multiple enters in some apps causes CellWriter to quit

[GoogleCodeExporter]

1) Please describe the problem.

Using gedit, I noticed CellWriter kept stopping/quitting unexpectedly upon
entering multiple newlines via its "Enter" button. It also happens in
Firefox' textedit box.

2) What steps will reproduce the problem?

1. launch gedit
2. Hit CellWriter`s "Enter" button numerous (1-5) times.
3. CellWriter disappears.

3) What version are you using? What Linux distribution?
version 1.3.5, Debian testing

installed versions of dependencies:
libatk1.0-0 (1.28.0-1), libbonobo2-0 (2.24.2-1), libc6 (2.10.2-2),
libcairo2 (1.8.8-2), libfontconfig1 (2.6.0-4.2), libfreetype6 (2.3.11-1),
libglib2.0-0 (2.22.3-1), libgnome2-0 (2.28.0-1), libgtk2.0-0 (2.18.3-1),
liborbit2 (1:2.14.17-2), libpango1.0-0 (1.26.1-1), libpopt0 (1.15-1),
libx11-6 (2:1.3.2-1), libxtst6 (2:1.0.3-1)

Original issue reported on code.google.com by sunshinekisses@gmail.com on 23 Dec 2009 at 10:32

[GoogleCodeExporter]

This error is reported right before cellwriter cleans up:

cellwriter: Fatal IO error 11 (Resource temporarily unavailable) on X server 
:0.0.

so it may be hardware reeated?

Original comment by sunshinekisses@gmail.com on 27 Dec 2009 at 7:04

[GoogleCodeExporter]

I also came up with this:

cellwriter: ../../src/xcb_io.c:249: process_responses: Assertion `(((long)
(dpy->last_request_read) - (long) (dpy->request)) <= 0)' failed.

which is from the last line of the function: process_responses(...); right 
above that
assertion is:

    if(xcb_connection_has_error(c))
        _XIOError(dpy);

which is  most likely where the other message is called from, c being a pointer 
to a
xcb_connection_t.

Original comment by sunshinekisses@gmail.com on 27 Dec 2009 at 10:10

[GoogleCodeExporter]

As far as I can tell this patch remedies this particular issue. I've installed 
this
updated version on my pc and will report back if I notice it breaking 
something. Of
course one could modify the key_event_send_enter() which, in hindsight, I 
perhaps
should have done as is it maybe less prone to break something.

Speaking of which, I`d like to know if anyone else can reproduce it on their own
hardware since it might be specific to my own.

Original comment by sunshinekisses@gmail.com on 27 Dec 2009 at 11:49

[GoogleCodeExporter]

Just adding a call to XSync() between key events appears to to solve the issue 
in a
much more elegant fashion.

Original comment by sunshinekisses@gmail.com on 30 Dec 2009 at 8:18

Attachments:

• cellwriter_issue_32_fix_01.diff

[GoogleCodeExporter]

This issue was closed by revision r21.

Original comment by risu...@gmail.com on 7 Jan 2010 at 12:05

• Changed state: Fixed

[GoogleCodeExporter]

I can't reproduce the bug so it must be specific to your particular hardware 
setup.
Adding that XSync event doesn't do any harm though so I committed the change. 
Thanks
for reporting it! Please check to make sure the unmodified SVN version works 
for you.

Original comment by risu...@gmail.com on 7 Jan 2010 at 12:07

• Added labels: For-1.3, Priority-Critical, Type-Bug
• Removed labels: Priority-Mediumxx, Type-Defect

[GoogleCodeExporter]

It works great! Thanx alot!

Original comment by sunshinekisses@gmail.com on 7 Jan 2010 at 2:37