[GHC] #8189: Default to infinite stack size?
GHC
ghc-devs at haskell.org
Sun Sep 8 21:11:57 CEST 2013
#8189: Default to infinite stack size?
-------------------------------+-------------------------------------------
Reporter: nh2 | Owner:
Type: bug | Status: new
Priority: normal | Milestone:
Component: Runtime | Version: 7.6.3
System | Keywords:
Resolution: | Architecture: Unknown/Multiple
Operating System: | Difficulty: Easy (less than 1 hour)
Unknown/Multiple | Blocked By:
Type of failure: Runtime | Related Tickets:
crash |
Test Case: |
Blocking: |
-------------------------------+-------------------------------------------
Comment (by simonmar):
There are a few things I'm not completely happy with here. By all means
turn off the default stack limit, but the business with `allocateFail()`
is misguided: `allocateFail()` will not return NULL. It only returns NULL
when there is a single request for memory greater than the maximum heap
size, and (a) by default there's no maximum heap size, and (b) even if
there was, it would be highly unlikely to be smaller than the stack chunk
size (32K). Did anyone test this? The `heapOverflow()` test in
`allocate()` is purely optional, see #1791.
The typical behaviour when someone writes a program that accidentally
blows the stack will be for the machine to grind to a halt swapping, until
the OS finally kills something (hopefully the right process). We have
this behaviour for space leaks now, so I guess it's no worse to have it
for stack leaks too. An actual "out of memory" error is rare, typically
it happens when you try to allocate an array larger than the memory size,
or something like that.
Austin: could you back this out please, and let's discuss it some more.
--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8189#comment:12>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
More information about the ghc-tickets
mailing list