1192834 - talos filters are a bit hacky - we can do better

Status: RESOLVED WORKSFORME

(Testing :: Talos, defect)

Description

[Joel Maher ( :jmaher ) (UTC -8)]

recently :parkouss did a great cleanup on talos filters, then in bug 1186076, I made it hacky again :(

This was for a good purpose (as all hacks are), but we should take the time in the near future to see if we can make this cleaner.

some things to consider here:
1) we have 2 types of filters, scaler and summarization filters- part of the hack is to apply the scaler signatures (by checking if ignore is in the filter name)
2) there are a few custom benchmarks that have either custom subtest summarization code, or custom suite summarization code, in the case of v8, we create special code which requires the test name and data to be added to the function, this in effect causes us to not call the function in the tuple (already prepared), but instead call the function in raw terms.

How can we make this better, so it is easy from the test configuration perspective as well as filter/results/output?  Would it be easier if we defaulted everything to the same filter and only had a few exceptions?  Right now it seems like we have a lot of choices and we could adjust some of our cycles, etc. to make this work.

Comment 1

[Joel Maher ( :jmaher ) (UTC -8)]

not much to do here- I think we have copied this code to raptor and simplified it where possible