Windows profiling: T11627b segfaults for WAY=prof_hc_hb

Reproducible with 7.10.3, 8.0.1 and today's HEAD ( afa6e830).

The test was added by @jme in #11627 (closed).

Trac metadata

Trac field	Value
Version	7.10.3
Type	Bug
TypeOfFailure	OtherFailure
Priority	normal
Resolution	Unresolved
Component	Profiling
Test case
Differential revisions
BlockedBy
Related
Blocking
CC	Phyx-, jme
Operating system
Architecture

changed weight to 5

added Tbug Trac import profiling labels

mentioned in commit 3fb9837f

I don't get a segfault, I get:

$ cat T11627b.comp.stderr
[1 of 1] Compiling Main             ( T11627b.hs, T11627b.o )

T11627b.hs:1:1: error:
    Failed to load interface for ‘Prelude’
    Perhaps you haven't installed the profiling libraries for package ‘base-4.9.0.0’?
    Use -v to see a list of the files searched for.

T11627b.hs:9:1: error:
    Failed to load interface for ‘GHC.Types’
    Perhaps you haven't installed the profiling libraries for package ‘ghc-prim-0.5.0.0’?
    Use -v to see a list of the files searched for.

T11627b.hs:10:1: error:
    Failed to load interface for ‘System.Mem’
    Perhaps you haven't installed the profiling libraries for package ‘base-4.9.0.0’?
    Use -v to see a list of the files searched for.

on a normal devel2 build. Do I need a different kind for prof?

You need the profiling libraries. Make sure BUILD_PROF_LIBS=YES, for example by setting it in mk/build.mk.

added Windows label

added broken test label

The test is to do with large objects. Perhaps there is something broken in the windows specific code for that but it has probably always been broken.

I don't quite understand what you mean with large objects...

You can see in the original patch. 8c61f125

The test creates a large enough ByteArray# such that it is the only closure in the block such that when it is shrunk there is slop left over at the end of block.

Ah like that. But that code doesn't have anything Windows specific in it. In any case, will have to see if this is still failing.

added Pnormal label

This unexpectedly passes under my Windows VM.

mentioned in commit 92759bf6

mentioned in commit ededd7ae

mentioned in commit 939295ea

mentioned in commit a16c7379

mentioned in commit a4f9e233

mentioned in commit 49327e24

mentioned in commit 411ba7ba