tests results win 32bits

[claudio canepa]

I'm spawing this thread from the one named 'How about a 1.2 release?'

http://groups.google.com/group/pyglet-users/browse_thread/thread/c93f238cc4a76741

---

I have problems in win xp sp3 with ati radeon 4650

They maybe are linked to issue 483 :   gl_info crash with 1.2dev but not 1.1.4 release

Seen with drivers Catalyst Version    11.5:

*  INFO:root:Testing app.EVENT_LOOP seems to take too much time to complete, around 50s. With pyglet 1.1.4 the time needed was imperceptible. This was seen in all runs, don't know if it is expected behavior.

*  Most of runs  will silently die a bit after MULTIPLE_WINDOW_OPEN

for this defect the last console output was:

------------------------------------------------------------------------------
Press return to begin test...
Test that multiple windows can be opened.

Expected behaviour:
Two small windows will be opened, one coloured yellow and the other
purple.

Close either window or press ESC to end the test.

[P]assed test, [F]ailed test: p
Running Test: window.CONTEXT_SHARE
Test that multiple windows share objects by default.

This test is non-interactive.

------------------------------------------------------------------------------
(end of console output)

And the last lines in the log:

INFO:root:User marked pass for window.MULTIPLE_WINDOW_OPEN
INFO:root:Testing window.CONTEXT_SHARE.
INFO:root:Begin unit tests for window.CONTEXT_SHARE
<- EOF

 

No OS message box or anything: test.y silently dies

* seen in one run that progressed more:

At the interactive multisample test (window.WINDOW_MULTISAMPLE),after user input m s s s s an normal,  passing test behavoir, when switching focus with the mouse from the pyglet driven window to the console (not sure  if focus from or to) I got a traceback, captured in the log:

INFO:root:Begin unit tests for window.WINDOW_MULTISAMPLE
ERROR:root:Error in window.WINDOW_MULTISAMPLE
ERROR:root:Traceback (most recent call last):
File "D:\hg_externals\pyglet_12dev\tests\..\tests\window\WINDOW_MULTISAMPLE.py", line 129, in test_multisample
self.render()
File "D:\hg_externals\pyglet_12dev\tests\..\tests\window\WINDOW_MULTISAMPLE.py", line 93, in render
self.win.switch_to()
File "D:\hg_externals\pyglet_12dev\tests\..\pyglet\window\win32\__init__.py", line 305, in switch_to
self.context.set_current()
AttributeError: 'NoneType' object has no attribute 'set_current'

The test session not aborted itself there, but later crashed on another test.

Full log for this session attached as pyglet.4_crash_win32_ati4650_drv_cat11.5.log 

At this point I switched to the last released stable drivers, but I will follow in another post

--

claudio

 

 

[claudio canepa]

On Fri, Oct 28, 2011 at 11:55 AM, claudio canepa <ccan...@gmail.com> wrote:

[snip] 

At this point I switched to the last released stable drivers, but I will follow in another post

--

claudio

 

Switching to the last stable ATI catalyst drivers: Catalyst Version    11.9

* the long waiting time at test begin continues

* I notice a difference in behavior:

  - with 11.5 drivers: no window visible except the ones spawned to do interactive tests

  - with 11.9 drivers: a small window pops up just after the initial long wait and remains along all tests

* failure at test window.CONTEXT_SHARE (traceback, see log), and now we have three pyglet windows showing: the small initial and two more, probably the one than the test failed to close. The next 2 test (vsync and multexture) performs fine, but got another traceback after that. (and the first style showed, but the second style window don't showed)

At this point I quit the session.

log attached as pyglet.11._win32_ati4650_drv_cat11.9.log

It will be useful to see results from other people on win32, any hardware.

--

claudio 

[claudio canepa]

win xp sp3, python 2.6.5, nvidia 6200

There are also failures.

for the log look at the attachments in Issue 539:     Usability issues with tests

--

claudio