[MLton-devel] -profile time and the native codegen

Matthew Fluet fluet@CS.Cornell.EDU
Mon, 20 Jan 2003 11:07:40 -0500 (EST)

> > I'm getting segfaults when trying to write the mlmon.out file at the end
> > of the program.  I assume that what is checked into the source tree works
> > without exhibiting that bug.
> I looked at this again real quickly.  I'm getting segfaults (due to an
> invalid MLton.Profile.Data.t being passed to MLton_Profile_Data_write)
> during the atExit code, with the sources currently checked into CVS.  Odd,
> because I don't think anything has changed in the CVS tree since I updated
> yesterday morning and profiled the new IO.  Can anyone else confirm this
> behavior?

I looked again at this with a cleaned source tree, and I can't reproduce
it.  I probably just got my libmlton.a out of whack.  (Since the linux
cluster is much faster than my office machine, I tend to do "big"
make nj-mlton's on the cluster and then synchronize with my office
machine.  But, if it's small, I'll do it locally.  It's quite plausible
that the latest checked in version of gc.c got compiled on one machine
with one version of gcc and got archived with the rest of the runtime
(compiled on the other machine with a different version of gcc) on the
other machine.  One would hope that .o is an unambigous format, but I've
never had much luck mixing different gcc versions.)

Anyways, I don't see any problems now with either the current CVS tree or
my modifications from yesterday.  I'll be checking them in momentarily.

This SF.NET email is sponsored by: FREE  SSL Guide from Thawte
are you planning your Web Server Security? Click here to get a FREE
Thawte SSL guide and find the answers to all your  SSL security issues.
MLton-devel mailing list