Meta issue for FIXMEs that reference closed issues

95aaaaf
Opened by Oli Scherer at 2020-01-07 16:13:33

@Eh2406 compiled this awesome list. I think it should be a meta issue. So here it is

https://gist.github.com/Eh2406/b92f034f8adc193894482e19c79d80f1

  1. The good part, this was made with an automated script. I can keep it up-to-date by rerunning as often as we want.

    The bad part, it was made with an automated script. Some of these are false positives, like the first one, where the script was too dumb to realize said it's an issue on a different Repository.

    In general fees can be fixed by:

    1. Doing what the fix me said to do and seeing if the pr passes
    2. Updating the fix me to refer to a more current issue number
    3. Removing the fix me has the issue was closed won't fix

    In the past couple of days I have been reviewing some of these, and if I'm not sure what to do I'v leaving a comment on the clothes issue with a link to the fix me asking for advice.

    Jacob Finkelman at 2017-09-06 13:39:42

  2. Thanks @nivkner! Let me know when I should rerun the script.

    Jacob Finkelman at 2017-10-06 20:05:22

  3. This kind of metabug seems like something that should be tackled continuously, as new FIXMEs are added over time, so if possible this would be better run routinely. Otherwise there really isn't much of a difference between running the script or marking FIXMEs as complete.

    Since I wouldn't ask you to dedicate so much time to this one issue, I'd like to know what is your opinion on turning this script into a github bot like "rust-highfive" or "bors". Then whenever an issue is closed, this "FIXME bot" could update this issue.

    nivkner at 2017-10-07 08:37:37

  4. It'd be helpful if the code links pointed to specific commit IDs, so that the line numbers remain correct.

    varkor at 2018-01-20 00:33:30

  5. @varkor thank you for the suggestion. I changed the script to use the sha in the links, and updated the gist to master.

    @nivkner It looks like my reply to you never posted, I am very sorry for not replying for so long. I don't have the experience to set up a bot like "rust-highfive" or "bors". So I posted the script I am using so someone with more experience can do that if they like. In the meantime it is no bother for me to run it as needed. I had written up a response like this but apparently it did not post.

    Jacob Finkelman at 2018-01-20 20:51:29

  6. Thanks for replying @Eh2406, Could you please post a link here to the script? I can't find it in your github repositories. If you don't mind, then could you also post the output of the script, just so we'll know where we are standing after all this time.

    nivkner at 2018-01-20 21:11:05

  7. The gist is at: https://gist.github.com/Eh2406/b92f034f8adc193894482e19c79d80f1 it has the current (as of today) output and the code. I posted the code on Oct 14, 2017, the same day I failed to reply hear.

    Jacob Finkelman at 2018-01-20 21:21:49

  8. Just found one not picked up by my script at: https://github.com/rust-lang/rust/blob/a86f556ee37c13047f966c35effab8b334b7de67/src/libstd/collections/hash/map.rs#L141-L143

    Jacob Finkelman at 2018-06-07 17:19:54

  9. Sorry for the long break. I am back to running the script and updating the gist on the semi regular.

    Jacob Finkelman at 2019-01-21 01:16:12

  10. Actually it appears that many of those have already been removed; we could use another pass of that script.

    ljedrz at 2020-01-07 13:18:17

  11. The script is still available on the Gist https://gist.github.com/Eh2406/b92f034f8adc193894482e19c79d80f1#file-fixme-py

    I stopped running it (when I updated computers) after see this comment https://internals.rust-lang.org/t/compiler-steering-meeting/8588/49 that @pnkfelix and @oli-obk had a more official project.

    Maybe we can add this data to that project?

    Jacob Finkelman at 2020-01-07 15:32:22

  12. For reference, @oli-obk's project is here.

    varkor at 2020-01-07 16:13:33