httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Brad Nicholes" <BNICHO...@novell.com>
Subject Re: [PATCH - Showstopper] 1.3.17-dev mod_userdir
Date Thu, 25 Jan 2001 20:25:17 GMT
I agree with Bill, it would be nice to fix this problem without having to #ifdef the code even
more than it already is.  This is the patch that only affects NetWare.

Brad

>>> bill@wstoddard.com Thursday, January 25, 2001 1:24:58 PM >>>

> From: "Bill Stoddard" <bill@wstoddard.com>
> Sent: Thursday, January 25, 2001 12:39 PM
>
>
> > I would really prefer to get 1.3.17 out w/o this patch.  We can roll a new
> > release in a few weeks.
>
> I want to get 1.3.17 patched and tarred today.  I can't parse your comment.
> Do you object to the patch on it's merit, object on the potential for
introducing
> a problem, or do you object to the [potential] delay?
>
> To the patch: mod_userdir today is broken on os2/win32 if the user isn't
very, very
> precise about what they type.  This patch assures that win32 accepts the
directive
>
>   UserDir "C:\WinNT\profiles\*\My Documents\My Website"
>
> without complaint.  Today it doesn't, and mod_userdir on netware is
unusable.
>
> The patch reduces the ambiguity of the code.  The question; side effects?  I
will
> be done with my complete review in about an hour.  I eliminated the
potential side
> effects for unix and localized them to ':' containing file systems, only.

And that hits anyone accessing files on a DFS mount if I am not mistaken. We
do quite a bit of that around these parts mister :-)

> It's
> probably the right patch for unix, but we can remove the #if defined(...)
wrapper
> _after_ we roll 1.3.17 to avoid introducing a new problem.
>
> As for the other 'patch', normalizing sys:/foo vs. sys:foo, I'm waiting on a
> response, but I won't be waiting all that long.  We can roll without it, and
if it's
> provided this afternoon, that patch will affect -only- netware.
>

I am against adding anything new to the code that doesn;t fix a critical
problem because of the risk introducing new bugs. If you can work up a patch
that you guarantee only affects Netware, then I have no problems with that. I
have not looked closely at any of this code but some of your earlier comments
did not inspire me with confidence :-)

Bill




Mime
View raw message