Known Bugs

[tperry2x]

While it might take up to a month for me to get this mac, you don't want to wait that long.
So, I wondered if you can see what this stack gives you - it should help (hopefully) pin down why the dictionary won't load.

dictionary diagnostic.png (103.79 KiB) Viewed 3002 times

dictionary diagnostic.oxtstack

(3.69 KiB) Downloaded 70 times

[richmond62]

SShot 2024-11-15 at 19.03.38.png (132.83 KiB) Viewed 2996 times

-
And I'm off home: so no machine running MacOS 15 until Monday.

[tperry2x]

Thank you for testing, that screenshot might explain why -- for some reason you are missing this bit on your screenshot:

hmm.png (4.78 KiB) Viewed 2993 times

So that possibly suggests that the IDE can't read that file because of permissions or other OS-sandboxing.
But seems like it can read the index file, so that's odd.

[tperry2x]

Causing a script error in a modal dialog will crash the entire IDE & engine.
On my test here (linux), it also completely froze the entire machine. In Windows, the IDE then needs force-quitting with the task manager.

The script to trigger this is:

terr.png (12.8 KiB) Viewed 2966 times

Here's a demo stack (if you wanted to crash the IDE and experience this yourself!)

dialog.png (19.42 KiB) Viewed 2966 times

[richmond62]

Screenshot 2024-11-17 at 13.45.56.png (137.53 KiB) Viewed 2961 times

-
Let's see on MacOS 12.7.6 . . .
-
NO! did NOT crash the IDE.
-
However it left me stuck in BROWSE mode.

And unable to QUIT the IDE except by sending a KILL signal.

[tperry2x]

NO! did NOT crash the IDE.

Try again. (perhaps with additional stacks open). You'll find that nothing can be interacted with. And yes, like you say, you'll have to force quit it - like in Windows via task manager.
This is a bug in the engine.
It's actually detailed here, but I can't make sense of the formatting. (it's a bloody mess)

[richmond62]

nothing can be interacted with

No, it cannot.

BUT I would not describe that as a 'crash', I would describe it as a 'freeze'.

[tperry2x]

BUT I would not describe that as a 'crash', I would describe it as a 'freeze'.

Well, you get what I mean - a freeze on MacOS and Windows. A "hard lockup" on Linux.

[richmond62]

I get what you mean, but your use of 'crash' is incorrect.

[tperry2x]

I get what you mean, but your use of 'crash' is incorrect.

As you also chastised me for a single letter typo a moment ago, I can see what kind of mood you are in. So back *much* later.

[richmond62]

HP.jpg (79 KiB) Viewed 2940 times

-

[richmond62]

Screenshot 2024-11-17 at 16.16.58.png (212.63 KiB) Viewed 2928 times

-
That is still happening, unfortunately.

[tperry2x]

That is still happening, unfortunately.

Yes, that's because I only just found it - and hasn't been fixed. Please re-read the above.
This is a bug in the engine.
This won't be fixed in the mac version of the engine any time soon (won't compile currently).
Hasn't been included in my recompiles of the engine for Linux and Windows (the registration screen is the only thing I've changed).
Why not? - I can't actually make sense of the github edit - I can't follow what the amended files should actually read as.

[richmond62]

This is a bug in the engine.

Given the number of apparent bugs in the engine (and specifically the Mac engine) we can see that it is nearly buggered.

The Bug Reporting system "somewhere else" is a list as long as your arm, and is notorious for a very large number of those bugs being ignored for years.

[tperry2x]

Here's a PDF copy of that bug, just in case the Livecode engine repo was to 'disappear' any time in the near future.
Just like the previously archived downloads have, leaving the only source as the waybackmachine (archive.org).

bugfix-for-modal-crash.pdf

(714.99 KiB) Downloaded 71 times

and a PNG which actually keeps the formatting when you save the page

bugfix-for-modal-crash.7z

(784.55 KiB) Downloaded 69 times

[richmond62]

One of the most interesting things about that otherwise mind-blowingly boring PDF is the reference to loads of stuff with the
prefix 'MC' . . . this makes me think of how human embryos grow a tail at a certain stage of their development, but (most of them) lose it later on.

"LiveCode Community Edition no longer available as of August 31, 2021"

That's odd, as the Bug-fix was implemented on June 30 2021 . . . so, KM and Co. had already decided to dump the 'Community' version at that stage as that bug-fix was NOT rolled into the Open Source version.

It is marvellous that you were able to access that bug-fix 'stuff' so you could implement it in OXT Lite.

[tperry2x]

It is marvellous that you were able to access that bug-fix 'stuff' so you could implement it in OXT Lite.

Well, that would be the plan eventually - if I could actually decipher it.
If someone wants to give me amended:

Code: Select all

engine/src/debug.cpp
engine/src/debug.h

with the bits deleted in red, and the bits added in green.

I mean, even if I sit there and line-by-line do a search and replace to paste in the green bits over the red, where are the final two files with the rest of the lines? How do I know I've got the right / final version?

Also - (indents matter in C++ but these seem totally wrecked).

Anyway, that *may* be added in at the next recompile. (compiling is a slow process).

[tperry2x]

Whatever rubbish Apple may spew out, I would treat ALL MacOS releases as part of an ongoing rolling-release where nothing is ever finished.

Likewise with OXT Lite - I'd not treat any release as finished - as there's always something that needs fixing, changing or adding. It's only finished when the project is dead and receives no more updates.

[richmond62]

Oh quite . . . Devawriter Pro, and now I have gone and strung a new millstone round my neck: Sheba.

But the point I was trying to make was that when Apple suddenly go "Yabba-Dabba-do; MacOS 69", that is NOT really anything except one thing
in a long line of beta versions stretching all the way back to MacOS 10.0; and the numbers (and the slightly silly names) are just a way for them to manipulate the market so you rush out and buy a Macintosh machine with an ARM 57 processor and chuck your one with an ARM 56 processor in a landfill (Please bear in mind all the 'yakkety-yak' from Apple about the environment).

After all, MacOS 12 'Monterey' 12.6 was supposedly 'finished', then they released a series of twiddles, fixes, security updates, and so forth to the extent that made their earlier claims look like utter crap.

Not being very commercially minded I am never going to make any claims about any of my software beyond being, "What I have managed so far."

The simple fact that I can run the latest Debian system on my Dad's 20 year old, 32-bit, steam-driven laptop gives the lie to the need to endlessly upgrade.

[tperry2x]

I'd replied, but gone off topic... as I do, so moved that here.