Would it be viable to move the script error checking before texture compression within the bundle process or is there something about the process which requires texture compression to happen first?
The reason is texture compression takes significantly longer than script error checking, it’s a big bummer to have script errors show up after long texture compression times.
I’m not familiar enough yet with the code to answer this myself.
It would make sense to do the slowest tasks last. Could you please create a GitHub ticket for this?
Would it be even better to cache the texture compression somehow to avoid needing to repeat the task without any changes to the texture data or is that not possible? I’m still learning the engine so might be overlooking something.
It would make sense to do the slowest tasks last. Could you please create a GitHub ticket for this?
The scene graph is built using the dependencies between files. Atlases depend on their images for example.
We build the scene graph “bottom up”. I.e files that have no dependencies will be built first (i.e. atlases).
I’m not sure, but I think that scripts have resource dependencies, and are thus not at the bottom level.
So, it might not be as trivial as it might seem. I agree that it would be a good improvement though.
I think that given the inordinate amount of time texture compression actually takes, it’s not a bad idea.
I would probably start by looking if we’re cleaning out the textures first though.
If textures+profile+os matches could be hashed and cached so the same hash is grabbed from the cache instead of being compressed again that would be amazing for increasing bundling speeds on large projects for sure. I would gladly trade any amount of needed drive space for closer to instant repeated bundles.