GNOME Bugzilla – Bug 171659
gst-launch crashes on a pipeline for recording a vnc session to theora
Last modified: 2005-06-30 16:00:03 UTC
Steps to reproduce: 1. Run a vnc server on port 5900. (I run vino 2.10.0-0ubuntu1) 2. Run this pipeline: gst-launch-0.8 rfbsrc port=5900 ! ffmpegcolorspace ! theoraenc ! oggmux ! filesink location=out.ogg 3. It segfaults. Stack trace: Program received signal SIGSEGV, Segmentation fault.
+ Trace 57340
Thread NaN (LWP 25152)
Other information:
Created attachment 39264 [details] CLI output The complete CLI output.
Could you get a backtrace with debugging symbols?
Created attachment 39278 [details] Another CLI output The attached file includes a backtrace with debugging symbols. This was compiled from the gst-plugins 0.8.8 package.
I'd really love to see this one fixed so that I can do theora demos for the liveCD; am reopening since the requested information was provided.
I've started looking at this bug this evening. The problem is that rfbsrc works only if bpp is 8. I've added some code to handle 8/16/24 bpp and it seems to work fine. I only need to add some check for LE/BE system. I'm planning to finish this patch tomorrow.. will file here for review!
Created attachment 46509 [details] [review] add support for 8/16/32 bpp This patch adds support for 8/16/24 bpp. I've only tested it on a little endian system and it seems to work fine.. I'm not committing this because i would like a review (i'm not totally sure if my changes are right..). Can you please test and see if this work? thanks!
The patch is against current cvs of gst-plugins/gst/librfb. sigh, i can't change the status of the bug :( please set to needinfo..
I don't think you mean NEEDINFO, as that will cause us to ignore it. I'll review the patch soon.
I did a bunch of hacking on rfbsrc today. This bug (and the 10 other related bugs) are fixed.