This commit introduces "AtlasEngine", a new text renderer based on DxEngine.
But unlike it, DirectWrite and Direct2D are only used to rasterize glyphs.
Blending and placing these glyphs into the target view is being done using
Direct3D and a simple HLSL shader. Since this new renderer more aggressively
assumes that the text is monospace, it simplifies the implementation:
The viewport is divided into cells, and its data is stored as a simple matrix.
Modifications to this matrix involve only simple pointer arithmetic and is easy
to understand. But just like with DxEngine however, DirectWrite
related code remains extremely complex and hard to understand.
Supported features:
* Basic text rendering with grayscale AA
* Foreground and background colors
* Emojis, including zero width joiners
* Underline, dotted underline, strikethrough
* Custom font axes and features
* Selections
* All cursor styles
* Full alpha support for all colors
* _Should_ work with Windows 7
Unsupported features:
* A more conservative GPU memory usage
The backing texture atlas for glyphs is grow-only and will not shrink.
After 256MB of memory is used up (~20k glyphs) text output
will be broken until the renderer is restarted.
* ClearType
* Remaining gridlines (left, right, top, bottom, double underline)
* Hyperlinks don't get full underlines if hovered in WT
* Softfonts
* Non-default line renditions
Performance:
* Runs at up to native display refresh rate
Unfortunately the frame rate often drops below refresh rate, due us
fighting over the buffer lock with other parts of the application.
* CPU consumption is up to halved compared to DxEngine
AtlasEngine is still highly unoptimized. Glyph hashing
consumes up to a third of the current CPU time.
* No regressions in WT performance
VT parsing and related buffer management takes up most of the CPU time (~85%),
due to which the AtlasEngine can't show any further improvements.
* ~2x improvement in raw text throughput in OpenConsole
compared to DxEngine running at 144 FPS
* ≥10x improvement in colored VT output in WT/OpenConsole
compared to DxEngine running at 144 FPS
## Summary of the Pull Request
This replaces the `GridLines` enum in the renderers with a `til::enumset` type, avoiding the need for the various `WI_IsFlagSet` macros and flag operators.
## References
This is followup to PR #10492 which introduced the `enumset` class.
## PR Checklist
* [ ] Closes #xxx
* [x] CLA signed.
* [ ] Tests added/passed
* [ ] Documentation updated.
* [ ] 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.
## Validation Steps Performed
I've manually confirmed that all the different gridlines are still rendering correctly in both the GDI and DX renderers.
This commit adds the ability to interact with subtrees of panes.
Have you ever thought that you don't have enough regression testing to
do? Boy do I have the PR for you! This breaks all kinds of assumptions
about what is or is not focused, largely complicated by the fact that a
pane is not a proper control. I did my best to cover as many cases as I
could, but I wouldn't be surprised if there are some things broken that
I am unaware of.
Done:
- Add `parent` and `child` movement directions to move up and down the
tree respectively
- When a parent pane is selected it will have borders all around it in
addition to any borders the children have.
- Fix focus, swap, split, zoom, toggle orientation, resize, and move to
all handle interacting with more than one pane.
- Similarly the actions for font size changing, closing, read-only, clearing
buffer, and changing color scheme will distribute to all children.
- This technically leaves control focus on the original control in the
focused subtree because panes aren't proper controls themselves. This
is also used to make sure we go back down the same path with the
`child` movement.
- You can zoom a parent pane, and click between different zoomed
sub-panes and it won't unzoom you until you use moveFocus or another
action. This wasn't explicitly programmed behavior so it is probably
buggy (I've quashed a couple at least). It is a natural consequence of
showing multiple terminals and allowing you to focus a terminal and a
parent separately, since changing the active pane directly does not
unzoom. This also means there can be a disconnect between what pane is
zoomed and what pane is active.
## Validation Steps Performed
Tested focus movement, swapping, moving panes, and zooming.
Closes#10733
Implements the following keyboard selection non-configurable key bindings:
- shift+arrow --> move endpoint by character
- ctrl+shift+left/right --> move endpoint by word
- shift+home/end --> move to beginning/end of line
- ctrl+shift+home/end --> move to beginning/end of buffer
This was purposefully done in the ControlCore layer to make keyboard selection an innate part of how the terminal functions (aka a shared component across terminal consumers).
## References
#715 - Keyboard Selection
#2840 - Spec
## Detailed Description of the Pull Request / Additional comment
The most relevant section is `TerminalSelection.cpp`, where we define how each movement operates. It's basically a giant embedded switch-case statement. We leverage a lot of the work done in a11y to perform the movements.
## Validation Steps Performed
- General cases:
- test all of the key bindings added
- Corner cases:
- `char`: wide glyph support
- `word`: move towards, away, and across the selection pivot
- automatically scroll viewport
- ESC (and other key combos) are still clearing the selection properly
## Summary of the Pull Request
This adds a new setting `intenseTextStyle`. It's a per-appearance, control setting, defaulting to `"all"`.
