| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
|
@refack build started: https://ci.nodejs.org/blue/organizations/jenkins/node-test-pull-request-lite-pipeline/detail/node-test-pull-request-lite-pipeline/797/pipeline |
Sorry, something went wrong.
There was a problem hiding this comment.
We currently publish the pdb, e.g. https://nodejs.org/dist/v10.9.0/win-x64/node_pdb.zip so this is semver-major?
cc @nodejs/platform-windows @nodejs/diagnostics
Sorry, something went wrong.
There was a problem hiding this comment.
See also
Lines 372 to 378 in 4cea01a
which is only run on the CI for releases (not for regular PR's).
Sorry, something went wrong.
There was a problem hiding this comment.
Ohh. This generates a .pdb file but only a single one during linking.
Sorry, something went wrong.
There was a problem hiding this comment.
This will exhaust the memory available in some machines (including our CI). Ref 6cfedf0 and #12184 (comment) . I tested this extensively when I made this change, to find the best combination that uses a reasonable number of threads, and this was the best. CL will use one thread for each processor, but since some of our projects are quite small this helps a lot. If we did it the other way around, not limiting here and limiting in CL, V8 would keep compiling long after everything else finished.
Sorry, something went wrong.
There was a problem hiding this comment.
On my 4 CPU box this does not affect compilation time.
Sorry, something went wrong.
There was a problem hiding this comment.
I'll continue testing this, but meanwhile I'll remove this from this PR.
Sorry, something went wrong.
There was a problem hiding this comment.
I didn't test yet, but this looks quite nice. Is it possible to simply set a much lower warning level for deps?
Sorry, something went wrong.
There was a problem hiding this comment.
I've been thinking about that, the problem is common.gypi is used by node-gyp so I didn't want to make a big change without further testing.
Sorry, something went wrong.
There was a problem hiding this comment.
This was needed in L482 as well
Sorry, something went wrong.
There was a problem hiding this comment.
I didn't test yet, but this looks quite nice.
output is filled with:
d:\code\node\src\node_internals.h(321): warning C4244: 'argument': conversion from 'const uint64_t' to 'const NativeT', possible loss of data
Sorry, something went wrong.
Sorry, something went wrong.
|
Could you describe this a bit more? Why are these required to re-enable ninja builds? |
Sorry, something went wrong.
|
Sorry, something went wrong.
|
This looks mostly good. @refack just two notes about the commit messages:
|
Sorry, something went wrong.
|
https://ci.nodejs.org/job/node-test-commit-windows-fanned/20664/ @joaocgreis updated messages |
Sorry, something went wrong.
There was a problem hiding this comment.
Looks great, thanks @refack!
Sorry, something went wrong.
* Fixes ninja build PR-URL: nodejs#22698 Reviewed-By: João Reis <reis@janeasystems.com>
* previusly set to generate a .pdb file for each .obj file * enables clcache PR-URL: nodejs#22698 Reviewed-By: João Reis <reis@janeasystems.com>
* also unify warning exclusion directive PR-URL: nodejs#22698 Reviewed-By: João Reis <reis@janeasystems.com>
Sorry, something went wrong.
* Fixes ninja build PR-URL: #22698 Reviewed-By: João Reis <reis@janeasystems.com>
* previusly set to generate a .pdb file for each .obj file * enables clcache PR-URL: #22698 Reviewed-By: João Reis <reis@janeasystems.com>
* also unify warning exclusion directive PR-URL: #22698 Reviewed-By: João Reis <reis@janeasystems.com>
| Back | FazBrowse Home | New Git URL |
Checklist