mpickering (Matthew Pickering)Administrator
User

Projects

Today

  • Clear sailing ahead.

Tomorrow

  • Clear sailing ahead.

Friday

  • Clear sailing ahead.

Badges

User Since
Dec 18 2014, 1:25 PM (113 w, 5 d)
Roles
Administrator
Availability
Available

Recent Activity

Today

mpickering closed T297: Phabricator stopped sending email as "Resolved".

@austin tells me that this should now be fixed.

Wed, Feb 22, 7:40 AM

Yesterday

mpickering added a comment to D3139: Try to fix boolformula into parsing "foo, bar, baz" into "And [foo, bar, baz]"instead of "And [foo, And [bar, baz]]".

Yes you are right.

Tue, Feb 21, 5:29 AM
mpickering added a comment to D3167: Simon's early-inline patch, take 2.

It is really hard to see what is going on here. Perhaps you could put each commit up as a separate diff?

Tue, Feb 21, 4:31 AM

Sun, Feb 19

mpickering commandeered D3113: Eliminate ListSetOps from imp_trust_pkgs.
Sun, Feb 19, 3:59 AM

Sat, Feb 18

mpickering accepted rGHC3eb737ee3f90: Generalize CmmUnwind and pass unwind information through NCG.
Sat, Feb 18, 12:50 PM

Fri, Feb 17

mpickering awarded Contributor to recipient: jcarey.
Fri, Feb 17, 3:43 AM

Thu, Feb 16

mpickering accepted D3139: Try to fix boolformula into parsing "foo, bar, baz" into "And [foo, bar, baz]"instead of "And [foo, And [bar, baz]]".

Nearly there!

Thu, Feb 16, 4:46 AM
mpickering added a comment to D3139: Try to fix boolformula into parsing "foo, bar, baz" into "And [foo, bar, baz]"instead of "And [foo, And [bar, baz]]".

Does your patch actually compile? arc is working again now, please can you update the diff so I can see it more clearly?

Thu, Feb 16, 3:47 AM

Wed, Feb 15

mpickering requested changes to D3139: Try to fix boolformula into parsing "foo, bar, baz" into "And [foo, bar, baz]"instead of "And [foo, And [bar, baz]]".

I don't think a test is necessary.

Wed, Feb 15, 4:53 AM

Tue, Feb 14

mpickering removed a reviewer for D3138: Testing...: rwbarton.
Tue, Feb 14, 10:59 AM
mpickering accepted D3138: Testing....

Great revision.

Tue, Feb 14, 10:53 AM
mpickering requested changes to D3138: Testing....
Tue, Feb 14, 10:49 AM
mpickering committed rGHC392cec4da9a7: Update .mailmap [skip ci] (authored by mpickering).
Update .mailmap [skip ci]
Tue, Feb 14, 10:16 AM
mpickering awarded Contributor to recipient: amindfv.
Tue, Feb 14, 10:13 AM

Mon, Feb 13

mpickering raised a concern with rGHC3eb737ee3f90: Generalize CmmUnwind and pass unwind information through NCG.

This commit seems like it broke the OS X build.

Mon, Feb 13, 3:14 AM

Sun, Feb 12

mpickering awarded Contributor to recipient: bollu.
Sun, Feb 12, 4:49 AM

Fri, Feb 10

mpickering updated the diff for D3113: Eliminate ListSetOps from imp_trust_pkgs.

rebase

Fri, Feb 10, 10:46 AM
mpickering accepted D3124: Relax test TH_addCStub2 so it succeeds on travis..

This passes for me now on OS X.

Fri, Feb 10, 10:24 AM
mpickering added a comment to D3124: Relax test TH_addCStub2 so it succeeds on travis..

Like I said on the commit. The test also fails on OS X due to differences between clang and gcc I guess. I can try to test this locally so you are not waiting years for the builder to finish.

Fri, Feb 10, 9:22 AM
mpickering added a comment to rGHCb9bebd8cedcc: Implement addCStub in template-haskell..

