Mitigation for Video "Player" bug on Linux

[tperry2x]

I can't solve it, as I don't think I have the expertise.
However, I can mitigate the issue for linux users.

Please use the following patch:

linux-player-mitigation.7z

(250.58 KiB) Downloaded 62 times

What this does:

screenshot.png (65.8 KiB) Viewed 1203 times

It gives Linux users the option of disabling the "Player" object, as this will likely cause a crash if most Linux distros currently try and use the tool. I've made this an option rather than just deleting it from the IDE altogether, as I understand it works for Mac and Windows.

List of bugs, which I'll change as solved: (bug tracker if you like)
https://www.openxtalk.org/forum/viewtop ... 4441#p4441

You "might" have some luck installing the following dependencies (and probably better to get updated ones via your package management system), but these [kind of] make the player not crash. Still doesn't do much though.

linux-required-deps-for-video.7z

(3.5 MiB) Downloaded 73 times

[richmond62]

About 19 years ago (!!!) I was designing some EFL software for Linux (probably Ubuntu 4.0 'Kinky Kangaroo') And found that Runtime Revolution would only play sound files at their normal speed if they were first slowed down to half their normal speed using an external sound editor: obviously utter bollocks.

About once a year, until 9 years ago when I gave up, I would "try to be clever" with either a video file or a sound file and RR/LC on Linux, to no avail.

Now whatever the developers of LC may have claimed Linux comes a very long way down their shopping list, and what it offers is a subset of what LC can do on Macintosh or Windows.

Until recently they claimed they were testing their Linux versions of a 2014 version of Ubuntu, which really says all we need to know about how they view Linux.

[FourthWorld]

Until recently they claimed they were testing their Linux versions of a 2014 version of Ubuntu, which really says all we need to know about how they view Linux.

After Fraser and Peter left, there's no one there with a passion and depth of understanding for Linux.

I still use it on Linux, but as Linux moves forward without them each year I find another issue creeping in to make the experience less on par with their macOS build.

FWIW the most recent Release Notes still list compatibility with only Ubuntu 14 and 16.

I'd press them on this, but from a business standpoint it might almost make more sense for them to drop Linux altogether. That audience is small, and the percentage inclined to pay high prices for proprietary software much smaller.

[tperry2x]

After Fraser and Peter left, there's no one there with a passion and depth of understanding for Linux.
I still use it on Linux, but as Linux moves forward without them each year I find another issue creeping in to make the experience less on par with their macOS build.

I'd say from at least a developer point of view, the linux build of OXT lite is by far the best behaved of the three. On Windows I have UAC permissions issues with revCopyFile and windows defender flags this on windows 11 erroneously.
On MacOS I have an absolute pain each time it comes to update with Codesigning issues and Gatekeeper - so much so that I have to put special workaround instructions in the DMG. Far from ideal, but needs must.
Although as you rightly say, Linux is moving on, it provides more of a static reliable base and the UI is rendered a lot more consistently. There are issues there as well (the browser widget and the "player" tool), but aside from this - it's the much more pleasant platform to develop on.

FWIW the most recent Release Notes still list compatibility with only Ubuntu 14 and 16.

That might be a good thing, as there are plenty of people running older distros, so if this is seen as a minimum requirement then that's quite handy for a lot of folks.
The fact that OXT is pretty much a portable program (you don't really need any dependencies aside from what come bundled with anything capable of running a window manager), means it's more compatible for older 64-bit capable hardware too.

I'd press them on this, but from a business standpoint it might almost make more sense for them to drop Linux altogether. That audience is small, and the percentage inclined to pay high prices for proprietary software much smaller.

We shouldn't really be concerned with business standpoints as we are not making any money from this. That's not why I'm doing it, this is just a side-project for me. I don't really mind what LC do to be honest as if LC ceased to exist tomorrow, then we'd just carry on.

