| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
Sorry, something went wrong.
|
Should this also apply to C++ sources and/or Markdown for consistency? (I'm not saying it's a good idea, but I am curious if the arguments for doing this in JavaScript are really specific to JavaScript.) |
Sorry, something went wrong.
There was a problem hiding this comment.
lgtm
Sorry, something went wrong.
I'd prefer to treat each one separately so that decisions about one don't get complicated by considerations that only apply in one of the other contexts. It might also be good to treat this as an experiment. If anyone is hesitant about increasing the limit, then maybe they'd be comfortable with just trying it for a few months and then deciding if we've had to give a lot of line-length nits or not. But here's my take anyway on C++ and markdown:
|
Sorry, something went wrong.
|
Increasing the c++ limit to 120 works for me. |
Sorry, something went wrong.
Massively favor eliminating line breaks. In general, it seems that the ecosystem has adopted this as a standard, and it's extremely strange for us to be an outlier. |
Sorry, something went wrong.
On a side note, I suspect that most of the ecosystem doesn't need to backport documentation changes to multiple release lines with diverging commit histories, so they are probably less concerned about the increased number of conflicts and the reduced usefulness of git operations such as git blame. |
Sorry, something went wrong.
Sorry, something went wrong.
Sorry, something went wrong.
PR-URL: #41586 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: Mestery <mestery@protonmail.com> Reviewed-By: Matteo Collina <matteo.collina@gmail.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Beth Griggs <bgriggs@redhat.com> Reviewed-By: Tierney Cyren <hello@bnb.im>
PR-URL: #41586 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: Mestery <mestery@protonmail.com> Reviewed-By: Matteo Collina <matteo.collina@gmail.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Beth Griggs <bgriggs@redhat.com> Reviewed-By: Tierney Cyren <hello@bnb.im>
PR-URL: #41586 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: Mestery <mestery@protonmail.com> Reviewed-By: Matteo Collina <matteo.collina@gmail.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Beth Griggs <bgriggs@redhat.com> Reviewed-By: Tierney Cyren <hello@bnb.im>
PR-URL: #41586 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: Mestery <mestery@protonmail.com> Reviewed-By: Matteo Collina <matteo.collina@gmail.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Beth Griggs <bgriggs@redhat.com> Reviewed-By: Tierney Cyren <hello@bnb.im>
PR-URL: #41586 Reviewed-By: Michaël Zasso <targos@protonmail.com> Reviewed-By: Mestery <mestery@protonmail.com> Reviewed-By: Matteo Collina <matteo.collina@gmail.com> Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Beth Griggs <bgriggs@redhat.com> Reviewed-By: Tierney Cyren <hello@bnb.im>
The previous deploy (1cf105f) failed its own postbuild verification on Vercel — correctly. On Linux, nf3 nests version-conflicted deps as symlinks *inside* store packages (google-auth-library/node_modules/gaxios and 21 friends), and fs.cpSync({ dereference: true }) only dereferences the top-level path: nested links are recreated as links (nodejs/node#41586), so the "materialized" tree still dangled. Windows never showed it because nf3's junction layout keeps every link top-level, which is exactly why the local verification passed while Vercel's failed. The copy is now hand-rolled: copyReal() stats through symlinks at every depth, so the output contains none regardless of layout. Verified in a clean node:24 container from a tarball of this tree: npm ci + VERCEL=1 build exit 0, 19 links materialized, `find -type l` = 0, and the function artifact serves / at 200 and /api/available-models at 200 in the same container. Also makes the Sentry sourcemap upload non-fatal via errorHandler. The plugin's default is to throw and kill the build on any upload failure — a wrong org slug, an expired token, a Sentry outage. Observability must never take down a deploy. Exercised in the same container with deliberately invalid SENTRY_* values: three 401s from Sentry, three warnings in the log, build exit 0, and zero .map files left in the output. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
| Back | FazBrowse Home | New Git URL |
No description provided.