Closed Bug 1074585 Opened 10 years ago Closed 10 years ago

TEST-UNEXPECTED-FAIL | /builds/slave/test/build/tests/xpcshell/tests/mailnews/compose/test/unit/test_detectAttachmentCharset.js | "Shift_JIS" == "UTF-8" - See following stack:

Categories

(MailNews Core :: Internationalization, defect)

defect
Not set
normal

Tracking

(Not tracked)

RESOLVED FIXED
Thunderbird 35.0

People

(Reporter: jcranmer, Assigned: mkmelin)

References

Details

(Keywords: intermittent-failure, regression)

Attachments

(1 file)

Filing a new bug on this issue, but this is pursuant to the last comment I made. Thoughts on the correct way to go about dealing with this?
Typo in dep bug number?
Flags: needinfo?(Pidgeot18)
(In reply to Ed Morley [:edmorley] from comment #1) > Typo in dep bug number? Yes.
Depends on: 844115
No longer depends on: 884115
Flags: needinfo?(Pidgeot18)
(In reply to Joshua Cranmer [:jcranmer] from comment #0) > Thoughts on the correct way to go about dealing with this? If the contents of the file are valid UTF-8, label as UTF-8. Otherwise, attach as unlabeled.
Blocks: 844115
No longer depends on: 844115
Keywords: regression
Like this? At least the test passes.
Assignee: nobody → mkmelin+mozilla
Status: NEW → ASSIGNED
Attachment #8500143 - Flags: review?(Pidgeot18)
I can see that we will receive some bug reports saying 'Hey! Thunderbird does not set charset from version xx.'. BUT attachment 8500143 [details] [diff] [review] might be better than setting wrong charset on the current trunk.
Comment on attachment 8500143 [details] [diff] [review] bug1074585_test_detectAttachmentCharset.patch Review of attachment 8500143 [details] [diff] [review]: ----------------------------------------------------------------- I'm honestly surprised the fix is this simple, as I thought lacking a charset would cause our compose code to do things with Content-Transfer-Encodings it shouldn't be doing. But after running some manual tests, those bad things that I thought would happen didn't. The biggest "regression" is that this labels US-ASCII as UTF-8. Which might not be a bad thing, because it stops our code from trying to convert LF to CRLF line endings. On the downside, we also force all UTF-8 blobs to use Base64 encoding. I'm not going to ask you to change these, but we should keep a close eye on people filing regressions. ::: mailnews/base/util/nsMsgUtils.cpp @@ +2387,5 @@ > + NS_ENSURE_SUCCESS(rv, rv); > + > + bool isMore = true; > + bool isUTF8Compat = true; > + while (isMore && isUTF8Compat && EWHITESPACE
Attachment #8500143 - Flags: review?(Pidgeot18) → review+
Status: ASSIGNED → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
OS: Linux → All
Hardware: x86_64 → All
Target Milestone: --- → Thunderbird 35.0
Depends on: 1188306
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: