Login
Username:

Password:

Remember me



Lost Password?

Register now!

Sections

Who's Online
127 user(s) are online (66 user(s) are browsing Forums)

Members: 0
Guests: 127

more...

Headlines

Forum Index


Board index » All Posts (Fab)




Re: Odyssey 1.23 progress
Just popping in
Just popping in


@samo79

In fullscreen mode, you shouldn't be even able to click the zoom gadget, so it's a mui os4 issue to me.

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@kas1e

On MorphOS we have grunch (amiupdate-like) which reports when userscripts are outdated, so it's not really an issue there, users are notified when a new one is available.

But anyway i also planned to have autoupdate feature in odyssey userscripts at some point, this is why i already include version and url in these scripts.

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@tlosm

There's really no need to wipe clean the HD when updating MorphOS, since it's designed to be updated properly.
On the other hand, chrysalis pack may cause some weird things sometimes.

Anyway, good if it finally works as intended.

On a G5@2.3GHz, with Odyssey, Youtube 720p is playing fine in window mode, and 1080p is generally fine in overlay mode.

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@tlosm

Your test is most likely flawed for several reasons (and most likely outdated youtube_fullscreen_fix.js). Just repeat the steps above carefully.

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@kas1e

Sigh.... You really don't make it easy.

Youtube is a moving target, it's really complicated, because they have many different cases depending on videos, and unfortunately you're very confused about all this. So just PLEASE do use the same scenario and configuration as below.

Get latest scripts from http://fabportnawak.free.fr/owb/scripts/
http://fabportnawak.free.fr/owb/scripts/YouTube.js
http://fabportnawak.free.fr/owb/scripts/YouTube_Fullscreen_Fix.js
Copy them to odyssey/scripts/ directory.
In scripts manager, ADD these two files (and remove them before if they're already there).

Then go to http://www.youtube.com/html5 and enable HTML5 player

Then either restart Odyssey or create a new tab/window before proceeding.

Now we will see four different cases:
1. http://www.youtube.com/watch?v=hUoJBerFDsA
HTML5 mode is not available with this link, but youtube.js works just fine, and so does overlay (which ALWAYS works with builtin player, as long as video is available).

Resized Image

Resized Image


2. http://www.youtube.com/watch?v=dQiimhYI-8c
HTML5 is available, Youtube HTML5 player just works, and so does overlay mode.

Resized Image

Resized Image



3. http://www.youtube.com/watch?v=BBtXL9NAPXk
You have to disable youtube.js userscript and spoof as chrome (menu -> settings -> spoof as -> chrome 32)
I get an annoying ad, and then the video plays, and overlay works just fine when pressing fullscreen button.

With OWB 1.23, you might not see player controls, because webp wasn't enabled, and when spoofing as chrome, webp is needed.

Resized Image

Resized Image

Resized Image


4. http://www.youtube.com/watch?v=yR6A-Bk9eZQ
HTML5 mode is not available and spoofing as Chrome won't even help in this case.
In this case, the easiest alternative is to spoof as Ipad. Fullscreen button will work fine in this case too.

Resized Image

Resized Image



Case 1 and 2 are the most common by far.

Case 3 (which is a subset of case 4) is rather new to me.

Case 4 might be related to VEVO links. In this case, the direct link extracted by the script just don't work and return an empty 403 result (i.e http://r8---sn-vg5obx-25ge.googlevide ... 93d7994&ms=au&sparams=gcr,id,ip,ipbits,itag,ratebypass,source,upn,expire&source=youtube&fexp=931320,941509,930102,916624,937417,913434,936910,936913&ratebypass=yes&ipbits=0&mt=1391777825&key=yt5&ip=90.59.141.178&upn=2uVBODzrv7M&fallback_host=tc.v21.cache3.googlevideo.com&quality=medium&itag=18 )
But spoofing as ipad works in these case. I might investigate what happens exactly and try to modify youtube.js to handle this case, but it's related to the signature/upn parameters, and it might not be so easy.


Read this post many times before just saying "it doesn't work". The situation is a bit more subtle as you can see.
And about comments about non working overlay,i just reject them, plain and simple. Overlay DOES work, when scripts are enabled AND video is available. That's it.

