[4948] in testers
Re: Why panel is bouncing
daemon@ATHENA.MIT.EDU (Mitchell E Berger)
Thu Jul 5 17:32:23 2001
Message-Id: <200107052132.RAA05256@byte-me.mit.edu>
To: Greg Hudson <ghudson@MIT.EDU>
cc: Mitchell E Berger <mitchb@MIT.EDU>, testers@MIT.EDU
In-Reply-To: Your message of "Thu, 05 Jul 2001 14:09:12 EDT."
<200107051809.OAA24164@egyptian-gods.MIT.EDU>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Thu, 05 Jul 2001 17:32:16 -0400
From: Mitchell E Berger <mitchb@MIT.EDU>
> So, it turns out there is a five-day window between the release of
> sawfish 0.37.3 (the last version we imported) and the comment from
> George Lebl on February 24. (Our version of gnome-core is
> significantly newer than February 24, but perhaps the bugfix is
> dependent on both pieces, or perhaps it only went into the mainline
> and not the 1.4 branch of panel development.) A little changelog
> grovelling revealed a likely candidate for the real fix.
>
> So I'm going to punt on moving the sawfish startup back before the
> panel startup, in hopes that we can squish the bug the right way.
OK. I would have to do more testing to be more confident that it's related,
but I'll be curious to see if the newer sawfish also has the same unexpected
side-effects that the patch I submitted seems to have had, namely:
- Since installing my new xsession, I haven't gotten the wrong color
background on an 8-bit display. If the bouncing bug gets fixed with the new
sawfish but this one doesn't (and if I can confidently claim correlation), I'd
still like to offer the sawfish/panel order switch for consideration.
- Since installing my new xsession, login time seems to have become quicker.
This likely won't be noticed too much on normal machines, but I think
something was definitely screwy as login time on undocumented-feature (my
silly 9.0 Classic) dropped from 7 minutes 24 seconds to 5 minutes 16 seconds.
Perhaps this is simply because panel repaints take an excruciatingly long time?
Mitch
