[GHC] #11300: outofmem RTS testcase fails on ARM

GHC ghc-devs at haskell.org
Sun Dec 27 16:31:35 UTC 2015


#11300: outofmem RTS testcase fails on ARM
----------------------------------+-------------------------------
        Reporter:  bgamari        |                Owner:  bgamari
            Type:  bug            |               Status:  new
        Priority:  normal         |            Milestone:  8.0.1
       Component:  Compiler       |              Version:  7.10.3
      Resolution:                 |             Keywords:
Operating System:  Linux          |         Architecture:  arm
 Type of failure:  Runtime crash  |            Test Case:
      Blocked By:                 |             Blocking:
 Related Tickets:                 |  Differential Rev(s):
       Wiki Page:                 |
----------------------------------+-------------------------------
Changes (by thomie):

 * cc: Phyx (added)


Comment:

 Fails on Windows as well.

 See this addition of `testsuite/tests/rts/outofmem.stdout-i386-unknown-
 mingw32` in https://phabricator.haskell.org/D1340?vs=on&id=4578#788ffc72.

 And this discussion on irc:
 {{{
 "15-10-21T18:56:56"<       thomie > Phyx-: the only comment I have now is,
 are you sure about updated output for outofmem? Last time I checked, I
 didn't quite understand what that test was supposed to do.
 "15-10-21T18:57:38"<        Phyx- > thomie: I am, but I was actually going
 to undo it and see if I can't het it to output the expected error code.
 "15-10-21T18:58:13"<        Phyx- > thomie: the problem is that when the
 exception is thrown, it's hitting the VEH handlers, which call
 stg_exit(1);
 "15-10-21T18:58:32"<        Phyx- > so it never enters main again.
 "15-10-21T18:58:54"<       thomie > ok, I'll look at it later.
 "15-10-21T18:59:37"<        Phyx- > it's a bit unrelated to the patch I
 changed it in, so i was going to make anothe diff for it. and see what the
 exception i get is and see how to handle it in the VEH stuff.
 "15-10-21T19:00:02"<        Phyx- > I don't know how the errors are
 handled under linux, but I find it odd that it can return to main..
 "15-10-21T19:08:52"<        Phyx- > thomie: Basically, the test is
 checking to see if this line is hit
 https://github.com/ghc/ghc/blob/master/rts/RtsMain.c#L80 but on Windows it
 just won't get there as it exits here
 https://github.com/ghc/ghc/blob/master/rts/win32/veh_excn.c#L65
 }}}

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11300#comment:2>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler


More information about the ghc-tickets mailing list