Implements an `Appearances` xaml object and an `AppearanceViewModel` in the SettingsEditor project. Updates `Profiles` to use these new objects for its default appearance.
This is the first step towards getting `UnfocusedAppearance` into the SUI.
Adds try-catch blocks to the parts where we layer a fragment onto a profile and create a new profile from a fragment. This allows us to continue looping over the remaining fragments after failing to use a badly-formed one, instead of aborting prematurely.
## PR Checklist
* [x] Closes#10590
## Validation Steps Performed
Non-badly formed fragments get loaded even if there is a badly formed one somewhere
Sets the working directory of the terminal when invoked from the shell extension. This ensures that new tabs opened with a starting directory of `.` open in the directory that the terminal was invoked from.
Closes#8933
## Validation Steps Performed
Manually tested - default PowerShell profile set to use home directory, Windows PowerShell profile set to use current directory. Launched via the shell extension and the default profile opened in the explorer directory, as did a new Windows PowerShell tab.
Change accessibility notifier creation so we do not create one when we're in PTY mode. (Guard all call sites to skip math/event work when the notifier is null.) MSAA events are legacy events that are registered for globally and used by some screen readers to find content in the conhost window. The PTY mode is not responsible for hosting the display content or input window, so it makes sense for it to not broadcast these events and delegate the accessibility requirement to the connected terminal.
## References
- #10537
## PR Checklist
* [x] Closes#10568
* [x] I work here
* [x] Manual test launches passed.
Currently, when the user changes the console codepage (manually or via a script) the GDI engine tries to find and set the "best possible" font. The "best possible" here is charset-wise, it doesn't mean that the font is actually better or more eye-candy for the user.
Example:
- Open cmd
- Set the font to Consolas
- Enter `chcp 932`
- Suddenly, a wild MS Gothic appears!
This kind of makes sense (*"if I'm changing the codepage I probably want to see the national characters"*) but it doesn't happen anywhere else - all other apps just substitute the missing glyphs.
After #10472 / #10478 this magic should finally work here as well. So, do we still need to change the whole font? Terminal doesn't do that after all.
## Validation Steps Performed
Download [932.cmd.txt](https://github.com/microsoft/terminal/files/6697577/932.cmd.txt),
rename to 932.cmd, run it, check if the output is still readable.
Closes#10497
## Summary of the Pull Request
Adds a feature flag `Feature_EditableActionsPage` that controls whether the Actions page in the Settings UI is read-only vs editable. The editable version is disabled for `Release` builds and enabled everywhere else (i.e. Dev, Preview, etc...).
Validated using `<stage>` `AlwaysEnabled` and `AlwaysDisabled`.
## References
#6900 - Actions Page Epic
## PR Checklist
Closes#10578
<!-- Enter a brief description/summary of your PR here. What does it fix/what does it change/how was it tested (even manually, if necessary)? -->
## Implementation of #10477 - handle surrogate pairs in GDI renderer.
<!-- Please review the items on the PR checklist before submitting-->
## PR Checklist
* [ ] Closes#10477
* [x] CLA signed.
* [ ] Tests added/passed
* [ ] Documentation updated.
* [ ] Schema updated.
* [x] I've discussed this with core contributors already. If not checked, I'm ready to accept this work might be rejected in favor of a different grand plan. Issue number where discussion took place: #10477
<!-- Provide a more detailed description of the PR, other things fixed or any additional comments/features here -->
## Detailed Description of the Pull Request / Additional comments
### Why not let Windows draw surrogate pairs? It can do that.
Basically, the comment says everything:
c90de69250/src/renderer/gdi/paint.cpp (L346-L347)
However, handling things above U+FFFF doesn't really require extra effort. It's enough to:
- Put *all* characters to the output buffer
- Set the first width to cluster width and the rest to 0
- Sit back and relax while Windows does the rest
<!-- Describe how you validated the behavior. Add automated tests wherever possible, but list manual validation steps taken as well -->
## Validation Steps Performed
```CMD
@echo off
chcp 65001
echo 𠜎𠝹𠱓𠱸𠲖𠳏𠳕𠴕𠵼𠵿𠸎
echo 👨👩👧👦
```
Save this as a UTF-8 cmd file and run.
