sjakobi (Simon Jakobi)
User

Projects

User does not belong to any projects.

User Details

User Since
Feb 18 2018, 12:08 PM (31 w, 2 d)

Recent Activity

Tue, Sep 11

sjakobi added a comment to D4924: Add a strict version of foldMap to Foldable.

The test failures look unrelated to me.

Tue, Sep 11, 4:03 PM

Mon, Sep 10

sjakobi updated the diff for D4924: Add a strict version of foldMap to Foldable.

Rebase & changelog entry

Mon, Sep 10, 9:14 PM

Wed, Aug 29

sjakobi added a comment to D4913: Make :doc work for the ghc library.

@bgamari: Can we get this into ghc-8.6 too?

Wed, Aug 29, 10:31 AM

Aug 23 2018

sjakobi added a comment to D5067: [WIP] Hi Haddock: Enable haddock to generate docs from .hi-files.
  • mkMaps and other Haddock survivors. Some of the nasty bits from haddock moved into GHC with this patch (and the prior one). We should look for an alternative to store and extract documentation. Simon suggest putting documentation into IfaceDecl directly. Rendering would then have to traverse the IfaceDecls only.
Aug 23 2018, 10:37 AM

Aug 22 2018

sjakobi added inline comments to D5067: [WIP] Hi Haddock: Enable haddock to generate docs from .hi-files.
Aug 22 2018, 4:21 AM

Aug 20 2018

sjakobi added inline comments to D5067: [WIP] Hi Haddock: Enable haddock to generate docs from .hi-files.
Aug 20 2018, 6:41 PM
sjakobi added a comment to D5067: [WIP] Hi Haddock: Enable haddock to generate docs from .hi-files.

Please mention the Trac ticket number in the summary (click "Edit revision"). Create a ticket first if it doesn't exist yet.

Aug 20 2018, 5:07 PM
sjakobi 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, 1:44 PM

Aug 16 2018

adamse awarded D5067: [WIP] Hi Haddock: Enable haddock to generate docs from .hi-files a Mountain of Wealth token.
Aug 16 2018, 5:12 PM

Aug 14 2018

sjakobi updated the summary of D5067: [WIP] Hi Haddock: Enable haddock to generate docs from .hi-files.
Aug 14 2018, 8:40 AM
sjakobi updated the summary of D5067: [WIP] Hi Haddock: Enable haddock to generate docs from .hi-files.
Aug 14 2018, 8:39 AM
sjakobi created D5067: [WIP] Hi Haddock: Enable haddock to generate docs from .hi-files.
Aug 14 2018, 8:15 AM

Aug 3 2018

sjakobi added a comment to D4852: --show-iface: Qualify all non-local names.

Bump

Aug 3 2018, 11:14 PM
sjakobi added a comment to D4924: Add a strict version of foldMap to Foldable.

@bgamari: Bump

Aug 3 2018, 11:13 PM
sjakobi abandoned D4996: Update haddock to state of upstream ghc-head branch.
Aug 3 2018, 11:11 PM

Jul 20 2018

sjakobi created D4996: Update haddock to state of upstream ghc-head branch.
Jul 20 2018, 9:34 AM
sjakobi created D4995: Update haddock to state of upstream ghc-8.6 branch.
Jul 20 2018, 9:30 AM

Jul 18 2018

sjakobi added a comment to D4913: Make :doc work for the ghc library.

It would be interesting to know how much of an effect this has on interface file sizes. I doubt it's much , but it would be a good measurement to do. I'll try to characterise this as I merge.

Jul 18 2018, 4:42 PM

Jul 17 2018

sjakobi abandoned D4973: Add functions for splitting sigma types while preserving type synonyms.

Any reason you can't put these functions in Haddock itself?

Jul 17 2018, 3:32 AM

Jul 16 2018

sjakobi created D4973: Add functions for splitting sigma types while preserving type synonyms.
Jul 16 2018, 6:30 PM

Jul 7 2018

sjakobi updated the diff for D4924: Add a strict version of foldMap to Foldable.
  • Bump version in @since annotation
Jul 7 2018, 1:25 AM

Jul 4 2018

sjakobi updated the diff for D4924: Add a strict version of foldMap to Foldable.

Rebase

Jul 4 2018, 11:12 AM
sjakobi planned changes to D4924: Add a strict version of foldMap to Foldable.

Oops. Looks like I lost my first commit that in fact added the foldMap' method.

