This comprises a massive rewrite of the UI for browsing captures. We use
the SysprofAid class to scan capture files for content and then auto-
matically add visualizers and details pages.
To avoid breaking things incrementally, we just land this as a very large
commit. Not necessarily ideal, but given the amount of stuff that could
break, this is easier.
As part of this process, we're removing a lot of the surface API so that
we can limit how much we need to maintain in terms of ABI.
The goal here is to be able to do a duplicate recording to the previous
with a quick key-combination like Ctrl+R.
We still need to extract the metadata from the capture file and setup
a new profiler, but this gets the mechanics in place.
