What the title says. 😄
Rather than relying on the potentially overloaded `!=` or `==` operators
when checking for null, now we'll use the `is` expression (possibly
combined with the `not` operator) to ensure correct checking. Probably
overkill for many of these classes, but decided to err on the side of
consistency. Would matter more on classes that may be inherited or
extended.
Using `is` and `is not` will provide us a guarantee that no
user-overloaded equality operators (`==`/`!=`) is invoked when a
`expression is null` is evaluated.
In code form, changed all instances of:
```c#
something != null
something == null
```
to:
```c#
something is not null
something is null
```
The one exception was checking null on a `KeyChord`. `KeyChord` is a
struct which is never null so VS will raise an error when trying this
versus just providing a warning when using `keyChord != null`. In
reality, we shouldn't do this check because it can't ever be null. In
the case of a `KeyChord` it **would** be a `KeyChord` equivalent to:
```c#
KeyChord keyChord = new ()
{
Modifiers = 0,
Vkey = 0,
ScanCode = 0
};
```
related to some #40113 work
The extension SDK shouldn't rely on a preview version of the Windows
SDK. It should use the stable one.
Also moves some messages around that we didn't need
A regression from #40132
We need to set the value of `_includeApps` in the `MainListPage` ctor.
Without it, if the user doesn't have any extensions installed, the value
is never updated.
## Summary of the Pull Request
Self-refresh of `MainListPage` introduced in #40132 causes unnecessary
spawning of tasks by `ReapplySearchInBackground` and pushing the code
down the scenic route instead of taking shortcut.
This drop-in fix introduces a single-worker coalescing refresh loop to
eliminate thread-pool churn and syncs state in early-return paths.
## PR Checklist
- [x] Closes: #40916
- [x] **Communication:** I've discussed this with core contributors
already. If the work hasn't been agreed, this work might be rejected
- [ ] **Tests:** no change
- [ ] **Localization:** nothing
- [ ] **Dev docs:** nothing
- [ ] **New binaries:** none
- [ ] **Documentation updated:** nothing
<!-- 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
<!-- Describe how you validated the behavior. Add automated tests
wherever possible, but list manual validation steps taken as well -->
## Validation Steps Performed
## Summary of the Pull Request
Changes the placeholder in the filter box only on the main list page to
"Search for apps, files and commands...":
<img width="786" height="473" alt="image"
src="https://github.com/user-attachments/assets/844d27ae-61cf-42c9-a7f6-ae78817e928c"
/>
The default value remains unchanged as "Type here to search..." for all
other pages (both built-in and internal), unless the author overrides
it:
<img width="786" height="473" alt="image"
src="https://github.com/user-attachments/assets/aeb3500b-9e36-4e35-8dd7-3bd226336823"
/>
<!-- Please review the items on the PR checklist before submitting-->
## PR Checklist
- [x] Closes: #40763
- [x] **Communication:** I've discussed this with core contributors
already. If the work hasn't been agreed, this work might be rejected
- [ ] **Tests:** Added/updated and all pass
- [x] **Localization:** All end-user-facing strings can be localized
- [x] **Dev docs:** Added/updated
- [x] **New binaries:** none
- [x] **Documentation updated:**
## Detailed Description of the Pull Request / Additional comments
## Validation Steps Performed
Closes#40781
Filters out TopLevelCommands whose Id matches an app coming from the
`AllAppsCommandProvider.Page.GetItems()`.
Hate adding processing there, but without adding some type of `bool
HideMeOnSearch` to something low enough (like ICommandItem), I don't see
another way to distinguish these.
This entirely rewrites the shell page. It feels a lot more like the old
run dialog now.