### Before the change
### After the change
An example of a third party app working with surrogate pairs in a patched OpenConsole:
As discussed, this change doesn't claim to be the full support for surrogate pairs (there are still corner cases possible), but brings it on par with Terminal with minimal effort.
This PR is a small start in a broader "Minimize to Tray" feature (#5727).
This particular change is scoped only to the scenario when a quake window
is minimized. Currently the only way to bring back the quake window
when it's minimized is to press the global hotkey again. This gives another
option - to press the terminal icon in the tray.
Eventually though, minimize to tray will be available for any window, and
I'd like more time to flesh out the general porpoise scenarios and context
menus. Having just a bit in this PR also helps reviewers by keeping it small!
Delay load call SetThreadDescription to restore WPF renderer on Win7
## PR Checklist
* [x] Closes something @DHowett asked me to do.
* [x] I work here
* [x] I F5'd it on a version with this function and it still works
## Detailed Description of the Pull Request / Additional comments
I keep forgetting that anything in the WPF control needs to keep working on Win7. Or more specifically... I remember this fact for the DX renderer, but not for the render thread base. Oops. Turns out this particular convenience method to set thread descriptions for visibility inside the debugger (to make my life easier) only works down to 1607 (see https://docs.microsoft.com/windows/win32/api/processthreadsapi/nf-processthreadsapi-setthreaddescription). Since it's just a debugging convenience... skipping it entirely when the procedure is not found should be fine. Also I don't try to load `kernel32.dll` and just get the handle of the existing module (which per the remarks at https://docs.microsoft.com/windows/win32/api/libloaderapi/nf-libloaderapi-getmodulehandlew will not increment the module reference count) because `kernel32.dll` pretty much has to be there or we're already in hot water.
## Summary of the Pull Request
This adds the "add new" button to the actions page. It build on the work of #10220 by basically just adding a new list item to the top of the key binding list.
This also makes it so that if you click the "accept changes" button when you have an invalid key chord, we don't do anything.
## References
#6900 - Actions page Epic
#9427 - Actions page design doc
#10220 - Actions page PR - set action
## Detailed Description of the Pull Request / Additional comments
- `ModifyKeyBindingEventArgs` is used to introduce new key bindings. We just ignore `OldKeys` and `OldActionName` because both didn't exist before.
- `IsNewlyAdded` tracks if this is an action that was added, but has not been confirmed to add to the settings model.
- `CancelChanges()` is directly bound to the cancel button. This allows us to delete the key binding when it's clicked on a "newly added" action.
## Validation Steps Performed
- Cancel:
- Deletes the action (because it doesn't truly exist until you confirm changes)
- Accept:
- Adds the new action.
- If you attempt to edit it, the delete button is back.
- Add Action:
- Delete button should not be visible (redundant with 'Cancel')
- Action should be initialized to a value
- Key chord should be empty
- Cannot add another action if a newly added action exists
- Keyboard interaction:
- escape --> cancel
- enter --> accept
- Accessibility:
- "add new" button has a name
- Interaction with other key bindings:
- editing another action --> delete the "newly added" action (it hasn't been added yet)
- only one action can be edited at a time
This commit adds support for bold text in DxRenderer.
For now, bold fonts are always rendered using DWRITE_FONT_WEIGHT_BOLD
regardless of the base weight.
As yet, this behavior is unconfigurable.
References
Previous refactoring PRs: #9096 (DxFontRenderData) #9201 (DxFontInfo)
SGR support tracking issue: #6879Closes#109
The code in this file was adapted from the STL on the 2021-07-05.
It backports the following Windows 8 functions to Windows 7:
* WaitOnAddress
* WakeByAddressSingle
* WakeByAddressAll
These functions are used within `til`. This commit will allow `til` to be used in the conhost source code.
Validation
* [x] correct .dll loads on Windows 7
* [x] correct .dll loads on Windows 10
* [x] link line for PublicTerminalCore prefers this fake apiset over kernel32
Previously `TermControl::Close` destroyed all `ThrottledFunc`s to ensure they're not scheduling any callbacks on the UI thread, as the call to `Close` signals the point at which the `TermControl` isn't part of the UI thread anymore. `_CursorPositionChanged` tried to prevent access to the potentially deallocated `_tsfTryRedrawCanvas` by checking the `std::shared_ptr` for nullability, but since the deallocation happens on the UI thread and the nullability check on a background thread, this check introduced a race condition.
