Seen on today's (06/09) master buri build: Gecko-1a316f5 Gaia-deed49c STR 1. Receive a OMA CP message 2. Open it but do not process it completely 3. Press home key EXPECTED The user will be able to come to process that message later ACTUAL Once the message is open and although it has not been completely processed if user presses home key, the message is discarded without any advise. This issue was reported in the past for 1.3 and solved, please see bug 934354. Please see also comment #18 where it is described a better solution than the one already done to be implemented in the future if possible.
Adding QA Wanted to check 1.4.
Taking in 2.0 for regressions and cert issue
Comment on attachment 8447876 [details] Patch The patch implements the proposal of Gabriele in bug 934354 #comment 18
Comment on attachment 8447876 [details] Patch First of all thank you for your work here and sorry for the delay. The patch is mostly looking good save for a few issues I've highlighted on GitHub; I'm going to sum them up here: - wpm_clearNotifications() should do what it says: clear the notifications. Remove the code that also conditionally deletes the message and explicitly delete them instead in the right places. - mdb_clearByTag() should be called mdb_deleteByTimestamp() for consistency with the rest of the code and because the database has no concept of a tag (that's an abstraction used by the notifications) but knows about message timestamps instead. - Similarly pm_clear() should be called pm_delete(). - Finally the patch is missing unit-tests for the new functionality and I'm not 100% sure that it still passes the existing ones. Ensure that existing unit-tests are still working and modify them if needed then add new unit-tests covering the additional functionality. When you've adjusted the patch update your PR, and flag me for review again.
Comment on attachment 8447876 [details] Patch Changes from comment 5
Comment on attachment 8447876 [details] Patch The patch structure is good enough and I've left some comments only regarding the function names / organizations and a couple of typos. This is r+ with those comments addressed. Feel free to land this on a green Travis once you've updated the PR.
Also remember to add r=gsvelto at the end of the patch comment when updating the PR.
Changes done and travis green
(In reply to Albert [:albert] from comment #10) > Master > https://github.com/mozilla-b2g/gaia/commit/ > 0f922b40725998f100790512140f384be2a7d182 Dear Albert Could your patch be landed on v1.4? I think v1.4 also needs your patch. Thanks a great.
Dear Wei Gao, please ask for 1.4+ flag. Albert will take a look to 1.4 branch in the meantime.
(In reply to Beatriz Rodríguez [:brg] from comment #13) > Dear Wei Gao, please ask for 1.4+ flag. Albert will take a look to 1.4 > branch in the meantime. Ok, I see. I have added this flag. Thanks.
Not sure why this one was moved to 1.4. Please see comment 2
Created attachment 8456654 [details] Patch 1.4 I had to modify wappush_test.js in order to adapt the patch for 2.0 The rest of the patch for 2.0 remains the same for 1.4
(In reply to Preeti Raghunath(:Preeti) from comment #15) > Not sure why this one was moved to 1.4. Please see comment 2 It was a partner request, see bug 934354 comment 27. Since the fix for that bug was a v1.3t-only affair I suggested asking for approval for this one that applies cleanly to 1.4.
Comment on attachment 8456654 [details] Patch 1.4 Looks good to me, thanks.
Tested and working in: Flame 2.1 Gecko-329b2e4 Gaia-da11dba Flame 2.0 Gecko-2b27bec Gaia-de28796 Hamachi Gecko-d7d9838 Gaia-1144cdc Scenarios: STR 1. Receive a OMA CP message 2. Open it but do not process it completely 3. Press home key and *also* tested with call incoming EXPECTED The user will be able to come to process that message later