osa1 (Ömer Sinan Ağacan)
User

Projects

User does not belong to any projects.

User Details

User Since
Jun 6 2014, 2:22 AM (232 w, 2 d)

Recent Activity

Today

osa1 added a comment to D5324: Don't track free variables in STG syntax by default.

@sgraf Could you add the note @simonpj asked? I'm hoping to land this.

Mon, Nov 19, 12:01 AM

Yesterday

osa1 accepted D5351: Remove explicit recursion in retainer profiling (fixes #14758).
In D5351#147266, @osa1 wrote:

I just remembered the full command I was using as a test for this bug: ghc-stage2 --interactive +RTS -hr <<(echo 'sequence_ (replicate 1000000 (return ()))')

Thanks! will report once do that.

Sun, Nov 18, 11:28 PM
osa1 updated the summary of D5342: Fix heap corruption during stable name allocation.
Sun, Nov 18, 8:07 AM
osa1 added a comment to D5351: Remove explicit recursion in retainer profiling (fixes #14758).
In D5351#147264, @osa1 wrote:

Thanks! One thing that reliably triggered this problem in the past (at least on my system) was running profiled GHCi (build profiled GHC using the prof flavor, then run ghc-stage2 --interactive +RTS -hr). Could you give it a try without this patch to see if it still fails, and then try with this patch?

I've tried exactly that, but the program that I have used didn't reliably fail before the patch. Though it has always failed when run under rr.
With the patch it has never failed neither directly nor when run under rr. I have not added the program in the testsuite because it depends on
timeout and not always pinpoints the problem.

Sun, Nov 18, 8:05 AM
osa1 added a comment to D5351: Remove explicit recursion in retainer profiling (fixes #14758).

I just remembered the full command I was using as a test for this bug: ghc-stage2 --interactive +RTS -hr <<(echo 'sequence_ (replicate 1000000 (return ()))')

Sun, Nov 18, 7:54 AM
osa1 added a comment to D5351: Remove explicit recursion in retainer profiling (fixes #14758).

Thanks! One thing that reliably triggered this problem in the past (at least on my system) was running profiled GHCi (build profiled GHC using the prof flavor, then run ghc-stage2 --interactive +RTS -hr). Could you give it a try without this patch to see if it still fails, and then try with this patch?

Sun, Nov 18, 7:48 AM
osa1 committed rGHCecfc7b466567: Fix typo in docs (authored by chrisdone).
Fix typo in docs
Sun, Nov 18, 6:01 AM
osa1 added a comment to D5342: Fix heap corruption during stable name allocation.

allocate() looks a bit hacky to me and I was under the impression that it should be the last resort (when calling GC will cause too many problems), but perhaps it's OK in this context. Updated the diff. This validates and the test passes with -debug.

Sun, Nov 18, 5:57 AM
osa1 updated the diff for D5342: Fix heap corruption during stable name allocation.
  • Use allocate()
Sun, Nov 18, 5:55 AM

Fri, Nov 16

osa1 committed rGHC6ba9aa5dd0a5: Fix a typo in the description of -fabstract-refinement-hole-fits (authored by ethercrow).
Fix a typo in the description of -fabstract-refinement-hole-fits
Fri, Nov 16, 11:40 PM
osa1 committed rGHC9696282911a5: user's guide: typo in ViewPatterns example (authored by Ben Price <ben@brprice.uk>).
user's guide: typo in ViewPatterns example
Fri, Nov 16, 11:38 PM
osa1 committed rGHC6677235181ea: Some assertions and comments in scheduler (authored by osa1).
Some assertions and comments in scheduler
Fri, Nov 16, 11:37 PM
osa1 closed D5337: Some assertions and comments in scheduler.
Fri, Nov 16, 11:37 PM
osa1 added a comment to D5337: Some assertions and comments in scheduler.

Looking at the test suite output with -debug, there are lots of failures but most of them due to debug outputs. There's one actual failure, memo001, which is tracked in Trac #15906. Landing.

Fri, Nov 16, 11:35 PM
osa1 accepted D5315: More compact Outputable instance for `Uniq(D)Set`.

@dfeuer any objections to merging this?

Fri, Nov 16, 6:19 AM
osa1 added a comment to D5005: Remove completed threads from the gen->threads list eagerly.

Ping @simonmar .

Fri, Nov 16, 6:11 AM
osa1 updated the test plan for D5342: Fix heap corruption during stable name allocation.
Fri, Nov 16, 6:05 AM
osa1 created D5342: Fix heap corruption during stable name allocation.
Fri, Nov 16, 5:59 AM
osa1 committed rGHCa50a59a96034: More efficient, non-allocating unsafeLookupStaticPtr (authored by osa1).
More efficient, non-allocating unsafeLookupStaticPtr
Fri, Nov 16, 4:52 AM
osa1 closed D5333: More efficient, non-allocating unsafeLookupStaticPtr.
Fri, Nov 16, 4:52 AM
osa1 added a comment to D5341: bump haddock submodule.