This commit solves the issue by not deallocating any `ThrottledFunc`s anymore and instead checking the `_closing` flag inside the `ThrottledFunc` callback on the UI thread.
Additionally this commit cleans up some antipatterns around the use of `std::optional`.
## PR Checklist
* [x] Closes#10479
* [x] Closes#10302
## Validation Steps Performed
* Opening and closing tabs doesn't crash ✔️
* Printing long text doesn't crash ✔️
* Manual scrolling doesn't crash ✔️
* ^G / the audible bell doesn't crash ✔️
## Summary of the Pull Request
#7960 was caused by `UiaTextRangeBase::_blockRange` not being initialized, thus pointing to random memory. In most cases, we initialize it properly in `RuntimeClassInitialize`, however, the copying version of `RuntimeClassInitialize` doesn't actually copy it over, resulting in it still containing random memory.
NVDA (and other screen readers) occasionally use `Clone` (really just the copy initializer), resulting in this bug occurring randomly.
## PR Checklist
* [X] Closes#7960
* [X] Tests added/passed
## Validation Steps Performed
Test failed before the change, but passes after the change.
This introduces the ability to set the action for a key binding. A combo box is used to let the user select a new action.
## References
#6900 - Actions page Epic
#9427 - Actions page design doc
#9949 - Actions page PR
## Detailed Description of the Pull Request / Additional comments
### Settings Model Changes
- `ActionAndArgs`
- new ctor that just takes a `ShortcutAction`
- `ActionMap`
- `AvailableActions` provides a map of all the "acceptable" actions to choose from. This is a merged list of (1) all `{ "command": X }` style actions and (2) any actions with args that are already defined in the ActionMap (or any parents).
- `RegisterKeyBinding` introduces a new unnamed key binding to the action map.
### Editor Changes
- XAML
- Pretty straightforward, when in edit mode, we replace the text block with a combo box. This combo box just presents the actions you can choose from.
- `RebindKeysEventArgs` --> `ModifyKeyBindingEventArgs`
- `AvailableActionAndArgs`
- stores the list of actions to choose from in the combo box
- _Unfortunately_, `KeyBindingViewModel` needs this so that we can populate the combo box
- `Actions` stores and maintains this though. We populate this from the settings model on navigation.
- `ProposedAction` vs `CurrentAction`
- similar to `ProposedKeys` and `Keys`, we need a way to distinguish the value from the settings model and the value of the control (i.e. combo box).
- `CurrentAction` --> settings model
- `ProposedAction` --> combo box selected item
## Validation Steps Performed
- Cancel:
- ✔️ change action --> cancel button --> begin editing action again --> original action is selected
- Accept:
- ✔️ don't change anything
- ✔️ change action --> OK! --> Save!
- NOTE: The original action is still left as a stub `{ "command": "closePane" }`. This is intentional because we want to prevent all modifications to the command palette.
- ✔️ change action & change key chord --> OK! --> Save!
- ✔️ change action & change key chord (conflicting key chord) --> OK! --> click ok on flyout --> Save!
- NOTE: original action is left as a stub; original key chord explicitly unbound; new command/keys combo added.
Introduces `FontConfig`, an object that isolates font-related settings
in our profiles
Users can now define font settings in their json as so:
```
"font":{
"face": "Consolas",
"size": 12
}
```
Backwards compatible with the currently expected way of defining font
settings in the json, note however that upon hitting 'Save' in the SUI,
these settings **will be rewritten to the font-object style in the json
(as above)**.
## Validation Steps Performed
Existing functionality works, new functionality works
References #1790Closes#6049
This commit adds a "(requires relaunch)" suffix to the header of the language picker control.
