| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
Sorry, something went wrong.
|
Is this a candidate for fast-track? |
Sorry, something went wrong.
There was a problem hiding this comment.
lgtm
Sorry, something went wrong.
No, I'm not sure it should land as is since we (at least I) don't fully understand the implications. Ideally folks who wrote the original code or contributed to it could weigh in and say "yes, this was why" |
Sorry, something went wrong.
|
Generally when code looks "weird" in Node it either fixes a bug/edge case (though I'd expect a test in this case we're pretty diligent about making PRs include tests) or it's there for some precondition that is no longer true. |
Sorry, something went wrong.
|
Just that it does not get overseen because it is a review remark: 'use strict'
const { join } = require('path');
const { cp } = require('fs');
const net = require('net');
const sock = join(__dirname, `${process.pid}.sock`);
const server = net.createServer();
server.listen(sock);
cp(sock, __dirname, (err) => {
console.log(err.stack)
server.close();
});before (main): aras@aras-Lenovo-Legion-5-17ARH05H:~/workspace/node$ node test.js
SystemError [ERR_FS_CP_NON_DIR_TO_DIR]: Cannot overwrite non-directory with directory: cp returned ENOTDIR (cannot overwrite non-directory /home/aras/workspace/node/51128.sock with directory /home/aras/workspace/node) /home/aras/workspace/node
at new SystemError (node:internal/errors:256:5)
at new NodeError (node:internal/errors:367:7)
at checkPaths (node:internal/fs/cp/cp:98:13)
at async cpFn (node:internal/fs/cp/cp:65:17)after (this PR): aras@aras-Lenovo-Legion-5-17ARH05H:~/workspace/node$ ./node test.js
SystemError [ERR_FS_CP_NON_DIR_TO_DIR]: Cannot overwrite non-directory with directory: cp returned ENOTDIR (cannot overwrite non-directory /home/aras/workspace/node/51188.sock with directory /home/aras/workspace/node) /home/aras/workspace/node
at checkPaths (node:internal/fs/cp/cp:98:13)
at async cpFn (node:internal/fs/cp/cp:65:17) |
Sorry, something went wrong.
|
stackTraceLimit is definitely still in use here. I think the bug is that we should be calling Error.captureStackTrace(this, new.target) instead of Error.captureStackTrace(this). Passing new.target as the 2nd parameter will hide the two extra stack frames you are seeing for NodeError and SystemError. |
Sorry, something went wrong.
|
Actually... in captureLargerStackTrace Line 562 in f05d6d0 you need to change it to ErrorCaptureStackTrace(err, err.constructor);But I wonder if we actually should keep captureLargerStackTrace on the long run. |
Sorry, something went wrong.
|
I tested it. If I patch it with err.constructor it removes the SystemError frames, but basically stackTraceLimit is broken, and instead of getting 3 frames like in my test, i get all 10 frames. |
Sorry, something went wrong.
|
If #49990 lands, this PR should fit perfectly. |
Sorry, something went wrong.
Sorry, something went wrong.
Commit Queue failed- Loading data for nodejs/node/pull/49956 ✔ Done loading data for nodejs/node/pull/49956 ----------------------------------- PR info ------------------------------------ Title errors: fix stacktrace of SystemError, add more tests and benchmark (#49956) Author Aras Abbasi (@Uzlopak) Branch Uzlopak:improve-system-error -> nodejs:main Labels errors Commits 1 - errors: fix stacktrace of SystemError Committers 1 - uzlopak PR-URL: https://github.com/nodejs/node/pull/49956 Reviewed-By: Matteo Collina Reviewed-By: Benjamin Gruenbaum Reviewed-By: Stephen Belanger Reviewed-By: James M Snell ------------------------------ Generated metadata ------------------------------ PR-URL: https://github.com/nodejs/node/pull/49956 Reviewed-By: Matteo Collina Reviewed-By: Benjamin Gruenbaum Reviewed-By: Stephen Belanger Reviewed-By: James M Snell -------------------------------------------------------------------------------- ℹ This PR was created on Fri, 29 Sep 2023 15:08:46 GMT ✔ Approvals: 4 ✔ - Matteo Collina (@mcollina) (TSC): https://github.com/nodejs/node/pull/49956#pullrequestreview-1651669625 ✔ - Benjamin Gruenbaum (@benjamingr) (TSC): https://github.com/nodejs/node/pull/49956#pullrequestreview-1651740602 ✔ - Stephen Belanger (@Qard): https://github.com/nodejs/node/pull/49956#pullrequestreview-1669388744 ✔ - James M Snell (@jasnell) (TSC): https://github.com/nodejs/node/pull/49956#pullrequestreview-1795453904 ✘ Last GitHub CI failed ℹ Last Full PR CI on 2023-12-19T01:35:44Z: https://ci.nodejs.org/job/node-test-pull-request/56378/ - Querying data for job/node-test-pull-request/56378/ ✔ Last Jenkins CI successful -------------------------------------------------------------------------------- ✔ Aborted `git node land` session in /home/runner/work/node/node/.ncuhttps://github.com/nodejs/node/actions/runs/7309376899 |
Sorry, something went wrong.
PR-URL: #49956 Reviewed-By: Matteo Collina <matteo.collina@gmail.com> Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com> Reviewed-By: Stephen Belanger <admin@stephenbelanger.com> Reviewed-By: James M Snell <jasnell@gmail.com>
PR-URL: #49956 Reviewed-By: Matteo Collina <matteo.collina@gmail.com> Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com> Reviewed-By: Stephen Belanger <admin@stephenbelanger.com> Reviewed-By: James M Snell <jasnell@gmail.com>
PR-URL: #49956 Reviewed-By: Matteo Collina <matteo.collina@gmail.com> Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com> Reviewed-By: Stephen Belanger <admin@stephenbelanger.com> Reviewed-By: James M Snell <jasnell@gmail.com>
| Back | FazBrowse Home | New Git URL |
Before I wanted to tackle the performance of SystemError I wanted better test coverage for SystemError.
While I was working on the tests, I realized that the stacktrace for SystemError is wrong. I fixed it accordingly ;). It also has slightly performance gains.
So my proposal is:
Benefit of this separate PR is, that we can see by the changes in the tests, if the behaviour of SystemError got changed.