OXT Lite 0.94 Mac not working

richmond62 · Post by **richmond62** » Sat Nov 25, 2023 4:26 pm

Devawriter Pro is an interface and learning system for rendering Classical Indic languages in a format extremely close to the original manuscripts that has been developed over about 14 years with LiveCode.

https://richmondmathewson.owlstown.net/projects

I am now working on my 6th SDLC with OXT Lite 0.93 on MacOS 12 (mainly because 0.94 is NOT working), but I am worried about whether standalones generated from that code will function on end-users' machines (the current version, made with LC 963 does function), and so am reluctant to make a complete switch.

richmond62 · Post by **richmond62** » Sat Nov 25, 2023 4:31 pm

Once OXT is at a stage that I can make a complete switch this is something serious that OXT can use to advertise:
-

: info.jpg (212.67 KiB) Viewed 2838 times

-

Post by **tperry2x** » Sat Nov 25, 2023 6:05 pm

And don't forget Windows 10 too!

richmond62 wrote: ↑Sat Nov 25, 2023 4:26 pm ...because 0.94 is NOT working...am reluctant to make a complete switch.

I have plans regarding this, which I'll be able to test in MacOS 10.15...."I'll be back"

(with screenshots)

Post by **tperry2x** » Sat Nov 25, 2023 6:26 pm

Will you be a guinea pig please?
You'll need to do this on something that is NOT sonoma, as this test is an unpatched version.
If you can download this attachment and uncompress it. It's a pretty original 9.6.3 binary / engine.
Please uncompress and place on your desktop.
https://www.tsites.co.uk/sites/openxtal ... atched.zip

Next, if you can "Show Package contents" on both the 0.94 that isn't working, and "Show Package Contents" on the one you just downloaded (still called Livecode Community 9.6.3, but I dare not even change the name at this stage).

Please drag the "Tools" folder from the 0.94 build over to this test version, like so:

I'd like to know if that then runs please.

richmond62 · Post by **richmond62** » Sat Nov 25, 2023 6:43 pm

Right: 3 glasses of strong Bulgarian red, and I'm on it . . .

Silly pictures following . . .
-

: Screenshot 2023-11-25 at 20.46.24.png (57.28 KiB) Viewed 2833 times

-

: Screenshot 2023-11-25 at 20.47.02.png (58.9 KiB) Viewed 2832 times

-
No joy . . . MacOS 12.7

richmond62 · Post by **richmond62** » Sat Nov 25, 2023 6:50 pm

Will you be a guinea pig please?

Well, as I was the first 10th of a caterpillar in year 2 at Primary school . . .

Post by **tperry2x** » Sat Nov 25, 2023 6:58 pm

richmond62 wrote: ↑Sat Nov 25, 2023 6:43 pm No joy . . . MacOS 12.7

So, all that took to 'break it' was replacing the icons.
Fragile or what.

Post by **tperry2x** » Sat Nov 25, 2023 6:59 pm

What about if you take the app from LCC 9.6.3, drag to your desktop, then swap the "tools" folder over. Does that run?
It all runs absolutely no problem here, and 0.94 does too, so I don't understand why it's not working for you.

What about if you download it through Safari?

richmond62 · Post by **richmond62** » Sat Nov 25, 2023 7:09 pm

Why on earth would downloading a ZIP file through Safari result in anything different from downloading is through Brave (as I did, or any other browser for that matter?

Post by **tperry2x** » Sat Nov 25, 2023 7:23 pm

richmond62 wrote: ↑Sat Nov 25, 2023 7:09 pm Why on earth would downloading a ZIP file through Safari result in anything different from downloading is through Brave (as I did, or any other browser for that matter?

Only because Safari is Apple's default browser, and that screenshot knows it's come via Brave.
This type of interference by the OS is why I class MacOS as already depreciated.

richmond62 · Post by **richmond62** » Sat Nov 25, 2023 7:32 pm

Well . . . I downloaded that via Safari and it unzipped the thing without asking me . . . not to my taste.

ANd got exactly the same result aafter swapping that folder across.

Post by **tperry2x** » Sat Nov 25, 2023 7:34 pm

richmond62 wrote: ↑Sat Nov 25, 2023 7:32 pm Well . . . I downloaded that via Safari and it unzipped the thing without asking me . . . not to my taste.