## PR Checklist
* [x] I work here
* [x] Tests added/passed
Passing structures larger than the register size is very expensive
due to Microsoft's x64 calling convention. We could reduce the
overhead by passing the string-view by reference, but this forces us
to allocate the parameters as static string-views on the data
segment of our binary. I've found that passing them as classic
C-strings is more ergonomic instead and fits the need for
high performance in this particular code.
This improves performance for VT-heavy output by 15-20%.
## PR Checklist
* [x] I work here
* [x] Tests added/passed
This commit introduces a basic ApplicationState class, without being used for anything yet to aid reviewers. At a later point actual usages of this new class may be added separately.
## References
This commit is an initial step towards implementing #8324.
## PR Checklist
* [x] I work here
* [x] Tests added/passed
## Validation Steps Performed
* Creating a `state.json` with `{"generatedProfiles":["{53e75ed9-2b63-4118-856d-0510c4f6b97e}"]}` updates the ApplicationState, as observed through a debugger ✔️
* Deleting the "generatedProfiles" field sets the corresponding field back to nullopt ✔️
Try to throttle the cursor redrawing in the conhost world.
The motivation of this is the high CPU usage of `TriggerRedrawCursor` (#10393).
This can be seen as the conhost version of #2960.
This saves 5%~8% of the CPU time.
Supports #10462.
## Summary of the Pull Request
Updates the `closeTab` action to optionally take an index.
## PR Checklist
* [x] Closes#7180
* [x] CLA signed. If not, go over [here](https://cla.opensource.microsoft.com/microsoft/Terminal) and sign the CLA
* [ ] Tests added/passed
* [x] Documentation updated. If checked, please file a pull request on [our docs repo](https://github.com/MicrosoftDocs/terminal) and link it here: MicrosoftDocs/terminal#347
* [x] Schema updated.
* [ ] I've discussed this with core contributors already. If not checked, I'm ready to accept this work might be rejected in favor of a different grand plan. Issue number where discussion took place: #xxx
## Validation Steps Performed
Added the following configuration to `settings.json` and validated both key combinations behaved as expected. Also opened the command palette and ensured that the actions were displayed.
```json
{ "command": "closeTab", "keys": "ctrl+shift+delete" },
{ "command": { "action": "closeTab", "index": 0 }, "keys": "ctrl+shift+end" }
```
Noticed the json schema was listing the option as invalid even though it's accepted by WT. So added it to schema to remove the error.
## PR Checklist
* [x] CLA signed. If not, go over [here](https://cla.opensource.microsoft.com/microsoft/Terminal) and sign the CLA
* [x] Schema updated.
## Validation Steps Performed
No longer shows as invalid in VSCode.
This commit introduce three new `til` features:
* "til/latch.h": A std::latch clone, until we're on C++20.
* "til/mutex.h": A safe mutex wrapper, which only allows you access to the protected data after locking it. No more forgetting to lock mutexes!
* "til/throttled_func.h": Function invocation throttling used to be available as the `ThrottledFunc` class already. But this class is vastly more efficient and doesn't rely on any WinRT types.
This PR also adds a `til::ends_with` string helper which is `til::starts_with` counterpart.
## Validation Steps Performed
* Scrollbar throttling still works as it used to ✔️
* No performance regressions when printing big.txt ✔️Closes#10393
This PR Introduces `DxFontInfo` to simplify the logic in
`DxFontRenderData`.
`DxFontInfo` aims to be the DWrite equivalent of `FontInfo` &
`FontInfoBase` in GDI. It encapsulates the needed information to
represent a displayable font face. It also provides the ability to
resolve a font face based on the available fonts on the system.
## References
This is a follow-up of #9096.
Initial Italic support was introduced by #8580.
The motivation behind this is to support bold & bold-italic text in
Windows Terminal.
Fix startup race of resizing ConPTY
- Depending on what the timing and ordering is of the message coming in
from the signal thread, it may be applied to the startup structure
after the I/O thread has begun initializing the console buffer
structures but before it has signaled that it is done and the signal
thread is ready to make changes directly. This likely happens because
the end of the I/O thread setup has a weird unlock/lock jog for the
input thread and the signal thread might have been scheduled in the
middle of it.
- My resolution here is to ensure that the signal thread just keeps
storing the latest resize message until it is told that everything is
initialized. Whomever comes in to tell the signal thread this
information (under lock) will pickup and run the resize if one came in
before everything was ready. This should resolve the race.
