[prev in list] [next in list] [prev in thread] [next in thread]
List: helix-video-dev
Subject: [Video-dev] Re: [Player-dev] CR-Client: Fix testing blocker.
From: Peter Fawcett <pfawcett () real ! com>
Date: 2005-03-14 18:01:19
Message-ID: ace1df7b4cf1c25133a2c4828897244a () real ! com
[Download RAW message or body]
Can do for the bug.
I'll be in all day today and tomorrow if you want to stop by for a
repro or want to run debug bits. If you have anything you'd like me to
try, let me know.
Thanks,
Peter
On Mar 10, 2005, at 3:48 PM, Greg Wright wrote:
> Peter Fawcett wrote:
> [snip]
>> - I still get REALLY bad frame rates for each video after the
>> first. Perhaps due to the multi window issue? First clip plays
>> really well. Remainder of the play the black nothing for about
>> 5-10s, then they will blit one or two frames and the either freeze
>> again, or just go back to black. Often once it's in a frozen state
>> and or bilitting frames, I can move the window and the view stays in
>> place with the window frame moving about it. If it blits again, it
>> will put the next frame in the window but then freeze again.
>
> This definatly sounds like a site problem, or at least how the site
> is being used. I will stop by and take a look, but I think you can
> go ahead and log a bug on just this one part of the problem.
>
> --greg.
>
>
>> - I got one seg fault, first run. But I have not been able to
>> repro it and thus I don't have a catchsegv file for you in this
>> case. I'll keep trying to get another seg fault.
>> Other than the video issue and the multiple video windows, this is
>> looking pretty good from our perspective.
>> Peter.
>> On Mar 10, 2005, at 9:29 AM, Greg Wright wrote:
>>> Peter Fawcett wrote:
>>> [snip]
>>>
>>>> Bad News: First video clip to play, plays very well and looks as
>>>> good as the Win32 client (high frame rates and so on).
>>>> Second and subsequent video blits nothing but black for usually the
>>>> first 5-10 seconds and then only blits one or two frames.
>>>> When this is going on, there is a very bright green field in the
>>>> old player windows (it's spawning a new window each time (though
>>>> it's supposed to use the same player each time (on Windows this
>>>> case would only spawn one window for all tests to use)))
>>>> If I mouse over the video while it is playing I get really ugly
>>>> green fields around the cursor.
>>>> I see output to the term that's not part of my standard out debug
>>>> code so I'm assuming it's from the site. It says:
>>>> "Class = TrueColor, Depth = 24"
>>>
>>>
>>>
>>> The green stuff you see is because you were using a debug build of
>>> the vidsite, it happens to be the overlay color-key. In release
>>> builds
>>> we use a color close to black, in debug builds I change the color-key
>>> to tell what color depth your screen is running at:
>>>
>>> Red: 16 bit
>>> Green: 24 bit
>>> Blue: 32 bit
>>>
>>> Anyway, that is why you are seeing green. Well the reason you are
>>> seeing green is because something has gone wrong with the overlay
>>> video, it should be drawn right on top of that green square. There is
>>> always a chance that my local debug build of vidsite will not work
>>> with your bits (I built it off of HEAD). So, the first thing I would
>>> try is a build with the fix checked in (I checked it in yesterday).
>>> If it is still broken you will see the same stuff except the green
>>> will be black, 0x01020304 actually.
>>>
>>> [snip]
>>>
>>>> Very close though. Greg can I do anything to help? Would you like
>>>> to stop by some time tomorrow and I'll show you what I'm seeing?
>>>> (I'm meeting free tomorrow other than a automation status meeting
>>>> at 11am).
>>>
>>>
>>> Try a build from last night first so that everything is build from
>>> the same branch. If it still doesn't work, send me a link to the
>>> build
>>> and I will install it and check it out. Is there an easy way to repro
>>> the bug with all the python testing stuff? If not, I may need help
>>> setting up that stuff.
>>>
>>> --greg.
>>>
>>>
>>>> Thanks for your assistance in getting this up and running.
>>>> Peter
>>>> On Mar 9, 2005, at 10:29 AM, Greg Wright wrote:
>>>>
>>>>> This fix is now checked into HEAD and 150Cay.
>>>>> --greg.
>>>>>
>>>>>
>>>>> Peter Fawcett wrote:
>>>>>
>>>>>> Sweet. Will do, I'll send out results once I give it a go.
>>>>>> Thanks Greg.
>>>>>> Peter
>>>>>> On Mar 9, 2005, at 8:13 AM, Greg Wright wrote:
>>>>>>
>>>>>>> You can try this real fast:
>>>>>>>
>>>>>>> ftp://172.23.129.253/pub/vidsite.so
>>>>>>>
>>>>>>> --greg.
>>>>>>>
>>>>>>>
>>>>>>> Peter Fawcett wrote:
>>>>>>>
>>>>>>>> This does repro with the current internal branch.
>>>>>>>> I have a fully setup box if any one wants to test bits.
