444d357: Add instances for Data.Ord.Down (authored by adamse).
Tue, Apr 25, 8:12 PM
bgamari committed rGHC244602697c30: Document mkWeak# (authored by bgamari).
Document mkWeak#
Tue, Apr 25, 8:12 PM
bgamari committed rGHCda792e47981f: Only pretty-print binders in closed type families with -fprint-explicit-foralls (authored by RyanGlScott).
Only pretty-print binders in closed type families with -fprint-explicit-foralls
Tue, Apr 25, 8:12 PM
bgamari closed D3498: Document mkWeak# by committing rGHC244602697c30: Document mkWeak#.
Tue, Apr 25, 8:12 PM
bgamari committed rGHC71c3cea60c74: Add backup url and sync support for Win32 tarball script (authored by Phyx).
Add backup url and sync support for Win32 tarball script
Tue, Apr 25, 8:12 PM
bgamari closed D3497: Don't pretty-print binders in closed type families without -fprint-explicit-foralls by committing rGHCda792e47981f: Only pretty-print binders in closed type families with -fprint-explicit-foralls (authored by RyanGlScott).
Tue, Apr 25, 8:12 PM
bgamari committed rGHC89a3241f7085: PPC NCG: Implement callish prim ops (authored by trommler).
PPC NCG: Implement callish prim ops
Tue, Apr 25, 8:12 PM
bgamari closed D3488: Add backup url and sync support for Win32 tarball script by committing rGHC71c3cea60c74: Add backup url and sync support for Win32 tarball script (authored by Phyx).
Tue, Apr 25, 8:12 PM
bgamari committed rGHC9373994acaf1: configure: Kill off FP_ARG_WITH_* (authored by bgamari).
configure: Kill off FP_ARG_WITH_*
Tue, Apr 25, 8:12 PM
bgamari closed D2973: PPC NCG: Implement callish prim ops by committing rGHC89a3241f7085: PPC NCG: Implement callish prim ops (authored by trommler).
Tue, Apr 25, 8:12 PM
bgamari committed rGHC661088645406: Revert "Remove special casing of Windows in generic files" (authored by bgamari).
Revert "Remove special casing of Windows in generic files"
Tue, Apr 25, 8:12 PM
bgamari closed D3499: configure: Kill off FP_ARG_WITH_* by committing rGHC9373994acaf1: configure: Kill off FP_ARG_WITH_*.
Tue, Apr 25, 8:12 PM
bgamari added a reverting commit for rGHC48385cb2fc29: Remove special casing of Windows in generic files: rGHC661088645406: Revert "Remove special casing of Windows in generic files".
Tue, Apr 25, 8:11 PM
bgamari committed rGHC688272ba31df: Don't describe tuple sections as "Python-style" (authored by chris-martin).
Don't describe tuple sections as "Python-style"
Tue, Apr 25, 8:11 PM
bgamari accepted D3500: Add instances `Num`, `Functor`, `Applicative`, `Monad`, `Semigroup` and `Monoid` for `Data.Ord.Down` (#13097)..

Looks good to me.

Tue, Apr 25, 4:37 PM
bgamari created D3499: configure: Kill off FP_ARG_WITH_*.
Tue, Apr 25, 3:59 PM
bgamari requested changes to D3448: [linker] fix armv7 & add aarch64.

What is the status of this, @angerman?

Tue, Apr 25, 11:54 AM
bgamari requested changes to D3469: [iserv] fix loadDLL.

Can we have a comment explaining this logic?

Tue, Apr 25, 11:39 AM
bgamari requested changes to D3447: [linker] Add ocInit/ocDeinit for ELF.

There are currently some build issues here.

Tue, Apr 25, 11:38 AM
bgamari accepted D3497: Don't pretty-print binders in closed type families without -fprint-explicit-foralls.

Indeed the previous behavior was indeed a bit confusing.

Tue, Apr 25, 11:37 AM
bgamari requested changes to D3453: WIP: Remove new name logic from renamer.

Bumping out of review queue.

Tue, Apr 25, 11:35 AM
bgamari accepted D3488: Add backup url and sync support for Win32 tarball script.

Sure, sounds reasonable.

Tue, Apr 25, 11:34 AM
bgamari requested changes to D3487: Some cleaning up in TyCoRep and friends.

Bumping out of review queue.

Tue, Apr 25, 11:32 AM
bgamari accepted D3495: Add regression test for #12104.

Thanks Ryan!

Tue, Apr 25, 11:32 AM
bgamari accepted D3383: Implement split-sections support for windows..

Yay! This looks good to me.

Tue, Apr 25, 11:32 AM
bgamari updated the Trac tickets for D3498: Document mkWeak#.
Tue, Apr 25, 10:47 AM
bgamari created D3498: Document mkWeak#.
Tue, Apr 25, 10:47 AM
bgamari added a comment to D3496: Ignore ANN pragmas with no TH and no external interpreter..

Also, this could really use a comment explaining why this logic is necessary.

Tue, Apr 25, 10:36 AM
bgamari requested changes to D3349: [libffi] use master..
Tue, Apr 25, 8:54 AM