harpocrates (Alec Theriault)
User

Projects

User does not belong to any projects.

User Details

User Since
Oct 3 2017, 5:55 PM (50 w, 6 d)

Recent Activity

Today

harpocrates updated the diff for D5173: Allow (unparenthesized) kind signatures.
  • 'type' -> 'kind'
Mon, Sep 24, 12:13 PM
harpocrates added inline comments to D5173: Allow (unparenthesized) kind signatures.
Mon, Sep 24, 11:39 AM
harpocrates updated the diff for D5173: Allow (unparenthesized) kind signatures.
  • Fix annotation bug around parens
  • Add another test case
Mon, Sep 24, 11:34 AM
harpocrates added a comment to D5173: Allow (unparenthesized) kind signatures.

I think the grammar/deriving issues are solved, but I've uncovered an annotation issue that I really don't know how to fix - perhaps someone could advise?

Mon, Sep 24, 3:24 AM
harpocrates updated the diff for D5173: Allow (unparenthesized) kind signatures.
  • More conservative grammar
Mon, Sep 24, 3:00 AM

Yesterday

harpocrates planned changes to D5173: Allow (unparenthesized) kind signatures.

This still needs some work. Looks like some a handful of the failing tests are legitimate. Two types of problems stand out:

Sun, Sep 23, 1:13 PM
harpocrates updated the diff for D5173: Allow (unparenthesized) kind signatures.
  • Address nits
  • accept test case
Sun, Sep 23, 1:13 PM
harpocrates added inline comments to D5172: Don't show constraint tuples in errors (#14907).
Sun, Sep 23, 10:42 AM
harpocrates updated the diff for D5172: Don't show constraint tuples in errors (#14907).
  • Reuse tuple constraint sanitization
Sun, Sep 23, 10:42 AM
harpocrates updated the diff for D5167: Improve generated `GHC.Prim` docs.
  • -fno-warn-deprecations -> -Wno-deprecations
Sun, Sep 23, 9:40 AM
harpocrates created D5173: Allow (unparenthesized) kind signatures.
Sun, Sep 23, 4:39 AM
harpocrates updated the diff for D5167: Improve generated `GHC.Prim` docs.
  • Actually add the test
Sun, Sep 23, 3:24 AM
harpocrates created D5172: Don't show constraint tuples in errors (#14907).
Sun, Sep 23, 2:49 AM

Sat, Sep 22

harpocrates planned changes to D5057: [WIP] Remove Haddock tokens from the grammar.

Apologize for letting this languish for so long - I expect to pick this work back up in the coming weeks.

Sat, Sep 22, 10:52 PM

Fri, Sep 21

harpocrates updated the diff for D5167: Improve generated `GHC.Prim` docs.
  • Enable '-fno-warn-deprecations' in GHC.PrimopWrappers
Fri, Sep 21, 11:45 AM
harpocrates added a comment to D5167: Improve generated `GHC.Prim` docs.

I really appreciate the suggestion regarding the docs around (->)'s levity polymorphism.

Fri, Sep 21, 10:02 AM
harpocrates updated the diff for D5167: Improve generated `GHC.Prim` docs.
  • Address nits
Fri, Sep 21, 9:59 AM
harpocrates added inline comments to D5167: Improve generated `GHC.Prim` docs.
Fri, Sep 21, 8:54 AM
harpocrates updated the diff for D5167: Improve generated `GHC.Prim` docs.
Fri, Sep 21, 3:59 AM

Thu, Sep 20

harpocrates updated the summary of D5167: Improve generated `GHC.Prim` docs.
Thu, Sep 20, 9:45 AM
harpocrates created D5167: Improve generated `GHC.Prim` docs.
Thu, Sep 20, 9:45 AM

Wed, Sep 19

harpocrates updated the diff for D5158: Document the list data type.
  • Examples use h4, not h3
Wed, Sep 19, 11:15 PM

Tue, Sep 18

harpocrates updated the diff for D5158: Document the list data type.
  • Grammar
Tue, Sep 18, 3:36 PM
harpocrates updated the diff for D5158: Document the list data type.
  • Fix missing links
Tue, Sep 18, 3:36 PM
harpocrates added a comment to D5158: Document the list data type.

