When 1.0 ships...

[OpenXTalkPaul]

Some things, like that systemVersion, can't be fixed even in the Engine because it's actually a bug from Apple. That bug is present when using JavaScript in web browser too, but OXT's Mac Native lib has a function macOSVersion() that uses different method to get an accurate version number result. Another one is 'Darkmode' of menus and window-frames in Windows 10, its not going to happen (without some hackery using 'private' MS libraries) because Windows 10 didn't really have a darkMode API that developers could use.

I wasn't aware of 'date' syntax problems on Sonoma, that seems odd. What could Apple have changed that screwed up the engines method of retrieving it? What odd method did LC use? There's a few ways to get the date from macOS.

Code: Select all

shell(date) 

should do it (should work on Linux too).

What can't be fixed entirely without recompiling the Engine leans towards UI related things, like support for Wayland + Whatevernew Desktop in Linux when the Linux engine was built for for X11/GTK/GNOME as they were most common 10+ years ago.

Anyone use the 'server' engine? That should be one of the easiest engines to compile from source because it's a bit like OpenXION in that it runs in the shell, sans UI, so I would think it has far less dependencies. Of course we want an xCard toolkit along with our xTalk, but IMO it's much more important that we have a compilable engine to execute xTalk scripts. Without the core language none of the rest of it matters.

Mark Wielder contacted me last Summer about successfully compiling the Linux engine from source, and he also hooked us up with the missing 'pre-built binaries'. So thanks again Mark.
I'd like it if he would share the work he did back then. I'm planning to give that a go myself soon. Even if I just manage to remove that damn embedded registration stack from the new binary it would be worth the effort.

[overclockedmind]

I'm sorry that I haven't got to do much with beta testing. I'm non-platform-wars too, the only reason my MBA has Monterey on it right now is, well, it did from the factory.

I think the question is, assume temporarily that we had more people for testing, on that front. Or more me. As it stands, can we right (write) the engine "well enough" without more testers?

That's kinda the question, I think, and I do plan more work on that front. Be that as it may...

[overclockedmind]

I'm sorry that I haven't got to do much with beta testing. I'm non-platform-wars too, the only reason my MBA has Monterey on it right now is, well, it did from the factory.

I think the question is, assume temporarily that we had more people for testing, on that front. Or more me. As it stands, can we right (write) the engine "well enough" without more testers?

That's kinda the question, I think, and I do plan more work on that front. Be that as it may...

OK, so this had connotations looking back upon it that weren't intended. It was intended to say "more of our currrent bunch of beta testers, and more me."

I meant no disrespect. Also, I can't edit that post... dunno why. The MacBook Air, with Monterey, should be up in about a day, or so.

[overclockedmind]

For the moment, I'm on OXT Lite 1.0 in Win64. Will be back in Monterey rather quickly.

[overclockedmind]

Yes, RC does seem a much better idea.

How do we differentiate what is currently OpenXTalk Lite with Paul's RC version, if they are both then to be called RC?
"Do you mean OpenXTalk Lite RC1, or do you mean OpenXTalk RCx?" a bit needlessly confusing.

I will also point out, at this point, the Mac version should come with an alert if you open it on anything past MacOS 11 Big Sur. Something along the lines of "Warning: OpenXTalk may not be 100% stable under this version of MacOS. You may proceed anyway."

I'm very tempted to put this in, in light of needing further bug fixing / testing by people on MacOS 12+
Until we can say that it's stable, I think the 'Preview' tag fits exactly where we are with it.
Linux and Windows can have a 1.0 release, certainly. MacOS - I don't feel that it's stable enough with the legacy engine we've been left, quite frankly.

OXTL/OXTP ? Open source stuff happens like this all the time; someone grabs an abandoned project and barely changes the name.