This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Re: 0.9.4 check failure
- To: Rodney Sparapani <rsparapa at mcw dot edu>
- Subject: Re: 0.9.4 check failure
- From: Brian Gough <bjg at network-theory dot co dot uk>
- Date: Fri, 19 Oct 2001 13:18:46 +0100 (BST)
- Cc: gsl-discuss at sources dot redhat dot com
- References: <263734E8-C436-11D5-A753-003065D59DC8@mcw.edu>
Rodney Sparapani writes:
> I had one GSL 0.9.4 check failure for Mac OS 10.1 w/ GCC 2.95.2. I combed
> through the output and I only found one suspect, although, it actually
> said PASS:
>
> PASS: filip gsl_fit_multilinear cov11 (313354 observed vs 313353 expected)
> ...
> FAIL: test
> ===================
> 1 of 1 tests failed
> ===================
Hmmm, a pass is a pass, the numbers don't need to identical, just
within a tolerance. There may be a segmentation fault which is not
visible when running from make check. Can you run the ./test program
under the debugger, thanks.
> On another note, I discovered what the problem was with the truncated file
> names. It turns out that gzip and tar are fine, but StuffIt Expander 6.5
> does not recognize names longer than 31 characters. This is the default
> application for most compressed file types, so you have to be careful.
> Finally, as was the case with 10.0 and 10.04, you must use
> --disable-shared with configure. CFLAGS set to -Wno-long-double will save
> you alot of useless warning messages as well.
Good. I'll mention avoiding Stuffit Expander in the installation notes.
Brian