Re: [Websec-users] Anybody here?

[Trevor Boicey]

Baruch Even wrote:
> Can you document what layout changes were needed? I only knew the Debian
> package and thus it's what I always had. I'd like to document the needed
> changes so that other users will have it easier.

  I can only document them roughly...

  For 1.3.4 I just installed everything in /home/tboicey/websec, 
including my URL list.

  When run out of that directory, it seemed to do all it's work in that 
directory. The archives stayed there, presumably index.html and 
retrieve.html were made there, and as long as websec and webdiff were 
both in that directory they would find each other.

  For 1.4.0, it seemed to make/want a ~/.websec directory, and the 
archives went there. (Probably better for a multiuser system)

  It also stopped looking in the current directory for webdiff, so it 
had to be put in a common area in the PATH.

  And finally, that weird "cannot find index.html" problem that just 
happens sometimes.

  It *MIGHT* have something to do with websec getting a 404, because it 
stops the whole list processing dead in it's tracks but one time I 
noticed I did get one item on my list checked and it returned a 404.

  So perhaps whatever websec uses to grab the URL doesn't make 
index.html when it gets an HTTP error code like 404, and then webdiff 
fails, and then the whole script craps out?

> The error messages that you give are not in the sources that I have, so
> maybe you have an old version of webdiff lying around in somce obscure
> $PATH?

  I may not have typed them word for word. The actual cron output 
(which I get by email) is:

Cannot open /home/tboicey/.websec/retrieve.html: No such file or directory
Cannot open /home/tboicey/.websec/index.html: No such file or directory

  ...I'm pretty sure I didn't have an old webdiff around. I checked now 
and I don't, but also with 1.3.4 I never had to move webdiff out of the 
websec directory for it to be found...

> You may want to file this bug report on the savannah web page, there is
> a bug tracker there and it would help to keep bug reports in it.

   I can copy and paste this email into a bug reporting page if you 
think it might help. I don't feel I have a lot of specifics, it only 
failed 2 or 3 times in a list of 40 URLs that runs a few times a day.

  The main annoyance is that the whole URL list stops processing when 
it does fail, which made me have to go back to 1.3.4 without being able 
to fully debug it.


--
Trevor Boicey, P. Eng.
Ottawa, Canada, address@hidden
ICQ #17432933 http://www.brit.ca/~tboicey/
Nothing protects dry skin like a Sith Lord.