Ah, so that worked?
(not meaning to sound amazed, but...) does the 0.94 version of OpenXTalk Lite (the original 0.94 I posted) open if you were to download it via Safari?

EDIT: damn. Commented too soon. Got my hopes up

I think the answer is "switch to linux"

richmond62 · Post by **richmond62** » Sat Nov 25, 2023 7:35 pm

does the 0.94 version

If I could just find that download link after all the other well confusing ones . . .

richmond62 · Post by **richmond62** » Sat Nov 25, 2023 7:36 pm

Tangentially, I have been prodding Matthias Rebbe . . . and he might just.

Post by **tperry2x** » Sat Nov 25, 2023 7:55 pm

richmond62 wrote: ↑Sat Nov 25, 2023 7:35 pm If I could just find that download link after all the other well confusing ones . . .

Please just use the ones mentioned either on this forum,
viewtopic.php?f=16&p=4646#p4646
or on the main page of this forum:
https://openxtalk.org/OXTDownloads.html
Or the one directly on my site:
https://www.tsites.co.uk/sites/openxtalk/changes/
(All points to the same thing ultimately)

...and here's all the info regarding re-codesigning it again:
https://www.openxtalk.org/forum/viewtop ... 4133#p4133

(none of this evidently works though 100% of the time).
I don't currently feel I can release a Mac version with confidence that it'll actually function for everyone. So I have a plan C (I think that's what we are up to now).

I could really do with someone else to confirm it's broken for them on the mac too, no disrespect intended, just so I know it's not an isolated issue with your mac and want to confirm it's affecting other mac users too. (I can't duplicate the issue, as it works fine for me).

Edit: Have just tried this on another Mac (Big Sur) that has never encountered OpenXTalk before. In fact, no modifications have been made to it after it's clean install of the OS.

(sorry about the reduced colourset in the screenshots, it's just that this is a gif limited to 256 colours).

: frames.gif (303.55 KiB) Viewed 2791 times

It works without error, so I'm confused.
Worth noting, this was a mac without any form of internet connection at all though.
(whether that has any bearing upon proceedings remains to be seen).

My plan C as alluded to above, is taking some of the information from here:
https://stackoverflow.com/questions/274 ... er-program
But scripting the creation of the keychain in a MacOS installer program, then code-signing it for development on the user's machine. If I can script that, this could be an approach to negate all the codesigning issues in future.

Using those instructions, here's the certificate:

: cert.png (75.88 KiB) Viewed 2748 times

openxtalk.p12.zip: (2.55 KiB) Downloaded 203 times

Double click the uncompressed certificate, and add to the keychain

: add 1.png (26.98 KiB) Viewed 2741 times

No password if prompted

: nopassword.png (15.99 KiB) Viewed 2741 times

You should then get it appearing in your keychain (search for openxtalk if you can't find it).

: import-success.png (36.17 KiB) Viewed 2741 times

With all this now in place, we should be able to use this for codesigning the app under MacOS using the 'deep' (recursive) flag.
1. Clear all extended arributes in the OpenXTalk 0.9.4 app:

Code: Select all

xattr -cr ~/Desktop/"OpenXTalk Lite 0.94.app"

2. codesign using this certificate:

Code: Select all

codesign -fs [path-to-unzipped-p12-file] --deep ~/Desktop/"OpenXTalk Lite 0.94.app"

richmond62 · Post by **richmond62** » Sun Nov 26, 2023 12:43 pm

I think the answer is "switch to linux"

Sorry, while I DO use Linux (Xubuntu, Debian) as both development and content delivery platforms, do MOST of my work on MacOS 10.7, 12 and 14, and what I develop needs to produce standalones that run on 10.9 through to 14 and on Windows 7-up (an operating system that I, personally, dislike, but one cannot 'kick against the pricks' so . . . ).

This morning woke to 2 feet of snow, no electricity, and as a result of that, no internet. Later, the electricity having been returned, it turned out that a tree had fallen on the line that delivers our internet to the lamp-post across the road, and so nuffin was coming up the optical cable to my WiFi modem.
-

: lanternWaste.jpg (103.17 KiB) Viewed 2733 times

-
Just restored 30 minutes ago . . . [that's 2pm here in Bulgaria].

