| commit | f0c0961a1d26e6c010af4d003c6d7008157e73a2 | |
| author | Lucian Wischik <ljw@fb.com> | |
| Tue, 14 Sep 2021 06:39:51 +0000 (13 23:39 -0700) | ||
| committer | Facebook GitHub Bot <facebook-github-bot@users.noreply.github.com> | |
| Tue, 14 Sep 2021 06:42:26 +0000 (13 23:42 -0700) | ||
| tree | 46e9f4513e8ef95b28efd17c9dd8fbebbcf69799 | treesnapshot (tar.gz zip) |
| parent | 49167bb47c33d6afb2c77471c512b35a936724d7 | commitdiff |
I think env should be preserved out of do_naming
Summary:
Error streaming was introduced in D29454477 (https://github.com/facebook/hhvm/commit/6805e3fbe5e5ced2cd3ecbf1ab49418839d81f7f). A comment on that diff says:
> This is to mirror the signature of do_naming, do_typing, do_decling and so on. These never access env.errorl. env.errorl is read at the beginning of the outer function, is threaded through do_parsing, do_naming, do_typing, etc., then is put back into env.errorl.
But looking at the code, I think that env is in fact NOT being threading out of do_naming. The result would be that env isn't aware we've already pushed out some "duplicate name" errors.
Reviewed By: CatherineGasnier
Differential Revision: D30882983
fbshipit-source-id: fc1acf88dc21ab48730100db25fba6da0e13089e
Summary:
Error streaming was introduced in D29454477 (https://github.com/facebook/hhvm/commit/6805e3fbe5e5ced2cd3ecbf1ab49418839d81f7f). A comment on that diff says:
> This is to mirror the signature of do_naming, do_typing, do_decling and so on. These never access env.errorl. env.errorl is read at the beginning of the outer function, is threaded through do_parsing, do_naming, do_typing, etc., then is put back into env.errorl.
But looking at the code, I think that env is in fact NOT being threading out of do_naming. The result would be that env isn't aware we've already pushed out some "duplicate name" errors.
Reviewed By: CatherineGasnier
Differential Revision: D30882983
fbshipit-source-id: fc1acf88dc21ab48730100db25fba6da0e13089e
| hphp/hack/src/server/serverTypeCheck.ml | diffblobblamehistory |