Jul 4 2018, 11:05 AM

Jul 3 2018

sjakobi updated the diff for D4924: Add a strict version of foldMap to Foldable.
  • Use (<>) instead of mappend
Jul 3 2018, 5:45 AM
sjakobi updated the diff for D4924: Add a strict version of foldMap to Foldable.
  • Remove bang pattern – we're using foldl'!
Jul 3 2018, 5:13 AM

Jul 2 2018

sjakobi created D4924: Add a strict version of foldMap to Foldable.
Jul 2 2018, 7:51 PM
sjakobi added a comment to D4913: Make :doc work for the ghc library.

Sorry, I wasn't clear. This change should be useful on both master and ghc-8.6.

Jul 2 2018, 12:25 PM
sjakobi accepted D4847: UniqFM: Use foldl' instead of fold.

LGTM. Using foldl' instead of foldl seems less confusing/surprising to me.

Jul 2 2018, 10:53 AM

Jul 1 2018

sjakobi added a comment to D4913: Make :doc work for the ghc library.

I've realized that I've forked this branch off master although I want this change to go onto the ghc-8.6 branch. Do I need to change something?

Jul 1 2018, 7:25 AM

Jun 30 2018

sjakobi created D4913: Make :doc work for the ghc library.
Jun 30 2018, 7:09 PM

Jun 17 2018

sjakobi accepted D4848: Add flag to show docs of valid hole fits.

LGTM.

Jun 17 2018, 6:18 PM
sjakobi requested changes to D4848: Add flag to show docs of valid hole fits.
Jun 17 2018, 5:29 PM
sjakobi updated the diff for D4852: --show-iface: Qualify all non-local names.
  • Add Note [Name qualification with --show-iface]
Jun 17 2018, 9:05 AM

Jun 15 2018

sjakobi accepted D4736: Change to improve Eq, Ord, Float and Double documentation (trac issue #15078)..

Thanks! :)

Jun 15 2018, 1:06 PM
sjakobi created D4852: --show-iface: Qualify all non-local names.
Jun 15 2018, 12:17 PM
sjakobi added inline comments to D4848: Add flag to show docs of valid hole fits.
Jun 15 2018, 10:17 AM
sjakobi added a comment to D4848: Add flag to show docs of valid hole fits.

I appreciate the changes to getDocs but I have some qualms about the import of InteractiveEval from typecheck/.

Jun 15 2018, 9:36 AM
sjakobi added a comment to D4848: Add flag to show docs of valid hole fits.

An alternative method to get the documentation, would be to use extractDocs on the existing TcGblEnv.

Jun 15 2018, 7:06 AM
sjakobi added a comment to D4848: Add flag to show docs of valid hole fits.

@sjakobi @alexbiehl do you know what's going on with the test errors? They all seem to be variants of

Bad interface file: T11274.hi
    T11274.hi: openBinaryFile: does not exist (No such file or directory)

or

Compile failed (exit code 1) errors were:
[1 of 2] Compiling ValidHoleFits    ( ValidHoleFits.hs, ValidHoleFits.o )
[2 of 2] Compiling Foo              ( valid_hole_fits.hs, valid_hole_fits.o )
attempting to use module ‘main:Foo’ (valid_hole_fits.hs) which is not loaded

Neither of which seem to be related to how I'm using the getDocs function.

Jun 15 2018, 6:08 AM

Jun 14 2018

sjakobi created D4847: UniqFM: Use foldl' instead of fold.
Jun 14 2018, 10:17 AM

Jun 13 2018

sjakobi abandoned D4806: --show-iface: Always qualify names.

I have created Trac #15269.

Jun 13 2018, 7:15 AM

Jun 8 2018

sjakobi updated the diff for D4816: Fix deserialization of docs (#15240).
  • Add reference to stableNameCmp
Jun 8 2018, 5:04 PM
sjakobi updated the diff for D4816: Fix deserialization of docs (#15240).
  • Testsuite: Accept corrected output for DocsInHiFile1
Jun 8 2018, 4:59 PM
sjakobi created D4816: Fix deserialization of docs (#15240).
Jun 8 2018, 8:45 AM

Jun 7 2018

sjakobi added a comment to D4806: --show-iface: Always qualify names.

I'm not sure I follow why this is an improvement. It seems like this is only introducing noise. Perhaps I'm missing some context?

Jun 7 2018, 9:49 AM
sjakobi created D4806: --show-iface: Always qualify names.
Jun 7 2018, 7:41 AM

Jun 5 2018

sjakobi retitled D4749: [WIP] Lex and rename docstrings from [DON'T MERGE] Lex docstrings and include them in .hi-files to [WIP] Lex and rename docstrings.
Jun 5 2018, 10:39 AM
sjakobi updated the diff for D4749: [WIP] Lex and rename docstrings.