* When set to `"all"` or `["bold", "bright"]`, then we'll render text as both **bold** and bright (1.10 behavior)
* When set to `"bold"`, `["bold"]`, we'll render text formatted with `^[[1m` as **bold**, but not bright
* When set to `"bright"`, `["bright"]`, we'll render text formatted with `^[[1m` as bright, but not bold. This is the pre 1.10 behavior
* When set to `"none"`, we won't do anything special for it at all.
## references
* I last did this in #10648. This time it's an enum, so we can add bright in the future. It's got positive wording this time.
* ~We will want to add `"bright"` as a value in the future, to disable the auto intense->bright conversion.~ I just did that now.
* #5682 is related
## PR Checklist
* [x] Closes#10576
* [x] I seriously don't think we have an issue for "disable intense is bright", but I'm not crazy, people wanted that, right? https://github.com/microsoft/terminal/issues/2916#issuecomment-544880423 was the closest
* [x] I work here
* [x] Tests added/passed
* [x] https://github.com/MicrosoftDocs/terminal/pull/381
## Validation Steps Performed
<!--  -->
Yea that works. Printed some bold text, toggled it on, the text was no longer bold. hooray.
### EDIT, 10 Aug
```json
"intenseTextStyle": "none",
"intenseTextStyle": "bold",
"intenseTextStyle": "bright",
"intenseTextStyle": "all",
"intenseTextStyle": ["bold", "bright"],
```
all work now. Repro script:
```sh
printf "\e[1m[bold]\e[m[normal]\e[34m[blue]\e[1m[bold blue]\e[m\n"
```
This PR adds conhost support for downloadable soft fonts - also known as
dynamically redefinable character sets (DRCS) - using the `DECDLD`
escape sequence.
These fonts are typically designed to work on a specific terminal model,
and each model tends to have a different character cell size. So in
order to support as many models as possible, the code attempts to detect
the original target size of the font, and then scale the glyphs to fit
our current cell size.
Once a font has been downloaded to the terminal, it can be designated in
the same way you would a standard character set, using an `SCS` escape
sequence. The identification string for the set is defined by the
`DECDLD` sequence. Internally we map the characters in this set to code
points `U+EF20` to `U+EF7F` in the Unicode private use are (PUA).
Then in the renderer, any characters in that range are split off into
separate runs, which get painted with a special font. The font itself is
dynamically generated as an in-memory resource, constructed from the
downloaded character bitmaps which have been scaled to the appropriate
size.
If no soft fonts are in use, then no mapping of the PUA code points will
take place, so this shouldn't interfere with anyone using those code
points for something else, as along as they aren't also trying to use
soft fonts. I also tried to pick a PUA range that hadn't already been
snatched up by Nerd Fonts, but if we do receive reports of a conflict,
it's easy enough to change.
## Validation Steps Performed
I added an adapter test that runs through a bunch of parameter
variations for the `DECDLD` sequence, to make sure we're correctly
detecting the font sizes for most of the known DEC terminal models.
I've also tested manually on a wide range of existing fonts, of varying
dimensions, and from multiple sources, and made sure they all worked
reasonably well.
Closes#9164
The `_invalidMap` size is dependent on both `clientSize` as well
as `glyphCellSize` and must be resized when either changes.
## PR Checklist
* [x] Closes#10855
* [x] I work here
## Validation Steps Performed
* Changing font size with Ctrl+Mousewheel in fullscreen works ✔️
This commit fixes the UseDx mode for conhost.
In order to add support for UseDx without calling `SetWindowSize`,
responsibility for resizing `_invalidMap` has been moved to occur
only when the renderer itself recognizes a new size. Furthermore
`InvalidateAll` is now the central point to invalidate `_invalidMap`.
## Validation Steps Performed
* Enabling `UseDx` enables the DxEngine for conhost ✔️
* Resizing windows in conhost works ✔️
* Resizing windows in WT works ✔️Closes#5455
Adds support for users to be able to set font features and axes (see the spec for more details!)
## Detailed Description
**CustomTextLayout**
- Asks the `DxFontRenderData` for the font features when getting glyphs
- _If any features have been set/updated, we always skip the "isTextSimple" shortcut_
- Asks the `_formatInUse` for any font axes when mapping characters in `_AnalyzeFontFallback`
**DxFontRenderData**
- Stores a map of font features (initialized to the [standard feature list])
- Stores a map of font axes
- Has methods to add font features/axes to the map or update existing ones
- Has methods to retrieve the font features/axes
- Sets the font axes in the `IDWriteTextFormat` when creating it
## Validation Steps Performed
It works!
[standard feature list]: ac5aef67d1/DrawableObject.ixx (L802)
Specified in #10457
Related to #1790Closes#759Closes#5828
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
This pull request converts four of our existing `#ifdef` (or `#ifndef`)
`INSIDE_WINDOWS` blocks to til::features:
* Attempting to establish a handoff session (inside Windows only)
* The ability to *receive* a handoff session (outside Windows only)
* The DX engine (outside Windows only) and shaders (also outside only)
* Whether we use numpad event synthesis for clipboard/conpty (inside
Windows only)
Most of these are using the preprocessor verison of til::feature, only
because it is more difficult to gate the inclusion of headers on
constant expressions. I'd love to prefer the compile time version.
