1417035 - Stop using <plaintext> in JSON Viewer

Status: RESOLVED FIXED

(DevTools :: JSON Viewer, enhancement)

Description

[Oriol Brufau [:Oriol]]

Using a <plaintext> element to switch the tokenizer to plaintext state seemed a good idea for bug 1368899.
However,
 - It's an obsolete element
 - It replaces U+0000 with U+FFFD (replacement character)
 - The current approach sends UTF-8 as is and converts UTF-16 into UTf-8. The latter automatically removes the BOM because the ignoreBOM parameter of TextDecoder is not supported (bug 1102679). Then the frontend removes the BOM again. Could be solved by checking the encoding, but doing different things depending on the encoding is getting messy.

So it would be better to always use a TextDecoder to obtain the data, and then insert it manually instead of sending it as part of the stream.

Comment 1

[Oriol Brufau [:Oriol]]

Attachment: Bug 1417035 - Stop using <plaintext> in JSON Viewer

Review commit: https://reviewboard.mozilla.org/r/199388/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/199388/

Comment 2

[Jan Honza Odvarko [:Honza] (always need-info? me)]

Comment on attachment 8928145 [details]
Bug 1417035 - Stop using <plaintext> in JSON Viewer

https://reviewboard.mozilla.org/r/199388/#review205424

Looks reasonable, R+ assuming try is green.

Honza

Comment 3

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/25f0af5e17cf
Stop using <plaintext> in JSON Viewer r=Honza

Comment 4

[Andrei Ciure[:aciure]]

https://hg.mozilla.org/mozilla-central/rev/25f0af5e17cf