AND, climbing into bed with one operating system is plain daft.

richmond62 · Post by **richmond62** » Sun Nov 26, 2023 12:54 pm

Sorry, old chap: back to the niceties of Sanskrit and Prakriti, and my Devawriter version 5.0, which will, unfortunately, be hived off at a standalone using LC 963:
-

: Screenshot 2023-11-26 at 14.52.27.png (140.3 KiB) Viewed 2732 times

-

Code: Select all

on mouseEnter
   lock screen
   set the coloroverlay["color"] of me to blue
   set the coloroverlay["opacity"] of me to "75"
   unlock screen
   -- deprecated code removed
   set the text of fld "fDECODE" to numToCodePoint(2305)
end mouseEnter

on mouseLeave
   set the outerGlow of me to empty
   set the colorOverlay of me to empty
   put empty into fld "fDECODE"
set the ink of me to srcCopy
end mouseLeave

on mouseDown
   set the ink of me to notsrcor
end mouseDown

on mouseUp
   set the ink of me to srcCopy
   if fld "WARN" is empty then
      --do nix--
      else
      put empty into fld "fBUILD"
   put empty into fld "WARN"
  put empty into fld "fBUILDW"
  put empty into fld "fBUILT"
   put empty into fld "fRRR"
      end if
   -- deprecated code removed 
   if fld "fBUILT" & fld "fRRR" is empty then
      set the text of the selectedText to numToCodePoint(2305)
else
      --MAHA--
   -- deprecated code removed
put 0 into DDROP
switch the text of fld "fBUILT"
   ---MAY 2011---
   ---biconsonantals added April 2011----
----n 2339----
case (numToCodePoint(2339) & numToCodePoint(2325))  
set the text of fld "fPROC" to numToCodePoint(61953)
put 2 into DDROP
break
case (numToCodePoint(2339) & numToCodePoint(2326))  
set the text of fld "fPROC" to numToCodePoint(61954)
put 2 into DDROP
break
case (numToCodePoint(2339) & numToCodePoint(2327))  
set the text of fld "fPROC" to numToCodePoint(61955)
break
case (numToCodePoint(2339) & numToCodePoint(2328))  
set the text of fld "fPROC" to numToCodePoint(61956)
break
case (numToCodePoint(2339) & numToCodePoint(2329))  
set the text of fld "fPROC" to numToCodePoint(61957)
break
case (numToCodePoint(2339) & numToCodePoint(2330))  
set the text of fld "fPROC" to numToCodePoint(61958)
break
case (numToCodePoint(2339) & numToCodePoint(2331))  
set the text of fld "fPROC" to numToCodePoint(61959)
break
case (numToCodePoint(2339) & numToCodePoint(2332))  
set the text of fld "fPROC" to numToCodePoint(61960)
break
case (numToCodePoint(2339) & numToCodePoint(2333))  
set the text of fld "fPROC" to numToCodePoint(61961)
put 3 into DDROP
break
case (numToCodePoint(2339) & numToCodePoint(2334))  
set the text of fld "fPROC" to numToCodePoint(61962)
break
case (numToCodePoint(2339) & numToCodePoint(2335))  
set the text of fld "fPROC" to numToCodePoint(61963)
put 2 into DDROP
break
case (numToCodePoint(2339) & numToCodePoint(2336))  
set the text of fld "fPROC" to numToCodePoint(61964)
put 2 into DDROP
break
case (numToCodePoint(2339) & numToCodePoint(2337))  
set the text of fld "fPROC" to numToCodePoint(61965)
break
case (numToCodePoint(2339) & numToCodePoint(2338))  
set the text of fld "fPROC" to numToCodePoint(61966)
break
case (numToCodePoint(2339) & numToCodePoint(2339))  
set the text of fld "fPROC" to numToCodePoint(61967)
break
case (numToCodePoint(2339) & numToCodePoint(2340))  
set the text of fld "fPROC" to numToCodePoint(61968)
put 2 into DDROP
break
case (numToCodePoint(2339) & numToCodePoint(2341))  
set the text of fld "fPROC" to numToCodePoint(61969)
break
case (numToCodePoint(2339) & numToCodePoint(2342))  
set the text of fld "fPROC" to numToCodePoint(61970)
put 2 into DDROP
break
case (numToCodePoint(2339) & numToCodePoint(2343))  
set the text of fld "fPROC" to numToCodePoint(61971)
break
case (numToCodePoint(2339) & numToCodePoint(2344))  
set the text of fld "fPROC" to numToCodePoint(61972)
break
case (numToCodePoint(2339) & numToCodePoint(2346))  
set the text of fld "fPROC" to numToCodePoint(61973)
break
case (numToCodePoint(2339) & numToCodePoint(2347))  
set the text of fld "fPROC" to numToCodePoint(61974)
put 2 into DDROP
break
case (numToCodePoint(2339) & numToCodePoint(2348))  
set the text of fld "fPROC" to numToCodePoint(61975)
break
case (numToCodePoint(2339) & numToCodePoint(2349))  
set the text of fld "fPROC" to numToCodePoint(61976)
break
case (numToCodePoint(2339) & numToCodePoint(2350))  
set the text of fld "fPROC" to numToCodePoint(61977)
break
case (numToCodePoint(2339) & numToCodePoint(2351))  
set the text of fld "fPROC" to numToCodePoint(61978)
break
case (numToCodePoint(2339) & numToCodePoint(2352))  
set the text of fld "fPROC" to numToCodePoint(61979)
break
case (numToCodePoint(2339) & numToCodePoint(2354))  
set the text of fld "fPROC" to numToCodePoint(61980)
break
case (numToCodePoint(2339) & numToCodePoint(2355))  
set the text of fld "fPROC" to numToCodePoint(61981)
break
case (numToCodePoint(2339) & numToCodePoint(2357))  
set the text of fld "fPROC" to numToCodePoint(61982)
break
case (numToCodePoint(2339) & numToCodePoint(2358))  
set the text of fld "fPROC" to numToCodePoint(61983)
put 2 into DDROP
break
case (numToCodePoint(2339) & numToCodePoint(2359))  
set the text of fld "fPROC" to numToCodePoint(61984)
break
case (numToCodePoint(2339) & numToCodePoint(2360))  
set the text of fld "fPROC" to numToCodePoint(61985)
put 2 into DDROP
break
case (numToCodePoint(2339) & numToCodePoint(2361))  
set the text of fld "fPROC" to numToCodePoint(61986)
put 2 into DDROP
break
---k---
case (numToCodePoint(2325) & numToCodePoint(2327))  
set the text of fld "fPROC" to numToCodePoint(61987)
break
case (numToCodePoint(2325) & numToCodePoint(2328))  
set the text of fld "fPROC" to numToCodePoint(61988)
break
case (numToCodePoint(2325) & numToCodePoint(2329))  
set the text of fld "fPROC" to numToCodePoint(61989)
break