Set keyword flags on all events so those sharing a provider with
telemetry do not fire unless tracing is enabled
## PR Checklist
* [x] Closes#10093
* [x] I work here
* [x] Tests passed
* [x] Documentation added in `til.h` about how keywords work and at the
only other site of keywords we define in the Host project tracing
files.
## Detailed Description of the Pull Request / Additional comments
I initially thought that we would need to split providers here to
accomplish this... but @DHowett helped me realize that might be a lot of
additional metadata and bloat binary size. So with help from a friend
from fundamentals, I realized that we could use Keywords to
differentiate here. We can no longer define 0 keywords as that
represents an any/all scenario. Every `TraceLoggingWrite` event now
needs a keyword. When our events have a keyword, they're not included in
any trace. Additionally, when we have an explicit keyword to check that
is different from the ones used for the telemetry pipeline, we can
ensure that we only do "hard work" to generate debug trace data when an
"ALL" type listener like TraceView or Windows Performance Recorder with
our profiles is listening to these providers for ALL keyworded events.
## Validation Steps Performed
- [x] - Built with full release build config to confirm performance is
worse than dev builds BECAUSE of the telemetry event collector camping
our provider and triggering full trace event generation on shared
providers.
- [x] - Built with full release build config to enable statistics
collection and validated trace event collection is excluded and trace
event short-circuits work with this change.
- [x] - Checked that TraceView still sees both telemetry and tracing
events
- [x] - Checked that WPR with our .wprp profile sees both telemetry and
tracing events
## Summary of the Pull Request
Brace yourselves, it's finally here. This PR does the dirty work of splitting the monolithic `TermControl` into three components. These components are:
* `ControlCore`: This encapsulates the `Terminal` instance, the `DxEngine` and `Renderer`, and the `Connection`. This is intended to everything that someone might need to stand up a terminal instance in a control, but without any regard for how the UX works.
* `ControlInteractivity`: This is a wrapper for the `ControlCore`, which holds the logic for things like double-click, right click copy/paste, selection, etc. This is intended to be a UI framework-independent abstraction. The methods this layer exposes can be called the same from both the WinUI TermControl and the WPF control.
* `TermControl`: This is the UWP control. It's got a Core and Interactivity inside it, which it uses for the actual logic of the terminal itself. TermControl's main responsibility is now
By splitting into smaller pieces, it will enable us to
* write unit tests for the `Core` and `Interactivity` bits, which we desparately need
* Combine `ControlCore` and `ControlInteractivity` in an out-of-proc core process in the future, to enable tab tearout.
However, we're not doing that work quite yet. There's still lots of work to be done to enable that, thought this is likely the biggest portion.
Ideally, this would just be methods moved wholesale from one file to another. Unfortunately, there are a bunch of cases where that didn't work as well as expected. Especially when trying to better enforce the boundary between the classes.
We've got a couple tests here that I've added. These are partially examples, and partially things I ran into while implementing this. A bunch of things from #7001 can go in now that we have this.
This PR is gonna be a huge pain to review - 38 files with 3,730 additions and 1,661 deletions is nothing to scoff at. It will also conflict 100% with anything that's targeting `TermControl`. I'm hoping we can review this over the course of the next week and just be done with it, and leave plenty of runway for 1.9 bugs in post.
## References
* In pursuit of #1256
* Proc Model: #5000
* https://github.com/microsoft/terminal/projects/5
## PR Checklist
* [x] Closes#6842
* [x] Closes https://github.com/microsoft/terminal/projects/5#card-50760249
* [x] Closes https://github.com/microsoft/terminal/projects/5#card-50760258
* [x] I work here
* [x] Tests added/passed
* [n/a] Requires documentation to be updated
## Detailed Description of the Pull Request / Additional comments
* I don't love the names `ControlCore` and `ControlInteractivity`. Open to other names.
* I added a `ICoreState` interface for "properties that come from the `ControlCore`, but consumers of the `TermControl` need to know". In the future, these will all need to be handled specially, because they might involve an RPC call to retrieve the info from the core (or cache it) in the window process.
* I've added more `EventArgs` to make more events proper `TypedEvent`s.
* I've changed how the TerminalApp layer requests updated TaskbarProgress state. It doesn't need to pump TermControl to raise a new event anymore.
* ~~Something that snuck into this branch in the very long history is the switch to `DCompositionCreateSurfaceHandle` for the `DxEngine`. @miniksa wrote this originally in 30b8335, I'm just finally committing it here. We'll need that in the future for the out-of-proc stuff.~~
* I reverted this in c113b65d9. We can revert _that_ commit when we want to come back to it.
* I've changed the acrylic handler a decent amount. But added tests!
* All the `ThrottledFunc` things are left in `TermControl`. Some might be able to move down into core/interactivity, but once we figure out how to use a different kind of Dispatcher (because a UI thread won't necessarily exist for those components).
* I've undoubtably messed up the merging of the locking around the appearance config stuff recently
## Validation Steps Performed
I've got a rolling list in https://github.com/microsoft/terminal/issues/6842#issuecomment-810990460 that I'm updating as I go.
