[17073] in Athena Bugs
Re: empty netscape preferences
daemon@ATHENA.MIT.EDU (John Carr)
Thu Aug 19 07:27:46 1999
Message-Id: <199908191127.HAA07704@m56-129-30.mit.edu>
To: "t. belton" <tbelton@MIT.EDU>
Cc: bugs@MIT.EDU
In-Reply-To: Your message of "Mon, 16 Aug 1999 12:51:13 EDT."
<Pine.GSO.3.96L.990816124418.17098D-100000@iphigenia.mit.edu>
Date: Thu, 19 Aug 1999 07:27:36 -0400
From: John Carr <jfc@MIT.EDU>
> However it worries me that you say the new preferences.js specified
> ~/.netscape/cache, as that is possibly the main thing we want to NOT do.
> One of the primary reasons we customize preferences.js at all is so that
> we can make sure the user caches into /var/tmp/.netscape-cache-USER/. This
> has been the case through several versions of Netscape, and this is the
> first I've heard of it not working.
>
> I'm wondering if you are having some other unusual situation here, since
> your preferences.js file should not be zero-length in the first place. Any
> other information you can give me would be extremely helpful.
I can't think of anything more. I was able reproduce the behavior --
getting a new preferences.js with the wrong cache directory -- by
truncating my preferences.js and starting netscape on an SGI.
I don't know how the file got truncated. Usually I quit netscape before
logout, but it's possible that I forgot on one occasion and it lots
tokens before it finished. I have nonstandard dotfiles so the timing of
events during logout may be different in my case. It wouldn't be the
first time I triggered a race condition by doing I/O during logout (see
bugs 1842, 4024).