delta/haskell.git, branch wip/openbsd-missing-const gitlab.haskell.org: ghc/ghc.git rts: Add missing const in HEAP_ALLOCED_GC 2019-11-03T17:01:43+00:00 Ben Gamari ben@smart-cactus.org 2019-11-03T16:58:50+00:00 f8db305589abc3e4c86623451cbc61a1a42f9d3e This was previously unnoticed as this code-path is hit on very few platforms (e.g. OpenBSD). 
This was previously unnoticed as this code-path is hit on very few
platforms (e.g. OpenBSD).
Add +RTS --disable-delayed-os-memory-return. Fixes #17411. 2019-11-02T03:12:17+00:00 Niklas Hambüchen mail@nh2.me 2019-10-29T11:54:10+00:00 9980fb58f613ee3363c7e4cb86453e542c6c69aa Sets `MiscFlags.disableDelayedOsMemoryReturn`. See the added `Note [MADV_FREE and MADV_DONTNEED]` for details. 
Sets `MiscFlags.disableDelayedOsMemoryReturn`.

See the added `Note [MADV_FREE and MADV_DONTNEED]` for details.
doc: Fix backticks 2019-11-02T03:12:17+00:00 Niklas Hambüchen mail@nh2.me 2019-10-31T23:37:29+00:00 01006bc79582616c9bbc842b397e85437a57ac18 






Make CSE delay inlining less
2019-11-02T03:11:37+00:00

Simon Peyton Jones
simonpj@microsoft.com

2019-10-29T09:19:14+00:00

4a6d3d68f2ce97c1f231eede3819299a0fd2b9c8

CSE delays inlining a little bit, to avoid losing vital
specialisations; see Note [Delay inlining after CSE] in CSE.

But it was being over-enthusiastic.  This patch makes the
delay only apply to Ids with specialisation rules, which
avoids unnecessary delay (#17409).





CSE delays inlining a little bit, to avoid losing vital
specialisations; see Note [Delay inlining after CSE] in CSE.

But it was being over-enthusiastic.  This patch makes the
delay only apply to Ids with specialisation rules, which
avoids unnecessary delay (#17409).







testsuite: Add test for #17423
2019-11-01T19:50:29+00:00

Ben Gamari
ben@smart-cactus.org

2019-10-31T22:38:40+00:00

cabafe34156acb80cd9a918363b2a342fb0c8e66












hadrian: Make runtest invocation consistency with Make
2019-11-01T19:49:53+00:00

Ben Gamari
ben@smart-cactus.org

2019-10-31T17:54:58+00:00

a4ce26e0dbbd736b4106c3d30979ae0058922a06

Use True/False instead of 0/1. This shouldn't be a functional change but
we should be consistent.





Use True/False instead of 0/1. This shouldn't be a functional change but
we should be consistent.







template-haskell: require at least 1 GADT constructor name (#17379)
2019-11-01T19:49:16+00:00

Adam Sandberg Eriksson
adam@sandbergericsson.se

2019-10-31T07:28:21+00:00

c9236384e5aaa50e188f2b8549d61c0a20d1af86












Describe optimisation of demand analysis of noinline
2019-11-01T19:48:37+00:00

Ben Gamari
ben@smart-cactus.org

2019-09-25T21:26:56+00:00

dab12c8780fe1e6e3c2adb1c9565e2a43aa207db

As described in #16588.





As described in #16588.







testsuite: Make ExplicitForAllRules1 more robust
2019-11-01T19:48:37+00:00

Ben Gamari
ben@smart-cactus.org

2019-09-23T16:11:14+00:00

c751082ca0946f8686453fa2a775809fd027c9c7

Previously the test relied on `id` not inlining. Fix this.





Previously the test relied on `id` not inlining. Fix this.







base: Ensure that failIO isn't SOURCE imported
2019-11-01T19:48:37+00:00

Ben Gamari
ben@smart-cactus.org

2019-09-16T16:15:28+00:00

51067194e03849eb0c04e1dbaf829972ea64be53

failIO has useful information in its demand signature (specifically that
it bottoms) which is hidden if it is SOURCE imported, as noted
in #16588. Rejigger things such that we don't SOURCE import it.

Metric Increase:
    T13701





failIO has useful information in its demand signature (specifically that
it bottoms) which is hidden if it is SOURCE imported, as noted
in #16588. Rejigger things such that we don't SOURCE import it.

Metric Increase:
    T13701