Don't get me wrong, I agree with you. Linux is always seen as an afterthought and I would love this to not be the case. It's probably merely a happy coincidence that OXT works as well as it does on Linux (more like a miracle) when you consider all the different distros out there.
Many of these issues should be solved at engine level, but this would be why we'd need support for the engine which might involve money.

Just as the IDE and the ENGINE have two separate repos in GitHub, I think we need to treat both separately. Keep OXT free with no mention of us asking for funds, but then in the engine, and these forums, perhaps be open to donations and/or offers of help regarding the engine.

[richmond62]

a business standpoint

Surely that is of concern for the people at LiveCode, but not for OXT, unless, like LibreOffice and so on, we aim to attract some revenue.

I believe that when some people become aware that OXT is;

1. FREE.

2. Can offer 98% of what LC can offer (i.e. no code protection).

3. Retains the LC 7-9 GUI (not everyone is going to go for the LC 10 'thing' (especially for use of computers with a single monitor)).

There might be quite an uptake.

At that point we have to work out something like the Mozilla foundation to handles monies and so on.

BUT, our first concern should NOT be business-based.

---------------------

Recently I had a "set to" with a slightly foolish mother of a wee girl who started coming to classes in my ESL school fresh from Primary school somewhere in England.

Yon wain starts greeting because the classes "are not all fun like the school classes in England", and the mother (who knows nothing about either learning or teaching a foreign language to wains) starts complaining.

I pointed out that, at my EFL school we do have a lot of fun, but as a side-effect of the way we study the language, not fun in and of itself: after all parents, paradoxically enough, are paying me to get English inside their children's noddles, and NOT run a sort of 'Zoo for children'.

I also pointed out that the British Government was currently holding a probe into English Primary Education as there was some suspicion it had swung too far away from the didactic way classes were conducted in the 1950s and 1960s to a free for all, fun at all costs way that was almost as counter-productive as the didactic way.

The child remains at my school, and the mother gulped and waited (as I suggested) a further 3 weeks [FFS how can you judge a classroom setting on 2 classes?] and then sent me a bottle of wine.

--------------------

So, too OXT: anything that is attractive (in a business fashion) about OXT should be a side effect of all the work that is being put into it, not the core value of it.

[OpenXTalkPaul]

You "might" have some luck installing the following dependencies (and probably better to get updated ones via your package management system), but these [kind of] make the player not crash. Still doesn't do much though.
linux-required-deps-for-video.7z

This is the main reason that I think that AppImage format is such a great idea for Linux(es) apps: if we can collect all of the exact library dependencies that are needed then we can include them in the AppImage, including the exact versions of those that are required, theoretically that would run properly on any distro.

[tperry2x]

This is the main reason that I think that AppImage format is such a great idea for Linux(es) apps...run properly on any distro.

This certainly has my vote.

[richmond62]

And mine.

It is a distinct pain in the bum if, having installed one's distro of choice on one's 'heap of junk', that one finds that:

1. OXT won't behave itself. Or

2. OXT will only behave itself after you have jumped through however many hoops installing libraries amidst dependency hell.

[OpenXTalkPaul]

Here is the macOS X86_64 build of OpenXION as a standalone 'native' executable: https://github.com/PaulMcClernan/openxi ... cOS_X86-64

I will try to install GraalVM / Nativefier on Linux and Windows when I can find the time to, and then I'll upload Linux/Win native binaries to that same repo.

Basically the process, once GraalVM and Natifier package are installed, is you just run a command, point it to the .jar file and it will create a binary that embeds any relevant parts of the Java VM (GraalVM) into a 'standalone' binary. I then took that binary and transferred it to a Mac with a fresh OS install (macOS v13), with no Java installed in order to test it out and it worked as expected.