This is my attempt to isolate all the dwrite font related thing by
introducing a new layer - `DxFontRenderData`. This will free
`DxRenderer` & `CustomTextLayout` from the burden of handling fonts &
box effects. The logic is more simplified & streamlined.
In short I just moved everything fonts-related into `DxFontRenderData`
and started from there. There's no modification to code logic. Just pure
structural stuff.
SGR support tracking issue: #6879
Initial Italic support PR: #8580
This PR adds support for the VT line rendition attributes, which allow
for double-width and double-height line renditions. These renditions are
enabled with the `DECDWL` (double-width line) and `DECDHL`
(double-height line) escape sequences. Both reset to the default
rendition with the `DECSWL` (single-width line) escape sequence. For now
this functionality is only supported by the GDI renderer in conhost.
There are a lot of changes, so this is just a general overview of the
main areas affected.
Previously it was safe to assume that the screen had a fixed width, at
least for a given point in time. But now we need to deal with the
possibility of different lines have different widths, so all the
functions that are constrained by the right border (text wrapping,
cursor movement operations, and sequences like `EL` and `ICH`) now need
to lookup the width of the active line in order to behave correctly.
Similarly it used to be safe to assume that buffer and screen
coordinates were the same thing, but that is no longer true. Lots of
places now need to translate back and forth between coordinate systems
dependent on the line rendition. This includes clipboard handling, the
conhost color selection and search, accessibility location tracking and
screen reading, IME editor positioning, "snapping" the viewport, and of
course all the rendering calculations.
For the rendering itself, I've had to introduce a new
`PrepareLineTransform` method that the render engines can use to setup
the necessary transform matrix for a given line rendition. This is also
now used to handle the horizontal viewport offset, since that could no
longer be achieved just by changing the target coordinates (on a double
width line, the viewport offset may be halfway through a character).
I've also had to change the renderer's existing `InvalidateCursor`
method to take a `SMALL_RECT` rather than a `COORD`, to allow for the
cursor being a variable width. Technically this was already a problem,
because the cursor could occupy two screen cells when over a
double-width character, but now it can be anything between one and four
screen cells (e.g. a double-width character on the double-width line).
In terms of architectural changes, there is now a new `lineRendition`
field in the `ROW` class that keeps track of the line rendition for each
row, and several new methods in the `ROW` and `TextBuffer` classes for
manipulating that state. This includes a few helper methods for handling
the various issues discussed above, e.g. position clamping and
translating between coordinate systems.
## Validation Steps Performed
I've manually confirmed all the double-width and double-height tests in
_Vttest_ are now working as expected, and the _VT100 Torture Test_ now
renders correctly (at least the line rendition aspects). I've also got
my own test scripts that check many of the line rendition boundary cases
and have confirmed that those are now passing.
I've manually tested as many areas of the conhost UI that I could think
of, that might be affected by line rendition, including things like
searching, selection, copying, and color highlighting. For
accessibility, I've confirmed that the _Magnifier_ and _Narrator_
correctly handle double-width lines. And I've also tested the Japanese
IME, which while not perfect, is at least useable.
Closes#7865
## References
* See also #8617
## PR Checklist
* [x] Supports #3075
* [x] I work here.
* [x] Manual test.
## Detailed Description of the Pull Request / Additional comments
### Window Title Generation
Every time the renderer checks the title, it's doing two bad things that
I've fixed:
1. It's assembling the prefix to the full title doing a concatenation.
No one ever gets just the prefix ever after it is set besides the
concat. So instead of storing prefix and the title, I store the
assembled prefix + title and the bare title.
2. A copy must be made because it was returning `std::wstring` instead
of `std::wstring&`. Now it returns the ref.
### Dirty Area Return
Every time the renderer checks the dirty area, which is sometimes
multiple times per pass (regular text printing, again for selection,
etc.), a vector is created off the heap to return the rectangles. The
consumers only ever iterate this data. Now we return a span over a
rectangle or rectangles that the engine must store itself.
1. For some renderers, it's always a constant 1 element. They update
that 1 element when dirty is queried and return it in the span with a
span size of 1.
2. For other renderers with more complex behavior, they're already
holding a cached vector of rectangles. Now it's effectively giving
out the ref to those in the span for iteration.
### Bitmap Runs
The `til::bitmap` used a `std::optional<std::vector<til::rectangle>>`
inside itself to cache its runs and would clear the optional when the
runs became invalidated. Unfortunately doing `.reset()` to clear the
optional will destroy the underlying vector and have it release its
memory. We know it's about to get reallocated again, so we're just going
to make it a `std::pmr::vector` and give it a memory pool.
The alternative solution here was to use a `bool` and
`std::vector<til::rectangle>` and just flag when the vector was invalid,
but that was honestly more code changes and I love excuses to try out
PMR now.
Also, instead of returning the ref to the vector... I'm just returning a
span now. Everyone just iterates it anyway, may as well not share the
implementation detail.