Also failing on OS X

Fri, Feb 10, 3:42 AM

Thu, Feb 9

mpickering added a comment to D3117: bufWrite: Save extra syscall when data fills handle buffer completely..

What command were you running?

Thu, Feb 9, 7:11 PM

Tue, Feb 7

mpickering updated the Trac tickets for D3103: Pass -v0 to ghc-pkg to reduce noise in build ouput.
Tue, Feb 7, 5:13 PM
mpickering created D3103: Pass -v0 to ghc-pkg to reduce noise in build ouput.
Tue, Feb 7, 5:12 PM
mpickering abandoned D3100: Pass -v0 to ghc-pkg when invoked by ghc-cabal.

This isn't the right fix it seems.

Tue, Feb 7, 4:33 PM
mpickering updated the Trac tickets for D3100: Pass -v0 to ghc-pkg when invoked by ghc-cabal.
Tue, Feb 7, 4:30 PM
mpickering created D3100: Pass -v0 to ghc-pkg when invoked by ghc-cabal.
Tue, Feb 7, 4:29 PM
mpickering retitled D2822: Allow type defaulting for multi-param type classes with ExtendedDefaultRules from Update: ExtendedDefaultRules to default multiparameter typeclasses to Allow type defaulting for multi-param type classes with ExtendedDefaultRules.
Tue, Feb 7, 10:43 AM

Mon, Feb 6

mpickering awarded Contributor to recipient: arybczak.
Mon, Feb 6, 1:46 PM
mpickering retitled D3057: Correct behaviour of push_bang_into_newtype when the pattern match has no arguments from Do not push bangs through newtypes when there are no arguments to Correct behaviour of push_bang_into_newtype when the pattern match has no arguments.
Mon, Feb 6, 12:52 PM
mpickering requested changes to D2822: Allow type defaulting for multi-param type classes with ExtendedDefaultRules.

Also the differential description needs to be updated to describe the feature as this will become the commit message.

Mon, Feb 6, 12:40 PM
mpickering added a comment to D2822: Allow type defaulting for multi-param type classes with ExtendedDefaultRules.

@amindfv Please can you upload you SSH key and update the diff (using arc diff --update D2822) so that the changes are pushed to the staging repo and built by harbormaster?

Mon, Feb 6, 12:40 PM

Sat, Feb 4

mpickering updated the diff for D3057: Correct behaviour of push_bang_into_newtype when the pattern match has no arguments.
  • Simon's comments and update existing test to test strictness properties
Sat, Feb 4, 8:32 AM
mpickering awarded Contributor to recipient: anniecherkaev.
Sat, Feb 4, 7:49 AM
mpickering added a comment to D3064: Don't return empty initial uncovered set for an unsat context.

I don't really think the pattern match checker is the right place to issue an inaccessibility warning so I think this is done for now.

