810172 - session restore corrupts non ASCII quick filter data

Status: RESOLVED FIXED

(Thunderbird :: General, defect)

Description

[Makoto Kato [:m_kato]]

We should save/read session.json by UTF-8.  Also when saving session.json, we should use NetUtil.asyncCopy.

Step
1. input non-ASCII character (あああ) to quick filter on main Window.
2. quit Thunderbird
3. launch Thunderbird

Result
value of quick filter isn't 1's character.

Comment 1

[Makoto Kato [:m_kato]]

Attachment: fix

- fix character encoding
- Use services.jsm for session restore code
- Use async store

Comment 2

[:Irving Reid (No longer working on Firefox)]

Comment on attachment 680523 [details] [diff] [review]
fix

Review of attachment 680523 [details] [diff] [review]:
-----------------------------------------------------------------

This looks good. I wonder if there are other configuration and state files that don't handle non-ASCII data correctly.

Comment 3

[Makoto Kato [:m_kato]]

https://hg.mozilla.org/comm-central/rev/5296b0401e94

If using JSON file for any data with non-ASCII, this same problem will occur.

Comment 4

[Makoto Kato [:m_kato]]

backout due to mozmill failure.
https://hg.mozilla.org/comm-central/rev/bcf2893ee19a

https://tbpl.mozilla.org/php/getParsedLog.php?id=17096940&tree=Thunderbird-Trunk

TEST-UNEXPECTED-FAIL | /home/cltbld/talos-slave/test/build/mozmill/session-store/test-session-store.js | test-session-store.js::test_periodic_session_persistence_simple
TEST-UNEXPECTED-FAIL | /home/cltbld/talos-slave/test/build/mozmill/session-store/test-session-store.js | test-session-store.js::test_periodic_nondirty_session_persistence
TEST-UNEXPECTED-FAIL | /home/cltbld/talos-slave/test/build/mozmill/session-store/test-session-store.js | test-session-store.js::test_single_3pane_periodic_session_persistence
TEST-UNEXPECTED-FAIL | /home/cltbld/talos-slave/test/build/mozmill/session-store/test-session-store.js | test-session-store.js::test_multiple_3pane_periodic_session_persistence

Comment 5

[Makoto Kato [:m_kato]]

humm, I change to async I/O code, but mozmill test expects sync'ed I/O...  It depends on timing...

Comment 6

[Wayne Mery (:wsmwk)]

(In reply to Makoto Kato from comment #5)
> humm, I change to async I/O code, but mozmill test expects sync'ed I/O... 
> It depends on timing...

there's another recent bug about this, no?

Comment 7

[Makoto Kato [:m_kato]]

(In reply to Wayne Mery (:wsmwk) from comment #6)
> (In reply to Makoto Kato from comment #5)
> > humm, I change to async I/O code, but mozmill test expects sync'ed I/O... 
> > It depends on timing...
> 
> there's another recent bug about this, no?

No, this problem is test code...

Comment 8

[Wayne Mery (:wsmwk)]

will this also fix bug 657478?

Comment 9

[Makoto Kato [:m_kato]]

(In reply to Wayne Mery (:wsmwk) from comment #8)
> will this also fix bug 657478?

That's right! That bug should be dup to this.

Comment 11

[alta88]

see also Bug 894793.

Comment 12

[alta88]

fixed in Bug 894793.