-- and so it goes for some 5100 lines of code in a single object

"Your message contains 205251 characters.
The maximum number of allowed characters is 60000."

Post by **tperry2x** » Sun Nov 26, 2023 4:53 pm

richmond62 wrote: ↑Sun Nov 26, 2023 12:54 pm Sorry, old chap: back to the niceties of Sanskrit and Prakriti,

Righto. Suffice to say, I'll try these codesigning shenanigans with 0.95 soon

richmond62 · Post by **richmond62** » Sun Nov 26, 2023 6:49 pm

I am looking forward to 0.95 and so on, as I really would very much like to turn Devawriter Pro 100% over to OXT; both as an advertising thing for OXT,
and also because, while the people in Edinburgh have sung the praises of all the commercial stuff that people have achieved with LC, they have singularly demonstrated how they have always looked at the Open Source strand as a way to get people to do clever things for them for nothing, rather than seeing that by 'making much' of what can be done with the Open Source strand that can also advertise xTalk.

After 14-15 years banging away at Devawriter I should also like to make some sort of blog about how one can make a thing to work with complex writing systems . . . Indian writing systems are fairly 'bloody-minded' in this respect (i.e. write a vowel expressed before a load of consonants after those consonants (Sanskrit), split 2 parts of a vowel expressed after a load of consonants into 2 components; one written before, and one writetn after the consonants (Bengali), and so on).

Of course this could be used to show off a lot of xTalk's capabilities that have NOT been properly advertised yet . . .

I have a fairly heavy fantasy about Egyptian Heiroglyphs . . .

Post by **tperry2x** » Mon Nov 27, 2023 9:58 am

0.95 being uploaded as I type this...
The link will be on the download section, but I'll come back here and put a link in too.
TerryL will need to update his bookmarks on his site if linking to the release too.
https://www.openxtalk.org/forum/viewtop ... 4646#p4646