### UTF-8 conversions
When testing with Terminal and looking at the `conhost.exe`'s PTY
renderer, it spends a TON of allocation time on converting all the
UTF-16 stuff inside to UTF-8 before it sends it out the PTY. This was
because `ConvertToA` was allocating a string inside itself and returning
it just to have it freed after printing and looping back around again...
as a PTY does.
The change here is to use `til::u16u8` that accepts a buffer out
parameter so the caller can just hold onto it.
## Validation Steps Performed
- [x] `big.txt` in conhost.exe (GDI renderer)
- [x] `big.txt` in Terminal (DX, PTY renderer)
- [x] Ensure WDDM and BGFX build under Razzle with this change.
Teaches renderer base to keep thread alive if engine requests it.
`DxEngine` now requests it if shaders are on.
- The render engine interface now has a true/false to return whether the
specific renderer wants another frame to immediately follow up. The
renderer base will ask for this information as it ends the paint on
any particular engine (which is the time where invalid regions are
typically cleaned up) and just poke the render thread the same as if
an invalidation request came in from outside of render-land. That will
trigger the render thread to just keep moving in the same way as any
other invalidation.
## Validation Steps Performed
- [x] Actually built it
- [x] Actually try it
I promised this in #8994
Correctly sets the time input on the pixelShaderSettings struct, which was previously hard-coded to `0.0f`.
## PR Checklist
* [x] Closes#8935
* [x] CLA signed. If not, go over [here](https://cla.opensource.microsoft.com/microsoft/Terminal) and sign the CLA
* [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: #8935
## Detailed Description of the Pull Request / Additional comments
I added a private field to `DxEngine` to store the timestamp for when a custom shader is first loaded. The field is initialized in `_SetupTerminalEffects()`, and the calculated time value (seconds since the timestamp) passed to the actual shader is set in `_ComputePixelShaderSettings()`.
There remains an issue with with jerky animation due to the renderer not repainting when the window contents are not updated (see discussion in the original issue).
This is basically my first time writing C++; constructive review is enthusiastically welcomed 🙂
## Validation Steps Performed
I manually tested using a variety of simple shaders that rely on time input for animation.
This PR adds support for the ANSI _italic_ graphic rendition attribute,
which is enabled by the `SGR 3` escape sequence.
For the GDI renderer, I've just created an additional italic variant of
the font, and then the `UpdateDrawingBrushes` method selects the
appropriate font variant into the device context based on the requested
text attributes.
It's a bit more complicated in the DX renderer, because we need both an
italic variant of the font, and a variant of the text format object. The
`CustomTextLayout` class also had to be updated to hold the two font and
format instances, and decide which of the variants to use based on a
`useItalicFont` property in the drawing context, initially set in the
`UpdateDrawingBrushes` method.
## Validation Steps Performed
I've created some test content using a range of different character sets
(e.g. CJK, block characters, emoji, etc.), then applied the italic
attribute mixed with various other SGR attributes to see how they
interact. The output isn't always perfect, but I think it seems
reasonable given the constraints of a cell-based terminal renderer.
Closes#5461
Co-authored-by: mrange <marten_range@hotmail.com>
I loved the pixel shaders in #7058, but that PR needed a bit of polish
to be ready for ingestion. This PR is almost _exactly_ that PR, with
some small changes.
* It adds a new pre-profile setting `"experimental.pixelShaderPath"`,
which lets the user set a pixel shader to use with the Terminal.
- CHANGED FROM #7058: It does _not_ add any built-in shaders.
- CHANGED FROM #7058: it will _override_
`experimental.retroTerminalEffect`
* It adds a bunch of sample shaders in `samples/shaders`. Included:
- A NOP shader as a base to build from.
- An "invert" shader that inverts the colors, as a simple example
- An "grayscale" shader that converts all colors to grayscale, as a
simple example
- An "raster bars" shader that draws some colored bars on the screen
with a drop shadow, as a more involved example
- The original retro terminal effects, as a more involved example
- It also includes a broken shader, as an example of what heppens
when the shader fails to compile
- CHANGED FROM #7058: It does _not_ add the "retroII" shader we were
all worried about.
* When a shader fails to be found or fails to compile, we'll display an
error dialog to the user with a relevant error message.
- CHANGED FROM #7058: Originally, #7058 would display "error bars"
on the screen. I've removed that, and had the Terminal disable the
shader entirely then.
* Renames the `toggleRetroEffect` action to `toggleShaderEffect`.
(`toggleRetroEffect` is now an alias to `toggleShaderEffect`). This
action will turn the shader OR the retro effects on/off.
`toggleShaderEffect` works the way you'd expect it to, but the mental
math on _how_ is a little weird. The logic is basically:
```
useShader = shaderEffectsEnabled ?
(pixelShaderProvided ?
pixelShader :
(retroEffectEnabled ?
retroEffect : null
)
) :
null
```
and `toggleShaderEffect` toggles `shaderEffectsEnabled`.
* If you've got both a shader and retro enabled, `toggleShaderEffect`
will toggle between the shader on/off.
* If you've got a shader and retro disabled, `toggleShaderEffect` will
toggle between the shader on/off.