[OpenXTalkPaul]

Oh and there's some other stuff in that OpenXION repo above, some of my simple experiments in creating some GUI for it, and a 'OpenXION Terminal' demo stack, that displays its docs (if placed in the correct folder), and has buttons for loading .xn files (OpenXION's scripts files)... its sort of a very basic IDE.

https://github.com/PaulMcClernan/openxi ... e.oxtstack

[tperry2x]

Here is the macOS X86_64 build of OpenXION as a standalone 'native' executable: https://github.com/PaulMcClernan/openxi ... cOS_X86-64

That's excellent, thank you Paul.
I will happily test this out and see what I can make with it.
The only way I could find to download it was through DownGit, https://minhaskamal.github.io/DownGit/# ... cOS_X86-64
and
https://minhaskamal.github.io/DownGit/# ... e.oxtstack

[OpenXTalkPaul]

Here is the macOS X86_64 build of OpenXION as a standalone 'native' executable: https://github.com/PaulMcClernan/openxi ... cOS_X86-64

That's excellent, thank you Paul.
I will happily test this out and see what I can make with it.
The only way I could find to download it was through DownGit, https://minhaskamal.github.io/DownGit/# ... cOS_X86-64
and
https://minhaskamal.github.io/DownGit/# ... e.oxtstack

You can download individual files from GitHub Repos by using the 'View RAW' buttons on the Web UI, or download the entire repo from the main page of any repo.

Untitled.jpg (121.44 KiB) Viewed 1104 times

downloading a complete repo:

DownWHoleRepo.jpg (143.18 KiB) Viewed 1103 times

[FourthWorld]

a business standpoint

Surely that is of concern for the people at LiveCode, but not for OXT

Yes. I had included "them" for clarity.

It would be silly to work on an open source tool without strong support for the one ecosystem that values open source the most.

[OpenXTalkPaul]

It would be silly to work on an open source tool without strong support for the one ecosystem that values open source the most.

I may not be a normal case, but that seems like a strange opinion to me because some of my favorite FOSS either has serious commercial competition or the opposite; not really viable as commercial product for various reasons (emulators for example).

For me the act of working on OXT, in order to make it a better, more easily user-modifiable software, for me as a user, has been rewarding enough.

I think that having goals of a commercial product leads to chasing the latest buzzwords like 'AI' & Low-Code/No-Code, concentrating on Web Dev over Desktop Apps, using lots of JSON, concentrating on database connectivity, and other sorts of things that I tend to find boring, instead of concentrating on being a user-modifiable 'Hyper-Media' creativity tool that is as close to what I fell in love with back 1987/8 (while leveraging contemporary APIs) as I can get, which is what I've been looking for since some time after 2004 (when HyperCard was finally no longer for sale from Apple)..

But for people who have commercial aspirations, and/or need help/hand-holding AND have the money to subscribe, LiveCode is the best commercial option for xTalk-based commercial app building as a service, and so I'm going to be honest about that. Of course that could change if SuperCard or some other commercial xTalk were to spring back to life (although I've never been a fan of SuperCard's IDE UI).

[OpenXTalkPaul]

I can't try this at the moment, but I wonder if setting 'the videoClipPlayer' global property on Linux could have any effect on the usability of the player control on the problematic Linux distros, or does it only apply to playing media with the 'play [vc/ac] [file]' command?

Code: Select all

set the videoClipPlayer to "/usr/bin/some_other_player_probably_FFMPLAY"

[FourthWorld]

It would be silly to work on an open source tool without strong support for the one ecosystem that values open source the most.

I may not be a normal case, but that seems like a strange opinion to me because some of my favorite FOSS either has serious commercial competition or the opposite; not really viable as commercial product for various reasons (emulators for example).

How many viable open source projects do not run on Linux?

I'm sure there are a few. But very few.

Linux is a culture that understands and embraces open source. Most Mac users don't even know what GPL is, nor need they. But Linux is GPL, quite adamantly so.

When LC had an open source edition, Linux was arguably essential. Few others even understand how open source is different from "free trial".

And that cultural difference appeared to be present at LC itself. They treated their FOSS edition like a marketing come-on, and were unsurprisingly disappointed by their own results.

[richmond62]

And that cultural difference appeared to be present at LC itself. They treated their FOSS edition like a marketing come-on, and were unsurprisingly disappointed by their own results.

