Mariano,
Actually, I bought a laptop for two reasons: (1) they
"went and got cheap" while I wasn't looking, so there was absolutely no reason
for me not to have one at these prices and capabilities; (2) to put Ubuntu on it
so I'd have a Linux box with me.
The bottom line: a segmentation fault. I saved
the text to a file and will check it for anything that I dare not release
(passwords, etc.) and then can post it if it will help.
There has been no debug log that I can find, at least
not associated with these crashes - like I said, they're ugly. The two
worst things, IMHO, that a computer can do are freeze, and a close second
is to simply quit w/o saying why.
Understood about a bug in ODBC, but that does not explain why the earlier
image has no troubles. I might have not loaded things correctly, maybe
there is a new version of the package that breaks something. However,
Pharo should give me some clues vs. just rolling over and disappearing from RAM
w/o a trace.
Respectfully, I do think it is a Pharo/FFI problem, if only that
something can crash it so easily and completely with no diagnostic
data.
More
to come, and I like Indian food, so if I have to eat crow, I'd appreciate a good
hot vindaloo :)
Minor
disagreements aside, THANKS for your reply. The command-line idea was very
helpful (I should have thought of it, and will next time).
Bill
2009/10/2 Schwab,Wilhelm K
<bschwab@anest.ufl.edu>
I moved my code into 09.09.3 web, and it is
crashing badly - the vm just dies, no indication (that I have found) as to
why. Reminds me of working in C on 16 bit systems
=:(
If you are in Linux, run it from command line so that you cansee the
output on it ? If you are in Windows, you can see the dump file
?.
what about the PharoDebug.log ?
By sneaking up on the crash, I see that it
is quitting when asking ODBCConnection to
connect.
Please, don't understand me wrong, but probably you have a "bug" in
your ODBC driver. I have A LOT of segmentation faults while developing
SqueakDBX. And ALWAYS was mine problem (wrong order of calls, or whatever) or an
OpenDBX (ODBC C library in your case) problem.
I don't think it is
a FFI or Pharo related problem.
At one point, just inspecting a
handle is enough to blow away the vm. Suggestions would be
appreciated. I might set up a new image with just ODBC to try to
reproduce it.
Yeah. Take a clean and latest PharoCore and try to load it there.
Best,
Mariano
_______________________________________________
Pharo-project
mailing list
Pharo-project@lists.gforge.inria.fr
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project