Re: [htdig-dev] more regex problems in 3.1.6

[Geoff H. <ghu...@ws...>]

On Fri, 25 Jan 2002, Gilles Detillieux wrote:

Well, this is why we have pre-releases... 

> > I think to make this all work, we need to rename the bundled regex.h to
> > something like htregex.h to avoid conflicts, as well as put some hooks
> > in the bundled regex.c code to disable it all if you need to use the
> > C library code instead.  What do you think, Geoff?

I remember seeing this, but it's also about the same time I got swamped
here, so I'm not surprised it disappeared in my mailbox.

> I don't think I ever got a response to that.

Obviously renaming to gregex.h and cleaning up the #includes in
EndingsDB.cc is a good idea. As I stated, there are also minor problems
with the current configure test: it doesn't check that there *is* a
rxposix.h header (never trust the user) or that the user didn't supply
--without-rx (ditto). So I still have to refine it.

> I ran htfuzzy/htfuzzy endings, and it ran further but still segfaulted.

What's the segfault--this obviously could be a different error
entirely. After all, I'm assuming that broken regex support would show up
in other programs. Do you have a backtrace?

> So, problem 2: we need a better test for regex, that will work if
> htfuzzy works, and will fail if htfuzzy fails.  Here is the relevant

We do need to improve the test, though if we rename the included header to
gregex.h and use this in the test #include, this should solve the issue
with using the bundled code with the system's header.

I'll try to get all of this in the snapshot so people can give it a try.

-Geoff