In Bug 827786 we added a small string to warn user about downloading on 2G/3G network. But we're showing this also when we have no connection at all. I think we should show another text when we have no connection at all (ie no data connection and no wifi connection), because we know this will ultimately fail.

note: We should not download from mobile data connection at all; we should remove warning as the ability mobile network.

(In reply to Naoki Hirata :nhirata from comment #1) > note: We should not download from mobile data connection at all; we should > remove warning as the ability mobile network. Not what was decided in Berlin...

Not actionable until the UX is sorted out.

(In reply to Naoki Hirata :nhirata from comment #1) > note: We should not download from mobile data connection at all; we should > remove warning as the ability mobile network. Per Etienne, there was a big discussion about this in Berlin, and the plan is to allow downloads over mobile networks, albeit with prominent warnings to the user about the associated risks. (In reply to Julien Wajsberg [:julienw] from comment #0) > Created attachment 701742 [details] > screenshot > > In Bug 827786 we added a small string to warn user about downloading on > 2G/3G network. But we're showing this also when we have no connection at all. > > I think we should show another text when we have no connection at all (ie no > data connection and no wifi connection), because we know this will > ultimately fail. Good catch. I thought we already had something for this, actually. Since that's apparently not the case, we should allow the user to open the prompt, display a small "can't connect" message in-line, and change the two buttons to a single "Cancel" button. I will attach a screenshot in my follow up.

I don't mind knocking this out unless someone else wants it.

UX node: This pull request implements a message when the user has no wifi or network. It continues to use the 'Later' messaging as that what was existing in the markup. It's trivial to change this to use 'Cancel' as specified in the new wire, just let me know.

Comment on attachment 702239 [details] Github pull request pointer Need to make one change first..

Comment on attachment 702239 [details] Github pull request pointer Ok Julien - this should be good to go now. I don't have a sim card on me that can test 3g here, so will look for someone in the office who can test these cases for me.

There is a locale change so adding late-l10n. You'd need to ask for a review from :stas when mine will be over too.

Comment on attachment 702239 [details] Github pull request pointer mostly nits, except that we should use window's online and offline events if these work. Asking a review from :stas for the locale bit.

Julien - Have made the requested changes and everything appears to work fine. Please let me know if there's anything else!

This is late-l10n and is not a critical user issue from our read of the bug - the only reason we would keep this on the list would be if no "connection unavailable" or similar error message came up when the user attempted to download. Geo will test and re-nominate as necessary.

I can say that we definitely get an appropriate error when we start the download in this situation.

Comment on attachment 702239 [details] Github pull request pointer r=me thanks !

Comment on attachment 702239 [details] Github pull request pointer This was tef+, but no longer so asking for approval to land. [Approval Request Comment] Bug caused by (feature/regressing bug #): New feature User impact if declined: Improper message when trying to download update (edge case). Testing completed: Manual testing of all steps, in addition to unit tests. Risk to taking this patch (and alternatives if risky): Possible, but unlikely, impact with the update download screen.

Given that this is late-l10n and tef-, I suggest to not approve landing this. We have to wrap up.

Just a warning - not fixing this bug means we'll still have to fix the translation. Just noting. See screenshot - the pt-BR translation is cut off. If the l10n guys feel we should fix the string, then let's fix the string and file a different bug. If not, then we'll have to fix this and modify the strings. What do you guys want to do?

Stealing from Geo since I tested this by caveat doing packaged app update testing.

Pike> that was vingtetun's thought too. Jason> this is another bug.

Landed in master: https://github.com/mozilla-b2g/gaia/commit/d020661210d3345ba95e1461d7992c22bd050139

Comment on attachment 702239 [details] Github pull request pointer Given Axel's comments, we don't want to uplift this late-l10n patch, it can go on master and ride the trains.

Verified fixed in 2013-02-05-07-02-01 pvt nightly b2g18 build