Apps disappeared from store

2 weeks ago when I uploaded another app I added edge1050 and it worked. Looks like it's an edge case (well it is an edge1050 case ;) Probably Brandon will be able to explain it once they found the bug in the scanner.

I´m back to eclipse to upload with no problem

It has absolutely nothing to do with the text editor you use. As I wrote it only happens in one of my apps and only 1 out of 271 devices.

flocsy said:

It has absolutely nothing to do with the text editor you use

That's not entirely true. The eclipse plugin (assuming it still works at all) passes slightly different options to the compiler. I've forgotten exactly what the difference was, but there was a thread a year or so ago where someone reported behavior that no one else could repro; then I noticed that I could repro by passing an extra argument to the compiler, and it turned out that the eclipse plugin was passing that by default.

So its just possible (but I agree unlikely) that building with eclipse could change the code generation enough to avoid the problem.

markw65 said:

The eclipse plugin (assuming it still works at all) passes slightly different options to the compiler.

The difference I've noticed is that when you run an app in Eclipse, it passes the -c {target_api_level} option to the compiler, where {target_api_level} is a string like 2.4.0 and comes from the Run As > Connect IQ App dialog which looks like this:

I always assumed that option was effectively useless, especially since:

- it made no apparent difference at runtime

- it's never been used by the VS Code extension

- it now produces the following build warning:

BUILD: WARNING: The command line option '-c' has been deprecated.

Ofc it's possible that option actually does something (whether it's noticeable or not), or did something in the past. I have no idea whether it was ever used during export though.

This is the current help text for -c:

-c,--api-level <arg>              API Level to target (deprecated)

I think that was it, and I think it was affecting whether "has" expression were being evaluated at compile time or not. But I can't seem to find the discussion...

Yeah I seem to vaguely recall something like that. (My main takeaway was that I was wrong about my assumption that -c does nothing haha.)

EDIT: after reading the comment under mine, I'm starting to doubt my recollection and I'm back to thinking "-c" never did anything. After all, compile-time evaluation of "has" is a very new feature that came a long time after the Eclipse plugin was deprecated in favor of VS Code. Even if "-c" had the effect stated above, what, if anything, did it do in the Eclipse-only era?

No, that's a new option

-disable-api-has-check-removal

jim_m_58 said:

No, that's a new option

It's true that there's a new option. But so what? We're talking about whether eclipse vs vscode could cause different code generation.

And the answer is yes (at least in the sense that eclipse and vscode pass different options by default, which is all I was claiming): https://forums.garmin.com/developer/connect-iq/i/bug-reports/bug-with-has-or-inline-symbol-on-venusq-and-maybe-other-devices?CommentId=528480de-eef6-419c-8f30-f66d4ac58138

You can see that (a year ago at any rate), --api-level did affect whether a "has" was resolved at compile time, and what value it was resolved to.