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@Fairdinkem, MickJT

Well, it was edited since... It was something like "I raised the issue of randomly working videos during the bounty". But since this forum doesn't even show when a post was edited, you can go unnoticed and edit your posts anytime.

@Fairdinkem

I'm not sure i want to go into details again about how these 2 scripts work, but the bottom line is: HTML5 video works, and overlay mode works. In between, there can be Youtube trying to break things in any way they can, and when it happens, the two userscripts need to be updated.
For instance, they recently prevented the scripts from working when clicking the related videos links, and now, it seems they also won't work when coming from search results. So I just updated these two scripts:
http://fabportnawak.free.fr/owb/scripts/YouTube.js
http://fabportnawak.free.fr/owb/scripts/YouTube_Fullscreen_Fix.js

Update them, and in scripts manager window, you REMOVE and ADD these two scripts again, so that the URL whitelist gets updated.

But to make it simple for you, just spoof as Chrome 32 (just for youtube in URL settings or globally, as you prefer), and just enable the YouTube_Fullscreen_Fix.js userscript. It should work in all cases then (with ads).

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@kas1e

Look, if the *builtin* player doesn't go fullscreen when you click fullscreen button on MorphOS (assuming there's a video playing), you have a serious setup issue, or you already have a video layer opened in another tab or application. It just works for everyone out there (provided there's overlay support of course), so it's on your side (especially the freezing issue), and i'd just suggest you boot from a plain morphos cd.

Now about Youtube direct links failing, it's very well possible they act up randomly, but again, it's not related to odyssey. That's just Youtube's way, and they might want to protect themselves from automatic downloaders in various ways, depending on videos, servers, time, whatever.

About your particular link ( https://www.youtube.com/watch?v=BBtXL9NAPXk ), it seems to be something new they have still not widely deployed, and it seems the point is to have ads even in HTML5 mode. The good news is that if you just spoof as Chrome (with no script enabled) it will work just fine, and you'll get a commercial ad, ...

About the userscript status icon it's like that at the bottom, and it will show which userscripts are running for a given page when the mouse is over the icon:

Window mode
Resized Image


Overlay mode (which can't be grabbed)
Resized Image


@Fairdinkem

I suggest you refrain from commenting on something you don't know about. Odyssey supports HTML5 media just fine (and even better than some other browsers since it supports all codecs combinations).
The problem is YouTube is just not very HTML5 compliant and likes to break its support every now and then. The problem would happen similarly in any other browser as well (FireFox, Safari, Opera, ...), but of course you don't even notice it since Flash is generally used on these platforms.

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@kas1e

It's kinda obvious the fullscreen button (be it in Youtube HTML5 player or builtin HTML5 player) has no effect on OS4, since you haven't ported the overlay code yet (and it might require significant changes, considering the p96 pip API which seems rather limited). So, again, the fullscreen button won't do anything before the overlay code is adapted.

Also, Youtube HTML5 player needs the youtube_fullscreen_fix.js script for another reason, but i won't explain it since the basics aren't really clear for you yet.

About this "other player"... In Flash replacement mode, when Youtube HTML5 player is not available, the HTML5 builtin player appears, and it works just fine (and quite faster than Youtube's HTML5 player even). So if it doesn't work for you (even on MorphOS), it's most likely a problem with your network or with the Youtube servers you connect to (which is nothing surprising actually). I'm fairly sure about this, because the youtube.js script basically just retrieves direct links (the same way as getvideo or smtube would do) and it creates a html5 video object with this URL as source.
You can see the video direct URL in several places (messages window or contextmenu copy media location, webinspector and so on...). If it doesn't work, it means the video URL returns empty contents, which is obviously not Odyssey's fault.

PS: it seems you have a bug or maybe you forgot to include some data files, because when a userscript is running, a small status icon should appear in the status bar, which is not the case in your screenshots (or you didn't add the youtube.js userscript?)

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


You can enable autoscrolling by doubleclicking middle mouse button. In that case, a "move" cursor is shown.

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@kas1e

The related code setting the path from settings probably got erased by a merge at some point in curl backend file. Trivial to fix.

Anyway, with defaults settings in CURL, it works fine, which is what matters most.

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@kas1e

The crt path option of course works too in Odyssey. It may be another issue in your curl build. But first we had to rule out other potential issues by setting default path properly.

