This reverts commit 2054f37a.

* benchmark updating many child layers

* Revert "[flutter_tools] Put a heap size limit on the frontend_server (#58039)"

* Adds support for material state mouse cursor to InkWell and InkResponse
* Set their default cursors to be MaterialStateMouseCursor.clickable.

* Do not assume imageCache is created when handleMemoryPressure is called

* comment for test

This reverts commit 8adaea90.

This reverts commit 39d1e4b7.

Remove copying from the shutdown stage since that seems risky. If the tool copies the first compilation there will still be a decent dill for initialization.

If a dependency specified uses-material-design: true and the main pubspec specifies uses-material-design: false, then the MaterialIcons font would be included in the font manifest, but not in the AssetManifest or final bundle. Remove it from the FontManifest if this occurs

removes globals from devFS operations and cleans up testing to be context free. Removes unused and unnecessarily temp directory in testing.

Work-around large timeline data killing devicelab machines by requesting data in 1 second intervals and combining at the end. Non-breaking change to the driver API.

This reverts the flutter command to use shlock when flock isn't available. It seems that the mkdir method isn't as reliable as we want. I think that this is because the trap isn't always be executed, which is why I think that shlock uses PIDs to help it be more reliable. Unfortunately, that means that we're back to not working over network shares (which is where things were before I moved to the mkdir method, so not really a regression). I did leave in the mkdir method for platforms that have neither flock nor shlock (which should be very few and far between, but still), so at least we'll do some locking there now.

This reverts commit cd593dae.