search

rust-itertools / itertools

Extra iterator adaptors, iterator methods, free functions, and macros.
https://docs.rs/itertools/
Apache License 2.0
2.64k stars 299 forks source link

Fuse iterators in InterleaveShortest #872

Open phimuemue opened 5 months ago

phimuemue commented 5 months ago

In my book, InterleaveShortest should not produce more items than Interleave, but so far it did for unfused source iterators.

This PR brings InterleaveShortest in line with Interleave. The documentation already (incorrectly) stated that InterleaveShortest is fused.

Related to #533.

codecov[bot] commented 5 months ago

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Comparison is base (6814180) 94.38% compared to head (d8f7f59) 94.28%. Report is 2 commits behind head on master.

Additional details and impacted files ```diff @@ Coverage Diff @@ ## master #872 +/- ## ========================================== - Coverage 94.38% 94.28% -0.11% ========================================== Files 48 48 Lines 6665 6665 ========================================== - Hits 6291 6284 -7 - Misses 374 381 +7 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

Philippe-Cholet commented 5 months ago

I started an investigation ("blame" in the web pages of our code is even more powerful than I thought (for each line, we can have navigate its history)): In #31 it was added and was actually fused. Then the commit that is interesting us here is https://github.com/rust-itertools/itertools/commit/beaf3955b51f5a5a2c5861d34ebd591736377b84 (by bluss, he missed one comment in this "unfuse" commit that you found).

phimuemue commented 5 months ago

I started an investigation ("blame" in the web pages of our code is even more powerful than I thought (for each line, we can have navigate its history)): In #31 it was added and was actually fused. Then the commit that is interesting us here is beaf395 (by bluss, he missed one comment in this "unfuse" commit that you found).

I'm fine with either unfused or fused, but I'd vote for uniform behavior. Not sure if we have a general guideline regarding fusedness, though.

Philippe-Cholet commented 5 months ago

I'm fine with either unfused or fused, but I'd vote for uniform behavior. Not sure if we have a general guideline regarding fusedness, though.

I agree. I think it would be great to have a clear guideline about fusedness. And maybe better tests about (un)fusedness. I'm not sure all the iterators behave correctly in that regard (and have correct documentation). It's not very high on my TODO list though as I'd rather triage/complete old issues/PRs and work on fold specializations and maybe_/Try to eventually move to use const-generics.

Philippe-Cholet commented 5 months ago

I'm gonna make an issue about fusedness to see opened issues and PRs related to it, clarify the situation (like make a guideline) and eventually make a TODO list about all this.