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@kas1e

You have to specify the bundle location, it won't default to "PROGDIR:curl-ca-bundle.crt" magically. IIRC, it should be CURL_CA_BUNDLE define in config.h

I don't know about the configure options, i made my own makefile/config.h for curl, i don't need their configure crap.

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@kas1e

Mmh, no idea, i built Odyssey with Curl 7.34 and OpenSSL 1.0.1f, and i don't have such issue.

I'd say your curl build doesn't find the proper ca bundle file anymore, especially since it works if you specify it with an option.

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@kas1e

Well, did you change all the functions as i wrote above?

I remember icu had fun replacing : with /, which is why i had to disable that silly code.

Quote:

And finally, in udata.cpp, around line 1137, you might want to disable that code:
if (U_FILE_SEP_CHAR != U_FILE_ALT_SEP_CHAR) /* Windows: try "foo\bar" and "foo/bar" */
{
/* remap from alternate path char to the main one */
CharString altSepPath;
...
}

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@kas1e

I don't have to call this function, since i set the proper data directory in ICU itself (in putils.cpp).
And that's u_getDataDirectory() that sets it itself with the default directory define, anyway (path=U_ICU_DATA_DEFAULT_DIR).

About this timezone stuff, i also had to undef as you did, but it's unrelated to your data files issue.

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@kas1e

I compile ICU myself with the configure line and changes i gave above. And it produces a very small libicudata.a (856 bytes, with just stubdata.ao:
00000000 R icudt49_dat as symbol).

Just in case, check what the file icu/source/config/mh-unknown looks like for you. In my case, i had to copy it from mh-linux, but it doesn't seem related to datafile package in any case, anyway.




Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@kas1e

It's obvious snoopy reports access to the icu data files, since i explicitly check for them a startup. But what matters is if they're actually found by ICU when it looks for them. And it looks for them at runtime, when you access a page and render some text, not right at startup. So just check for snoopy when you actually access a page.
You can disable the function in icufiles.cpp for now if it makes it easier for you to understand what's going on.

And regarding build options, i used something like
./configure --disable-tests --disable-samples --host=ppc-morphos --disable-threads --disable-shared --enable-static --disable-dyload --enable-release --with-data-packaging=files

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@tlosm

It just doesn't make *any* sense at all. But what about checking if you have framedrop option enabled? If you do, just disable it.

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@tlosm

The issue is most likely not MPlayer or its decoding library. But to be 100% sure about it, give us a link to a video file exposing this issue.

But i repeat it again, on my peg2, on my mac mini, on my powerbook, on my pmac g5, a movie will take as much CPU as it needs to be decoded (not more, not less). So if a hd movie only uses 30% of your CPU on your peg2, you HAVE a SERIOUS issue, quite specific to your hardware/software driver combination. And you're the only one to have ever reported that, ever. So, what about trying again with your radeon 9250 instead, as a test?

And no, a "new decoding library" won't help in any way, since your issue is most likely unrelated to MPlayer/ffmpeg at all...

And the fact your videos work on PS3, PC, Rasp is entirely irrelevant (especially for the machines that have accelerated hardware decoding...). I'll add one: my videos work great on my peg2, doh.

Go to top


Re: Odyssey 1.23 progress
Just popping in
Just popping in


@tlosm

ffmpeg is used for almost everything in mplayer (and thus smtube) and odyssey. Be it mkv, webm, mp4 or whatever you want.

Regarding your desynchronized issue, well, i asked you for a specific file that would expose that issue, because it's clearly not the case with the vast majority of files i've tested.

But if you have the same "performance" on an efika and on a pegasos2, you obviously have a very serious issue with your peg2. It's of course not a decoding issue. Maybe you have an issue with this radeon for some reason. What about trying your 9250 in your Peg2?

I don't get what you mean with your G5 reference. My main system is a pegasos2, and i have absolutely no such issue on this system. Of course a G5 will perform better and allow 1080p h264 playback, but a Peg2 will still happily play perfectly synchronized 480p contents (or even 720p/1080p for less demanding codecs).

Go to top



TopTop
« 1 2 (3) 4 5 6 ... 10 »




Powered by XOOPS 2.0 © 2001-2023 The XOOPS Project