Haddock doesn't currently do anything with documentation of special items like [] (see https://github.com/haskell/haddock/issues/368), but I plan on changing this shortly (see https://github.com/haskell/haddock/pull/943). Plus, extra documentation won't hurt anyone in the meantime.

Tue, Sep 18, 1:23 PM
harpocrates created D5158: Document the list data type.
Tue, Sep 18, 1:23 PM

Aug 23 2018

harpocrates added a comment to D5100: Expose 'moduleToPkgConfAll' from 'PackageState'.

Some background: https://github.com/haskell/haddock/issues/916.

Aug 23 2018, 11:41 PM
harpocrates created D5100: Expose 'moduleToPkgConfAll' from 'PackageState'.
Aug 23 2018, 11:41 PM

Aug 22 2018

harpocrates added inline comments to D5067: [WIP] Hi Haddock: Enable haddock to generate docs from .hi-files.
Aug 22 2018, 10:11 AM

Aug 21 2018

harpocrates added inline comments to D5067: [WIP] Hi Haddock: Enable haddock to generate docs from .hi-files.
Aug 21 2018, 10:08 PM

Aug 20 2018

harpocrates requested changes to D5067: [WIP] Hi Haddock: Enable haddock to generate docs from .hi-files.
Aug 20 2018, 5:40 PM
harpocrates added a comment to D5057: [WIP] Remove Haddock tokens from the grammar.

I suspect this might fix https://ghc.haskell.org/trac/ghc/ticket/15029 too?!

Aug 20 2018, 2:13 PM

Aug 17 2018

harpocrates updated the diff for D5057: [WIP] Remove Haddock tokens from the grammar.

Really fix base commit

Aug 17 2018, 12:16 AM
harpocrates updated the diff for D5057: [WIP] Remove Haddock tokens from the grammar.

Rebased and fixed base commit.

Aug 17 2018, 12:16 AM

Aug 12 2018

harpocrates added a comment to D5057: [WIP] Remove Haddock tokens from the grammar.

I rebased, did some work, then did arc diff. Unfortunately, it looks like the base commit of this diff is the still the pre-rebase one. How would I go about fixing this?

Aug 12 2018, 10:17 PM
harpocrates updated the diff for D5057: [WIP] Remove Haddock tokens from the grammar.
  • Add some tests
Aug 12 2018, 10:01 PM

Aug 10 2018

harpocrates updated the summary of D5055: Fix ambiguous/out-of-scope Haddock identifiers.
Aug 10 2018, 3:37 PM
harpocrates added a comment to D5057: [WIP] Remove Haddock tokens from the grammar.

There are still some pretty big TODOs.

Aug 10 2018, 11:02 AM
harpocrates created D5057: [WIP] Remove Haddock tokens from the grammar.
Aug 10 2018, 10:55 AM

Aug 9 2018

harpocrates added a comment to D5055: Fix ambiguous/out-of-scope Haddock identifiers.

I've been playing with a branch of Haddock that prints out line numbers beside the warnings, so fixing this was actually not too bad. And it cuts down a fair bit on the amount of warnings that scroll by every time Haddock runs on base.

Aug 9 2018, 8:04 PM
harpocrates created D5055: Fix ambiguous/out-of-scope Haddock identifiers.
Aug 9 2018, 8:04 PM

Jul 23 2018

harpocrates added a comment to D5003: Explicitly tell 'getNameToInstances' mods to load.

Sold. Do I understand correctly, that this is as precise as your previous code but more memory efficient?

Jul 23 2018, 4:16 PM
harpocrates added a comment to D5003: Explicitly tell 'getNameToInstances' mods to load.

Haddock isn't going to build until https://github.com/haskell/haddock/pull/893 gets merged.

