* docs: improve Windows build variable documentation
Documents an additional and possibly unclear variable, `DRIVE:`, in the example global variables for building with Windows
* docs: update formatting in build-instructions-windows.md to align with styleguide
In #20829, we fixed compositor recycling when switching between
BrowserViews, but it turns out that there is one additional case that we
need to handle. When we create a completely new BrowserView instance, it
starts of as visible (even when it hasn't been added to the window),
which means that it will need its own compositor instead of using the
recycled compositor.
To fix this, lets make BrowserViews hidden by default until they're
added to the window. See also #19988. This is a potentially breaking
change given that the initial value of `document.visibilityState` will
now be `hidden`, but given the experimental status of BrowserViews, I
think this is a fine change to make. The old behavior can be restored
with `webPreferences: { show: true }`.
Notes: Fix compositor recycling when creating new BrowserView
* build: cache the out directory for exact deps hash matches
* chore: generate a target based depshash discriminator
* fix: hash on gn args
* build: share logic on the mac builds
* build: ensure that the mksnapshot binary is built before stripping before zipping
* build: attach the workspace on macOS
* build: optimize the macOS checkout path for testing
* build: fix mksnapshot zip generation
* build: make the mac src cache restore work
* build: v2 out cache
* build: macOS cache restore is just stupidly slow
* build: strip more binaries
* build: attach the out cache to the workspace for macOS builds
* build: allow linux boxes to restore darwin out caches
* build: cat the deps hash target file
* build: ensure that the deps target hash matches on the linux box
* build: do not use host arch in target key
* build: force undefined in the target hash file
* build: only restore out cache when it isn't in the workspace
* build: fix the macOS cache workspace trick
* build: do not double restore
* build: remove the big stuff from the out dir
* build: workaround layer issue
* build: try it back on macOS again but with smaller thingy
* build: macOS needs the out cache now
* build: clean up for omptimal macOS path
* build: use old docker image
* build: idek at this point
* build: we need a deps hash
* build: yeah we need a checkout too
* chore: use testing env on save cache job
* chore: well that should fix the cache key thing
* chore: handle cross-OS path mismatch for src cache restore
* build: use a /portal directory to transfer the src cache appropriately
* build: use the correct docker image
* build: super perms for /portal
* build: increment out cache number
* build: ensure target hash is correct for args + disable pre-compiled headers on macOS
* build: wipe the cross-arch libffmpeg before building Electron
* fix(urlrequest): allow non-2xx repsponse results
- closes#21046
* test(net): add test cases to verify non-2xx body
* test(session): update spec to match clientrequest behavior
* test(net): update test cases to match clientrequest behavior
* spec: clean up async net spec
* test: update DOM storage quota limits test
* fix: update dom_storage_limits.patch (fixes#13465)
The previous version of this patch did not include
changes required to circumvent the quota enforcement
performed by StorageAreaImpl. Consequently when
the quota was exceeded, things still "appeared to
work" at first but then would later fail silently.
That is, the cache would be updated but the backing
store would not.
This could be fixed by disabling the code below
(from `content/browser/dom_storage/storage_area_impl.cc`)
```
// Only check quota if the size is increasing, this allows
// shrinking changes to pre-existing maps that are over budget.
if (new_item_size > old_item_size && new_storage_used > max_size_) {
if (map_state_ == MapState::LOADED_KEYS_ONLY) {
receivers_.ReportBadMessage(
"The quota in browser cannot exceed when there is only one "
"renderer.");
} else {
std::move(callback).Run(false);
}
return;
}
```
However, since this seems to have some unintended side-effects
(see updated notes in dom_storage_limits.patch) it seems
more prudent to simply increase the quota to a larger
yet still reasonable size rather than attempt to circumvent
the storage quota altogether.
