257854 2023-06-08 09:01:48 -0700 WPT export should not use the bug URL in the commit subject (title) 2026-05-23 18:32:15 -0700 1 1 1 Unclassified WebKit Tools / Tests WebKit Nightly Build Unspecified Unspecified NEW https://bugs.webkit.org/show_bug.cgi?id=274079 https://bugs.webkit.org/show_bug.cgi?id=297919 InRadar P2 Normal --- 1 annevk gsnedders bfan2 gsnedders webkit-bug-importer youennf zcorpan oldest_to_newest 1960560 0 annevk 2023-06-08 09:01:48 -0700 That's not great context for the upstream project. 1961008 1 webkit-bug-importer 2023-06-12 05:24:30 -0700 <rdar://problem/110638179> 1961009 2 youennf 2023-06-12 05:28:21 -0700 Pull request: https://github.com/WebKit/WebKit/pull/14859 2019689 3 ntim 2024-03-08 12:47:52 -0800 *** Bug 270349 has been marked as a duplicate of this bug. *** 2072571 4 gsnedders 2024-11-04 12:16:16 -0800 From the dupe: (In reply to Sam Sneddon [:gsnedders] from comment #0) > Currently Chromium and Mozilla exports contain the full commit message, with > some metadata. e.g. > https://github.com/web-platform-tests/wpt/commit/ > 9aa0f67e426fac988de1625eb6c8926e225df602 (Chromium) or > https://github.com/web-platform-tests/wpt/commit/ > 91dd2eb6f5a512dc4329a0d334e17a498c249f24 (Mozilla). > > By comparison, our exporter produces a commit message like "WebKit export of > https://bugs.webkit.org/show_bug.cgi?id=269971" or "Export made from a > WebKit repository". > > These are much less useful, and provide much less context for the change > being made (especially in the latter case!). Which is arguably broader in scope than just the summary line. 2124031 5 gsnedders 2025-06-16 11:52:56 -0700 I do wonder if this is better fixed by using `git-am` rather than `git-apply`; but we should almost certainly add more metadata (e.g., linking to where it was reviewed as is true for Chromium and Firefox). That also then leaves the commit in a state where if there's a merge conflict you can run `git-am --continue` and get the right commit message and author etc. 2138858 6 gsnedders 2025-08-26 10:56:26 -0700 I'd suggest, based on our commit messages: * We move bugs away from the summary line (because "Limit the scope of invalidation traversal for :has(> .changed) .subject https://bugs.webkit.org/show_bug.cgi?id=297893 rdar://159173631" is a ridiculous summary line, when you fold whitespace) into `Bug` trailers * Rename our `Canonical Link` "trailer" to `WebKit-Canonical-Link` * Possibly add a `Reviewed-on: [PR]` trailer * Maybe move the "Reviewed by…" line to something more like Chromium's `Reviewed-by: [name] <[email]>` or Mozilla's `gecko-reviewers: [username]`, so that the start of the commit message is the actual content? * Maybe drop the changelog from the commit message? Brianna's code to parse commit messages and extract sections would be helpful here… 2138912 7 ntim 2025-08-26 15:13:35 -0700 *** Bug 297925 has been marked as a duplicate of this bug. *** 2213573 8 gsnedders 2026-05-23 18:32:15 -0700 Currently, the title is a load-bearing part of how wpt-pr-bot detects WebKit exports, given it has to handle exports from arbitrary users (to mitigate the risk of a malicious actor getting review approved by claiming to be a WPT export): https://github.com/web-platform-tests/wpt-pr-bot/blob/3348953311e68835f1a2d03ca3384ee37b6fa199/lib/metadata/index.js#L113-L133 This needs to be changed prior to us landing any change here.