Yeah let's test and land this. Unless someone beats me to it I'll validate in an hour and then land.

Fri, Nov 16, 2:47 AM

Thu, Nov 15

osa1 updated the Trac tickets for D5341: bump haddock submodule.
Thu, Nov 15, 10:47 PM
osa1 accepted D5324: Don't track free variables in STG syntax by default.

This looks good to me, thanks! Let's wait to hear from @simonpj before landing.

Thu, Nov 15, 11:17 AM
osa1 accepted D5339: Remove an ancient occurrence analysis hack from StgFVs.
Thu, Nov 15, 11:04 AM
osa1 added a comment to D5324: Don't track free variables in STG syntax by default.

and counted instructions

Thu, Nov 15, 11:04 AM
osa1 added a comment to D5324: Don't track free variables in STG syntax by default.

Omitting the hack had no effect on nofib allocations (and counted instructions) whatsoever. Remove it?

Thu, Nov 15, 10:57 AM
osa1 added a comment to D5324: Don't track free variables in STG syntax by default.

moving the occurrence hack from CoreToStg to CoreFVs means that all STG-to-STG transformations don't have the generated dead case binder info available

Thu, Nov 15, 9:30 AM

Wed, Nov 14

osa1 added inline comments to D5337: Some assertions and comments in scheduler.
Wed, Nov 14, 11:08 PM
osa1 updated the test plan for D5337: Some assertions and comments in scheduler.
Wed, Nov 14, 11:04 PM
osa1 created D5337: Some assertions and comments in scheduler.
Wed, Nov 14, 11:03 PM
osa1 committed rGHCeb46345d37ee: Fix a bug in SRT generation (#15892) (authored by simonmar).
Fix a bug in SRT generation (#15892)
Wed, Nov 14, 9:36 PM
osa1 closed D5334: Fix a bug in SRT generation (#15892).
Wed, Nov 14, 9:36 PM
osa1 accepted D5334: Fix a bug in SRT generation (#15892).

Confirmed locally that this validates. Also tried the original (larger) reproducer which this fixes.

Wed, Nov 14, 9:08 AM

Tue, Nov 13

osa1 added a comment to D5333: More efficient, non-allocating unsafeLookupStaticPtr.

CI is completely broken.. this validates on my desktop.

Tue, Nov 13, 12:27 AM
osa1 created D5333: More efficient, non-allocating unsafeLookupStaticPtr.
Tue, Nov 13, 12:19 AM

Sun, Nov 11

osa1 added a comment to D5315: More compact Outputable instance for `Uniq(D)Set`.

Could you add an example of how this was printed previously and how it's printed now, in the summary?

Sun, Nov 11, 11:16 PM
osa1 accepted D5326: Fix #12525: Remove derived bindings from the TyThings from getBindings.

Thanks!

Sun, Nov 11, 11:03 PM
osa1 added a comment to D5324: Don't track free variables in STG syntax by default.

Thanks! I didn't review it line by line but the idea sounds good to me.

Sun, Nov 11, 10:58 PM
osa1 committed rGHCd30352add1da: Remove StgBinderInfo and related computation in CoreToStg (authored by osa1).
Remove StgBinderInfo and related computation in CoreToStg
Sun, Nov 11, 10:17 PM
osa1 closed D5232: Remove StgBinderInfo and related computation in CoreToStg.
Sun, Nov 11, 10:17 PM
osa1 accepted D5318: Calling gcc: Pass optc flags as last options (#14452).

Thanks!

Sun, Nov 11, 10:12 PM
osa1 added a comment to D5232: Remove StgBinderInfo and related computation in CoreToStg.

Sorry everyone, I was on vacation last week and didn't have access to a computer for 8 days! I now rebased the branch and will push shortly after a validate.

Sun, Nov 11, 9:50 PM

Thu, Oct 25

osa1 added a comment to D5221: Lower precedence for {-# UNPACK #-}.

I realized something that we should've perhaps discussed in the proposal (btw, could you link to the proposal in the description? The comments will get lost but the description will be the commit message).

Thu, Oct 25, 2:29 AM
osa1 committed rGHC49c7a51ef614: Fix some broken links (#15733) (authored by fangyizhou).
Fix some broken links (#15733)
Thu, Oct 25, 2:20 AM
osa1 closed D5257: Fix some broken links (#15733).
Thu, Oct 25, 2:20 AM
osa1 accepted D5262: Fix some broken links (#15733).
Thu, Oct 25, 2:07 AM
osa1 accepted D5260: Fix some broken links (#15733).
Thu, Oct 25, 2:06 AM

Mon, Oct 22

osa1 updated subscribers of D5232: Remove StgBinderInfo and related computation in CoreToStg.
Mon, Oct 22, 2:41 AM
osa1 updated the diff for D5232: Remove StgBinderInfo and related computation in CoreToStg.
  • Revert removed dead case bndr update
Mon, Oct 22, 2:40 AM
osa1 added a reviewer for D5249: [#14037] Allow fusion for zip7 and related: dfeuer.

