[Bf-committers] Fwd: Bug#486878: blender does not process autosaves correctly
kibi at debian.org
Wed Jun 18 22:41:46 CEST 2008
please find below a bug regarding autosaves that has been opened in the
Debian Bug Tracking System. I've merge the messages from the submitter,
but you can also have a look at the whole thread at .
----- Forwarded message from Mikko Rasa <tdb at tdb.fi> -----
From: Mikko Rasa <tdb at tdb.fi>
To: 486878 at bugs.debian.org
Subject: Bug#486878: blender does not process autosaves correctly
Reply-To: Mikko Rasa <tdb at tdb.fi>, 486878 at bugs.debian.org
Date: Wed, 18 Jun 2008 23:22:13 +0300
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:220.127.116.11) Gecko/20080313 Iceape/1.1.9 (Debian-1.1.9-5)
Message-ID: <48596E75.3090307 at tdb.fi>
Occasionally, blender enters a select() with a NULL timeout, waiting for
events from the X server. If this happens when it's inactive for a
prolonged time, autosaves get "queued up" and are processed at once when
the select finishes (due to blender getting an event). If it has been
frozen for several days, the hundreds of autosaves will take a
There seems to be a good amount of variation in what series of events
cause this bug to appear and what don't. Changing workspaces appears to
be the most reliable way to see it.
In my opinion, the problem is not in the infinite-timeout select but in
the fact that blender seems to increment the next autosave time instead
of setting it to current time plus interval.
I took a look at the source and found that I was indeed correct in my
initial analysis. The increment happens in file
GHOST_TimerManager::fireTimer, lines 146-149. I don't know enough about
the architecture to make a patch for it though.
----- End forwarded message -----
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 197 bytes
Desc: Digital signature
Url : http://lists.blender.org/pipermail/bf-committers/attachments/20080618/a5fef0a2/attachment.pgp
More information about the Bf-committers