Thanks for working on this! I tried to compare OOPIF to non-OOPIF cases on a page like http://csreis.github.io/tests/cross-site-iframe.html, but I'm not sure I see what's missing in the OOPIF case. Can you describe repro steps and what part needs to be fixed?
I'm mainly curious if we should treat this as a launch blocking issue. We're currently targeting M67 for Site Isolation. Thanks!
OOPIFs are not shown on Performance panel at all :) We explicitly filter to a single process which contains main frame. To repro it, just record performance run while reloading a page with OOPIFs.
The plan is to show multiple tracks (one per process) and somehow indicate where each frame goes.
Since 67 is in two weeks, I am positive we'll have it by 67.
alph@: Just checking in since r563032 seems to have stuck (as of 69.0.3447.0). Is this bug fixed now?
I tried to verify the fix, but I didn't notice a difference in the performance tab (with respect to CPU profile info) between 67.0.3396.87 and 69.0.3469.2. Maybe you can help me understand where to look? I was visiting http://csreis.github.io/tests/cross-site-iframe.html, starting a recording in the DevTools performance panel, clicking on "Go cross-site (complex page)," and stopping the recording.
Just trying to determine if it's worth merging r563032 or if it's ok to let that part of the fix go out in M69. Thanks!
Comment 1 by bugdroid1@chromium.org
, Mar 15 2018