1507950 - Allow calling ReadableStream controller.error() more than once

Status: RESOLVED FIXED

(Core :: JavaScript Engine, enhancement, P2)

Description

[Jason Orendorff [:jorendorff]]

The spec has changed since we implemented it. We're too strict now; this causes some test failures in testing/web-platform/tests/streams/readable-streams/bad-underlying-sources*.html.

Comment 1

[Jason Orendorff [:jorendorff]]

This is probably also causing the failures in testing/web-platform/tests/streams/readable-streams/garbage-collection*.html .

Comment 2

[Jason Orendorff [:jorendorff]]

Attachment: Bug 1507950 - Allow calling controller.error() when the stream is not readable. r?arai

Comment 3

[Jason Orendorff [:jorendorff]]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=6458c22043577b1620853d0e60c20868f3a50522

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/bc83a2fe5c17
Allow calling controller.error() when the stream is not readable. r=arai

Comment 5

[Pulsebot]

Backout by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/7eb42458e2d8
Backed out 5 changesets (bug 1507950, bug 1503012, bug 1507943, bug 1512050, bug 1512008) for spidermonkey and jit failures

Comment 6

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/a6eb6e3732ae
Allow calling controller.error() when the stream is not readable. r=arai

Comment 7

[Arthur Iakab [arthur_iakab]]

https://hg.mozilla.org/mozilla-central/rev/a6eb6e3732ae

Comment 8

[Chris Mills [:cmills]]

Note to MDN writers:

I'm not convinced this desparately needs to be added to the release notes, given that readable streams have only been enabled by default anyway. I think this just needs us to check whether the readable stream controller docs report the .error() behavior accurately.

Comment 9

[Chris Mills [:cmills]]

I've documented this by adding a note to the error() page:
https://developer.mozilla.org/en-US/docs/Web/API/ReadableStreamDefaultController/error

I didn't think it was worth calling this out specifically in the compat data because it was only enabled by default in Firefox 65 anyway.

Let me know if that's OK. thanks!