According to Geoff Hutchison:
> At 3:47 AM -0700 7/14/01, Joe R. Jah wrote:
> > > ssl.4 296 omitted
> >
> >This is the second most popular patch in the site; I am sure there will be
> >scores of requests for it. If you haven't backported the ssl code from
> >3.2, I suggest to port this patch to 3.1.6.
>
> There's absolutely no way anyone wants to backport the HTTPS/SSL code
> from 3.2--this would require backporting all the new Transport class
> code as well as the modifications to the URL, Retriever, Document,
> URLRef, etc. classes.
>
> As to excluding the ssl.4 patch from 3.1.6, I'd generally give it a
> -1. It's not that the patch is bad or buggy. But the code prior to
> the 3.2 split (which happened shortly after 3.1.0 came out) is simply
> too heavily tied to HTTP.
>
> I think it's a great patch. But I wouldn't run it on htdig.org--the
> base code was changed for 3.2 and patching the duct tape is not a
> good idea for a "maintenance release."
Based on the problem reports I've seen on the list from ssl.4 users, I'm
concerned about it's reliability and thoroughness. I'm also very concerned
about its portability, and its dependence on an external library. None of
this is the sort of thing we want to inflict on a whole population of users
who are depending on a STABLE release. I don't care how popular the ssl.4
patch is, it's still used by a small minority of htdig users.
> > > new.pl 74 omitted
> >This is the working version of whatsnew.pl; why is it omitted?
>
> It might be worth updating the contrib/ directory--but it's not
> really a patch. At least that would be my interpretation. Gilles?
No, it's not a patch, but it depends on some other code, and possibly
on another patch. I did a few quick tests with it and failed to get it
working on my system. I also didn't think the docs/comments it came with
explained well enough what you need to make this work. I'd be happy to
reverse my position on this if it were made clearer to everyone how this
script goes in.
> >Have you by any chance backported duplicate prevention procedure form 3.2
> >code, or plan to do so? If not I urge you to port this patch to 3.1.6.
>
> Hmm. I don't think the 3.2 duplicate detection code is very robust
> yet, so I certainly wouldn't suggest it be backported. As for the
> patch that deals with local_urls, I'd give the same reasoning I did
> for not including it in previous 3.1.x releases--it doesn't help with
> HTTP indexing. And I think it's in the HTTP indexing that there are
> more problems.
>
> Just my $0.02. I'm going to let Gilles make the call unless he asks otherwise.
> -Geoff
I'm with you on this too, Geoff. This is a maintenance release, so I'm
only backporting stuff that's small and easy to port, and adds to the
stability of the package. I will not backport large and inadequately
tested add-ons in 3.2.
That's it from me for now. I may try to catch up on the list a bit next
week, before I leave again for 3 weeks.
--
Gilles R. Detillieux E-mail: <[EMAIL PROTECTED]>
Spinal Cord Research Centre WWW: http://www.scrc.umanitoba.ca/~grdetil
Dept. Physiology, U. of Manitoba Phone: (204)789-3766
Winnipeg, MB R3E 3J7 (Canada) Fax: (204)789-3930
_______________________________________________
htdig-dev mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/htdig-dev