## Validation Steps Performed
- o-sdn-o confirms this resolves their issue
Closes#10400
Replace PolyTextOutW with ExtTextOutW to allow substitution of missing
glyphs from other fonts.
Why not let Windows substitute the glyphs that are missing in the
current font? Currently the GDI renderer of conhost/OpenConsole uses
`PolyTextOutW` for drawing. `PolyTextOutW` doesn't try to substitute
any missing glyphs, so the only way to see, say, Hiragana is to change
the _whole font_ to something like MS Gothic (which is eye-bleeding, to
be honest).
A trivial replace `PolyTextOutW` -> `ExtTextOutW` does the trick.
Switch to `PolyTextOutW` happened in Windows 7 along with introduction
of conhost.exe. Substitution worked in previous Windows versions, where
internal NT interfaces were used.
# Before the change:
# After the change:
Closes#10472
The CursorStyle enum is declared as being of type `uint` on the C# side,
but as `size_t` on the C++ side. There's a C# size_t impostor we could
use, System.UIntPtr, but I don't want to risk changing the public API of
TerminalTheme and I don't know if it can be used as a base type for an
enum.
Anyway, since we don't have more than four billion cursor types I chose
to narrow the field to a uint32_t and unpack it in TerminalSetTheme.
Fixes#10485
Retrieved from https://microsoft.visualstudio.com os.2020 OS official/rs_wdx_dxp_windev 40df712b33a40e76aeeac87f823bbb028d2e3972
Related work items: MSFT-32007459
Fixes the `Invoke-CodeFormat` and `Invoke-OpenConsoleTests` functions in `OpenConsole.psm1` so that they can be run directly from PowerShell.
Addresses the issues found when creating #10447.
`Invoke-CodeFormat` did not work when invoked directly from PowerShell due to a relative path being passed into the .NET function `[IO.File]::WriteAllLines()`. The working directory for .NET objects does not change when you change directory in PowerShell, so the paths were being treated as relative to the initial working directory of the shell - which was not the terminal git repo.
`Invoke-OpenConsoleTests` had 3 issues:
1. The path to `TestHostApp` was wrong.
2. It would attempt to run the "in host app" tests both in the host app and not in the host app.
3. The test configuration in `tests.xml` wasn't in sync with the `runABC.cmd` files, so the remoting and control unit tests didn't run.
## Validation Steps Performed
1. Ran `Invoke-CodeFormat` and `runformat.cmd` from multiple directories and didn't see errors.
2. Ran `Invoke-OpenConsoleTests` and didn't see errors.
This update brings some significant changes to the Cascadia family:
* Arabic and Hebrew support
* Italics (the new ones, not the cursive ones)
* Tweaked letterforms and fixed interpolation values for the upright
faces.
Since we now have four font files, this commit also relocates them to a
much more reasonable place (res/fonts/) and tidies up the build and
exclude rules to make them more extensible in the future.
This commit introduces localization for the "Open in Windows Terminal"
menu item and differentiates it based on compile-time branding (rather
than runtime detection!).
@leonMSFT's tray icon pull request had the excellent idea to use the
TerminalApp's resource compartment for auxiliary resources for projects
that can't otherwise be localized the same way. Doing localization in
the shell extension (or WindowsTerminal.exe) would require us to use
MUIRCT and split the build process up to support mui files. That's a
huge amount of work... but this is *not* a huge amount of work.
Fixes#6112
## Summary of the Pull Request
#10297 found a bug in `ActionMap` where the `ToggleCommandPalette` key chord could not be found using `GetKeyBindingForAction`.
This was caused by the following:
- `AddAction`: when adding the action, the `ActionAndArgs` is basically added as `{ToggleCommandPalette, ToggleCommandLineArgs{}}` (Note the default ctor used for the action args)
- `GetKeyBindingForAction`: we're searching for an `ActionAndArgs` structured as `{ToggleCommandPalette, nullptr}`
- Since these are _technically_ two different actions, we are unable to find it.
