OpenXTalk Forums

I want to ask this question on the parent forum but my registration has not yet been blessed. It appears to me that Livecode Create (LCC) may have issues running legacy stacks. I refer to this thread https://discourse.livecode.com/t/my-fir ... -result/55 from September 2023. In it user Jehale writes

one of the first things to hit you when moving from the desktop IDE to LC Create is that fact that all the controls are widgets. So the “field” control is not the same as as a field in the IDE but is a widget.
So you can’t put “xyz” into the field, you need to set the appropriate property of the widget
to put “wyz” into your field (called say “f1”)
set the text of widget “f1” of me to “xyz”

Does anyone know if this is still the case? If true then it implies that even if one subscribes to LCC there will be quite a lot of work to get stack files written in Livecode to run.

Well . . .

"The parent forum" seems a bit funny when this forum might be better described as a partner recovering from a messy divorce.

And WTF 'that lot' are doing with 'Create' is not really relevant to OXT.

OXT (both flavours) is the direct descendent of LC 9.6.3 'community', while 'Create' increasingly looks less so.

AND IF all controls/objects are to be widgets in 'Create' LC are not going to make many friends amongst those who expect backwards compatibility.

Personally I would focus 99% of the time on OXT, just casting an occasional 'weather eye' at LiveCode to confirm I made the right choice in a large number of ways.

As LiveCode have NOT released a demo of any sort of 'Create', and I for 1 am not going to fork out $99 for a beta version 'sight unseen', this is just my fairly uninformed speculation.

richmond62 wrote: ↑Tue Aug 13, 2024 12:08 pm As LiveCode have NOT released a demo of any sort of 'Create',..., this is just my fairly uninformed speculation.

But you are right; by not having a demo - what else can anyone do but speculate?
That's always been my issue with software that you subscribe to without seeing or playing around with first - how do you know what you are actually spending your money on, or even if it's the right product if you can't do a test-drive of it first?

Interesting that it's throwing up compatibility issues - and makes sense, as you don't interact with widgets in any of the same way you would with classic controls. This is an argument for keeping them firmly in place for compatibility.

As LiveCode have NOT released a demo of any sort of 'Create', and I for 1 am not going to fork out $99 for a beta version 'sight unseen', this is just my fairly uninformed speculation.

The only thing I disagree with you on is the price as I have been quoted a larger figure.

Having watched the video and in an attempt to answer my own question Legacy or Classic stack will load and run but they are running in a version of Livecode that is part of LCC. It appears that the code completion and new widgets are only part of LCC so the true answer appears to be no and yes.

The other takeaway is that many features are still only planned and will be coming later; how much later no one is saying.

For me the biggest issue is the licensing but it appears that I am in a minority.

So the bad news is that I will be hanging around here

Skids wrote: ↑Tue Aug 13, 2024 10:19 am on the parent forum

Do you mean on ex-MotherShip's forums? I think they have a separate forum for people that have been test using LC Create now. Your account looks activated here.

It appears to me that Livecode Create (LCC) may have issues running legacy stacks. I refer to this thread https://discourse.livecode.com/t/my-fir ... -result/55 from September 2023. In it user Jehale writes

one of the first things to hit you when moving from the desktop IDE to LC Create is that fact that all the controls are widgets. So the “field” control is not the same as as a field in the IDE but is a widget.
So you can’t put “xyz” into the field, you need to set the appropriate property of the widget
to put “wyz” into your field (called say “f1”)
set the text of widget “f1” of me to “xyz”

Does anyone know if this is still the case? If true then it implies that even if one subscribes to LCC there will be quite a lot of work to get stack files written in Livecode to run.

This is actually what I've been thinking about myself lately, why not stop bothering with even trying to style 'classic controls' to be look-a-like to 'naive' controls, or messing around with 'Native Layer' actual native controls for simple things like buttons, and instead use Widgets with properties set equivalent to it's 'classic' counterpart, but also with a theming system (macOS-11,windows-10,winows-11,android-ics,etc.). I believe that's where LC was headed with that too, way back around 2015.

For any Text-widget control I would want it to behave as close as possible to the classic text field control. Such a text widget should be click-to-edit text (or at least use a simple click-to-input in an 'Ask' style dialog), not to have to edit the text in property inspector (unless maybe the PI was given much better editing tools). And of course it MUST able to render Styled RTFText and HTMLText besides plain text, and be used for List-Select fields, just like the classic field controls do.

As far as cross-IDE compatibility, it's like Richmond said we can't worry about what they're doing 'over-there', we're on our own. Yes, a widget based theme-able set of new-'classic controls'-as-Widgets, could be added to OXT and theoretically those could be crafted to match a properties set of some other set of controls from somewhere else, HOWEVER, there is an identifier used with widget controls that is it's 'kind' property. OXT uses org.openxtalk.widget.whatever while LC has used com.livecode.widget.whatever, so such a widget would have to be deleted and then the other version replacing it and given the same property attributes and script of the former.

If the lc file format remains the same, and widgets that are used in such an 'LC Create' stack are of the same 'kind', which it looks like they have retained widgets that were introduced as part of the LC Community Edition, for example the SVGPath icon widget, Header Bar, etc. then those should still work fine in OXT since the identifiers of pre-existing widgets were left as-is, with their original LC identifiers intact in OXT.

Basically as long as you can load a widgets module (lcm file) in the open source version of LC Community 9.6.3, then it should work fine installed in OXT as well. That means it must be open-source, it can't be encrypted like LC commercial editions extensions/widgets are.

I've already been treating some of the 'classic controls', such as the slider/scrollbars, as if there are deprecated because their default attributes aren't good with macOS darkMode, and they don't really match native controls anymore, and SVG Icon libraries and Widgets like Tree-view (and dare I say my Piano Widget), can be so much more useful to me.

But just to be clear, I have no intention of removing any legacy classic controls or anything like that, and in fact I'm good with (and have) added new styles for 'classic controls.', such as 'filled roundRect button' or 'square bevel button', but when I finish Tools palette redo then maybe I spend some time getting a base set collected, updated and ready for prime-time.

Do you mean on ex-MotherShip's forums?

I was referring to the dedicated Livecode Create forum on discourse. I wanted to ask a few questions on the forum but my account needed blessing by RunRev and this took thirty six hours to be actioned. This was unfortunate as they have set a deadline for present Livecode users to upgrade to Create at a special price. This is a decision that I/we are suppose to make based only on their marketing campaign. I watched the demo video to try and get some idea of what I might be missing and have now have formed an opinion that I probably should not voice in public.

This is actually what I've been thinking about myself lately, why not stop bothering with even trying to style 'classic controls' to be look-a-like to 'naive' controls, or messing around with 'Native Layer' actual native controls for simple things like buttons, and instead use Widgets with properties set equivalent to it's 'classic' counterpart, but also with a theming system (macOS-11,windows-10,winows-11,android-ics,etc.). I believe that's where LC was headed with that too, way back around 2015.

With the caveat that I have no idea how difficult it is to do either it appears to me that using widgets is more flexible as long as Xcode builder is robust enough.

For any Text-widget control I would want it to behave as close as possible to the classic text field control. Such a text widget should be click-to-edit text (or at least use a simple click-to-input in an 'Ask' style dialog), not to have to edit the text in property inspector (unless maybe the PI was given much better editing tools). And of course it MUST able to render Styled RTFText and HTMLText besides plain text, and be used for List-Select fields, just like the classic field controls do.

From the viewing the recent video by RunRev I would suggest this is exactly what they are doing. Kevin demonstrated a number of specialist text entry widgets such as a password and date picker fields. I was going to try are duplicate them as custom controls but I may try to make them the subject of my first widget.

As far as cross-IDE compatibility, it's like Richmond said we can't worry about what they're doing 'over-there', we're on our own.

I never intended to imply any compatibility and apologise if I came across that way. Instead I was seeking other opinions and views on what is a major change of direction for a piece of software that I have used since circa 2009. There is no free and frank discussion allowed on runRev's forums or the use group and I wonder if I have missed something. For example in yesterdays promotional video I lost count of how many new features are not yet implemented and will be coming soon yet despite this the hearts and hand clapping icons kept floating up the screen. Was I the only one thinking vapourware?

However, I have now made my decision and neither Livecode Classic or Create are my concern. My plan at present is to use OXT and maybe take a look at Python. The reason for looking at Python is that I am using an ARM powered mac so am relying on Rosetta2 to run OXT and any built applications. I plan to stay using a version of MacOS that allows Rosetta2 for as long as possible but I need an escape plan should it become unavailable for any reason.

I've already been treating some of the 'classic controls', such as the slider/scrollbars, as if there are deprecated because their default attributes aren't good with macOS darkMode, and they don't really match native controls anymore, and SVG Icon libraries and Widgets like Tree-view (and dare I say my Piano Widget), can be so much more useful to me.

All sounds good to me.

best wishes

Simon

Skids wrote: ↑Wed Aug 14, 2024 7:16 am I plan to stay using a version of MacOS that allows Rosetta2 for as long as possible but I need an escape plan should it become unavailable for any reason.

Just so you know, there's also the Whiskey project for MacOS.
(download link) - so that will allow you to run windows intel x86 apps on Arm macs - so you could run the windows version of OXT Lite via Whiskey on an Arm mac perfectly well.

There was also this, but hasn't been updated in 3 years

While I'm on the subject of Rosetta 2, can someone with an arm mac make a backup of it using these steps, just in case the installer 'goes away' at some point.
Edit: Actually, can probably do this myself - by right-clicking and downloading the file here.