XFree86-xserv-4.3.0-32 tests

XFree86-xserv-4.3.0-32 tests

[Øyvind Harboe]

Rejoice!

Here are the tests I did that all worked fine:

- Copy piece of text under Linux, paste in Windows.
- Copy openoffice formula under Linux, paste in Linux
- Copy openoffice cell range under Linux, paste in Linux Evolution HTML
email
- Copy norwegian letters from Linux, paste in Windows

This was a bit surprising:

1. Copy cell range under Linux openoffice
2. Copy text under Windows 
3. Paste using scrolling wheel under Linux -> text shows
4. Paste in Evolution HTML email under Linux using ctrl-v openoffice
cell range, cell range showed.

This failed:

- Copy Norwegian letters from evolution email, paste under Windows.
Paste in Windows shows "\x{00D8}\x{00C5}\x{00C6}"

Inconclusive:

- I've had a problem in the past that when I specify the -clipboard
option, XWin.exe will sometimes crash with a NULL pointer access
violation. This sometimes happened during normal operation, but more
often when I exited XWin.exe via the taskbar icon. After about 10-15
minutes of testing, I haven't been able to reproduce this problem.


Øyvind

RE: XFree86-xserv-4.3.0-32 tests

[Andrew Braverman]

Just to follow up on this, I found this morning that this does not happen
all the time.  If I start the server and the new code is working, selecting
text with the mouse leaves data selected (inverse video) and none of these
things happen (there is still some debug logging, but much less).  At other
times, selected text has the inverse video go away (like used to happen) and
the three things I listed here happen.  I expect, based upon Harold's
comments about fixing a bunch of things, that this will go away in his next
release, but figured I would mention it as a data point anyway.

- Andy

> -----Original Message-----
> From: Andrew Braverman [mailto:abraverman@itms.com]
> Sent: Tuesday, January 06, 2004 1:45 PM
> To: 'cygwin-xfree@cygwin.com'
> Subject: XFree86-xserv-4.3.0-32 tests
>
>
> I tried this version and it solves the problem that I saw
> yesterday with -31.  I have, however, found a few issues that
> I thought you should at least be aware of.
>
> 	1. XWin is taking a lot more system resources than it
> usually does.  This may be related to
> 	2. XWin is doing a lot of debug logging to XWin.log.  I
> understand that this may be on purpose, but figured I would
> mention it.
> 	3. I do not know if it was this version or something
> just before, but my autorepeat does not seem to work (even
> though xset q shows it on).
>
> Thanks again for the wonderful work you are doing.
>

XFree86-xserv-4.3.0-32 tests

[Andrew Braverman]

I tried this version and it solves the problem that I saw yesterday
with -31.  I have, however, found a few issues that I thought you should at
least be aware of.

	1. XWin is taking a lot more system resources than it usually does.  This
may be related to
	2. XWin is doing a lot of debug logging to XWin.log.  I understand that
this may be on purpose, but figured I would mention it.
	3. I do not know if it was this version or something just before, but my
autorepeat does not seem to work (even though xset q shows it on).

Thanks again for the wonderful work you are doing.

Re: XFree86-xserv-4.3.0-32 tests

[Harold L Hunt II]

Øyvind,

I wouldn't worry too much about it... I am in the middle of completely 
reworking the clipboard shutdown code.  The crashes will be fixed when I 
am through with it, so I don't particularly need details about any 
current crashes in -32 anymore.

Thanks for testing,

Harold

Øyvind Harboe wrote:
> I've now seen the xwinclip: xwin.exe crash on exit with 
> 4.3.0-32 as well. It was an access violation to address 1.
> 
> Is this a known problem?
> 
> If not, does anyone have any pointers on how to create a useful bug report?
> 
> One idea I had was to somehow invoke GDB once the OK Cancel access violation requester
> pops up, but I've no idea on how to go about this?
> 
> A fixed sequence of steps to reproduce the problem has eluded me, hence my 
> interest in a postmortem approach.
> 
> 
> Øyvind
> 
> 
> 
> 

XFree86-xserv-4.3.0-32 tests

[Øyvind Harboe]

I've now seen the xwinclip: xwin.exe crash on exit with 
4.3.0-32 as well. It was an access violation to address 1.

Is this a known problem?

If not, does anyone have any pointers on how to create a useful bug report?

One idea I had was to somehow invoke GDB once the OK Cancel access violation requester
pops up, but I've no idea on how to go about this?

A fixed sequence of steps to reproduce the problem has eluded me, hence my 
interest in a postmortem approach.


Øyvind