* It's got icons for files & exes
* it can handle network paths
* it can handle `commands /with args...`
* it'll suggest files in that path as you type
* it handles `%environmentVariables%`
* it handles `"Paths with\spaces in them"`
* it shows you the path as a suggestion, in the text box, as you move
the selection
References:
Closes#39044Closes#39419Closes#38298Closes#40311
### Remaining todo's
* [x] Remove the `GenerateAppxManifest` change, and file something to
fix that. We are still generating msix's on every build, wtf
* [x] Clean-up code
* [x] Double-check loc
* [x] Remove a bunch of debug printing that we don't need anymore
* [ ] File a separate PR for moving the file (indexer) commands into a
common project, and re-use those here
* [x] Add history support again! I totally tore that out
* did that in #40427
* [x] make `shell:` paths and weird URI's just work. Good test is
`x-cmdpal://settings`
### further optimizations that probably aren't blocking
* [x] Our fast up-to-date is clearly broken, but I think that's been
broken since early 0.91
* [x] If the exe doesn't change, we don't need to create a new ListItem
for it. We can just re-use the current one, and just change the args
* [ ] if the directory hasn't changed, but we typed more chars (e.g.
`c:\windows\s` -> `c:\windows\sys`), we should cache the ListItem's from
the first query, and re-use them if possible.
_targets #40504_
Major refactoring for #40113
This moves a large swath of the codebase to a `.Core` project. "Core"
doesn't have any explicit dependencies on "extensions", settings or the
current `MainListPage`. It's just a filterable list of stuff. This
should let us make this component a bit more reusable.
This is half of a PR. As I did this, I noticed a particular bit of code
for TopLevelVViewModels and CommandPaletteHost that was _very rough_.
Solving it in this PR would make "move everything to a new project" much
harder to review. So I'm submitting two PRs simultaneously, so we can
see the changes separately, then merge together.
This adds settings to each provider to allow us to control if individual
fallback items are enabled or not, regardless of the provider being
enabled.
This is relevant to _all the threads where disabling fallback commands
came up_
This just adds another section to each provider's settings page, with a
list of the fallback commands.
This also has nothing to do with the "top-level apps search", which is
not really a fallback command - it's its own thing.
Ref #38288. Doesn't close that, because this only controls
enable/disable, not ranking.
From here, we should be able to add a dedicated page in the SUI that
shows all the fallbacks across all providers. That's where we'll enable
the ordering.
<!-- 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
Prevents installed applications from appearing in the top-level search
when Installed Apps extension is disabled.
Previously, application commands were still returned in the global
search results even when the *Installed Apps* extension was turned off.
To match user expectations, the search now respects the extension’s
enabled state.
- Added `IsActive` property to `CommandProviderWrapper` to indicate
whether the provider is both valid and enabled by the user in the
settings.
- Updated `MainListPage` to verify that the provider for `AllApps` is
active before including apps in filtered results.
<!-- Please review the items on the PR checklist before submitting-->
## PR Checklist
- [x] **Closes:** #39937
- [x] **Communication:** I've discussed this with core contributors
already. If the work hasn't been agreed, this work might be rejected
- [x] **Tests:** Added/updated and all pass
- [x] **Localization:** All end-user-facing strings can be localized
- [x] **Dev docs:** nothing to update
- [x] **New binaries:** none
- [x] **Documentation updated:** nothing to update
<!-- 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
<!-- Describe how you validated the behavior. Add automated tests
wherever possible, but list manual validation steps taken as well -->
## Validation Steps Performed
Verified that the Installed app entries are shown in the top-level
search only when the Installed apps extension is enabled. Verified that
turning the Installed apps extension on or off has an immediate effect,
and that the behavior persists after an application restart.
## Problem
The "Reload" command in the Command Palette was only showing up when
searching with a lowercase 'r' (e.g., "reload") but not with an
uppercase 'R' (e.g., "Reload"). This was inconsistent with the
documentation which references a "Reload" command.
## Solution
Fixed the case-sensitivity issue in `FallbackReloadItem.UpdateQuery()`
by changing the string comparison from case-sensitive to
case-insensitive:
```csharp
// Before
_reloadCommand.Name = query.StartsWith('r') ? "Reload" : string.Empty;
// After
_reloadCommand.Name = query.StartsWith("r", StringComparison.OrdinalIgnoreCase) ? "Reload" : string.Empty;
```
This change makes the Reload command visible when typing either "reload"
or "Reload" in the Command Palette, improving the user experience for
extension developers.
