I had a few minutes to spare while waiting for something else to fail, so I
installed stunnel and tried the tests. 306 fails for me, but *only* if I run
305 as well. So, I juggled things around a bit (yes, I was bored, OK?):

    ./runtests.pl 304 306 - all OK

    ./runtests.pl 305 306 - 306 fails

    ./runtests.pl 300 301 302 303 304 306 - all OK

    ./runtests.pl 306 305 - all OK

    ./runtests.pl 305 304 - 304 fails

    ./runtests.pl 306 304 - all OK

These are just the combinations I tried. Probably there are others that
interact like this.

Dunno if it's relevant, but sometimes even the smallest hint helps fix stuff...
:) I'd have a bash at it myself, but I never really got into perl, so I don't
know where to start.

Oh, BTW, when 306 or 304 fails, it leaves a core file from 'sws'.

Rather than fix whatever's going on in there, would just re-ordering the tests
would be good enough?? Is there any particular reason for taking them in this
order, or could you just switch the numbers on 305 & 306??

runtests.pl says this about my system:
********* System characteristics ********
* curl 7.10.2 (i686-pc-linux-gnu)
* libcurl/7.10.2 OpenSSL/0.9.6g zlib/1.1.4
* Host: lorien
* System: Linux lorien 2.4.19-gentoo-r7 #1 Sat Jul 13 07:29:15 CDT 2002 i686
Pentium III (Katmai) GenuineIntel GNU/Linux
* Memory debugging: OFF
* Netrc debugging: OFF
* HTTPS server: ON
* FTPS server: ON
* libcurl SSL: ON
*****************************************

and 'stunnel -V' says this:

stunnel 3.22 on i686-pc-linux-gnu PTHREAD+LIBWRAP with OpenSSL 0.9.6g 9 Aug 2002

Ralph Mitchell

Daniel Stenberg wrote:

> [snip...]
>
> I don't think there were any problems left with stunnel. I do know that there
> are two test cases that usually fail for me as well, but that's because some
> lameness in the test system that make them fail when run after the previous
> test case as they always run fine when run stand-alone.
>
> > Do you need more information from me?
>
> Which OS do you run? Which stunnel version do you have installed? Can you use
> curl even though it failed this much?
>
> Also, if you 'cd tests' and run the failing tests one by one it will help us
> to sort out what fails and what to do about it. Run them like this:
>
> ./runtests [num]

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Received on 2002-11-22