Can't build distribution because of large data sets

[FKATan]

My game contains some extremely large data sets (dozens of arrays containing over 500,000 characters each) and it causes renpy to crash when attempting to build a distribution during the scanning phase (I have tried across multiple machines). Is there any way to manually build out distributions so I can try to avoid this?

[rames44]

What about moving the data out of rpy files and over into pure Python files (.py)? That might eliminate the need for Ren’py to try to swallow all that data.

[FKATan]

What about moving the data out of rpy files and over into pure Python files (.py)? That might eliminate the need for Ren’py to try to swallow all that data.

Unfortunately that did not make any difference

[rames44]

Bummer. Well, I guess you could PM PyTom and see if there’s a setting somewhere that is constraining how much memory the packager can use, and whether it’s possible to tweak it. Past that, IDK...

[rames44]

Ok, here’s one more thought. It’s ugly, but it might work. Suppose that you put the data sets into non-executable files and loaded the data at runtime. (Delay in the launch process, I know.) Then build the game WITHOUT those files included. Unzip the result, add the files to the game directory, re-zip.

Or, if the files get handled ok when you’re running in development mode, exclude them from the packaging options, build the game, than manually add the .rpyc files the same way.

Grasping at straws here...

Last stupid question - are the files getting packaged into an rpa file? Maybe it’s that process that’s dying, and things would work if those files were not rpa’d?

[FKATan]

This has been solved with an update

https://github.com/renpy/renpy/issues/2110