References #6191
References #7058Closes#7013Closes#3930 "Add setting to retro terminal shader to control blur
radius, color"
Closes#3929 "Add setting to retro terminal shader to enable drawing
scanlines"
- At this point, just roll your own version of the shader.
I added `enum class` to one thing and decided that that was quite enough
before disabling the `enum class` warning.
Looks like 16.8 made more map/vector operations noexcept, so we have to
re-annotate to remain compliant.
This pull request switches up the treatment we use for pattern-detected
links and OSC 8 hyperlinks:
* Links generated via OSC 8 have a sparse dotted underline instead of a
thick dashed one
* Links generated by pattern detection _are not underlined until they've
hovered_
* This papers over a visual glitch that is a result of us updating
the pattern matches every ~500ms (on change)
Closes#8123
Adds the ability to manually handle the terminal renderer resizing
events by allowing different render size and WPF control size. This is
done by adding an `AutoFill` property to the control that prevents the
renderer from automatically resizing and tells the WPF control to fill
in the extra space with the terminal background as shown below:
This PR adds the following:
- Helper method in the DX engine to convert character viewports into
pixel viewports
- `AutoFill` property that prevents automatic resizing of the renderer
- Tweaks and fixes that automatically fill in the empty space if
`AutoFill` is set to false
- Fixes resizing methods and streamlines their codepath
## Validation Steps Performed
Manual validation with the Visual Studio Integrated Terminal tool
window.
- Render hyperlinks with a dashed underline
- Render hovered hyperlinks with a solid underline
- Show URI tooltip on hover
TermControl now has a canvas that contains a tiny border to which a
tooltip is attached. When we hover over hyperlinked text, we move the
border to the mouse location and update the tooltip content with the
URI.
Introduced a new underline type (HyperlinkUnderline), supports rendering
for it, and uses it to render hyperlinks. HyperlinkUnderline is usually
a dashed underline, but when a link is hovered, all text with the same
hyperlink ID is rendered with a solid underline.
References #5001
This commit removes our local copy of clang-format 8 and replaces it
with a newly-built nuget package containing clang-format 10.
This resolves the inconsistency between our version of clang-format and
the one shipped in Visual Studio.
A couple minor format changes were either required or erroneously forced
upon us--chief among them is a redistribution of `*`s around SAL
annotations in inline class members of COM classes. Don't ask why; I
couldn't figure it out.
We had some aspirational goals for our formatting, which were left in
but commented out. Enabling them changes our format a little more than
I'm comfortable with, so I uncommented them and locked them to the
format style we've been using for the past year. We may not love it, but
our aspirations may not matter here any longer. Consistent formatting is
better than perfect formatting.
Adds the ability to set the selection background opacity when setting the
selection background. This also exposes the selection background and alpha
through the terminal WPF container.
New warnings were added in VS 16.7 and `std::map::erase` is now `noexcept`.
Update our code to be compatible with the new enforcement.
## PR Checklist
* [x] Closes broken audit in main after Agents updated over the weekend.
* [x] I work here.
* [x] Audit mode passes now
* [x] Am core contributor.
## Validation Steps Performed
* [x] Ran audit mode locally
This PR adds support for the ANSI _doubly underlined_ graphic rendition
attribute, which is enabled by the `SGR 21` escape sequence.
There was already an `ExtendedAttributes::DoublyUnderlined` flag in the
`TextAttribute` class, but I needed to add `SetDoublyUnderlined` and
`IsDoublyUnderlined` methods to access that flag, and update the
`SetGraphicsRendition` methods of the two dispatchers to set the
attribute on receipt of the `SGR 21` sequence. I also had to update the
existing `SGR 24` handler to reset _DoublyUnderlined_ in addition to
_Underlined_, since they share the same reset sequence.
For the rendering, I've added a new grid line type, which essentially
just draws an additional line with the same thickness as the regular
underline, but slightly below it - I found a gap of around 0.05 "em"
between the lines looked best. If there isn't enough space in the cell
for that gap, the second line will be clamped to overlap the first, so
you then just get a thicker line. If there isn't even enough space below
for a thicker line, we move the offset _above_ the first line, but just
enough to make it thicker.
The only other complication was the update of the `Xterm256Engine` in
the VT renderer. As mentioned above, the two underline attributes share
the same reset sequence, so to forward that state over conpty we require
a slightly more complicated process than with most other attributes
(similar to _Bold_ and _Faint_). We first check whether either underline
attribute needs to be turned off to send the reset sequence, and then
check individually if each of them needs to be turned back on again.
## Validation Steps Performed
For testing, I've extended the existing attribute tests in
`AdapterTest`, `VTRendererTest`, and `ScreenBufferTests`, to make sure
we're covering both the _Underlined_ and _DoublyUnderlined_ attributes.
I've also manually tested the `SGR 21` sequence in conhost and Windows
Terminal, with a variety of fonts and font sizes, to make sure the
rendering was reasonably distinguishable from a single underline.
Closes#2916
This is a refactoring of the grid line renderers, adjusting the line
widths to scale with the font size, and optimising the implementation to
cut down on the number of draw calls. It also extends the supported grid
line types to include true underlines and strike-through lines in the
style of the active font.
