Possibility to skip indexing large files?

[edvreu]

I am working with a repo that has huge .erl files with generated protocol code (up to 10 megabytes or 100 000 lines of code).
Since it is generated code I rarely navigate in these files, so indexing them just wastes time (at least 20 min) and memory.
Would it be possible to add a size limit (2 MB?) for files that will be indexed?

[edvreu]

Support was added for this shortly after my question about it. I can't find anything in the configuration documentation, but add this to the erlang_ls.config to get it working:

indexing:
  skip_generated_files: true
  generated_files_tag: "Generated"

[robertoaloi]

The reason why this is undocumented is that, shortly after implementing the skip functionality, we switched indexing to an on-demand model, which should make the skipping not so relevant anymore.

[edvreu]

Ok, I still feel it makes a difference (just by looking at how much memory the erlang_ls process uses) so please keep the functionality.