The only thing I can think of, is that there might be some 'weird' invisible character' somewhere.

Comment line 3 of your global variable declarations and hit ctrl+L to see if the error goes away. If so, retype that line from scratch.

Hello George,

Another approach would be to declare an nvo app_manager as a Global application variable and then in that nvo you would have your of_set() and of_get() functions to return and set your global environment variables that were declared as PRIVATE in the app_manager nvo. Just my 2% of a buck.

Warm Regards,

Greg

Two developers have this issue, we've both restarted powerbuilder and our PCs. We tried commenting line 3 and also leaving a blank line. In my 3rd screenshot, you can see line 1-3 commented out AND pushed down so line 3 is blank, but still there's an error on line 3. We've also tried re-typing with no luck. I've edited the source in a hex editor and couldn't see any strange characters. 

I'm convinced that this is a mismatch/sync issue between the ws_objects folder and the pbls. Powerbuilder is not compiling the same script as it's displaying. 

When I run in the debugger, gs_inifile gets initialized with c:\pcscale\pcscale.ini, which is from a previous version of this object. 

I didn't mention this yet. Someone else wrote these 3 lines in 1993. There's nothing wrong with the code. This has to be pb & git messing up. 

This makes me think this is a git issue. when I try to switch branches, I get garbage like this. Once this happens, the local repo is unusable. 

Regenerate throws errors. My code is now unusable. 

First three lines in edit source:

forward
global type pcscale from application
end type

when I use CTRL-L in pb, it puts me on the declaration of gs_version. 

What if you make a copy of everything, connect to that copied workspace and do a full build without any source control connected?

Hi George,

Judging from the described problem, it is not sure that it is related to Git. The three objects pcscale.sra, w_ini_maintain.srw, and w_pc_lw_main.srw in the screenshot after switching branches also failed. You can check the code of these three objects and see if they are related.

The error message prompted by PB's syntax detection may point to the wrong location, and the real problem may lie elsewhere (the error syntax hint pointed to by Edit source is inaccurate).

It is recommended to check all the code of the problematic objects first. If the problem cannot be solved, can you provide us with a test case for debugging?

Regards,

Kitty

This resulted in many errors during the full build. From a non-source controlled copy. All of this code has been working for years. There's nothing wrong with the code. 

Line 30. 

I'm working to submit the pbl with a case. I need approval internally to share the data...