Jul 23 2018, 3:56 PM
harpocrates created D5003: Explicitly tell 'getNameToInstances' mods to load.
Jul 23 2018, 3:52 PM

Jun 8 2018

harpocrates added a comment to rGHC0361fc038e11: Move 'HsBangTy' out in constructor arguments.

The CI failures here look legit: the STDERR output haddockA038.stderr should have two trailing newlines. I reckon my git probably swallowed those up.

Jun 8 2018, 10:46 AM

May 31 2018

harpocrates updated the diff for D4727: Move 'HsBangTy' out in constructor arguments.
  • Refer to Trac ticket in test cases
May 31 2018, 3:16 PM
harpocrates updated the diff for D4727: Move 'HsBangTy' out in constructor arguments.
  • Better comment
May 31 2018, 10:56 AM

May 24 2018

harpocrates created D4727: Move 'HsBangTy' out in constructor arguments.
May 24 2018, 1:34 AM

Mar 8 2018

harpocrates added inline comments to D4217: Support adding objects from TH.
Mar 8 2018, 4:36 PM
harpocrates abandoned D4064: Support adding objects from TH (fixes #14298).

It looks like https://phabricator.haskell.org/D4217 is going to be the way forward for this.

Mar 8 2018, 4:36 PM
harpocrates updated the diff for D4217: Support adding objects from TH.
  • Update deprecation version
Mar 8 2018, 4:18 PM

Mar 5 2018

hsyl20 awarded D4217: Support adding objects from TH a Like token.
Mar 5 2018, 9:11 AM

Mar 3 2018

harpocrates updated the diff for D4217: Support adding objects from TH.
  • Rename TH foreign file API
    • Typo in comment
    • Add back addForeignFile
    • Get rid of warning
Mar 3 2018, 2:29 PM

Feb 20 2018

harpocrates added a comment to D4217: Support adding objects from TH.

@angerman ping!

Feb 20 2018, 7:59 PM

Feb 8 2018

harpocrates added a comment to D4217: Support adding objects from TH.

@angerman ping!

Feb 8 2018, 9:11 AM

Jan 28 2018

harpocrates added a comment to D4217: Support adding objects from TH.

I don't really want this to slip through the cracks, so I've gone ahead and rebased this onto today's HEAD. As I understand, the last time I left this ticket 2 months ago, @angerman had mentioned wanting to test this out in conjunction with D3608: Extend the Quasi Monad but progress on that front has stalled for the moment.

Jan 28 2018, 1:50 PM
harpocrates updated the diff for D4217: Support adding objects from TH.
  • Rename TH foreign file API
  • Typo in comment
  • Add back addForeignFile
  • Get rid of warning
Jan 28 2018, 1:42 PM

Jan 22 2018

harpocrates updated the diff for D4336: Option for LINE pragmas to get lexed into tokens.
  • Carelessly missed definition
Jan 22 2018, 7:32 AM
harpocrates added a comment to D4336: Option for LINE pragmas to get lexed into tokens.

There is something going slightly wrong. The comment on use_pos_prags doesn't seem to match the implementation?

Jan 22 2018, 6:54 AM
harpocrates updated the diff for D4336: Option for LINE pragmas to get lexed into tokens.
  • Update comment...
Jan 22 2018, 6:50 AM
harpocrates added a comment to D4336: Option for LINE pragmas to get lexed into tokens.

Most third-party uses of GHC's lexer I've seen use getTokenStream/getRichTokenStream. It is known that this is somewhat problematic in the presence of CPP (see https://ghc.haskell.org/trac/ghc/ticket/8265), but there are reliable workarounds (again see the thread). I recently stumbled on something much nastier: Haskell sources that use {-# LINE ... #-} and {-# COLUMN ... #-} pragmas will send getRichTokenStream into an infinite loop.

Jan 22 2018, 12:16 AM

Jan 21 2018

harpocrates added a reviewer for D4336: Option for LINE pragmas to get lexed into tokens: alexbiehl.
Jan 21 2018, 11:50 PM
harpocrates created D4336: Option for LINE pragmas to get lexed into tokens.
Jan 21 2018, 11:50 PM

Jan 13 2018

harpocrates abandoned D4094: Support constructor Haddocks in more places.

