|
| 1 | +# 11/28/2024 ESLint TSC Meeting Transcript |
| 2 | + |
| 3 | +**mdjermanovic:** Hi! |
| 4 | + |
| 5 | +**fasttime:** Hi! |
| 6 | + |
| 7 | +**mdjermanovic:** nzakas will not be present today, so we can start |
| 8 | + |
| 9 | +**fasttime:** Yes, just the two of us today |
| 10 | + |
| 11 | +**mdjermanovic:** Let's start with statuses. I made a follow-up refactor PR that adds missing properties in `meta.defaultOptions` of core rules and simplifies getting options in several rules. Also added new `ignoreComputedKeys` option in the `sort-keys` rule and enabled `eslint-plugin-eslint-plugin` in `@eslint/*` language plugins. |
| 12 | + |
| 13 | +**fasttime:** I've been adding type tests for the language plugins and triaging several issues. I'm also working on preparing an RFC for multithread linting. |
| 14 | + * 👍 @mdjermanovic |
| 15 | + |
| 16 | +**mdjermanovic:** RFC Duty schedule |
| 17 | + |
| 18 | +**mdjermanovic:** This week: @fasttime |
| 19 | +December 2: @nzakas |
| 20 | +December 9: @mdjermanovic |
| 21 | +December 16: @fasttime |
| 22 | + * 👍 @fasttime |
| 23 | + |
| 24 | +**mdjermanovic:** We had an action item from the meeting before the previous one, to compare performances starting from ESLint v9.11.1 |
| 25 | + |
| 26 | +**fasttime:** Ah yes |
| 27 | + |
| 28 | +**mdjermanovic:** I did some testing: https://github.com/eslint/eslint/pull/19042#issuecomment-2500523800 |
| 29 | + |
| 30 | +**mdjermanovic:** Some observations: |
| 31 | + |
| 32 | +**mdjermanovic:** 1. Per the "Multiple Individual Files" test made for that PR, the config caching problem introduced in v9.12.0 has been fixed in v9.14.0. |
| 33 | + |
| 34 | +**mdjermanovic:** 2. Interestingly, on the other hand, all of our standard performance tests, which are Loading, Single File (one big file), and Multi Files (450 files matched by a glob) show notable performance improvements in v9.12.0. This was unexpected, but welcome 🙂 |
| 35 | + |
| 36 | +**mdjermanovic:** 3. Surprisingly, the effects of Node.js compile cache, enabled in ESLint v9.13.0, aren't noticeable in test results. |
| 37 | + |
| 38 | +**mdjermanovic:** 4. There seems to be a small but noticeable degradation in Loading and Multiple Individual Files tests in v9.15.0 |
| 39 | + |
| 40 | +**fasttime:** I thought the compile cache was only enabled when launching eslint from the CLI |
| 41 | + |
| 42 | +**mdjermanovic:** I'm not overly confident in the validity of my testing as 2-4 were unexpected, so I'll doublecheck. |
| 43 | + * 👍 @fasttime |
| 44 | + |
| 45 | +**mdjermanovic:** Yes, doesn't the Loading test run CLI? |
| 46 | + |
| 47 | +**fasttime:** So mocha tests would not count unless they run `eslint` as a command |
| 48 | + |
| 49 | +**fasttime:** I think tests in `tests/lib/cli.js` do |
| 50 | + |
| 51 | +**mdjermanovic:** I was mostly expecting to see the effects in the Loading test, as in the PR that enabled the Node.js compile cache: https://github.com/eslint/eslint/pull/19012#issue-2580716644 |
| 52 | + |
| 53 | +**mdjermanovic:** Anyways, I'll repeat the tests and try to figure out if I'm doing something wrong |
| 54 | + |
| 55 | +**fasttime:** I'm not sure what the reason could be. You could try to run a profiler if you haven't yet. But well, perhaps the effect of the compile cache is just not that noticeable for our setup. |
| 56 | + * 👍 @mdjermanovic |
| 57 | + |
| 58 | +**fasttime:** Are you planning to open a new issue to track your results? |
| 59 | + |
| 60 | +**mdjermanovic:** I'll post new results on the same PR, and if something doesn't seem right I'll open an issue (or issues) |
| 61 | + |
| 62 | +**fasttime:** Sounds good, thanks. |
| 63 | + |
| 64 | +**mdjermanovic:** We don't have any issues/PRs tagged for this meeting. Are there any issues/PRs you would like to discuss today? |
| 65 | + |
| 66 | +**fasttime:** Nothing in particular from my side I guess. |
| 67 | + |
| 68 | +**mdjermanovic:** Nothing from my side as well. |
| 69 | + |
| 70 | +**mdjermanovic:** So, we can talk about release |
| 71 | + |
| 72 | +**fasttime:** I could do the release tomorrow. |
| 73 | + |
| 74 | +**mdjermanovic:** Thanks! |
| 75 | + |
| 76 | +**fasttime:** It looks like we will have just `@eslint/js` and `eslint` this time. |
| 77 | + * 👍 @mdjermanovic |
| 78 | + |
| 79 | +**mdjermanovic:** I think the same |
| 80 | + |
| 81 | +**fasttime:** I will be probably ignoring browser test errors because those are very unreliable lately, and I think it's not useful to retry running failed jobs until the browser test happens to pass. |
| 82 | + |
| 83 | +**mdjermanovic:** Yeah, we can check locally and if it works then we can consider it successful (interesting that it usually works locally, at least for me, but not in CI) |
| 84 | + |
| 85 | +**fasttime:** For me, the browser tests works locally on Windows. It fails always on MacOS. |
| 86 | + |
| 87 | +**fasttime:** But yes, seems good to run the test locally. |
| 88 | + * 👍 @mdjermanovic |
| 89 | + |
| 90 | +**fasttime:** Especially if doesn't pass on CI. |
| 91 | + * 👍 @mdjermanovic |
| 92 | + |
| 93 | +**fasttime:** Okay, I don't see any issues that should be included in tomorrow's release. Do you have any? |
| 94 | + |
| 95 | +**mdjermanovic:** I don't see nothing in particular either |
| 96 | + |
| 97 | +**fasttime:** Alright. Anything else we should discuss? |
| 98 | + |
| 99 | +**mdjermanovic:** Nothing in particular for today from my side |
| 100 | + |
| 101 | +**fasttime:** Nothing from me either. We can call it a meeting 🙂 |
| 102 | + |
| 103 | +**mdjermanovic:** Yes 🙂 |
| 104 | + |
| 105 | +**mdjermanovic:** Thanks! 👋 |
| 106 | + |
| 107 | +**fasttime:** Thanks! And thanks for the notes @sam3k_ |
| 108 | + * 👍 @mdjermanovic |
0 commit comments