Sat, Feb 4, 5:41 AM
mpickering updated the diff for D3064: Don't return empty initial uncovered set for an unsat context.
  • Choose right base commit
Sat, Feb 4, 5:37 AM
mpickering updated the diff for D3065: Don't panic when printing match with RecUpd context.
  • Update with new base commit
Sat, Feb 4, 5:36 AM
mpickering updated the diff for D3064: Don't return empty initial uncovered set for an unsat context.
  • Accept new test output
Sat, Feb 4, 5:35 AM
mpickering committed rGHC283acec1d730: Make split sections by default work again (authored by rwbarton).
Make split sections by default work again
Sat, Feb 4, 4:34 AM
mpickering closed D3070: Make split sections by default work again by committing rGHC283acec1d730: Make split sections by default work again (authored by rwbarton).
Sat, Feb 4, 4:34 AM
mpickering committed rGHC4d31880a0bfb: Fix comment (old filename '.lhs') in libraries/ (authored by takenobu).
Fix comment (old filename '.lhs') in libraries/
Sat, Feb 4, 4:34 AM
mpickering closed D3060: Fix comment (old filename '.lhs') in libraries/ by committing rGHC4d31880a0bfb: Fix comment (old filename '.lhs') in libraries/ (authored by takenobu).
Sat, Feb 4, 4:34 AM
mpickering committed rGHC8d60d739489b: Fix comment (old file names) in compiler/ (authored by takenobu).
Fix comment (old file names) in compiler/
Sat, Feb 4, 4:34 AM
mpickering closed D3076: Fix comment (old file names) in compiler/ by committing rGHC8d60d739489b: Fix comment (old file names) in compiler/ (authored by takenobu).
Sat, Feb 4, 4:34 AM
mpickering committed rGHC31bb85ffc4b6: Fix comment (old file names) in rts/ (authored by takenobu).
Fix comment (old file names) in rts/
Sat, Feb 4, 4:34 AM
mpickering closed D3075: Fix comment (old file names) in rts/ by committing rGHC31bb85ffc4b6: Fix comment (old file names) in rts/ (authored by takenobu).
Sat, Feb 4, 4:34 AM
mpickering committed rGHC9984024a58a2: Fix comment (old file names) in includes/ (authored by takenobu).
Fix comment (old file names) in includes/
Sat, Feb 4, 4:34 AM
mpickering closed D3074: Fix comment (old file names) in includes/ by committing rGHC9984024a58a2: Fix comment (old file names) in includes/ (authored by takenobu).
Sat, Feb 4, 4:34 AM
mpickering accepted D3070: Make split sections by default work again.

Good job you spotted this Reid.

Sat, Feb 4, 4:18 AM
mpickering awarded Contributor to recipient: takenobu.
Sat, Feb 4, 4:13 AM
mpickering requested changes to D3073: Do not drop dead code in the desugarer.

This solution feels very indirect. It is obtuse to keep dead bindings by making them live bindings in an unrelated part of the code upstream and potentially very confusing.

Sat, Feb 4, 4:09 AM

Fri, Feb 3

mpickering abandoned D3071: Testing with different base commit how diff shows up.

TEST COMPLETE

Fri, Feb 3, 4:53 PM
mpickering retitled D3071: Testing with different base commit how diff shows up from to Testing with different base commit how diff shows up.
Fri, Feb 3, 4:12 PM
mpickering updated D3069: Two new knobs for GHC API users.
Fri, Feb 3, 3:13 PM
mpickering raised a concern with rGHC266a9dc4cd34: Don't use the splitter on Darwin.

@rwbarton has raised concerns about this patch in Trac #13230

Fri, Feb 3, 3:03 PM
mpickering updated the diff for D3065: Don't panic when printing match with RecUpd context.

TEst

Fri, Feb 3, 2:41 PM
mpickering added a comment to D3066: Implement the basics of hex floating point literals.

@yav The commit message is generated from the description field. If in future a user needs to search the commit logs then it is useful to include a detailed commit message rather than a pointer. For example, a user might want to know which commit introduced HexFloatLiterals extension, which they could find out by searching the commit logs for HexFloatLiterals.

Fri, Feb 3, 11:40 AM
mpickering added a comment to D3024: Introduce GHC.TypeNats module, change KnownNat evidence to be Natural.

I think I agree with @christiaanb about this. What is the benefit of adding natural literals to core? Trac #13186 provides no motivation and this patch provides no motivation for changing the non-core representation to be Natural. It seems much easier to keep the internal representation as Integer and convert it once to a natural in a library somewhere if that is the evidence you want to provide.

I think there is an argument that can be made either way here. I'm personally rather agnostic on the question and would have been happy to continue using an Integer representation but I can see why @dfeuer would object that we should be more precise in our types. I think David would say that the fact that Nat was backed by an Integer should be viewed as a historical mistake and we should be writing the code that we want, not the code that we happened to stumble our way into.

