Re: New set of patches for 0.6.0

Ville Skyttä <ville.skytta@iki.fi> wrote:

>>The two files do not exist in HEAD as they were first checked in on the
>>validator-0_6_0 branch. They'll move into place when I merge.
>
>Hmm, I might be missing something here, but cvsweb shows them in the
>Attic even when looking at the validator-0_6_0 branch.  Maybe a cvsweb
>bug?

<URL:http://www.cvshome.org/docs/manual/cvs_2.html#SEC15>

A peculiarity of CVS is all.


>On a side note, I see dev.w3.org is running a quite old version of
>cvsweb and cvs.

Trust me, I've been bugging Gerald about this, but getting them upgraded is
apparently Not Done Lightly. Presumably, this is the server where _all_ W3C
documents exist, including RECs and DateSpace[tm] etc. so they may just be
really paranoid about it. I really wish they'd upgrade CVSWeb though;
Sourceforge is using a really nice Python reimplementation and there's lots
of goodies in newer versions of Hen's modified version (which IIRC is what
FreeBSD uses).


>The cvs is a minor nuisance, but the current version doesn't produce
>diff's that are understood by patch, but one has to manually tweak
>them.

I apply most patches "by hand" anyway, so it isn't such a big issue for me,
but it's annoying, yes.


>>>- check-paths.patch: Portability patch.
> [...]
>Anyway, post-0.6.0 stuff if you like.

Ok. Remind me if I forget?


>>>I tried the absolute URI fixup a few months ago to make validator
>>>easier to install locally, and found it generally possible, but the
>>>currently used SSI's are a problem.
>>
>>EXPN? What is the problem?
>
>[SNIP problem with paths in SSI]

I hadn't really thought of that -- I was thinking of paths in the CGI -- so
I haven't formulated any sane opinion yet. I'll mull it over...



>>There is a problem with /check/referer?
>
>This may be a partial brain fart, but checking for "$q->path_info() eq
>'/referer'" might not be enough if the validator can be installed in
>whatever path below a server root.

CGI path_info is relative to nothing but the URI for the CGI script; it can
get translated internally to something relative to configuration, but we
use it only as a trigger. This, BTW, is _deprecated_ usage as far as I'm
concerned, but Gerald doesn't like "uri=referer" so it stays for now. :-)


-- 
Editor's note: in the last update,   we noted that Larry Wall would "vomment"
on existing RFCs. Some took that to be a cross between "vomit" and "comment."
We are unsure of whether it was a subconscious slip or a typographical error.
We are also unsure of whether or not to regret the error.      -- use.perl.org

Received on Monday, 1 April 2002 07:32:27 UTC