Re: [htdig-dev] ExternalTransport patch, and URL format

[Gilles D. <gr...@sc...>]

According to Lachlan Andrew:
> I've been trying to use htDig://'s ExternalTransport  to 
> index KDE's "kurl"s, such as  help:/ man:/ info:/ etc.
> 
> As the code stands (or stood, circa June) it seg faults 
> when it dereferences _access_time immediately after 
> deleting it in Reset().  This can be fixed by the trivial 
> patch below, but I'm surprised it hasn't caused havoc 
> earlier.  Has anyone else used (tested?) ExternalTransport?
> (I haven't yet checkied if this patch introduces a 
> memory leak.)

Where did you get your code in June?  I fixed this problem in the CVS tree
back on January 11/02.  You must have had an older snapshot.

> The other problem is the format of "kurl"s, which use a 
> single '/' after the ':'.  This complies with RFC 1738, 
> since they don't "involve the direct use of an IP-based 
> protocol to a specified host on the Internet".  However 
> htDig:// parses the url and returns "host not known" before 
> calling  ExternalTransport.  Moreover, it automatically 
> translates the  help:/  into  help:// .  Technically, the 
> simplest hack is to make KDE accept arbitrarily many '/'s 
> (but getting the KDE team to accept them may be harder :)
> but it may be more elegant to fix  htDig://.  Comments or 
> pointers would be welcome.

The proper fix would be in htcommon/URL.cc, where it handles parsing of
URLs.  I think it has to deal with exceptions like this on a case by case
basis.

-- 
Gilles R. Detillieux              E-mail: <gr...@sc...>
Spinal Cord Research Centre       WWW:    http://www.scrc.umanitoba.ca/
Dept. Physiology, U. of Manitoba  Winnipeg, MB  R3E 3J7  (Canada)