Fri, Feb 3, 10:02 AM
mpickering raised a concern with rGHC1fcede43d2b3: Introduce GHC.TypeNats module, change KnownNat evidence to be Natural.

See some more discussion now on D3024

Fri, Feb 3, 4:46 AM
mpickering added a comment to D3024: Introduce GHC.TypeNats module, change KnownNat evidence to be Natural.

I think I agree with @christiaanb about this. What is the benefit of adding natural literals to core? Trac #13186 provides no motivation and this patch provides no motivation for changing the non-core representation to be Natural. It seems much easier to keep the internal representation as Integer and convert it once to a natural in a library somewhere if that is the evidence you want to provide.

Fri, Feb 3, 4:43 AM
mpickering requested changes to D3066: Implement the basics of hex floating point literals.

Please can you update the description to describe the feature and link to a relevant trac ticket?

Fri, Feb 3, 3:13 AM

Thu, Feb 2

mpickering added a comment to D3064: Don't return empty initial uncovered set for an unsat context.

I think this is better than not issuing any warnings as otherwise users will fix this warning and recompile and then have to fix the warnings which we didn't bother to show them. If someone complains later (which I doubt they will) we can implement something else.

Thu, Feb 2, 5:55 PM
mpickering retitled D3065: Don't panic when printing match with RecUpd context from to Don't panic when printing match with RecUpd context.
Thu, Feb 2, 4:49 PM
mpickering added a comment to D3064: Don't return empty initial uncovered set for an unsat context.

I am not 100% sure this is the right fix but we will see what George says.

Thu, Feb 2, 4:45 PM
mpickering retitled D3064: Don't return empty initial uncovered set for an unsat context from to Don't return empty initial uncovered set for an unsat context.
Thu, Feb 2, 4:43 PM
mpickering committed rGHC5728f4b83a06: Remove INLINE pragma on loopbreaker (authored by mpickering).
Remove INLINE pragma on loopbreaker
Thu, Feb 2, 4:35 PM
mpickering retitled D3062: Support abs as a primitive operation on floating point numbers. from compiler: fix trac issue #13212 to Support abs as a primitive operation on floating point numbers..
Thu, Feb 2, 10:44 AM
mpickering awarded Contributor to recipient: idontgetoutmuch.
Thu, Feb 2, 10:43 AM
mpickering requested changes to D3062: Support abs as a primitive operation on floating point numbers..
Thu, Feb 2, 10:43 AM

Wed, Feb 1

mpickering added a comment to D3058: Make interface loading for COMPLETE pragmas lazy.

I think it is fairly self-evident removing this will cause bad things to happen but I can add a test if you so desire.

Wed, Feb 1, 5:53 PM
mpickering retitled D3058: Make interface loading for COMPLETE pragmas lazy from to Make interface loading for COMPLETE pragmas lazy.
Wed, Feb 1, 4:33 PM
mpickering abandoned D3056: Add OS X specific allocations for T12234.
Wed, Feb 1, 3:48 PM
mpickering planned changes to D3057: Correct behaviour of push_bang_into_newtype when the pattern match has no arguments.
Wed, Feb 1, 11:43 AM
mpickering retitled D3057: Correct behaviour of push_bang_into_newtype when the pattern match has no arguments from to Do not push bangs through newtypes when there are no arguments.
Wed, Feb 1, 11:42 AM
mpickering added a comment to D3056: Add OS X specific allocations for T12234.

I don't know why they are different. How do people usually fix this?

