Partner reported that memory leak happened during testing. After checking, the issue was because of the usage navigator.mozApps.mgmt.getAll(). Gaiatest already release a patch to fix it. Since there are still many places calling it. Could anyone help to check if there is a chance to have the memory leak. Code contains the call: About.js (gaia\apps\settings\js): navigator.mozApps.mgmt.getAll().onsuccess = function gotApps(evt) { aboutApps.js (gecko\mobile\android\chrome\content): let request = navigator.mozApps.mgmt.getAll(); Appinfo.js (gecko\b2g\chrome\content\engmode\projectmenu\js): navigator.mozApps.mgmt.getAll().onsuccess = function mozAppGotAll(evt) Applications.js (gaia\apps\system\js): navigator.mozApps.mgmt.getAll().onsuccess = function mozAppGotAll(evt) { Apps.js (gaia\apps\settings\js): navigator.mozApps.mgmt.getAll().onsuccess = function mozAppGotAll(evt) { Icc_worker.js (gaia\apps\system\js): navigator.mozApps.mgmt.getAll().onsuccess = function gotApps(evt) { Permissiontest.js (gaia\test_apps\testpermission\js): navigator.mozApps.mgmt.getAll().onsuccess = function mozAppGotAll(evt) { Runapp.js (gecko\b2g\chrome\content): this._req = navigator.mozApps.mgmt.getAll(); Test_clear_browser_data.html (gecko\dom\tests\mochitest\localstorage): navigator.mozApps.mgmt.getAll().onsuccess = function() { UpdatePrompt.js (gecko\b2g\components): let all = window.navigator.mozApps.mgmt.getAll();

Hi Thinker, Could you provide your comments here?

(In reply to Brian Huang [:brianhuang] from comment #1) > Hi Thinker, > > Could you provide your comments here? You probably want Fabrice's input here. He is the primary developer on this API.

Hi Tim, Could you provide your comments here?

Please refer to here to know the memory leak (bug 924565).

I don't understand why we have a new bug here. There are known issues with object lifetime (which are not necessarily leaks), especially when used in windows that never go away. I would: 1) apply the gaiatest patch. 2) wait for bug 924565 to land.

Duping this over to bug 924565

Hi Fabrice, (In reply to Fabrice Desré [:fabrice] from comment #5) > I don't understand why we have a new bug here. There are known issues with > object lifetime (which are not necessarily leaks), especially when used in > windows that never go away. I would: > 1) apply the gaiatest patch. > 2) wait for bug 924565 to land. I am not open another case, since there are many places called this function. I just want to confirm if there will have chance to cause memory leak if code did not handle well. Like Gaiatest, it was modified.