Fixes#39769.
---
💡 You can make Copilot smarter by setting up custom instructions,
customizing its development environment and configuring Model Context
Protocol (MCP) servers. Learn more [Copilot coding agent
tips](https://gh.io/copilot-coding-agent-tips) in the docs.
---------
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: zadjii-msft <18356694+zadjii-msft@users.noreply.github.com>
## Summary of the Pull Request
The Command Palette allows users to create extensions with dashes in their names, which is invalid for C# namespaces. This causes projects to fail during build because the name cannot be used as a valid namespace.
## Changes
- Updated the validation regex in `NewExtensionForm.cs` to only allow valid C# identifiers
- Now only allows names starting with a letter or underscore, followed by letters, numbers, or underscores
- Explicitly prevents dashes, spaces, and other special characters
- Improved the error message to clearly explain the requirements for valid C# identifiers
## Before
Previously, the input only validated that no spaces were used:
```csharp
"regex": "^[^\\s]+$"
```
## After
Now the input properly validates for C# namespace compatibility:
```csharp
"regex": "^[a-zA-Z_][a-zA-Z0-9_]*$"
```
This ensures that users cannot create extension projects with names that would fail to build.
## PR Checklist
- [ ] **Closes:** https://github.com/microsoft/PowerToys/issues/38522
* starting AoT flag push
* Few more
* bookmarks
* Really? The VM project compiles?
* Disable publish AOT before we really testing it.
---------
Co-authored-by: Mike Griese <migrie@microsoft.com>
Co-authored-by: Yu Leng (from Dev Box) <yuleng@microsoft.com>
The problem:
> * we need to go update all the Fallback commands. (these are ones that extensions can use to react to the search text - basically, "what the user typed wasn't found immediately, but here's something they can fall back on"
> * this is wacky, because the way I had it, I update each item, and if it "changes visibility", then we need to update the main list, because we've already removed it from the list. So we need to re-update the list to account for that
> * you missed it reading that (and i missed it writing it) but that basically means we re-populate the list F={num fallbacks} times, because each one sends the "do it again" message
> * That results in us basically creating (F+1)*(N=num items+apps) view models, initializing them, and not needing most of them
The crux here being a single thread, to update all the fallback items,
that then only raises _one_ items changed at the very end.
I don't love this, one misbehaving fallback could stop all the others. In theory, we should do a parallel update of all these things, with a like, 1s timeout on each leg.
But it has gotta be faster till we can do #38140 (or similar)
Closes: (not sure I filed one). But the first typed character _felt_ slow.
This is a much tidier solution. Don't default _everything_ to a weight of 1 if the query is whitespace. Instead, do a simple string contains check (because FuzzySearch will beef it on just whitespace)
Closes#38133
I originally based this off of #38157, so I know these two won't collide
Both `TopLevelCommandItemWrapper` and `TopLevelViewModel` were really the same thing. The latter was from an earlier prototype, and the former is a more correct, safer abstraction. We really should have only ever used the former, but alas, we only used it for the SUI, and it piggy-backed off the latter, and that meant the latter's bugs became the former's.
tldr: I made the icon access safe in the SUI.
And while I was doing this, because we now have a cleaner VM abstraction here in the host, we can actually cleanly disable extensions, because the `CommandProviderWrapper` knows which `ViewModel`s it made.
Closes https://github.com/zadjii-msft/PowerToys/issues/426
Closes https://github.com/zadjii-msft/PowerToys/issues/478
Closes https://github.com/zadjii-msft/PowerToys/issues/577
Windows Command Palette ("CmdPal") is the next iteration of PowerToys Run. With extensibility at its core, the Command Palette is your one-stop launcher to start _anything_.
By default, CmdPal is bound to <kbd>Win+Alt+Space</kbd>.