Wed, Feb 1, 11:30 AM
mpickering updated the diff for D3056: Add OS X specific allocations for T12234.
  • Typo
Wed, Feb 1, 11:19 AM
mpickering retitled D3056: Add OS X specific allocations for T12234 from to Add OS X specific allocations for T12234.
Wed, Feb 1, 11:17 AM
mpickering committed rGHC4fa439e3ee28: Remove very broad ignore. [ci skip] (authored by Phyx).
Remove very broad ignore. [ci skip]
Wed, Feb 1, 4:21 AM
mpickering closed D3054: Remove very broad ignore. [ci skip] by committing rGHC4fa439e3ee28: Remove very broad ignore. [ci skip] (authored by Phyx).
Wed, Feb 1, 4:21 AM
mpickering accepted D3054: Remove very broad ignore. [ci skip].
Wed, Feb 1, 4:20 AM
mpickering created GHC (8.2).
Wed, Feb 1, 3:40 AM
mpickering changed the edit policy for GHC.
Wed, Feb 1, 3:39 AM
mpickering requested changes to D3052: Extensionalize if-then-else syntax, on by default in H98 and H2010.
Wed, Feb 1, 3:34 AM

Tue, Jan 31

mpickering committed rGHC748b79741652: Use top-level instances to solve superclasses where possible (authored by Daniel Haraj <dan@obsidian.systems>).
Use top-level instances to solve superclasses where possible
Tue, Jan 31, 5:00 PM
mpickering closed D2714: Use top-level instances to solve superclasses where possible by committing rGHC748b79741652: Use top-level instances to solve superclasses where possible (authored by Daniel Haraj <dan@obsidian.systems>).
Tue, Jan 31, 5:00 PM
mpickering updated D2714: Use top-level instances to solve superclasses where possible.
Tue, Jan 31, 4:25 PM
mpickering updated the diff for D2714: Use top-level instances to solve superclasses where possible.
  • Fix typo
Tue, Jan 31, 4:08 PM
mpickering updated the diff for D2714: Use top-level instances to solve superclasses where possible.
  • Respond to Simon's comments.
Tue, Jan 31, 3:59 PM
mpickering raised a concern with rGHCf41c27d3ffdd: Slighly clean up symbol loading error..

I think this change is causing two tests to fail on the travis build.

Tue, Jan 31, 9:37 AM
mpickering accepted D3050: Fix binary instance for SrcStrictness.

Ok, looks like all the tests pass. I am still intrigued whether this information is used in downstream modules at all.

Tue, Jan 31, 4:37 AM
mpickering added a comment to D3050: Fix binary instance for SrcStrictness.

Did you observe that this was causing a problem or just noticed it when reading the definitions?

Tue, Jan 31, 4:12 AM
mpickering requested changes to D3047: Implement cheapBuild.

Do you add a test or some performance numbers to show what difference this change makes?

Tue, Jan 31, 2:35 AM

Mon, Jan 30

mpickering requested changes to D3044: Turn libraries/integer-gmp/gmp/tarball into a submodule.

The builders are tripping up because of this change.

Mon, Jan 30, 3:11 AM

Fri, Jan 27

mpickering updated the diff for D3010: Add a flag to emit error messages as JSON.
  • Another test
Fri, Jan 27, 7:55 AM
mpickering updated the diff for D3010: Add a flag to emit error messages as JSON.
  • Don't initiase the logging action twice.
Fri, Jan 27, 7:40 AM
mpickering retitled D3032: Print COMPLETE pragmas in --show-iface from to Print COMPLETE pragmas in --show-iface.
Fri, Jan 27, 5:38 AM

Thu, Jan 26

mpickering added a comment to D3010: Add a flag to emit error messages as JSON.

ok! please review.

Thu, Jan 26, 5:20 PM
mpickering retitled D3010: Add a flag to emit error messages as JSON from WIP: Add a flag to emit error messages in JSON to Add a flag to emit error messages as JSON.
Thu, Jan 26, 5:08 PM
mpickering updated the diff for D3010: Add a flag to emit error messages as JSON.
  • Some docs
  • Don't introduce newlines when pprinting
  • Cleanup
Thu, Jan 26, 5:02 PM
mpickering added a reviewer for D3025: Fixes bug #11046: RyanGlScott.
Thu, Jan 26, 3:43 PM
mpickering updated the diff for D3010: Add a flag to emit error messages as JSON.
  • Updates
Thu, Jan 26, 3:42 PM