* Add support for capturing cpu profiles into tsc itself
* Accept baseline for new compiler option in showConfig
* Fix lints
* Support profiling build mode, only ever have one live profiling session
* Minor modification to enable/disable semaphore, accept re-cased baseline
* Add pid into autognerated cpuprofile path
* Rename to fix case
* Sanitize filepaths in emitted cpuprofile for easier adoption by enterprise people, add inspector to browser field
* Update LKG
* Add @types/node version bump
* Small paatches/casts to be compatible with latest node
* Accept API baseline update
* Make internal NodeBuffer compatable with latest Buffer
* Why do we even have an internal buffer type
* Sync up internal buffer type better
* Fix lint
* Readd mroe missing Buffer methods
1. Add a script to generate a sorted list of most costly tests. A tests'
cost is roughly `runtime% / number of edits`. A slow test that's only
been updated once is much less valuable than a slow test that has
been updated 20 times: the latter test is catching more changes in the
type system.
2. Check in the results of running this script. I want to make the
skipping behaviour deterministic and the same for everybody, even though
you may get slightly better performance by examining only *your* test
changes.
3. Add code to skip tests until it reaches a 5% chance of missing an
edit. Right now this provides a 38% speedup.
Still not done:
4. Make this value configurable.
5. Make the CI configuration specify a 0% chance of missing an edit.
* Remove jake (hopefully for real this time)
* Fix gulpfile non-lkg build, add sanity-check build to posttest on CI, accept older baseline style to go with lkgd build
* More docs/scripts jake -> gulp
* Some callbacks in watchUtilities werent being strictly checked due to the structural fallback
* Add direct dependeny on ms since mocha removed its impl
* Manually init stats collection on base runner like mocha.run now does