I don't think there was any real understanding of the theoretical underpinnings of Open Source at all on LC's part: they saw it as a quick-and-easy route to getting other people to do work for them as well as a sort of hook to get people to buy the commercial version.
-----------
This reminds me of a story I was told when I was about 15-16, that at disco-techs then (late 1970s) people would go round offering free heroin injections on the theory (scientifically dubious) that on the basis of a single heroin dose you'd be hooked.

Well, LiveCode, while being 'addictive' to work with, doesn't work like that (and nor does heroin).
-----------

What they certainly did NOT realise is that the crowd they attracted to the Open Source version were probably the crowd wo use GIMP, Inkscape, LibreOffice, Blender, and other Open Source software as their 'go-to' tools: tools which do not depend on end-users contributing code, nor having commercial variants.

The subsequent fund-raising efforts were indicative of that misunderstanding. Had they understood exactly how open source worked they, probably would have done what they had done previously, released a 'castrated' version (DreamCard, RevMedia) free instead.

However, WE should be extremely grateful for their misunderstanding as, owing to that we have their open source code on which to build our towers, castles, and so on.

[OpenXTalkPaul]

It would be silly to work on an open source tool without strong support for the one ecosystem that values open source the most.

I may not be a normal case, but that seems like a strange opinion to me because some of my favorite FOSS either has serious commercial competition or the opposite; not really viable as commercial product for various reasons (emulators for example).

How many viable open source projects do not run on Linux?

I'm sure there are a few. But very few.

Linux is a culture that understands and embraces open source. Most Mac users don't even know what GPL is, nor need they. But Linux is GPL, quite adamantly so.

I'm not sure where the macOS vs Linux in FOSS cred matchup is coming into it ("how many viable open source projects do not run on Linux"?)?
I would agree that it's more central to think about GPL licensing to builders of Linux distros and people writing apps for them, but I seriously doubt that most average desktop Linux users, like someone running SteamOS (Linux) or ChromeOS / Android (Linux) who probably don't even know there's Linux Kernel underneath, or even Ubuntu or PopOS! or many others regular users actually think much of anything about FOSS other than it means FREE software as in inexpensive, not so much as in Freedom. I've read a few tech articles about the polling that shows younger users care less and less about FOSS (in the same way that generation tends to care much less about ownership of physical media). I just don't think that the problem you're getting at is exclusive to macOS users.

But it still seems like a crazy opinion to me, am I really THAT abnormal of a Mac user?
How could I be when there is MacPorts, Homebrew, Fink, and other package managers for macOS that are chock full of useful FOSS? Have they kept those running for decades just for me? I know there's lots of popular Mac apps that are built around open source underneath. I was just using AUDACITY, VLC, and FFMPEG this morning, all FOSS. In fact the iMac I was using was even booted from a FOSS boot loader (that's much better looking than GRUB), OpenCore (with "legacy patches"). Seems like more than a few to me. It seems more like most every app ever created has benefited from 'open-source' in some way. As I look at the licenses now in the 'proprietary' prepress automation system I'm using (on Mac), I find it is built on on all sorts of FOSS open source libraries. But I guess it depends on your opinion of 'viable' and FOSS means in this context.

[OpenXTalkPaul]

I got side-tracked again a bit this weekend when I spent some time revisiting what nice work that guy Mark Smith left to the community, I never before looked at his Audio Wave form graphic generator, brilliant, and I'm converting it to a script only stack / library.

But while having a play, I noticed the player control on the Mac gets buggy on me, maybe because the 'videos' only contain audio track? It seems to go into a totally blocking mode when playing back the WAV files in my demo stack, I'm not sure why. And the Property Inspector needs a flushEvents("all") somewhere in the behavior for Player controls in Props Inspector. I think this is another reason I should build my own generic 'player/transport controller'. Which is something I've been thinking about trying for a while now.

[tperry2x]

...I should build my own generic 'player/transport controller'. Which is something I've been thinking about trying for a while now.

Would that be based on a VLC back-end? If so, could it be cross-platform-savvy?