Rebase on master

Jun 5 2018, 10:22 AM

Jun 4 2018

sjakobi requested changes to D4736: Change to improve Eq, Ord, Float and Double documentation (trac issue #15078)..
Jun 4 2018, 4:44 PM

Jun 3 2018

sjakobi added a comment to D4758: Serialize docstrings to ifaces, display them with new GHCi :doc command.

Maybe we could output a note that :doc is a tech preview and that ANSI codes and highlighting might come in later iterations.

Jun 3 2018, 10:48 AM
sjakobi updated the diff for D4758: Serialize docstrings to ifaces, display them with new GHCi :doc command.
  • HsDoc: Fix complexity of concatDocs
  • getDocs: Add TODO
  • Expand comment on why we keep the renamed syntax with -haddock
Jun 3 2018, 10:27 AM
sjakobi accepted D4736: Change to improve Eq, Ord, Float and Double documentation (trac issue #15078)..
Jun 3 2018, 3:04 AM
sjakobi added a comment to D4736: Change to improve Eq, Ord, Float and Double documentation (trac issue #15078)..

This is the first time I hear about exponential fields and division rings though, so it would be good if someone more knowledgeable about algebra could review this too.

Jun 3 2018, 3:04 AM

Jun 1 2018

sjakobi updated the diff for D4758: Serialize docstrings to ifaces, display them with new GHCi :doc command.
  • Fix concatenation of docstrings
Jun 1 2018, 9:04 AM

May 31 2018

sjakobi updated the diff for D4758: Serialize docstrings to ifaces, display them with new GHCi :doc command.
  • mk/config.mk.in: Add -haddock to GhcLibHcOpts
May 31 2018, 10:34 AM
sjakobi added a comment to D4758: Serialize docstrings to ifaces, display them with new GHCi :doc command.

Here is a little demonstration of the :doc command:

May 31 2018, 9:29 AM
sjakobi updated the diff for D4758: Serialize docstrings to ifaces, display them with new GHCi :doc command.
  • Remove outdated TODO
May 31 2018, 8:37 AM
sjakobi created D4758: Serialize docstrings to ifaces, display them with new GHCi :doc command.
May 31 2018, 8:37 AM

May 30 2018

sjakobi requested changes to D4736: Change to improve Eq, Ord, Float and Double documentation (trac issue #15078)..
May 30 2018, 7:15 PM
sjakobi added a comment to D4736: Change to improve Eq, Ord, Float and Double documentation (trac issue #15078)..

Thanks for your changes! I really like how this is turning out.

May 30 2018, 7:15 PM

May 29 2018

sjakobi added a comment to D4749: [WIP] Lex and rename docstrings.

BTW the build fails only because haddock hasn't been adapted yet.

May 29 2018, 1:18 PM
sjakobi added inline comments to D4749: [WIP] Lex and rename docstrings.
May 29 2018, 1:17 PM
sjakobi updated the diff for D4749: [WIP] Lex and rename docstrings.
  • Fix build with ghc-8.2
  • Fix line length issues
May 29 2018, 9:53 AM
sjakobi retitled D4749: [WIP] Lex and rename docstrings from [DON' MERGE] Lex docstrings and include them in .hi-files to [DON'T MERGE] Lex docstrings and include them in .hi-files.
May 29 2018, 7:34 AM
sjakobi created D4749: [WIP] Lex and rename docstrings.
May 29 2018, 7:34 AM

May 28 2018

sjakobi requested changes to D4736: Change to improve Eq, Ord, Float and Double documentation (trac issue #15078)..
May 28 2018, 4:09 PM

May 27 2018

sjakobi added a comment to D4743: Make HsDocString a newtype of ByteString.

Not sure what caused the timeout of T13410 in https://phabricator.haskell.org/harbormaster/build/46363/…

May 27 2018, 11:31 PM
sjakobi created D4743: Make HsDocString a newtype of ByteString.
May 27 2018, 9:03 PM

May 25 2018

sjakobi added a comment to D4736: Change to improve Eq, Ord, Float and Double documentation (trac issue #15078)..

Thanks for working on this! :)

May 25 2018, 1:41 PM

May 11 2018

sjakobi added a comment to D4679: RTS: Auto-size the allocation area depending on CPU cache size.

I don't see how perf.haskell.org will help you here. It just counts
instructions, so it's not going to tell you anything about the cache
effects you're after.

May 11 2018, 4:36 PM
sjakobi added a comment to D4679: RTS: Auto-size the allocation area depending on CPU cache size.

Thanks for the insightful comments! I agree that this method is very likely too naive to result in good performance in the majority of scenarios.

May 11 2018, 12:53 PM
sjakobi accepted D4687: Add note documenting refineDefaultAlt.

