https://mxr.mozilla.org/mozilla-central/source/dom/apps/src/Webapps.jsm#3467 is effectively dead code (unless I am really missing something) at somepoint between v1.2 and now we removed the addition of the flag but this logic is still needed so we can correctly detect when an app download was canceled vs some other network error

Was removed here https://github.com/mozilla/gecko-dev/commit/267ecd7ab1c2e0bb077417d8b31baa3cd3526808

Comment on attachment 8442652 [details] [diff] [review] Correctly cleanup after canceling a mozapp download I came up with this from a combination of reading what the offline (app cache) case does and digging up the past functionality from v1.2.

Comment on attachment 8442654 [details] [diff] [review] Correctly cleanup after canceling a mozapp download Review of attachment 8442654 [details] [diff] [review]: ----------------------------------------------------------------- Hm... so I'm pretty sure we had something similar in the past and ferjm removed it. CC-ing fernando. ::: dom/apps/src/Webapps.jsm @@ +2779,5 @@ > > + // save the current state of the app to handle cases where we may be > + // retrying a past download. > + yield DOMApplicationRegistry._saveApps(); > + yield DOMApplicationRegistry.broadcastMessage("Webapps:UpdateState", { no yield here.

try: https://tbpl.mozilla.org/?tree=Try&rev=d27d96b8bce1

More complete try run https://tbpl.mozilla.org/?tree=Try&rev=8e8352871d35 (looks like some of the apps tests do not run on b2g?)

Comment on attachment 8443659 [details] [diff] [review] Correctly cleanup after canceling a mozapp download Review of attachment 8443659 [details] [diff] [review]: ----------------------------------------------------------------- r=me with nit fixed. ::: dom/apps/src/Webapps.jsm @@ +2776,5 @@ > > // initialize the progress to 0 right now > oldApp.progress = 0; > > + // save the current state of the app to handle cases where we may be nit: Save

Comment on attachment 8443776 [details] [diff] [review] Correctly cleanup after canceling a mozapp download carying r+ over from fabrice... this fixes the review comment.

Pushed to b2g-inbound (with fix to the commit message with r=fabrice). https://hg.mozilla.org/integration/b2g-inbound/rev/59fda077a2ab

Comment on attachment 8443776 [details] [diff] [review] Correctly cleanup after canceling a mozapp download [Approval Request Comment] Bug caused by (feature/regressing bug #): User impact if declined: Not able to correctly resume app download in gaia (FxOS) Testing completed (on m-c, etc.): tests written (live in gaia) Risk to taking this patch (and alternatives if risky): low risk mostly adding back some removed code String or IDL/UUID changes made by this patch: none

Blocks a blocker (bug 1025782).

Testing follow-up landed: https://github.com/mozilla-b2g/gaia/commit/952b1e06fdb6eeade76f28aed78ee9eb070876e1

Sorry, that follow-up was meant for the blocking bug, bug 1023950.