This issue was fixed by making the `Hash(ActionAndArgs)` function smarter! If the `ActionAndArgs` has no args, but the `ShortcutAction` _supports_ args, generate the args using the default ctor.
By making `Hash()` smarter, everybody benefits from this logic! We can basically now enforce that `ActionAndArgs{ <X>, nullptr } == ActionAndArgs{ <X>, <default_ctor> }`.
## Validation Steps Performed
- Added a test.
- Tested this on #10297's branch and this does fix the bug
## Summary of the Pull Request
`ApplicationLanguages::PrimaryLanguageOverride` requires packaged activation.
This PR prevents any such application crashes, by skipping any calls to `PrimaryLanguageOverride`, as well as hiding the language selector in the settings UI.
## PR Checklist
* [x] I work here
* [x] Tests added/passed
* [ ] Documentation updated. If checked, please file a pull request on [our docs repo](https://github.com/MicrosoftDocs/terminal) and link it here: #xxx
* [ ] Schema updated.
## Validation Steps Performed
When WT is run unpackaged:
* Doesn't crash during start ✔️
* SUI doesn't show the language selector ✔️
Persist inbox conhost; delegate control activities to it via a pipe
## PR Checklist
* [x] Closes#10194 - WSL Debug Tap doesn't work
* [x] Closes#10134 - WSL Parameter is Incorrect
* [x] Closes#10413 - Ctrl+C not passed to client
* [x] Closes#10414 - Leftover processes on abrupt termination
* [x] Might help #10251 - Win+X Powershell sometimes fails to attach
* [x] I work here
* [x] Manually tested with assorted launch scenarios
## Detailed Description of the Pull Request / Additional comments
It turns out that there's a bit of ownership that goes on with the original inbox `conhost.exe` and the operating system/driver. The PID of that original `conhost.exe` is stowed when the initial connection is established and it is verified for several activities. This means that the plan of letting it go completely away and having the `OpenConsole.exe` take over all of its activities must be slightly revised.
I have tested the following two alternatives to keeping `conhost.exe` around and they do not work:
1. Replacing the original owner `conhost.exe` with `OpenConsole.exe` - A.) The driver does not allow this. Once the owner is registered, it cannot be replaced. B.) There's no way of updating this information inside the client process space and it is kept there too in the `kernelbase`/`conclnt` data from its initial connection.
2. Attempting to pick up the first packet (to determine headed/headless and other initial connection information that we use to determine whether handoff is appropriate or not) prior to registering any owner at all. - The driver doesn't allow this either. The owner must be registered prior to a packet coming through.
Put this mental model in your head:
CMD --> Conhost (inbox) --> OpenConsole (WT Package) --> Terminal (WT Package)
So since the `conhost.exe` needs to stick around, here's what I'm doing in this PR:
- `conhost.exe` in the OS will receive back the `OpenConsole.exe` process handle on a successful handoff and is expected to remain alive until the `OpenConsole.exe` exits. It's now waiting on that before it terminates itself.
- `conhost.exe` in the OS will establish a signal channel pipe and listen for control commands from `OpenConsole.exe` in a very similar fashion to how the `ConPTY` signal pipe operates between the Terminal and the PTY (provided by `OpenConsole.exe` in this particular example.) When `OpenConsole.exe` needs to do something that would be verified by the OS and rejected... it will instead signal the original `conhost.exe` to do that thing and it will go through.
- `conhost.exe` will give its own handle through to `OpenConsole.exe` so it can monitor its lifetime and cleanup. If the owner is gone, the session should end.
- Assorted handle cleanup that was leading to improper exits. I was confused between `.reset()` and `.release()` for some of the `wil::unique_any<T>` handling and it lead to leaked handles. The leaked handles meant that threads weren't aware of the other sides collapsing and wouldn't cleanup/terminate appropriately.
How does this fix things?
- For the WSL cases... WSL was specifically looking up the owner PID of the console session from the driver. That was the `conhost.exe` PID. If it exits, that PID isn't valid and is recycled. Thus the parameter is incorrect or other inappropriate WSL setup behaviors.
- Ctrl+C not passed... this is a signal the operating system rejects from a PID that is not the owner. This is now relayed through the original owner and it works.