The diff in the previous comment has been merged.

Jan 13 2018, 10:13 AM

Jan 12 2018

harpocrates added a comment to D4292: Support constructor Haddocks in more places.

@bgamari Ping? This is just https://phabricator.haskell.org/D4094 rebased, with all the tests passing.

Jan 12 2018, 1:15 PM

Jan 8 2018

harpocrates added a comment to D4290: Haddock needs to pass visible modules for instance filtering.

@harpocrates Have you found out why this still misses some instances?

Jan 8 2018, 11:31 AM
harpocrates updated the diff for D4290: Haddock needs to pass visible modules for instance filtering.
  • Better comment
Jan 8 2018, 11:02 AM

Jan 7 2018

harpocrates added a comment to D4094: Support constructor Haddocks in more places.

@bgamari I've created https://phabricator.haskell.org/D4292 which is just these changes squashed together and on HEAD. My tests are still running, but everything in haddock and parser passes.

Jan 7 2018, 2:29 PM
harpocrates created D4292: Support constructor Haddocks in more places.
Jan 7 2018, 2:27 PM
harpocrates added a comment to D4094: Support constructor Haddocks in more places.

@bgamari No problem. Should I go ahead and rebase then?

Jan 7 2018, 1:13 PM

Jan 6 2018

harpocrates added a comment to D4290: Haddock needs to pass visible modules for instance filtering.

Actually, on closer inspection, this fixes a bunch of missing instance issues in base (I imagine outside of base too, but checking that would involve compiling a lot of large libraries with HEAD, and I don't have the time/willpower to do this now).

Jan 6 2018, 12:25 PM
harpocrates updated subscribers of D4290: Haddock needs to pass visible modules for instance filtering.

Here is the Haddock PR: https://github.com/haskell/haddock/pull/724

Jan 6 2018, 11:43 AM
harpocrates created D4290: Haddock needs to pass visible modules for instance filtering.
Jan 6 2018, 11:35 AM

Jan 2 2018

harpocrates added a comment to D4094: Support constructor Haddocks in more places.

@bgamari Just checking in to see what the status is on merging this.

Jan 2 2018, 2:46 PM

Dec 7 2017

harpocrates added a comment to D4094: Support constructor Haddocks in more places.

I have absolutely no clue why there is still a failing test. The test is about the exitcode obtained after running a compiled test case, while this diff is just supposed to be surface-level changes to the parser.

Dec 7 2017, 4:37 PM

Dec 6 2017

harpocrates updated the diff for D4094: Support constructor Haddocks in more places.
  • Fix annotations
Dec 6 2017, 10:48 AM

Dec 4 2017

harpocrates added a comment to D4217: Support adding objects from TH.

I'm starting to also want a LangCmm constructor.

Dec 4 2017, 5:04 PM

Nov 30 2017

harpocrates added a comment to D4094: Support constructor Haddocks in more places.

Alright. I now understand the three API annotation related failures I get. The [GADT decl discards annotations] note in the parser explains how the type production adds an annotation twice in different places since in some of the uses of that production (GADT constructors), the usual annotation would be lost.

Nov 30 2017, 10:37 PM
harpocrates updated the diff for D4217: Support adding objects from TH.
  • Rename TH foreign file API
  • Typo in comment
  • Add back addForeignFile
  • Get rid of warning
Nov 30 2017, 7:38 PM
harpocrates updated the diff for D4094: Support constructor Haddocks in more places.
  • Rebase
Nov 30 2017, 3:26 PM

Nov 29 2017

harpocrates added a comment to D4094: Support constructor Haddocks in more places.

The change on the Haddock side is now complete. See the PR linked in the previous comment.

Nov 29 2017, 9:34 AM
harpocrates added a comment to D4217: Support adding objects from TH.

I finally got around to finishing my WIP implementation of an inline-rust package using this feature: https://github.com/harpocrates/inline-rust/.

