Feature request: live status on "Analyzing"

[catskul]

During >>> Analyzing commands used in @//... It can take quite a while before anything prints, or between things printing.

It might be helpful to have a status printout at the bottom of the screen which indicates that it's still working, and how many targets/total targets.

[cpsauer]

Hey man! I hear you, and thanks for taking the time to write in with feedback.

Currently we have this stuff from bazel shushed on the theory that it created a lot of noise, and that, probably, if it was taking a long time we should be coaching people to run after builds w/ the same flags so it can hit cache for the slow header search part--to run it on faster-running subsets of their code. Does that seem reasonable/applicable, or do you definitely want more logging?

I sadly don't have bandwidth for this one at the moment. Would you be interested in exploring? If so, there are two internal phases that are (sometimes) slow: (1) aquerying bazel (if v large) and (2) running header search (if can't hit cache from prior builds/runs of this tool). For (1) you could play with removing the silencing of bazel (go into the generated python and comment '--ui_event_filters=-info', and '--noshow_progress',. For (2) we could put TQDM or similar around the big threadpool.map, though we'd have to figure out how to get it to install nicely w/ Bazel. pip_parse and all that.

Thanks again,
Chris

[cpsauer]

We'd added rules_python, which would have made the tqdm install easy--and then had to revert because it had lots of issues :(

Reverted rules_python in 0b821b7.

Tracking restoration in #168, but that more or less blocks this.