Build standalone for iOS (8.1.2 RC2) First time- fails, second time it works

Sannyasin Brahmanathaswami brahma at hindu.org
Tue Dec 13 22:32:53 EST 2016


We are working our new mobile app in git hub. The challenge is that we need to use N number of binary files as "views" and at least one loader stack to run the build. Even though we have zero code in the stack script and zero code in the card script (the app is initialized with a behavior for the card… so that we use and external .livecodescript library… 

this is working pretty well, where different developers are the designated SOA fo any given module/some.livecode # binary stack  so for example in source we many have modules/color-meditation.livecode  and every one know that I am the SOA and they never touch it. so updates and merging are painless.

BUT  the main top level loader stack  is getting changes  where dev A will add a few more stack files, Dev B doesn’t have any iOS devices, so he turns off "unchecks" the iOS in the standalone builder and LC of course wants to save and the hash changes…ergo Git says the file was modified.  

Currently were looking at ways to externalize the stack file also to a text file in source that can be managed more easily by git and then loader stack reads it… so the loader stack will not have anyone fiddling directly with the stack files in the LC IDE for that stack… *but*  still there is the issue of changes to the standalone builder… i.e. that's the back story… 

So if I have to pull a copy of that stack and I have an iOS device , then I check that in the standalone builder and save… *now* I get the "missing ARM"  error messages.. but when I run it a second time, it goes through. 

So maybe (just maybe) the recipe is that the standalone builder's global props are not being updated when the user make changes in the standalone build GUI panel until *after* the first attempt to build runs… *then* that process updates something in the SA build profile, which is saved… and so now when you run the build, it goes through… at least that's how it feels at the moment.

BR

 

On 12/12/16, 9:10 AM, "use-livecode on behalf of Mike Kerner" <use-livecode-bounces at lists.runrev.com on behalf of MikeKerner at roadrunner.com> wrote:

    There is a bug report related to something similar, but I don't know if any
    of us have developed a reliable recipe, yet.  For me it's surprising and
    almost random when it happens, and I don't uncheck/check



More information about the Use-livecode mailing list