738298 - navigator.mozApps.mgmt.getAll() doesn't handle non-ASCII characters in the manifest

Status: VERIFIED FIXED

(Core Graveyard :: DOM: Apps, defect)

Description

[David Flanagan [:djf]]

When I use navigator.mozApps.mgmt.getAll() in the Gaia homescreen, and set the local to French, accented characters in the app name are garbled.  It appears that UTF-8 strings in the manifest are being treated as 7-bit ASCII and not properly converted into JavaScript UTF-16 strings.

Comment 1

[Jason Smith [:jsmith]]

As a point of similarity, this issue also occurs with desktop native installation (see bug 733482). Seeing that this happens in more than spot makes me think that this UTF-8 character issue could be more areas we have not looked at yet.

Comment 2

[Ian Bicking (:ianbicking)]

We should definitely include some non-ASCII manifests in our testing (right now all our test manifests are strictly ASCII).

Comment 3

[[e-mail] [:onecyrenus]]

Hey jason I had written a manual test with japanese characters. 

In the interim, could you add this test to your smoketest. 

https://github.com/mozilla/openwebapps/tree/develop/site/tests/apps/app3

Comment 4

[[e-mail] [:onecyrenus]]

To do that you need to run the node test server, and load appstore.html

the above app will be listed to be installed, and you should be able to install to the dashboard, view on dashboard..etc

Comment 5

[Jason Smith [:jsmith]]

Yup, I've added the UTF-8 character test on native install here on test case IA-15 (https://docs.google.com/spreadsheet/ccc?key=0AiZoGR-iOAlUdDdfMHo5aTI2WjVIeWxnNWxvV0ZVOWc#gid=0).

Comment 6

[Ian Bicking (:ianbicking)]

I you post the manifest somewhere I can add it to the /appdir/ list

Comment 7

[[e-mail] [:onecyrenus]]

ian that touches on another issue which is where / how to host test apps ......

I think the staging server would be a good spot for this, the question is whether we can get ops to open some firewall ports so we can run the node.js test server. 

i think that would be optimal

Comment 8

[Ian Bicking (:ianbicking)]

I think appengine or heroku might be better hosting options for samples.

Comment 9

[Jason Smith [:jsmith]]

Filed bug 738845 for the hosting of test apps. Let's pull the conversation on the test case applications over to that bug.

Comment 10

[[e-mail] [:onecyrenus]]

Jason, do you have an eta on this bug fix ?

Comment 11

[Jason Smith [:jsmith]]

Which bug are you referring to? This one? Bug 738845? If it's bug 738845, please comment in that bug.

Comment 12

[[e-mail] [:onecyrenus]]

I was referring to this bug.

Comment 13

[Jason Smith [:jsmith]]

This is an implementation bug - I'm not the appropriate person to fix this bug.

Comment 14

[[e-mail] [:onecyrenus]]

k, wanted to determine if you had more information about the bug than I did... 

Fabrice: do you have an eta on this bug fix ?

Comment 15

[[:fabrice] Fabrice Desré]

I'll probably have a fix by the end of next week

Comment 16

[Jason Smith [:jsmith]]

Nominating for k9o - this affects localized b2g apps

Comment 17

[[:fabrice] Fabrice Desré]

Attachment: patch

Tested by changing some manifest.webapp files in gaia to use utf-8 instead of the current escaping.

Comment 18

[Vivien Nicolas (:vingtetun) (:21) - (NOT reading bugmails, needinfo? please)]

Comment on attachment 647477 [details] [diff] [review]
patch

Review of attachment 647477 [details] [diff] [review]:
-----------------------------------------------------------------

\o/

Comment 19

[[:fabrice] Fabrice Desré]

https://hg.mozilla.org/integration/mozilla-inbound/rev/2bef20dfc8fb

Comment 20

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/2bef20dfc8fb

Comment 21

[Jason Smith [:jsmith]]

Testing on desktop, this looks okay outside of the regression that was caught on Android. I'll want to test this on Firefox OS though to be sure.

Comment 22

[[:fabrice] Fabrice Desré]

(In reply to Jason Smith [:jsmith] from comment #21)
> Testing on desktop, this looks okay outside of the regression that was
> caught on Android. I'll want to test this on Firefox OS though to be sure.

Which android regression?

Comment 23

[Jason Smith [:jsmith]]

(In reply to Fabrice Desré [:fabrice] from comment #22)
> (In reply to Jason Smith [:jsmith] from comment #21)
> > Testing on desktop, this looks okay outside of the regression that was
> > caught on Android. I'll want to test this on Firefox OS though to be sure.
> 
> Which android regression?

bug 779490

Comment 24

[Jason Smith [:jsmith]]

Verified on 8/6/2012 Firefox OS Daily Build with Ed Lee's app with Japanese characters.