Closed Bug 1405615 Opened 3 years ago Closed 3 years ago
_rs to 0 .7 .1
Bug 1405615 - encoding_rs 0.7.1: Correctly encode U+DC00 followed by another low surrogate from UTF-16. .
59 bytes, text/x-review-board-request
Quoting the git changeset message: Correctly encode U+DC00 followed by another low surrogate from UTF-16. `wrapping_sub()`-based high surrogate check was off by one due to error in copy and paste when defining the constant to compare against. That is, the subtraction that defines the constant was completely wrong but the result of the subtraction was only off by one, which is why the bug wasn't discovered immediately. This lead to the first low surrogate (U+DC00), and only the first low surrogate, getting accepted as a high surrogate. Discovered using cargo-fuzz.
Also contains a bit of SIMD refactoring.
(In reply to Henri Sivonen (:hsivonen) from comment #1) > Also contains a bit of SIMD refactoring. And making some uselessly-mutable variables immutable. I.e. small stuff that was in queue for a release, but the UTF-16 thing was what triggered a new release.
The Android failures on try look unrelated, but the base revision on inbound looks OK. I guess I should rebase and see if the failures go away.
Comment on attachment 8915075 [details] Bug 1405615 - encoding_rs 0.7.1: Correctly encode U+DC00 followed by another low surrogate from UTF-16. . https://reviewboard.mozilla.org/r/186332/#review192482
Attachment #8915075 - Flags: review?(VYV03354) → review+
> review+ Thanks. Triggering another try run after rebase to see if the unexplained oranges go away.
Pushed by firstname.lastname@example.org: https://hg.mozilla.org/integration/autoland/rev/7abbc56b3bb7 encoding_rs 0.7.1: Correctly encode U+DC00 followed by another low surrogate from UTF-16. r=emk.
You need to log in before you can comment on or make changes to this bug.