|
RE: eval_thunk_selector: strange selectee 29: msg#00022lang.haskell.glasgow.bugs
On Mon, 10 Nov 2003, Simon Marlow wrote: > > > I can confirm that this happens on my solaris machine as > > well. Moreover it > > crashes my computer! I have seen the same crash when working > > with the ffi > > on solaris but haven't been able to reproduce it. What > > happens is this: > > In the window I'm running ghci/my ffi application gets filled with > > prompts. I'm guessing something has started forking like crazy. Anyway > > this eats up all the computer resources and all I can do is > > restart the > > computer. > > > > It would be Very Nice if you could fix the crashing bug as > > this sometimes > > crashes programs running for example wxHaskell. I get very > > uneasy as soon > > as I start to program with the ffi. > > We've fixed the "strange selectee 29" bug, but I doubt that could be the > cause of your infinite forking behaviour. If you manage to get a > repeatable example, we can look into it. > Well, if the infinite forking behaviour doesn't depend on the strange selectee 29 bug then you should be able to use the same program to provoke ghci to start forking. Apart from that I attach a small program which exibits the forking behaviour. It's a wxHaskell application which might make things a little tricky, but the program itself is just "hello world" where I forget to supply the "start" function in the beginning. I perhaps should repeat again that the bug is not wxHaskell specific but the problem seems to be with the ffi as we've encountered the problem before when using the ffi. > To prevent it crashing your machine, you should set a limit on the > number of processes creatable by each user. To do it locally for > yourself, use the 'ulimit' shell command. Most OSs have a way to set > this globally or on a user-by-user basis too (can't remember exactly how > Solaris does it, sorry). > Hmmm, the problem is that I don't know either. ulimit doesn't do the job anyway. I've managed to close the window by pressing Ctrl-D in it now, which is nice because I don't have to restart my computer :) /Josef
Glasgow-haskell-bugs mailing list Glasgow-haskell-bugs@xxxxxxxxxxx http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
|
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| Previous by Date: | RE: type class problem / GHC bug, Brandon Michael Moore |
|---|---|
| Next by Date: | ghc -cpp on Mac OS X, Ashley Yakeley |
| Previous by Thread: | RE: eval_thunk_selector: strange selectee 29, Simon Marlow |
| Next by Thread: | RE: type class problem / GHC bug, Simon Peyton-Jones |
| Indexes: | [Date] [Thread] [Top] [All Lists] |
| News | FAQ | advertise |