- Leftover processes... I believe I explained this was both not-enough-monitoring of each others' process lifetimes coupled with mishandling of release/resetting handles and leaking them.
- Powershell sometimes fails to attach... my theory on this one is that it's a race that became upset when the `conhost.exe` disappeared while something about Powershell/.NET was still starting, much like the WSL one. I believe now that it is sticking around, it will be fine.
Also, this WILL require an OS update to complete improvement of functionality and I have revised the interface ID. This is considered an acceptable breaking change with no mitigation because we said this feature was an alpha preview.
## Validation Steps Performed
- Launched WSL with defapp set, it works
- Launched WSL with defapp set and the debug tap on, it works and opens in two tabs
- Launched CMD, ran ping, did Ctrl+C, it now receives it
- Launched Win+X powershell a ton of times. It seems fine now
- Launched cmd, powershell, wsl, etc. Killed assorted processes in the chain (client/conhost/openconsole/windowsterminal) and observed in Process Explorer (with a long delta timer so I could see it) that they all successfully tear down now without leftovers.
An exception was introduced for the 'Toggle Command Palette' action to **not** being dispatched. Otherwise the command palette that was just closed will become visible again.
## PR Checklist
* [x] Closes#10240
* [x] CLA signed. If not, go over [here](https://cla.opensource.microsoft.com/microsoft/Terminal) and sign the CLA
## Detailed Description of the Pull Request / Additional comments
- Selecting the `Toggle command palette` item in the command palette will now properly close the command palette.
- Opening and closing the Command Palette through shortcut keys is still working fine.
- Other command palette items are still working fine as well.
## Summary of the Pull Request
This commit introduces a copy constructor/operator for
`_CONSOLE_API_MSG`. The change is not trivial as the struct contains a
union of unnamed structs that cannot be copied using regular language
features. As such a copy operator using `memcpy` was implemented.
Additionally all access specifiers were removed, as those allow a C++
compiler to reorder struct members. This would break message passing.
This commit is a good opportunity to prevent such miscompilations
proactively.
## Validation Steps Performed
* Command prompts of WSL2 fish-shell and pwsh still work ✔️Closes#10076
<!-- Enter a brief description/summary of your PR here. What does it fix/what does it change/how was it tested (even manually, if necessary)? -->
## Summary of the Pull Request
Immediately cancelling the preview when the user is navigating back from a nested command.
<!-- Other than the issue solved, is this relevant to any other issues/existing PRs? -->
## References
<!-- Please review the items on the PR checklist before submitting-->
## PR Checklist
* [x] Closes#10165
* [x] CLA signed. If not, go over [here](https://cla.opensource.microsoft.com/microsoft/Terminal) and sign the CLA
* [ ] Tests added/passed
* [ ] Documentation updated. If checked, please file a pull request on [our docs repo](https://github.com/MicrosoftDocs/terminal) and link it here: #xxx
* [ ] Schema updated.
* [ ] I've discussed this with core contributors already. If not checked, I'm ready to accept this work might be rejected in favor of a different grand plan. Issue number where discussion took place: #xxx
<!-- Provide a more detailed description of the PR, other things fixed or any additional comments/features here -->
## Detailed Description of the Pull Request / Additional comments
Basically 2 changes are done here:
- Allow the click handler to run for the back button when the button has focus and user hits the enter key (similarly as hitting space now).
- Now immediately cancelling the preview when the user is navigating back. Felt nicer to do it immediately at that point then keeping the preview active until the user hits cancel to close the palette. So the preview is already cancelled at step **5** instead of 6 as mentioned in the reproduction steps here https://github.com/microsoft/terminal/issues/10165#issue-899838383. But of course let me know if you're not agreeing here 😀 .
<!-- Describe how you validated the behavior. Add automated tests wherever possible, but list manual validation steps taken as well -->
## Validation Steps Performed
- Open 'Color Scheme' and verified preview is still working fine when selecting different schemes.
- After tabbing back to the Back button verified that when hitting enter or space the preview is cancelled and the original color scheme is being used again.
- Then after going back to 'Color Scheme' previews are still working ok.
- After hitting Enter on one of the Color Schemes the scheme still becomes active as before.
