435877 - UTF-8 decoding problem breaks wcap calendar

Status: VERIFIED FIXED

(Calendar :: Provider: WCAP, defect)

Description

[Daniel Boelzle [:dbo]]

Due to recent changes to calWcapRequest (especially introducing strong UTF-8 decoding), some calendars break on loading their user prefs. CS doesn't encode proper UTF-8, thus the octet conversion to UTF-8 fails. Although this is for sure a cs bug, we need to be error tolerant on reading.

Comment 1

[Daniel Boelzle [:dbo]]

Attachment: moving back to unichar stream loader

Comment 2

[Philipp Preißer-Kewisch [:Fallen][☁️📆]]

Comment on attachment 322804 [details] [diff] [review]
moving back to unichar stream loader

>+ * Returns true if this is MOZILLA_1_8_BRANCH, else false.
>+ */
>+function isBranch() {
>+    if (isBranch.mIsBranch === undefined) {
>+        isBranch.mIsBranch = (Components.interfaces.nsIStreamLoader.number == "{31d37360-8e5a-11d3-93ad-00104ba0fd40}");
>+    }
>+    return isBranch.mIsBranch;
>+}
Components.classes["@mozilla.org/xre/app-info;1"].getService(Components.interfaces.nsIXULAppInfo).platformVersion.match(/^1.8/);


r=philipp

Comment 3

[Daniel Boelzle [:dbo]]

(In reply to comment #2)
> (From update of attachment 322804 [details] [diff] [review])
> >+ * Returns true if this is MOZILLA_1_8_BRANCH, else false.
> >+ */
> >+function isBranch() {
> >+    if (isBranch.mIsBranch === undefined) {
> >+        isBranch.mIsBranch = (Components.interfaces.nsIStreamLoader.number == "{31d37360-8e5a-11d3-93ad-00104ba0fd40}");
> >+    }
> >+    return isBranch.mIsBranch;
> >+}
> Components.classes["@mozilla.org/xre/app-info;1"].getService(Components.interfaces.nsIXULAppInfo).platformVersion.match(/^1.8/);

Hmm, I was thinking about that, too. But app-info is not available e.g. in unit tests, thus I moved over to testing some interface version that has changed from branch to trunk. Any objections against doing so?

Comment 4

[Daniel Boelzle [:dbo]]

Checked in on HEAD and MOZILLA_1_8_BRANCH => FIXED.

Philipp, if your have any objections w.r.t. comment #3, we can fix that later on.

Comment 5

[Daniel Boelzle [:dbo]]

repopening because of code regression

Comment 6

[Daniel Boelzle [:dbo]]

Attachment: fixes regression

Comment 7

[Daniel Boelzle [:dbo]]

Checked in fix on HEAD and MOZILLA_1_8_BRANCH =< FIXED again.

Comment 8

[Daniel Boelzle [:dbo]]

Attachment: fixes trunk

Comment 9

[Daniel Boelzle [:dbo]]

Checked in trunk patch on HEAD and MOZILLA_1_8_BRANCH => FIXED once again.

Comment 10

[Andreas Treumann]

Checked in lightning build 2008080603 and sunbird 20080805 -> VERIFIED