Re: [htdig] All search results end in .gif

[Gilles Detillieux]

I'm sorry, Rusty.  I just noticed what was the problem with my advice
below, and why the changes to the config file didn't work.  Wherever I
said "url_part_aliases", what I really meant was "common_url_parts".
Sorry about the confusion.  The compiled-in value of url_part_aliases is
empty, as it should be, but it's the compiled-in value of common_url_parts
that changed with the ssl.3 patch.  I think when my brain said
common_url_parts, my fingers typed url_part_aliases just out of habit.
The common_url_parts attribute almost never comes up in discussions on
this list, but questions about url_part_aliases come up almost every week.

In any case, using the patched htsearch is probably your best bet.

According to Rusty Nejdl:
> Gilles,
> 
> Thanks!  That was what was needed.  I think that when I upgraded htdig and use
> the ssl.3 patch, I kept my old version of htsearch.  I still need to clean up
> a few things with the image links being broken now, but thanks again for your
> help!
> 
> Rusty
> 
> Gilles Detillieux(grdetil) wrote:
> > According to Rusty Nejdl:
> > > I applies the simpler solution first, which was to add the url_part_aliases
> > > line as follows:
> > > 
> > > #URL Alias for Patching .gif bug
> > > url_part_aliases: https:// https://www. http:// http://www. ftp:// ftp://ftp.
> > > /pub/ .html .htm .gif .jpg .jpeg /index.html /index.htm .com/ .com mailto:
> > > 
> > > However, after rebuilding the search database, no change.  
> > 
> > You shouldn't have needed to rebuild the database, as the internal
> > setting of url_part_aliases for the patched htdig should already have
> > been this value.  This is, of course, assuming you used the "ssl.3" patch
> > from the patch archive, or some variant of it which had the same value.
> > You can confirm this by searching (with grep) for url_part_aliases in
> > the htcommon/defaults.cc source file.
> > 
> > The important thing it to have this setting in the config file that
> > htsearch uses, so it will decode correctly.  Are you sure you're using
> > the same config file for htsearch as for htdig?
> > 
> > > So, this is where I get to the depth of my knowledge.  I thought that I 
> > > applied the patch correctly, but I guess I didn't.  Where is the htsearch
> > > patch that I need to apply?  How do I apply it?  And, how do I know which patch
> > > to use?
> > 
> > There isn't a separate patch to apply for htsearch.  The ssl.3 patch changes
> > htcommon/defaults.cc, which sets the default attribute values for all the
> > ht://Dig programs, including htdig, htmerge and htsearch.
> > 
> > The problem, as far as I can tell from the information you've given, is
> > that the new htsearch binary, that would have been built after applying the
> > patch, is not installed in your cgi-bin (or wherever you put CGI programs
> > on your system).  Base on the behaviour you described, it sounds like
> > you're still running an older htsearch binary, which was installed from
> > an earlier build of ht://Dig before you applied the patch.  If you find
> > that binary and replace it with the newer one, built from the patched
> > source, it should clear up the problem, without needing to change the
> > config file.  The change to the config file is only needed if for some
> > reason you can't replace the older htsearch binary with the newer one,
> > e.g. if some parts of the web site are indexed with the unpatched htdig,
> > and therefore would need the unpatched htsearch.  If everything is indexed
> > with the same, patched htdig, then you should install the newer htsearch
> > so it will have the newer, patched setting of url_part_aliases built-in.
> > 
> > > Gilles Detillieux(grdetil) wrote:
> > > > According to Rusty Nejdl:
> > > > > Let me add that I don't have the url_part_aliases configed.  The only real 
> > > > > tweaking I did of the conf file is the following:
> > > > ...
> > > > > Also, I have found that ONLY .html files get renamed as .gif files.  Files 
>that
> > > > > have .cgi in them work fine and files that end in .pdf work fine.  That is, 
>the
> > > > > links print correctly.
> > > > 
> > > > I just had a look at the ssl.3 patch on the ftp.ccsf.org archive,
> > > > and it does indeed add a couple new items to the front of the list for
> > > > the compiled-in value of url_part_aliases.  So, the behaviour you're
> > > > describing is consistent with building a database with a patched htdig
> > > > and htmerge, but searching with an unpatched htsearch.  In the patched
> > > > code, .html is the 8th item in the list, while in the unpatched code
> > > > .gif is the 8th item in the list.  Either install the patched htsearch
> > > > on your system, or add the following to your search config file:
> > > > 
> > > > url_part_aliases: https:// https://www. http:// http://www. ftp:// ftp://ftp. 
>/pub/ .html .htm .gif .jpg .jpeg /index.html /index.htm .com/ .com mailto:
> > > > 
> > > > > "If it ain't broke, it doesn't have enough features yet."
> > > >  
> > > > Well, I guess it's got enough features!  :)


-- 
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-general mailing list <[EMAIL PROTECTED]>
Information: http://lists.sourceforge.net/lists/listinfo/htdig-general
FAQ: http://htdig.sourceforge.net/FAQ.html