itertools icon indicating copy to clipboard operation
itertools copied to clipboard

Published 20 hours ago • verified publisher icon rust-itertools

Maintenance

Open jplatte opened this issue 5 years ago • 3 comments

Hi! It seems to me that itertools is quite low on maintenance power – a fairly big amount of PRs have the waiting-on-review tag, some are untagged and some issues also haven't received any comments.

Is there anything I can do to help?

jplatte avatar Feb 19 '20 12:02 jplatte

@bluss, @phimuemue, @jswrenn could you clarify this?

Veetaha avatar Jul 13 '20 16:07 Veetaha

When we receive a PR for a new method, we need to resolve:

  • [ ] Without the addition, how would the operation be expressed with Itertools and the standard library? Is it worth adding to Itertools?
  • [ ] Is the name intuitive?
  • [ ] If it's an adaptor: ensure that constructing the adaptor on an infinite iterator will terminate.
  • [ ] Will the implementation ever panic? Can this be avoided? Is this documented?
  • [ ] Does it behave sensibly on edge-cases? Is its behavior documented?
  • [ ] Does it have illustrative doctests?
  • [ ] Can the implementation be simplified, or consolidated with other parts of Itertools?

If the contribution is an optimized specialization of an Iterator method (such as nth or fold), the specialization needs to be very thoroughly tested (preferably with quickcheck tests) to ensure it behaves identically (especially on edge cases, and with when it panics).

Reviews that resolve any of the above are very helpful!

The bar for inclusion in itertools is much lower than it is for the standard library, but we'd still like to know that additions will be helpful to more than just the person submitting it. If there's a PR proposing something you need, leave a note describing your use-case. This feedback is helpful for prioritizing PRs, too.

jswrenn avatar Jul 13 '20 16:07 jswrenn

Hm… I think I'm more concerned about things that aren't new features missing reviews (e.g. finally getting everything rustfmt'ed so I don't have to bypass my default format-on-save constantly). Of the 23 PRs that are currently open, only 8 are labeled 'waiting on review'. There's also lots of feature requests that have no answers from a maintainer. I usually ask whether a PR is welcome before writing the code. Should I not do that here? Should I just ping somebody if I don't get an answer after a while?

jplatte avatar Aug 26 '20 16:08 jplatte