Nov 29 2017, 7:16 AM

Nov 28 2017

harpocrates added a reviewer for D4094: Support constructor Haddocks in more places: alexbiehl.

The changes to Haddock are nearly complete: see https://github.com/haskell/haddock/pull/709. Please feel free to provide feedback or suggestions!

Nov 28 2017, 2:05 AM

Nov 22 2017

harpocrates added a comment to D4217: Support adding objects from TH.
In D4217#117811, @ljli wrote:

I'm quite happy with this. Thanks for doing this @harpocrates.
I tested it like this: https://gist.github.com/ljli/c5aca47ca4eb89bded7e5cd87144bd96
GHCi (with -fobject-code) works and it makes my life a lot easier in general.
Longterm, we probably want to disentangle TH, IO and the compiler internals, but for the time being this solves a real problem.

Nov 22 2017, 6:48 PM
harpocrates added a comment to D4217: Support adding objects from TH.

If it was up to me I'd only leave and deprecate the "main" function, addForeignFile, since we're reasonably sure that there are no users of the other.

Nov 22 2017, 12:39 PM
harpocrates updated the diff for D4217: Support adding objects from TH.
  • Add back addForeignFile
Nov 22 2017, 12:27 PM
harpocrates added a comment to D4217: Support adding objects from TH.

The breakage is felt by people using addForeignFile/qAddForeignFile, and the fix is to use addForeignSource instead. Does the author of inline-c @bitonic have an opinion on this matter?

Thanks for the ping. I have no problem with it, but generally I'd just leave addForeignFile there and deprecate it. No need to break the API needlessly.

Nov 22 2017, 9:43 AM

Nov 21 2017

harpocrates updated subscribers of D4217: Support adding objects from TH.

The breakage of addForeignFile is indeed concerning. I think retiring addForeignFile and naming the new function addForeignFilePath might be preferrable for this reason.

Nov 21 2017, 10:25 PM
harpocrates updated the diff for D4217: Support adding objects from TH.
  • Typo in comment
Nov 21 2017, 10:00 PM
harpocrates updated the diff for D4217: Support adding objects from TH.
  • Rename TH foreign file API
Nov 21 2017, 10:00 PM
harpocrates added a comment to D4217: Support adding objects from TH.

Some bikeshedding of names:

Nov 21 2017, 11:49 AM
harpocrates added a comment to D4064: Support adding objects from TH (fixes #14298).

I've opened https://phabricator.haskell.org/D4217 which is FilePath based variant of this. The API is pretty much exactly the one @hsyl20 suggested.

Nov 21 2017, 3:18 AM
harpocrates added a comment to D4217: Support adding objects from TH.

This is a variant of https://phabricator.haskell.org/D4064 which is FilePath based instead.

Nov 21 2017, 3:16 AM
harpocrates created D4217: Support adding objects from TH.
Nov 21 2017, 3:13 AM

Nov 9 2017

harpocrates added a comment to D4064: Support adding objects from TH (fixes #14298).

Is there still a chance this will make it into 8.4?

Nov 9 2017, 5:24 PM

Oct 19 2017

harpocrates added a comment to D4064: Support adding objects from TH (fixes #14298).

By the way, I'll implement whatever this thread ends up agreeing on, and I'll do it as soon as there is a consensus. :)

Oct 19 2017, 9:38 PM
harpocrates updated the diff for D4105: Fix a bug in 'alexInputPrevChar'.
  • Add a Note explaining unicode handling in Alex
Oct 19 2017, 12:52 PM
harpocrates added a comment to D4064: Support adding objects from TH (fixes #14298).

I definitely much prefer the FilePath based approach (in fact, that was the one I first implemented) for its simplicity and efficiency. As I understand it, here are the arguments for the two approaches:

Oct 19 2017, 12:17 PM

Oct 18 2017

harpocrates created D4105: Fix a bug in 'alexInputPrevChar'.
Oct 18 2017, 1:47 AM

Oct 16 2017

harpocrates updated the diff for D4064: Support adding objects from TH (fixes #14298).
  • Rebase
Oct 16 2017, 3:54 PM