Teo Camarasu activity https://gitlab.haskell.org/teo 2026-03-17T16:19:32Z tag:gitlab.haskell.org,2026-03-17:622703 Teo Camarasu commented on issue #26707 at Glasgow Haskell Compiler / GHC 2026-03-17T16:19:32Z teo Teo Camarasu

!15731

tag:gitlab.haskell.org,2026-03-17:622702 Teo Camarasu opened merge request !15731: Revert "Set default eventlog-flush-interval to 5s" at Glasgow Haskell Compiler / GHC 2026-03-17T16:19:19Z teo Teo Camarasu

See #26707 (comment 665037)

tag:gitlab.haskell.org,2026-03-17:622701 Teo Camarasu pushed new project branch wip/T26707-revert at Glasgow Haskell Compiler / GHC 2026-03-17T16:18:44Z teo Teo Camarasu

Teo Camarasu (bf184c31) at 17 Mar 16:18

Revert "Set default eventlog-flush-interval to 5s"

tag:gitlab.haskell.org,2026-03-17:622700 Teo Camarasu opened issue #26707: Enable eventlog-flush-interval by default when using the eventlog at Glasgow Haskell Compiler / GHC 2026-03-17T16:13:32Z teo Teo Camarasu tag:gitlab.haskell.org,2026-03-17:622699 Teo Camarasu commented on issue #26707 at Glasgow Haskell Compiler / GHC 2026-03-17T16:13:30Z teo Teo Camarasu

Reverting, then figuring out how to implement this without synchronisation and then re-applying this seems sensible to me. So, I'll make an MR

tag:gitlab.haskell.org,2026-03-17:622685 Teo Camarasu commented on issue #26707 at Glasgow Haskell Compiler / GHC 2026-03-17T16:00:42Z teo Teo Camarasu

I wasn't aware of there being a high cost to doing this. I'm happy to revert this if you think the previous default made more sense

tag:gitlab.haskell.org,2026-03-17:622684 Teo Camarasu commented on issue #26707 at Glasgow Haskell Compiler / GHC 2026-03-17T15:59:48Z teo Teo Camarasu

Good point! Created an issue: #27056

tag:gitlab.haskell.org,2026-03-17:622682 Teo Camarasu opened issue #27056: Document that eventlog-flush-interval can be disabled by setting it to 0 at Glasgow Haskell Compiler / GHC 2026-03-17T15:59:13Z teo Teo Camarasu tag:gitlab.haskell.org,2026-03-17:622624 Teo Camarasu opened issue #27050: Add safe variants of methods to Q's interface at Glasgow Haskell Compiler / GHC 2026-03-17T14:18:56Z teo Teo Camarasu tag:gitlab.haskell.org,2026-03-17:622623 Teo Camarasu approved merge request !440: Draft: Add test for ghc/ghc#14565 at Glasgow Haskell Compiler / head.hackage 2026-03-17T14:06:57Z teo Teo Camarasu

The test is supposed to check that GHC can compile the T14565 module without running out of memory.

Closes ghc#14565.

tag:gitlab.haskell.org,2026-03-17:622615 Teo Camarasu commented on issue #21738 at Glasgow Haskell Compiler / GHC 2026-03-17T13:37:38Z teo Teo Camarasu

This should be trivial to fix now. I'll tackle it. template-haskell is now just a regular package, and I have made it so that filepath no longer depends on template-haskell so we aren't risking loops anymore

tag:gitlab.haskell.org,2026-03-17:622559 Teo Camarasu commented on issue #27022 at Glasgow Haskell Compiler / GHC 2026-03-17T10:30:28Z teo Teo Camarasu

Note that there are other forms of global state such as adding top level declarations (which I'm pretty sure) don't roll back with either implementation.

tag:gitlab.haskell.org,2026-03-16:622432 Teo Camarasu pushed to project branch wip/hugepages at Glasgow Haskell Compiler / GHC 2026-03-16T10:27:20Z teo Teo Camarasu

Teo Camarasu (bd0444c1) at 16 Mar 10:27

rts: Implement support for 2MB hugepages

... and 41 more commits

tag:gitlab.haskell.org,2026-03-16:622415 Teo Camarasu commented on merge request !15582 at Glasgow Haskell Compiler / GHC 2026-03-16T08:26:00Z teo Teo Camarasu

Looks like T17912 which is marked as expected fail on windows is now passing? Maybe a retry will resolve it?

tag:gitlab.haskell.org,2026-03-15:622376 Teo Camarasu commented on merge request !15582 at Glasgow Haskell Compiler / GHC 2026-03-15T20:14:58Z teo Teo Camarasu

@mniip looks like this needs one last rebase (Sorry! it's my fault cause I merged something that also touched one of these files)

tag:gitlab.haskell.org,2026-03-14:622242 Teo Camarasu approved merge request !15709: Use explicit syntax rather than pure at Glasgow Haskell Compiler / GHC 2026-03-14T12:20:22Z teo Teo Camarasu

This is part of the approach to upstream independent bits from !15231 . The originating branch just grew too big focus on the all details.

I've chosen the reviewers of the big MR because they may already know this. However, anybody should be able to review this.


Using explicit list syntax [ ... ] makes the code a bit more readable, because one doesn't have to reason about concrete type of the pure call.

tag:gitlab.haskell.org,2026-03-14:622186 Teo Camarasu pushed to project branch wip/abstract-q at Glasgow Haskell Compiler / GHC 2026-03-14T01:10:05Z teo Teo Camarasu

Teo Camarasu (3fd2d50d) at 14 Mar 01:10

wip

... and 1 more commit

tag:gitlab.haskell.org,2026-03-13:622166 Teo Camarasu approved merge request !15704: base: fix redundant imports in GHC.Internal.Weak.Finalize at Glasgow Haskell Compiler / GHC 2026-03-13T21:54:58Z teo Teo Camarasu

This patch fixes redundant imports in GHC.Internal.Weak.Finalize that causes a regression in bootstrapping head from 9.14 with validate flavours. Fixes #27026.

tag:gitlab.haskell.org,2026-03-13:622122 Teo Camarasu closed issue #26985: Move parts of GHC.Internal.Weak.Finalize interface to base at Glasgow Haskell Compiler / GHC 2026-03-13T19:11:02Z teo Teo Camarasu

The GHC.Internal.Weak.Finalize interface contains a mixture of user facing and internal functions. The module also contains some source imports. If we move the bits that are only required by base into base then perhaps we can untangle these module cycles.

tag:gitlab.haskell.org,2026-03-13:622059 Teo Camarasu opened issue #27024: Where should template-haskell internals be exposed from at Glasgow Haskell Compiler / GHC 2026-03-13T14:18:39Z teo Teo Camarasu