Since the conversation so far and further analysis leads me to believe that: a) XUL Mini Cache is useful, even if it only reduces the amount of strings we localize at warm startup b) JSM is loaded eagerly right now, and making it load lazily is actually more work than just skip the non-proto-elements I suggest that we: 1) Land these patches 2) Re-introduce lazy-JSM in a new bug in which I'll move all state management to C++ 3) Work on Mini Cache for UA widgets in bug 1630743 4) Work on not-caching if localization is incomplete in bug 1629291
Bug 1627809 Comment 20 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
Since the conversation so far and further analysis leads me to believe that: a) XUL Mini Cache is useful, even if it only reduces the amount of strings we localize at warm startup b) JSM is loaded eagerly right now, and making it load lazily is actually more work than just skip the non-proto-elements c) This patchset doesn't impact performance (see comment 15) I suggest that we: 1) Land these patches 2) Re-introduce lazy-JSM in a new bug in which I'll move all state management to C++ 3) Work on Mini Cache for UA widgets in bug 1630743 4) Work on not-caching if localization is incomplete in bug 1629291