I get a stack overflow on graphs 3 and 4. I havn't investigated, but
perhaps
there's an infinite loop in your program? The next release of GHC will
at least report stack overflow rather than seg-faulting!
Simon
| -----Original Message-----
| From: Nick Williams [mailto:[EMAIL PROTECTED]]
| Sent: 05 June 2000 19:43
| To: [EMAIL PROTECTED]
| Subject: Segfault using code compiled with 4.06
|
|
| Hi there,
|
| I have some code here that causes ghc-4.06 to generate code that
| segfaults here on my Linux box. According to the ghc
| documentation, this is probably a bug in ghc.
|
| Actually, it either segfaults, bails out with a 'fatal error:
| scavenge: stack frame' message, or overflows a 200MB stack (if
| compiled with -O2). It's clear something is going wrong, anyway.
|
| The code is in a number of modules, which I have placed on the
| web, seeing as how I doubt very many people are interested in
| them, it seems rather wasteful to send them to the mailing list.
|
| The URLs follow:
|
| http://www.new.ox.ac.uk/~nick/ghc-probs/Main.hs
| http://www.new.ox.ac.uk/~nick/ghc-probs/Graphs.hs
| http://www.new.ox.ac.uk/~nick/ghc-probs/Dijkstra1.hs
| http://www.new.ox.ac.uk/~nick/ghc-probs/Dijkstra2.hs
|
| On my machine, it deals successfully with the first two data
| sets, and segfaults on the third and fourth:
|
| http://www.new.ox.ac.uk/~nick/ghc-probs/graph1
| http://www.new.ox.ac.uk/~nick/ghc-probs/graph2
| http://www.new.ox.ac.uk/~nick/ghc-probs/graph3
| http://www.new.ox.ac.uk/~nick/ghc-probs/graph4
|
| The program takes its input from the standard input.
|
| Any ideas?
|
| Cheers,
|
| Nick.
|
| --
|
| [ Nick Williams
| ]
| [ MSc Computation Mobile -
| 07957-138179 ]
| [ New College, Oxford
http://www.new.ox.ac.uk/~nick/ ]
[ [EMAIL PROTECTED] Finger for PGP public key ]
