It says in the gmlive documentation that dynamic expressions in switch statements are iffy.
But, to me it was unclear to which degree this was the case. Quote from the documentation:
"Needless to say, that does get a little nastier with fall-through cases, and GMLive currently
may not agree to compile a "case" if it's expression cannot be computed compile-time (things like ord("A") are fine though)."
I wrote some code for a state machine-like object. A very simplified example of the code would be something like this:
I was going to add an actual enum to the game later to use in the switch statement but, first i did it like this for convenience's sake.
Subsequently, gmlive stops live-updating the script for this object. It doesn't crash, it doesn't give any warnings, the server clock is still ticking. The documentation doesn't mention this sort of behavior. It was first when I changed the code to use the enum instead of the temporary dynamic variables that I made the connection between what was stated in the manual and what had happened.
The working code looks kinda like this now:
I saw someone who had made a thread here 18 days ago to write about how they were frustrated with gmlive not working for them. Seeing how they wrote about how they had followed each tutorial "to a T" got me to think that they may have tried to use gmlive in one of their already existing projects that may have code that runs into one of gmlive's slightly ambiguous limitations.
I think gmlive would benefit from having some sort of warning system for some of its limitations. This way it would be way clearer to me and other people when and where they might have screwed up.
Other than that I have greatly enjoyed my first week with gmlive and, I'm looking forward to be using it in future projects.
- Cheers