----
This brings the current preview version of CmdPal into the upstream PowerToys repo. There are still lots of bugs to work out, but it's reached the state we're ready to start sharing it with the world. From here, we can further collaborate with the community on the features that are important, and ensuring that we've got a most robust API to enable developers to build whatever extensions they want.
Most of the built-in PT Run modules have already been ported to CmdPal's extension API. Those include:
* Installed apps
* Shell commands
* File search (powered by the indexer)
* Windows Registry search
* Web search
* Windows Terminal Profiles
* Windows Services
* Windows settings
There are a couple new extensions built-in
* You can now search for packages on `winget` and install them right from the palette. This also powers searching for extensions for the palette
* The calculator has an entirely new implementation. This is currently less feature complete than the original PT Run one - we're looking forward to updating it to be more complete for future ingestion in Windows
* "Bookmarks" allow you to save shortcuts to files, folders, and webpages as top-level commands in the palette.
We've got a bunch of other samples too, in this repo and elsewhere
### PowerToys specific notes
CmdPal will eventually graduate out of PowerToys to live as its own application, which is why it's implemented just a little differently than most other modules. Enabling CmdPal will install its `msix` package.
The CI was minorly changed to support CmdPal version numbers independent of PowerToys itself. It doesn't make sense for us to start CmdPal at v0.90, and in the future, we want to be able to rev CmdPal independently of PT itself.
Closes#3200, closes#3600, closes#7770, closes#34273, closes#36471, closes#20976, closes#14495
-----
TODOs et al
**Blocking:**
- [ ] Images and descriptions in Settings and OOBE need to be properly defined, as mentioned before
- [ ] Niels is on it
- [x] Doesn't start properly from PowerToys unless the fix PR is merged.
- https://github.com/zadjii-msft/PowerToys/pull/556 merged
- [x] I seem to lose focus a lot when I press on some limits, like between the search bar and the results.
- This is https://github.com/zadjii-msft/PowerToys/issues/427
- [x] Turned off an extension like Calculator and it was still working.
- Need to get rid of that toggle, it doesn't do anything currently
- [x] `ListViewModel.<FetchItems>` crash
- Pretty confident that was fixed in https://github.com/zadjii-msft/PowerToys/pull/553
**Not blocking / improvements:**
- Show the shortcut through settings, as mentioned before, or create a button that would open CmdPalette settings.
- When PowerToys starts, CmdPalette is always shown if enabled. That's weird when just starting PowerToys/ logging in to the computer with PowerToys auto-start activated. I think this should at least be a setting.
- Needing to double press a result for it to do the default action seems quirky. If one is already selected, I think just pressing should be enough for it to do the action.
- This is currently a setting, though we're thinking of changing the setting even more: https://github.com/zadjii-msft/PowerToys/issues/392
- There's no URI extension. Was surprised when typing a URL that it only proposed a web search.
- [x] There's no System commands extension. Was expecting to be able to quickly restart the computer by typing restart but it wasn't there.
- This is in PR https://github.com/zadjii-msft/PowerToys/pull/452
---------
Co-authored-by: joadoumie <98557455+joadoumie@users.noreply.github.com>
Co-authored-by: Jordi Adoumie <jordiadoumie@microsoft.com>
Co-authored-by: Mike Griese <zadjii@gmail.com>
Co-authored-by: Niels Laute <niels.laute@live.nl>
Co-authored-by: Michael Hawker <24302614+michael-hawker@users.noreply.github.com>
Co-authored-by: Stefan Markovic <57057282+stefansjfw@users.noreply.github.com>
Co-authored-by: Seraphima <zykovas91@gmail.com>
Co-authored-by: Jaime Bernardo <jaime@janeasystems.com>
Co-authored-by: Kristen Schau <47155823+krschau@users.noreply.github.com>
Co-authored-by: Eric Johnson <ericjohnson327@gmail.com>
Co-authored-by: Ethan Fang <ethanfang@microsoft.com>
Co-authored-by: Yu Leng (from Dev Box) <yuleng@microsoft.com>
Co-authored-by: Clint Rutkas <clint@rutkas.com>