The main gist of the optimisation was to render the horizontal lines
with a single draw call, instead of a loop with lots of little strokes
joined together. In the case of the vertical lines, which still needed
to be handled in a loop, I've tried to move the majority of static
calculations outside the loop, so there is bit of optimisation there
too.
At the same time this code was updated to support a variable stroke
width for the lines, instead of having them hardcoded to 1 pixel. The
width is now calculated as a fraction of the font size (0.025 "em"),
which is still going to be 1 pixel wide in most typical usage, but will
scale up appropriately if you zoom in far enough.
And in preparation for supporting the SGR strike-through attribute, and
true underlines, I've extended the grid line renders with options for
handling those line types as well. The offset and thickness of the lines
is obtained from the font metrics (rounded to a pixel width, with a
minimum of one pixel), so they match the style of the font.
VALIDATION
For now we're still only rendering grid lines, and only the top and
bottom lines in the case of the DirectX renderer in Windows Terminal. So
to test, I hacked in some code to force the renderer to use all the
different options, confirming that they were working in both the GDI and
DirectX renderers.
I've tested the output with a number of different fonts, comparing it
with the same text rendered in WordPad. For the most part they match
exactly, but there can be slight differences when we adjust the font
size for grid alignment. And in the case of the GDI renderer, where
we're working with pixel heights rather than points, it's difficult to
match the sizes exactly.
This is a first step towards supporting the strike-through attribute
(#6205) and true underlines (#2915).
Closes#6911
We spend a lot of time invalidating in the DX Renderer. This is a
creative trick to not bother invalidating any further if we can tell
that the bitmap is already completely invalidated. That is, if we've
scrolled at least an entire screen in height... then the entire bitmap
had to have been marked as invalid as the new areas were "uncovered" by
the `InvalidateScroll` command. So further setting invalid bits on top
of a fully invalid map is pointless.
Note: I didn't use `bitmap::all()` here because it is significantly
slower to check all the bits than it is to just reason out that the
bitmap was already fully marked.
## Validation Steps Performed
- Run `time cat big.txt`. Checked average time before/after, WPR traces
before/after.
We were using std::basic_string_view as a stand-in for std::span so that
we could change over all at once when C++20 dropped with full span
support. That day's not here yet, but as of 54a7fce3e we're using GSL 3,
whose span is C++20-compliant.
This commit replaces every instance of basic_string_view that was not
referring to an actual string with a span of the appropriate type.
I moved the `const` qualifier into span's `T` because while
`basic_string_view.at()` returns `const T&`, `span.at()` returns `T&`
(without the const). I wanted to maintain the invariant that members of
the span were immutable.
* Mechanical Changes
* `sv.at(x)` -> `gsl::at(sp, x)`
* `sv.c{begin,end}` -> `sp.{begin,end}` (span's iterators are const)
I had to replace a `std::basic_string<>` with a `std::vector<>` in
ConImeInfo, and I chose to replace a manual array walk in
ScreenInfoUiaProviderBase with a ranged-for. Please review those
specifically.
This will almost certainly cause a code size regression in Windows
because I'm blowing out all the PGO counts. Whoops.
Related: #3956, #975.
This is a refactoring of the renderer color calculations to simplify the
implementation, and to make it easier to support additional
color-altering rendition attributes in the future (e.g. _faint_ and
_conceal_).
## References
* This is a followup to PRs #3817 and #6809, which introduced additional
complexity in the color calculations, and which suggested the need for
refactoring.
## Detailed Description of the Pull Request / Additional comments
When we added support for `DECSCNM`, that required the foreground and
background color lookup methods to be able to return the opposite of
what was requested when the reversed mode was set. That made those
methods unnecessarily complicated, and I thought we could simplify them
considerably just by combining the calculations into a single method
that derived both colors at the same time.
And since both conhost and Windows Terminal needed to perform the same
calculations, it also made sense to move that functionality into the
`TextAttribute` class, where it could easily be shared.
In general this way of doing things is a bit more efficient. However, it
does result in some unnecessary work when only one of the colors is
required, as is the case for the gridline painter. So to make that less
of an issue, I've reordered the gridline code a bit so it at least
avoids looking up the colors when no gridlines are needed.
## Validation Steps Performed
Because of the API changes, quite a lot of the unit tests had to be
updated. For example instead of verifying colors with two separate calls
to `LookupForegroundColor` and `LookupBackgroundColor`, that's now
achieved with a single `LookupAttributeColors` call, comparing against a
pair of values. The specifics of the tests haven't changed though, and
they're all still working as expected.
I've also manually confirmed that the various color sequences and
rendition attributes are rendering correctly with the new refactoring.
Pretty straightforward. `toggleRetroEffect` will work to toggle the
retro terminal effect on/off.
* Made possible by contributions from #6551, _and viewers like you_
This PR reimplements the VT rendering engines to do a better job of
preserving the original color types when propagating attributes over
ConPTY. For the 16-color renderers it provides better support for
default colors and improves the efficiency of the color narrowing
conversions. It also fixes problems with the ordering of character
renditions that could result in attributes being dropped.