I found it surprising that T9203 allocates -53% now, as it doesn't even use lists. Do you know why it's effected by these changes?

Mon, Oct 22, 1:24 AM

Oct 19 2018

osa1 added inline comments to D5232: Remove StgBinderInfo and related computation in CoreToStg.
Oct 19 2018, 1:14 PM
osa1 updated the summary of D5232: Remove StgBinderInfo and related computation in CoreToStg.
Oct 19 2018, 2:55 AM
osa1 updated the diff for D5232: Remove StgBinderInfo and related computation in CoreToStg.
  • Rebase
  • Remove more comments
Oct 19 2018, 2:54 AM

Oct 18 2018

osa1 added a comment to D5232: Remove StgBinderInfo and related computation in CoreToStg.

I found some relevant comments/code about this optimisation, and filed a ticket: Trac #15770

Oct 18 2018, 3:24 AM

Oct 17 2018

osa1 updated the diff for D5232: Remove StgBinderInfo and related computation in CoreToStg.
  • Remove more comments
Oct 17 2018, 6:23 AM
osa1 created D5232: Remove StgBinderInfo and related computation in CoreToStg.
Oct 17 2018, 6:21 AM

Oct 12 2018

osa1 created D5225: Fix cardinality change of fields in addDataConStrictness.
Oct 12 2018, 2:54 PM
osa1 committed rGHCfc4c3df046d7: Fix dataToTag# comment syntax again (primops.txt.pp) (authored by osa1).
Fix dataToTag# comment syntax again (primops.txt.pp)
Oct 12 2018, 1:17 PM
osa1 added a comment to D5223: WIP: Support profiling in ghc-heap.

What does it show if you print return value of getClosureData con? I'm asking because you don't know in this code that con points to a constructor without any indirections. For example, GHC can float con to the top-level and make it a CAF, in which case con might point to a IND_STATIC, not to a CONSTR_1_0.

Oct 12 2018, 7:13 AM

Oct 10 2018

osa1 committed rGHC2f693b3e496b: Re-enable test T14251 (authored by osa1).
Re-enable test T14251
Oct 10 2018, 1:58 PM
osa1 committed rGHCac977688523e: Fix dataToTag# argument evaluation (authored by osa1).
Fix dataToTag# argument evaluation
Oct 10 2018, 2:33 AM
osa1 closed D5201: Fix dataToTag# argument evaluation.
Oct 10 2018, 2:33 AM
osa1 updated the summary of D5201: Fix dataToTag# argument evaluation.
Oct 10 2018, 1:29 AM

Oct 9 2018

osa1 added a comment to D5218: Meaning-preserving parsing of SCC/GENERATED/CORE annotations.

Sorry if I misunderstood what this patch is doing -- I tried to build it and test it but it doesn't build. In the future it'd be helpful if you could say more about the patch in the summary.

Oct 9 2018, 3:39 PM
osa1 requested changes to D5218: Meaning-preserving parsing of SCC/GENERATED/CORE annotations.

I don't think we can do this, for several reasons. First, this is a breaking change of a user-facing feature so even if it makes sense it'd need to go through the GHC proposal process with a migration plan etc. Secondly, I think it makes sense that annotations have lower precedence than everything else, and I know a lot of code are currently relying on that. If you grep in GHC for SCC you'll find code like:

Oct 9 2018, 3:07 PM
osa1 added inline comments to D5201: Fix dataToTag# argument evaluation.
Oct 9 2018, 2:35 AM

Oct 8 2018

osa1 added inline comments to D5201: Fix dataToTag# argument evaluation.
Oct 8 2018, 1:05 PM
osa1 updated the diff for D5201: Fix dataToTag# argument evaluation.
  • Remove bogus comments in CorePrep
Oct 8 2018, 1:04 PM
osa1 added a comment to D5210: Add pprList method to Outputable for better String rendering.

It just means that every Outputable dictionary is now a boxed pair rather than a simple function. Maybe that doesn't matter.

Oct 8 2018, 8:33 AM
osa1 added inline comments to D5201: Fix dataToTag# argument evaluation.
Oct 8 2018, 8:06 AM
osa1 updated the diff for D5201: Fix dataToTag# argument evaluation.
  • Update Note [dataToTag#]
  • Add a reference to Note [dataToTag#]
Oct 8 2018, 7:44 AM
osa1 added inline comments to D5201: Fix dataToTag# argument evaluation.
Oct 8 2018, 7:04 AM
osa1 added a comment to D5210: Add pprList method to Outputable for better String rendering.

