As for the current version, well what we miss are the speelcheck library
You know for years we don't have sources of it, only if you already ask author and got them ? Try today, maybe you will be lucky.
Quote:
and the of course a fixed video support
Is there any 100% reproducable test-cases which show the problem without "sometime it crashes" or "in some sometime condition dunno when but from time to time" ?:))
@Petrol Quote:
Is code need to be optimised to have an altivec version of odyssey or is it a thing that could be done during compilation?
General optimisation such as -O3 are already in odyssey, but for altivec based optimisation need altivec based code, and that usually mean 2 different binaries and it most of time make no sense, only with case of video decoding maybe, but then, instead of worry about altivec, its better to add some radeonHd decoding features in, as altivec give a little, and today not all machines have it.
Quote:
A fix for the Odyssey media-player would be welcome ;)!
If you have some 100% reproducable test case, that will be a start.
@Trixie I do not remember, but didn't we discuss some years ago, that something not done in odyssey in terms of application library support, and you are willing to help with ? (or it was about some other application?)
only if you already ask author and got them ? Try today, maybe you will be lucky.
Ok, mail sent ... let's see if he reply ...
Quote:
Is there any 100% reproducable test-cases which show the problem without "sometime it crashes" or "in some sometime condition dunno when but from time to time" ?:))
There are no special condition ... sometimes it crash while streaming an HTML5 video, sometimes not, but it also drain the whole system memory and this is pretty obvious when, for example, you are listening a radio streaming ... after a while the whole system hang because Odyssey will drain the whole memory ! .. so the more phisical memory you have mounted in your system, more time you can listen a radio ...
Actually the only method is to close the browser and then reopen the page, doing this the memory will be freed and you can restart listening )
As for the HTML5 video in general, as far as i remember Odyssey used Cybergraphics from MorphOS, so probably we are actually emulating it via Picasso96, that means we doesn't have overlay, and we poor users with old Radeon 9200 equiped system need to survive with that sloooower video without fullscreen
Aniway, if you really need a testcase i vote for the radio first ... try to open a radio streaming source , for example one of these:
sometimes it crash while streaming an HTML5 video, sometimes not, but it also drain the whole system memory and this is pretty obvious when, for example, you are listening a radio streaming ... after a while the whole system hang because Odyssey will drain the whole memory ! .. so the more phisical memory you have mounted in your system, more time you can listen a radio ...
That what i expect sadly : no test case, but "sometime it crash" => how to fix it if i can't reproduce it on my machine ?
Quote:
As for the HTML5 video in general, as far as i remember Odyssey used Cybergraphics from MorphOS, so probably we are actuallt emulating it via Picasso96, that means we doesn't have overlay,
No, not emulating of cgx over picasso96 mean we don't have overlay, its we don't have overlays code at all there , and i do not have sources of the players they use on morphos which they wrote specially for, and which was writen not by Fab, and it wasn't shared.
Quote:
and we poor users with old Radeon 9200 equiped system need to survive with that sloooower video without fullscreen
I on x5000 also don't have fullscreen, and also it plays slow.
But you can be sure that i will not worry about writing new overlay code for old cards. Even and if i ever will do anything there, is adding radeonhd/warp3dnova support, but not spending limited time and motivation for supporting old gfx cards.
I’ve worked on the code in deadwood’s repo. Trying to gather all dependencies and build it. I’ve setup a docker container with a cross-compiler setup, SDK and a script that downloads and installs the dependencies. See the readme for instructions. Anyone with docker should be able to try it out.
For now there's a missing dependency, libicu, in my list, but I’ll try and fix that tonight.
I’ve merged the changes done by the Aros team, and my plan is to have it build, and then try and update to the next revision that the TenFourFox/leopard ppc-safari has a JIT-patch for. I’m however currently stuck on freetype/fontconfig dependency, which I don’t know if the latest libcairo by salass00 has support for (the port only mentions an Amiga font backend).
LOL I meant: yes sometimes i got DSI when i try to watch a youtube video, BUT ACTUALLY my MAIN problem is the usability ! ... for now that HTML5 video are really too slow to be usable, so when i need i use something else --> SMTube
That's why i was not so specific ... of course if you need i can send you some crashlogs aswell (as far as i have) but i don't consider that sporadic crash as a priority
Quote:
, and i do not have sources of the players they use on morphos which they wrote specially for, and which was writen not by Fab, and it wasn't shared.
Mmm so what is the player we are using now ? I though it was the one developed by Fab
Quote:
But you can be sure that i will not worry about writing new overlay code for old cards. Even and if i ever will do anything there, is adding radeonhd/warp3dnova support, but not spending limited time and motivation for supporting old gfx cards.
