Previously, AudioServer would deadlock when trying to play another audio
stream, i.e. creating a queue. The m_pending_cond condition was used
improperly, and the condition wait now happens independently of querying
the pending queue for new clients if the mixer is running.
To make the mixer's concurrency-safety code more readable, the use of
raw POSIX mutex and condition syscalls is replaced with Threading::Mutex
and Threading::ConditionVariable.
ConditionVariable is a thin wrapper over the pthread_cond_* APIs, just
as Mutex is a wrapper over pthread_mutex.
Because ConditionVariable might want to wait on a high-level Mutex, it
needs to be friends with it.
Across the entire audio system, audio now works in 0-1 terms instead of
0-100 as before. Therefore, volume is now a double instead of an int.
The master volume of the AudioServer changes smoothly through a
FadingProperty, preventing clicks. Finally, volume computations are done
with logarithmic scaling, which is more natural for the human ear.
Note that this could be 4-5 different commits, but as they change each
other's code all the time, it makes no sense to split them up.
If you saved a file under a different name and then went back to
the first file, then you had the same TextDocument buffer, and therefore
the same changes to the file as in the new one.
For example, this CSS previously produced a lot of log spam about the
`display` properties having invalid values:
```css
.foo {
display: -webkit-box;
display: -moz-box;
display: -ms-flexbox;
display: -webkit-flex;
display: flex;
}
```
Now, it just ignores them, because we don't need to know about it. :^)
Our debug logging when we fail to parse a legitimate selector, is less
useful when it's hidden among selectors that don't parse because they
contain vendor-prefixed pseudo-elements and classes. So, now we
specifically ignore these and don't produce a log message.
When a file deletion event happens, we now iterate over all views of the
FileSystemModel and remove any selection & cursor indices that hold
dangling references do the deleted filesystem node.
This fixes#9602.
getComputedStyle(element) now returns a ComputedCSSStyleDeclaration
object, which is a live view of the computed style of a given element.
This works by ComputedCSSStyleDeclaration being a wrapper around an
element pointer. When you ask it for a CSS property, it gets the latest
computed style values from the element and returns them as a
CSS::StyleProperty object.
This first cut adds support for computed 'color' and 'display'.
In case the element doesn't have a corresponding node in the layout
tree, we fall back to using specified style instead. This is achieved by
performing an on-the-fly style resolution for the individual element and
then grabbing the requested property from that resolved style.
This patch moves the CSS property+value storage down to a new subclass
of CSSStyleDeclaration called PropertyOwningCSSStyleDeclaration.
The JavaScript wrapper for CSSStyleDeclaration now calls virtual
functions on the C++ object.
This is preparation for supporting computed style CSSStyleDeclaration
objects which won't have internal property storage, but rather an
internal element pointer. :^)
The CI system often develops a significant backlog when we have a lot
of PRs in the queue, and folks are pushing to master directly, or
other PRs are getting merged. The individual pushes to master or PR
merges to master each end up creating a dedicated master CI build.
These builds complete for machines with the normal PR validation builds.
To aid with this, Azure DevOps has a feature where they allow the CI
builds to "batch" multiple changes together, instead of running
multiple builds for each change.
Azure DevOps defines batching as:
When a pipeline is running, the system waits until the run is
completed, then starts another run with all changes that have
not yet been built.
Documentation Reference:
https://docs.microsoft.com/en-us/azure/devops/pipelines/repos/github?view=azure-devops&tabs=yaml#batching-ci-runs
Prior this change, the button was updated on user selection change
in the file view.
This isn't quite right, as you could remove the text from the text box
or (even worse) start typing a filename there and the button state
wouldn't change.
The better fix is to have a linker script. We'll need this to set
the entry point to 0x80000 for bare-metal builds anyways. But I'd
like to get some UART output in qemu before I add this (otherwise
I can't check if the bare-metal version does anything), so put
in this temporary kludge for now.
Needed for functions that have local variables.
In time we need to share this between aarch64 and intel, but while
we figure out what exactly the aarch64 Prekernel should do, let's
duplicate this.
Else, function-local statics create calls to
__cxa_guard_acquire / __cxa_guard_release on aarch64, which we don't
(yet?) implement. Since Prekernel is single-threaded, just sidestep
that for now.
This is just a config file with the default options that PixelPaint
recognizes and reads so far. Adding this in since the options are
not really documented anywhere so at least the user can now know what
options are available.
In a previous commit we read default values from a commit file, this
commit now also writes back any changes to those settings made by
the user. Persistent settings always feel good :^)