The use case is when I have a String in a larger type, e.g. Maybe [String]. Now I want to pprDebug this, but by default it gives me bad output because string is rendered as a list of characters. I could try fmap (map text) x but that would not work because we don't have Outputable instance for SDoc (that is D2761). For the same reason text "..." also doesn't work (you may get an unrenderable type, e.g. Just [text "..."] is not renderable).

Oct 8 2018, 2:58 AM

Oct 7 2018

osa1 added a comment to D5201: Fix dataToTag# argument evaluation.

Was that with the libraries recompiled too?

Oct 7 2018, 3:06 PM
osa1 updated the diff for D5201: Fix dataToTag# argument evaluation.
  • Rebase
Oct 7 2018, 6:46 AM
osa1 added a comment to D5201: Fix dataToTag# argument evaluation.

Nofib results:

Oct 7 2018, 6:45 AM
osa1 updated the diff for D5201: Fix dataToTag# argument evaluation.
  • Update note again
  • Fix dataToTag# code gen
Oct 7 2018, 3:33 AM
osa1 added a comment to D5201: Fix dataToTag# argument evaluation.

I'd be really interested to see what effect this has on performance and code size too.

Oct 7 2018, 3:20 AM

Oct 6 2018

osa1 retitled D5210: Add pprList method to Outputable for better String rendering from RFC: Add pprList method to Outputable for better String rendering to Add pprList method to Outputable for better String rendering.
Oct 6 2018, 11:39 AM
osa1 created D5210: Add pprList method to Outputable for better String rendering.
Oct 6 2018, 11:39 AM
osa1 committed rGHC8be27c0371d5: Add a missing write barrier to small array writes (authored by osa1).
Add a missing write barrier to small array writes
Oct 6 2018, 9:55 AM
osa1 closed D5209: Add a missing write barrier to small array writes.
Oct 6 2018, 9:55 AM
osa1 updated the summary of D5209: Add a missing write barrier to small array writes.
Oct 6 2018, 5:54 AM
osa1 created D5209: Add a missing write barrier to small array writes.
Oct 6 2018, 5:53 AM

Oct 5 2018

osa1 updated the diff for D5201: Fix dataToTag# argument evaluation.
  • Remove more comments
  • Revert can_fail, put getTag argument back
Oct 5 2018, 11:04 AM
osa1 added a reviewer for D5201: Fix dataToTag# argument evaluation: dfeuer.
Oct 5 2018, 1:53 AM
osa1 retitled D5201: Fix dataToTag# argument evaluation from Simplify dataToTag# and getTag handling to Fix dataToTag# argument evaluation.
Oct 5 2018, 1:41 AM
osa1 updated the diff for D5201: Fix dataToTag# argument evaluation.
  • Reword
Oct 5 2018, 1:41 AM
osa1 updated the diff for D5201: Fix dataToTag# argument evaluation.
  • Update
Oct 5 2018, 1:18 AM

Oct 4 2018

osa1 added inline comments to D5201: Fix dataToTag# argument evaluation.
Oct 4 2018, 5:30 AM
osa1 added inline comments to D5201: Fix dataToTag# argument evaluation.
Oct 4 2018, 5:19 AM
osa1 created D5201: Fix dataToTag# argument evaluation.
Oct 4 2018, 2:46 AM

Oct 3 2018

osa1 added a comment to D2761: added "instance Outputable SDoc".

adding it would mean that the codebase may start growing superfluous ppr uses and we would have no way to catch them

Oct 3 2018, 10:59 AM

Oct 2 2018

osa1 updated the diff for D5196: Fix dataToTag# argument evaluation.
  • Update comments
Oct 2 2018, 3:24 PM
osa1 added inline comments to D5196: Fix dataToTag# argument evaluation.
Oct 2 2018, 2:55 PM
osa1 updated the diff for D5196: Fix dataToTag# argument evaluation.
  • Forgot to push stdout files
Oct 2 2018, 2:55 PM
osa1 added inline comments to D5196: Fix dataToTag# argument evaluation.
Oct 2 2018, 2:37 PM
osa1 updated the diff for D5196: Fix dataToTag# argument evaluation.
  • Whitespace, typo
Oct 2 2018, 2:09 PM
osa1 created D5196: Fix dataToTag# argument evaluation.
Oct 2 2018, 2:03 PM
osa1 added a comment to D4717: [WIP] CoreToStg: Try treating String unpackings as single-entry.

Could you say a few more words on reverting a CAF? An evaluated CAF becomes an IND_STATIC, right? Do you mean we could somehow save the old info ptr and somehow scavenge the CAF using the old info ptr? (rather than scavenging it as IND_STATIC?)

Oct 2 2018, 1:54 AM