>>>>>>>> Or feel free to send me a link to bits and I'll give them a
>>>>>>>> whirl...
>>>>>>>> Peter
>>>>>>>> On Mar 8, 2005, at 3:00 PM, Ryan Gammon wrote:
>>>>>>>>
>>>>>>>>> Automation testing should be running against our "current"
>>>>>>>>> branch, which is 150Cay.
>>>>>>>>>
>>>>>>>>> Looks good to me too.
>>>>>>>>>
>>>>>>>>> Greg Wright wrote:
>>>>>>>>>
>>>>>>>>>> Bug 3870: Seg Fault in video/site when playing
>>>>>>>>>> video with PyPlayer [automated testing bug]
>>>>>>>>>>
>>>>>>>>>> Synopsis:
>>>>>>>>>>
>>>>>>>>>> I think this will fix Bug 3870. I don't have PyPlayer but I
>>>>>>>>>> think I was able to repro the crash (this at least
>>>>>>>>>> certainly
>>>>>>>>>> fixes a 'splay' crash on playback end.
>>>>>>>>>>
>>>>>>>>>> Fix:
>>>>>>>>>>
>>>>>>>>>> Just check for NULL.
>>>>>>>>>>
>>>>>>>>>> Files Modified:
>>>>>>>>>>
>>>>>>>>>> platform/unix/unixsite.cpp
>>>>>>>>>>
>>>>>>>>>> Platforms affected:
>>>>>>>>>>
>>>>>>>>>> Anything that uses the Unix Site.
>>>>>>>>>>
>>>>>>>>>> Branch:
>>>>>>>>>>
>>>>>>>>>> HEAD & 150Cay only (I don't have any idea what branch the
>>>>>>>>>> bug was on).
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> --greg.
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Index: platform/unix/unixsite.cpp
>>>>>>>>>> ==============================================================
>>>>>>>>>> == == =
>>>>>>>>>> RCS file: /cvsroot/video/sitelib/platform/unix/unixsite.cpp,v
>>>>>>>>>> retrieving revision 1.11
>>>>>>>>>> diff -u -w -r1.11 unixsite.cpp
>>>>>>>>>> --- platform/unix/unixsite.cpp 9 Jul 2004 18:35:18 -0000
>>>>>>>>>> 1.11
>>>>>>>>>> +++ platform/unix/unixsite.cpp 8 Mar 2005 22:32:25 -0000
>>>>>>>>>> @@ -384,7 +384,9 @@
>>>>>>>>>>
>>>>>>>>>> void CHXUnixSite::_DestroyScrollButtons()
>>>>>>>>>> {
>>>>>>>>>> - Display* pDis = (Display*)GetWindow()->display;
>>>>>>>>>> + HXxWindow* pWin = GetWindow();
>>>>>>>>>> + Display* pDis = (Display*)(pWin?pWin->display:NULL);
>>>>>>>>>> +
>>>>>>>>>> if( m_ScrollHorizButtonL && pDis )
>>>>>>>>>> {
>>>>>>>>>> XLockDisplay(pDis);
>>>>>>>>>> @@ -394,6 +396,7 @@
>>>>>>>>>> XUnlockDisplay(pDis);
>>>>>>>>>> m_ScrollHorizButtonR = 0;
>>>>>>>>>> }
>>>>>>>>>> +
>>>>>>>>>> if( m_ScrollVertButtonT && pDis)
>>>>>>>>>> {
>>>>>>>>>> XLockDisplay(pDis);
>>>>>>>>>> @@ -759,11 +762,15 @@
>>>>>>>>>> void CHXUnixSite::_Destroy(HXxWindow* pWindow)
>>>>>>>>>> {
>>>>>>>>>> _DestroySliders();
>>>>>>>>>> - Display* pDis = (Display*)GetWindow()->display;
>>>>>>>>>> + HXxWindow* pWin = GetWindow();
>>>>>>>>>> + Display* pDis = (Display*)(pWin?pWin->display:NULL);
>>>>>>>>>> + if( pDis )
>>>>>>>>>> + {
>>>>>>>>>> XLockDisplay(pDis);
>>>>>>>>>> XDestroyWindow( pDis, (Window)pWindow->window );
>>>>>>>>>> XUnlockDisplay(pDis);
>>>>>>>>>> }
>>>>>>>>>> +}
>>>>>>>>>>
>>>>>>>>>> void CHXUnixSite::_SetSize(HXxSize size)
>>>>>>>>>> {
>>>>>>>>>>
>>>>>>>>>> _______________________________________________
>>>>>>>>>> Player-dev mailing list
>>>>>>>>>> Player-dev@helixcommunity.org
>>>>>>>>>> http://lists.helixcommunity.org/mailman/listinfo/player-dev
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> --
>>>>>>>>> Ryan Gammon
>>>>>>>>> rgammon@real.com
>>>>>>>>> Developer for Helix Player
>>>>>>>>> https://player.helixcommunity.org
>>>>>>>>>
>>>>>>>>>
>>>>>>>
>>>>>
>>>
>
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic