Version of the IPA-AdHoc at same level as iOS-Companion?

Testing my App I get everything to work well with the iOS Companion compiler but when I run it with an AdHoc IPA, new errors (too many) are occurring that are not present with Companion. In addition to the problems with the dimensions of the layouts, I now detect that none of the "Any component" blocks work, for me this is a "job stopper".

Mr. @ewpatton , when you plan to have the same compiler for Companion and for the Buid-IPAs. Is this possible? The current version of the Buid-IPA Compiler has many bugs that we do not have in the iOS-Companion and these bugs are important functions that we cannot overcome with botch work.

They are technically in sync at the moment as they are both built from the 2.64.4 version of the AIComponentKit library. As I believe I have mentioned in other posts, there seems to be an initialization bug related to screens that causes a host of issues between companion and compiled apps, but the underlying logic of every component is still the same otherwise.

1 Like

These initialization errors that you mention produce many new bugs in the IPAs that make their use unfeasible. Is very maddening when having all correct in the iOS-Companion tests after we have these secondary new bugs in the IPAs. They are not only issues of layout design dimensions but also basic errors, that do not allow us to continue (this is my case).

When will it be possible to synchronize both compilers? (estimated date will be good for me).

Hello Mr @ewpatton

I'm waiting for you answer about the planned date to have the IPA builder on the same level of the iOS-Companion builder.

I take this opportunity to tell you about an error in the IPAs that I do not know if it is also motivated by the Initialization problem that you have described previously.

I attach an .aia to show you that the IPA builder does not run with the "Any Component". In this example you can see that the iOS-Companion is running well for a "Direct-Label_1" and also for the "AnyComponent-Label_2". All right.

But, in the IPA exec you can see that the AnyComponent-Label_2 has an irritant mistake for native object.

This irritant mistake happens in the all "Any Component" elements.

This mistake is a "job stopper" for me because I don't find any "shortcut botch" for it.

nomisCP.aia (9.9 KB)

If you need more documentation about it, please, no hesitate to ask me.

The .aia blocks of teh test are these:

Hello @Anke ,
Would you be so kind as to verify and confirm this error.
Thank you so much.


Works with the IPA.

1 Like

Unfortunately I have yet to have the time to look at this with any depth, so that makes it pretty much impossible for me to give you a date at which the fix will be finished.

Uhauu¡ Runs well.
You are the Queen of the "Shortcut Botch".

Only a final question. I continue to receiving in the IPA the irritating error:
Is it possible to hide it? I think that Apple does not accept Apps with errors or error notices. What do you recommend I do about it?
Very Thanks

Only by avoiding such mistakes. :wink:

And how can do it?

I think this is part of the initialization problem. When an event is dispatched via the screen, the name of the component is looked up in an internal map of all of the components. If the component is not found, then false (#f) is returned. Erroneously, the scheme then tries to fire the event on #f, but since that's not a component instance it fails and triggers this error message. Normally, the return of #f would indicate the function needs to stop since it hasn't met a precondition but it doesn't seem to be doing that in this case. We will need to review the scheme code to fix it.


Hola @Anke ,
Based on his "sloppy shortcut" and with Mr. @ewpatton 's explanations regarding iOS builder initialization issues, I have improved his "workaround" so that now it runs well and I don't have the irritating error. I leave you a hard-copy of the AIA in case you can verify and validate it. The solutions is to define the list of the compnents inside the first ClockTimer because the inilizations components not run as Mr Patton said. Thank you so much for your help me.

I don't receive the irritant object error and all the "Any Component" runs well.

1 Like

Interesting – that means that initializing the variables before Screen.Initialize doesn't work either.

Not all. The single variables is it possible. I think it happens only for the "Any Compònents". I have other normal variables (included Lists) init before the initilization and they are running well. The problems may be happens only for variables with components.

A different doubt to favor the use of iOS with AI2. Do you think it could work for MIT to leave a generic AdHoc .mobileprovision file so that everyone can run IPAs on iPhones registered with Apple even if they cannot upload their apps to the Apple Store? I'm not sure that the permission Apple asks for should be personal or not.

On the IPA execution (not Companion) I receive this message error about JSON.
The App has the option JSON activated:

If I put the option = false, the message is the same.
Whith the Companion I have not the message.
How can I hidden this message in the IPA exec?

In the IPA exec I have not the app-icon (the app show only the AI2 icon). Is there some solution or workaround to solve this mistake?

@ewpatton I report you on the iOS platform that the SpeechRecognizer component only understand the english voice and not the local voices.

As far as the documentation from Apple, the speech recognizer is supposed to default to the user's Locale and fall back to something else if the language selected isn't supported by Siri. There may be something more subtle going on here though.