We should keep track of counts of frames constructed (excluding breaking!) and frames reflowed so that we can: * write tests that assert that these things have or have not happened * write developer tools that do the same I plan to use this for testing bug 1110277.

(I've tested the frame construction counts in my test for bug 1110277; I haven't actually tested the reflow counts.)

This counts are intended to be useful for: * tests * perhaps developer tools that need to determine whether frame construction or reflow has happened in a particular interval.

Comment on attachment 8541737 [details] [diff] [review] patch 1 - Store counts of frames constructed and reflowed on the pres context Review of attachment 8541737 [details] [diff] [review]: ----------------------------------------------------------------- I think you should use 64 bits. A page with 1000 frames reflowing 60 times per second for a day will overflow 32 bits.

(I've tested the frame construction counts in my test for bug 1110277; I haven't actually tested the reflow counts.)

This counts are intended to be useful for: * tests * perhaps developer tools that need to determine whether frame construction or reflow has happened in a particular interval.

remote: https://hg.mozilla.org/integration/mozilla-inbound/rev/d1510c5deff9 remote: https://hg.mozilla.org/integration/mozilla-inbound/rev/9fe58932c192

https://hg.mozilla.org/mozilla-central/rev/d1510c5deff9 https://hg.mozilla.org/mozilla-central/rev/9fe58932c192