No need to address my comment. My reading of refineDefaultAlt was strongly influenced by my understanding of combineIdenticalAlts.

May 11 2018, 12:34 PM
sjakobi added a comment to D4687: Add note documenting refineDefaultAlt.

Thanks for writing this up, Matt! Just one minor point.

May 11 2018, 12:10 PM

May 9 2018

Herald added a reviewer for D4679: RTS: Auto-size the allocation area depending on CPU cache size: erikd.
May 9 2018, 3:40 AM

Apr 20 2018

sjakobi abandoned D4542: Add -fcombine-most-common-alts optimization.

Could you add an entry to the release notes for this and address @simonpj's remarks on the comment, @sjakobi?

Apr 20 2018, 3:45 PM

Apr 11 2018

sjakobi added a comment to D4582: In Exitify, zap idInfo of abstracted variables (fixes #15005).

Great! Thanks for checking!

Apr 11 2018, 11:23 AM

Apr 10 2018

sjakobi added a comment to D4582: In Exitify, zap idInfo of abstracted variables (fixes #15005).

Does this possibly also fix the order-maintenance testsuite that I mentioned in https://ghc.haskell.org/trac/ghc/ticket/15005#comment:4?

Unclear. Can you check?

Apr 10 2018, 2:02 PM

Mar 30 2018

sjakobi requested review of D4542: Add -fcombine-most-common-alts optimization.
Mar 30 2018, 9:30 AM

Feb 25 2018

sjakobi created D4450: rts/win32: Assert that the IO manager has been initialised.
Feb 25 2018, 4:33 PM

Feb 22 2018

sjakobi updated the diff for D4441: Correct default -A value in RTS flag usage info.
  • Update example value
Feb 22 2018, 6:34 PM
sjakobi created D4441: Correct default -A value in RTS flag usage info.
Feb 22 2018, 6:31 PM

Feb 21 2018

sjakobi created D4435: doCorePass: Expand catch-all.
Feb 21 2018, 4:20 PM
sjakobi created D4433: Implement first exemplary primop for #11143.
Feb 21 2018, 2:17 PM

Feb 20 2018

sjakobi added inline comments to D4428: RTS: Remember to free some pointers.
Feb 20 2018, 11:43 AM
sjakobi added inline comments to D4428: RTS: Remember to free some pointers.
Feb 20 2018, 8:52 AM

Feb 19 2018

sjakobi added inline comments to D4428: RTS: Remember to free some pointers.
Feb 19 2018, 11:36 PM
sjakobi updated the diff for D4427: RTS: Remove unused retainer schemes.
  • Delete redundant comment
Feb 19 2018, 9:22 PM
sjakobi created D4428: RTS: Remember to free some pointers.
Feb 19 2018, 8:36 PM
sjakobi created D4427: RTS: Remove unused retainer schemes.
Feb 19 2018, 8:36 PM
sjakobi created D4426: driver/utils/dynwrapper.c: Remove unused variable.
Feb 19 2018, 8:36 PM
sjakobi added a comment to D4419: Combine the CoreAlts with the most common RHS.

In this build I just so break the upper max_bytes_used limit for T1969: 15.9% on a 15% limit.

Feb 19 2018, 6:53 PM
sjakobi updated the diff for D4419: Combine the CoreAlts with the most common RHS.
  • Update Note [Combine identical alternatives]
Feb 19 2018, 2:27 PM
sjakobi updated the diff for D4419: Combine the CoreAlts with the most common RHS.
  • Test: Remove redundant import
Feb 19 2018, 1:22 PM
sjakobi updated the diff for D4419: Combine the CoreAlts with the most common RHS.
  • Test: Also check for remaining alternatives
Feb 19 2018, 1:19 PM
sjakobi added inline comments to D4419: Combine the CoreAlts with the most common RHS.
Feb 19 2018, 1:03 PM

Feb 18 2018

sjakobi added inline comments to D4419: Combine the CoreAlts with the most common RHS.
Feb 18 2018, 9:06 PM
sjakobi updated the diff for D4419: Combine the CoreAlts with the most common RHS.
  • Check only that the combined DEFAULT alt has the right RHS
Feb 18 2018, 9:05 PM
sjakobi added inline comments to D4419: Combine the CoreAlts with the most common RHS.
Feb 18 2018, 4:26 PM
sjakobi added a comment to D4419: Combine the CoreAlts with the most common RHS.

This is just a first version that assumedly is too slow. I'll try to do some performance testing to check.

Feb 18 2018, 1:40 PM
sjakobi created D4419: Combine the CoreAlts with the most common RHS.
Feb 18 2018, 1:21 PM