480179 - X11 error BadIDChoice [crash]

Status: RESOLVED DUPLICATE of bug 458092

(Core :: Widget: Gtk, defect)

Description

[Robert Bradbury]

User-Agent:       Opera/9.63 (X11; Linux i686; U; en) Presto/2.1.1
Build Identifier: Mozilla Firefox 3.0.6 (BonEcho) - [from /usr/bin/firefox --version]

Gdk-ERROR **: The program 'firefox-bin' received an X Window System error.
This probably reflects a bug in the program.
The error was 'BadIDChoice (invalid resource ID chosen for this connection)'.
  (Details: serial 655192291 error_code 14 request_code 150 minor_code 4)
  (Note to programmers: normally, X errors are reported asynchronously;
   that is, you will receive the error a while after causing it.
   To debug your program, run it with the --sync command line
   option to change this behavior. You can then get a meaningful
   backtrace from your debugger if you break on the gdk_x_error() function.)
aborting...


Reproducible: Sometimes

Steps to Reproduce:
1. Run Firefox; open 65 windows; 640 tabs (restart of previous session though this is an error derived from a new window).
2. Open up the uhaul.com site in a new window (with NoScript active).
3. Request a quote on a route starting from Seattle; enable temporary scripting on the uhaul site (which had previously been blocked by NoScript); watch abort + core dump occur dumping the difficult to recreate "active" state of the browser (requires 15+ minutes to restart this many windows/tabs) -- and even then state information (e.g. workspaces for the windows will be lost).
Actual Results:  
Browser aborts.  Worst case for X errors there should be some communication between the browser and X to redraw everything (its only 65 windows) -- not to terminate the browser session entirely.

Expected Results:  
Firefox should catch errors, both critical errors within thread as well as errors taking place within cooperating programs such as the X server and tolerate them.  It is too expensive to restart "complex" states -- the program should be tolerant of faults within the web site programming or the Firefox code itself.

After all -- if X failed entirely I should lose *all* windows -- that is not the case.  I only lose all windows associated with Firefox (along with the Firefox core dump).  Have you ever heard of fault-tolerant programming?

Firefox should have, if necessary, reinitialized the X server, reset all windows and redrawn them -- presumably paying careful attention to data which may have become corrupted (and generating an X error).  This suggests that there may be an underlying error in the Firefox code of some kind of buffer overrun, random memory access, etc. which is leading to the X error.

I have seen this error at least 3 times (on different sites) so it is not a random problem.  I did not see this type of error under Firefox 2.0.X so it would appear to be that this is an inherent Firefox 3.0.X problem!

I am labeling this a Critical problem because (a) Firefox terminates and leaves a core dump which is non-debugable; (b) There are sites such as PubMed (where I normally may have dozens of tabs open) where even if one returns to the same URL one loses the state information about the session (thus Firefox cannot fully return one to the previous state even if one waits through the lengthy restore session process) -- you have to use PubMed to understand these problems; and (c) it is a bug which seems to have appeared in Firefox 3.0 which was not present in Firefox 2.0 -- thus one has taken the program backward rather than forward.

Comment 1

[Tyler Downer [He/Him]]

It is hard to figure what is at fault here, because you have noscript. Can you create a new profile and try there? Also, you seem to be using a distro build. Can you use an official build? That will give stacktraces and all that for debuging.

Comment 2

[Ben Bucksch (:BenB)]

I also just got this. Console said:

The program 'firefox-bin' received an X Window System error.
This probably reflects a bug in the program.
The error was 'BadIDChoice (invalid resource ID chosen for this connection)'.
  (Details: serial 208305651 error_code 14 request_code 156 minor_code 4)

I also had several windows with each several tabs open, probably about 100 tabs.
The crash happened IIRC when I middle-clicked on a link to open it in a new tab. I had IIRC done same with 1-2 other links seconds before (which I do a lot, but normally doesn't crash).

Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.2a1pre) Gecko/20090126 Minefield/3.2a1pre
Built myself from trunk

Comment 3

[timeless]

get a better xlib/xserver/distro :)