Yeah, i can understand this ... however that means 40% and more of OS4 users out
Aniway, for now let's forget these video problems .. if you agree let's testing atleast these radio streaming crap ..
Mmm so what is the player we are using now ? I though it was the one developed by Fab
They use 2 players :
one inbuild in webkit (as we use) which render through Cairo, and which is why it slow things even on x5000, i.e. when i play youtube on x5000 i also have about 15 fps. I do not know how to speed that one up, maybe only by writing harware accelerated cairo backend, dunno.
And another one, which is triggered when you press full view or fullscreen do not remember, and where they have cgx overlay, and that part is not writen by fab if i remember right, and wasn't shared.
So if i remember right (but that need to be cheked, or some morphos users can fix me), in the normal, usuall window mode, youtube also slow for morphos users, as use playback through cairo. But once they tick "fullscreen", then fullwindow open, and this one is "overlay", and then things go fast. But in window mode it the same suck for them as for us.
Quote:
if you agree let's testing atleast these radio streaming crap ..
If it just JS, then its probabaly just webkit issues or something, which i can't fix.
If it just JS, then its probabaly just webkit issues or something, which i can't fix.
Try this one, is for sure an HTML5 one --> https://www.muses.org/ The player is at the top-right side of the page, just press play and see if your memory decrease or not
Try this one, is for sure an HTML5 one --> https://www.muses.org/ The player is at the top-right side of the page, just press play and see if your memory decrease or not
Don't looks like html5 too. Need just normal page, with usuall html5 player (youtube ?) , which will show that once i play video, everything crashes.
Muses Radio Player - The HTML5 MP3/OGG/AAC Live Stream Player! https://www.muses.org/? Now the HTML5 version of Muses is the default option, and fallbacks to Flash if needed. It also adds more translations and implement many of the Flash version
And looked at the source to me it seems an HTML5 player
Quote:
Need just normal page, with usuall html5 player (youtube ?) , which will show that once i play video, everything crashes
No need imho and it will be even confusing because:
1 . You don't know if is the video to crash or if is it something else involved ..
I mean usually HTML5 video in youtube crashed (when they crash) after 1 or 2 minutes of playing, maybe a bit more but more or less this is the range of the time ... so it can't be the same problem of the radio ... with the radio i can stream a audio for 15 minutes, 30 minutes and even more time without any problem.. until they crash because memory ends ... with video involved instead you don't know what happened, sure maybe still also similar memory problem with videp, but i believe video crashed also for other reasons
2 . Even because the problem seems the memory drainage, eventually it crashed after a while (when the memory finished) and not instantanely
3 . To do a test using Youtube i will need to found a "video" that only contains an audio source, but not a video .. but in Youtube everything in reality is a video !
Wait more, 10 minutes is not enough ... in base of your memory (i presune you have 1 or 2 GB right ?) you need to wait atleast 1 or 2 hours ... Maybe try tonight, open the audio stream and leave it running till the morning
P.S. Of course if you a tool to monitorize the memory in real time would be better
Not sure if i want to dig into that to waiting hours for just trigger the bug: is too much for me today.
It's not only because a casual bug that may happens from time to time, this issue indicate that the player is almost broken by design because it costantly eats memory ... in this condition streaming something for some time become an hell because you will never know how much time you left until you need to quit OWB (or restart your machine because it fucked up)
Quote:
Just to clarify: your problem is constant decreasing of memory when you use html5 mediaplayer be it audio or video+audio, right ?
Yes. for "audio only" this problem always happen ... 100% tested Instead for "Audio + video" i don't know ... because in general i always open small videos for the reason you know, so i can be sure at 100% if we have the same memory decrease also with them ...
Kas1e, thank you very much providing an update on the situation and for your work on previous versions.
Regarding the built-in media players, rather than trying mess with the kludges within the browser, would it be possible to just turn off all Video & Audio playback within the browser and then we could re-route those to reliable external players like FFMpeg, MPlayer and AmigaAmp?
How about a on/off checkbox on the Media settings page?
That would avoid a ton of crashes and memory leakage.
I've tried resetting the MIME types, but they have been ineffective diverting video from playing within webpages.
OTOH, it would be fantastic if the security side of things could be updated. Is that built into webkit or separate and updatable?
News feed
OS4 Feedback forum
OS4Depot Feedback forum
Amigabounty forum
OpenAmiga forum
AmiCygnix forum
ABC shell forum
AmiKit forum
Cinnamon Writer forum
CodeBench forum
Digital Universe forum
Dopus 5 forum
E-UAE forum
Gnash forum
Ibrowse forum
JAmiga forum
Odyssey forum
OWB forum
Qt forum
SmartFileSystem forum
Timberwolf forum
TouchDevice forum
TuneNet forum
Unsatisfactory Software forum
Donate
)
