On Wed, 06 Jun 2007 12:20:44 -0700 (PDT)
Ben Rockwood <[EMAIL PROTECTED]> wrote:
> I've had a couple applications have assorted strange problems on recent
> Nevada builds. One of the applications was the LiteSpeed Web Server, and the
> response from LiteSpeed regarding the problem was that an ABI change broke
> compatibility with existing Solaris8/9 X86 32 bit binaries.
>
> Is this true or false? I can't find any evidence of such a change, but I
> might be overlooking something. Can anyone comment?
IN this specific case it sounds like it would be easy to get a truss of
this working on an older revision. I might do that. The comparison
could tell you a lot. Make sure you keep as many variables except for
the OS constant. So use the same config file, directory hierarchy, etc.
This truss by itself doesn't tell a very interesting story. The
application checks and opens a fd or two and then sucks up its
configuration (which is something to check). It then seems to go on a
hunt for something (I preseume lshttpd). Somewhere in there it does
the same stat/access as earlier but then doesn't appear to do much with
the info (guess: it did the stat/access before seeing if it already
loaded the config?) The paths that end in "lshttpd/" would worry me but
if that was really lshttp you would get at least one ENOTDIR. Then the
first process completely runs off the rails. The error code for having
been given a bogus configuration sounds like it might be buggy.
Anything special about the underlieing filesystem?
mph
>
> benr.
>
>
> This message posted from opensolaris.org
> _______________________________________________
> opensolaris-code mailing list
> [email protected]
> http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
_______________________________________________
opensolaris-code mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
