According to Geoff Hutchison:
> Aside from the problem Gabriele reported, I think we're pretty set. I
> wanted to put in code to call sort -m when merging databases for a speed
> boost, but it's not worth holding up the release.
>
> If people can send me documentation updates, I'd be psyched. Should I roll
> a pre-release with the docs slightly unfinished?
I for one would like to see a pre-release in the next day or two, to
test all the latest patches (or am I the only one still sending them
in? :) Give us at least a day, though, to get in any last bug fixes,
and documentation updates in the works right now.
> REPORTED SHOWSTOPPERS:
> * Problems between running update digs with or without -a.
The only explanation I could find for this problem, after looking through
the code, is that with the -a, the db.wordlist.work file doesn't have any
records in it beginning with "-" or "!", though I can't see anything in
htdig that would prevent those records from appearing. Can Gabriele
provide a dump of db.wordlist.work, after htdig but before htmerge,
to confirm this? This would tell us whether to focus on htdig or htmerge
as the cause of the problem.
> OTHER BUGS:
> (none outstanding)
Doesn't stop me from finding 'em! :) See below.
> ISSUES:
> * htdig usage message does not specify -l option for start and resume feature.
> (if -l is specified, on receiving a signal, htdig will write the urls it
> must process to the file config["url_log"] and resume from them if called
> with -l when restarted.)
Another issue is that the new --with-cgi-bin-dir stuff in configure
has broken my current RPM support. I'm going to try to figure out
a technique like Ric was suggesting earlier, which will require patches
to the Makefile.in's. I hope to get this done today or tomorrow, and I'd
like it in the final release if possible, but it's not a show-stopper.
If need be, I'll just include my own patches in my source RPMs.
> DOCUMENTATION:
> * htmerge for -m option to merge databases
> * htdig for -l option for stop and restart feature
> * attrs.html, cf_byname, cf_byprog for:
> url_log
> compression_level
> noindex_start
> noindex_end
> allow_in_form
> bad_querystr
> no_title_text
I was just looking through the no_title_text handling in
htsearch/Display.cc, and noticed something that got dropped when this
code was put in. It no longer handles the case where the URL has no
"/", and leaves str uninitialized before it's used. This should fix it:
--- ./htsearch/Display.cc.noslashbug Fri Jan 29 12:55:08 1999
+++ ./htsearch/Display.cc Wed Feb 3 11:41:51 1999
@@ -277,8 +277,10 @@
{
title++; // Skip slash
str = new String(form("[%s]", title));
- // URL without '/' ??
}
+ else
+ // URL without '/' ??
+ str = new String("[No title]");
}
else
// use configure 'no title' text
An unlikely event to occur, sure, but I don't like to see bits of
defensive programming being taken out, especially in light of all the
problems we've been dealing with that cause segmentation faults.
--
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
------------------------------------
To unsubscribe from the htdig3-dev mailing list, send a message to
[EMAIL PROTECTED] containing the single word "unsubscribe" in
the SUBJECT of the message.
