Hi David,
./configure with CFLAGS including -DBUFSIZ=225.
Then don't do that? :-)
Where your pioneering spirit? :-)
I suspect that m_getfld() might be involved. I wouldn't recommend
trying to fix it. If you want to do something, #error if BUFSIZ < 226
or something like that?
I've good news and bad news.
The good news is I know of no platform where BUFSIZ < 226; I was just
playing with the definitions to see what might break as it seemed to me
that shuffling and sliding data through two buffers, of potentially
different sizes, with m_getfld() involved, could be error prone.
But I thought this failure might mean a problem does exist and more
realistic buffer sizes could also make it occur, so I went to tweak
test/inc/test-eom-align that loops over email sizes, 221 of them, to try
and break things. I think the script has never worked. Instead of
generating 221 distinctly sized emails, running a couple of tests on
each, it comes up with one email and runs the two tests on it 221 times.
So I fixed the test script and it fails with the existing parameters.
That's the bad news.
I'm sure it's m_getfld() and friends. I did dig into quite a bit before
reporting the problem above, but then ran out of time so sent the email
instead. I'll try and return to it soon.
Anyone recall ctrace(1)? It used to take foo.c and produce a new one
with lots of calls added that output a fragment of source code as it
executed; running the code traced the program.
--
Cheers, Ralph.
https://plus.google.com/+RalphCorderoy
_______________________________________________
Nmh-workers mailing list
Nmh-workers(_at_)nongnu(_dot_)org
https://lists.nongnu.org/mailman/listinfo/nmh-workers