Originally the base renderer would calculate the RGB color values and
legacy/extended attributes up front, passing that data on to the active
engine's `UpdateDrawingBrushes` method. With this new implementation,
the renderer now just passes through the original `TextAttribute` along
with an `IRenderData` interface, and leaves it to the engines to extract
the information they need.
The GDI and DirectX engines now have to lookup the RGB colors themselves
(via simple `IRenderData` calls), but have no need for the other
attributes. The VT engines extract the information that they need from
the `TextAttribute`, instead of having to reverse engineer it from
`COLORREF`s.
The process for the 256-color Xterm engine starts with a check for
default colors. If both foreground and background are default, it
outputs a SGR 0 reset, and clears the `_lastTextAttribute` completely to
make sure any reset state is reapplied. With that out the way, the
foreground and background are updated (if changed) in one of 4 ways.
They can either be a default value (SGR 39 and 49), a 16-color index
(using ANSI or AIX sequences), a 256-color index, or a 24-bit RGB value
(both using SGR 38 and 48 sequences).
Then once the colors are accounted for, there is a separate step that
handles the character rendition attributes (bold, italics, underline,
etc.) This step must come _after_ the color sequences, in case a SGR
reset is required, which would otherwise have cleared any character
rendition attributes if it came last (which is what happened in the
original implementation).
The process for the 16-color engines is a little different. The target
client in this case (Windows telnet) is incapable of setting default
colors individually, so we need to output an SGR 0 reset if _either_
color has changed to default. With that out the way, we use the
`TextColor::GetLegacyIndex` method to obtain an approximate 16-color
index for each color, and apply the bold attribute by brightening the
foreground index (setting bit 8) if the color type permits that.
However, since Windows telnet only supports the 8 basic ANSI colors, the
best we can do for bright colors is to output an SGR 1 attribute to get
a bright foreground. There is nothing we can do about a bright
background, so after that we just have to drop the high bit from the
colors. If the resulting index values have changed from what they were
before, we then output ANSI 8-color SGR sequences to update them.
As with the 256-color engine, there is also a final step to handle the
character rendition attributes. But in this case, the only supported
attributes are underline and reversed video.
Since the VT engines no longer depend on the active color table and
default color values, there was quite a lot of code that could now be
removed. This included the `IDefaultColorProvider` interface and
implementations, the `Find(Nearest)TableIndex` functions, and also the
associated HLS conversion and difference calculations.
VALIDATION
Other than simple API parameter changes, the majority of updates
required in the unit tests were to correct assumptions about the way the
colors should be rendered, which were the source of the narrowing bugs
this PR was trying to fix. Like passing white on black to the
`UpdateDrawingBrushes` API, and expecting it to output the default `SGR
0` sequence, or passing an RGB color and expecting an indexed SGR
sequence.
In addition to that, I've added some VT renderer tests to make sure the
rendition attributes (bold, underline, etc) are correctly retained when
a default color update causes an `SGR 0` sequence to be generated (the
source of bug #3076). And I've extended the VT renderer color tests
(both 256-color and 16-color) to make sure we're covering all of the
different color types (default, RGB, and both forms of indexed colors).
I've also tried to manually verify that all of the test cases in the
linked bug reports (and their associated duplicates) are now fixed when
this PR is applied.
Closes#2661Closes#3076Closes#3717Closes#5384Closes#5864
This is only a partial fix for #293, but I suspect the remaining cases
are unfixable.
This saves an awful lot of construction/destruction and memory
allocation, especially with text that changes a lot (see: cacafire).
Three things:
1. Recycling the text layouts. This holds onto the `CustomTextLayout` so
all the things that don't change related to drawing targets and
whatnot aren't freed and recreated every frame.
2. Reordering the runs in place. This saves a vector
allocation/copy/delete every time OrderRuns is called. They can be
rearranged in place.
3. Only clip once per row. This reduces the clip push/pop to only one
time per row. Since we're always redrawing an entire row at a time,
this saves a lot of alloc/free of the clip frame, dramatically
reduces queued commands, and makes less work on the flush since
clipping requires staging the drawing and then bringing it back to
the main surface.
I was told that the DeviceContext version supercedes the RenderTarget
one. This moves us to it so we can gain access to a higher level of
control over the various pieces in our pipeline as we continue to evolve
the renderer.
The underlying motivation here is to potentially use a
`ID2D1CommandList` to batch our commands and run them all later outside
the lock. That can only really be done with this more granular level of
control over the pipeline. So this moves to that in a single step that
is easily findable in history should we have problems
I discussed this with @NiklasBorson of the Direct2D/DirectWrite team as
well as with @DHowett before doing it.
## Validation
- [x] Checked docs to make sure that these work on Windows 7 with
Platform Update
- [x] Manual smoke test real quick
- [ ] Try running on Win7 + Platform Update after change
- [x] Probably do more than just a smoke test manually or otherwise
Closes#6525
## Summary of the Pull Request
## PR Checklist
* [x] Closes#3927
* [x] I work here.
* [x] Tested manually.
* [x] Requires documentation to be updated: (generate doc bug here)
* [x] Am core contributor.
## Detailed Description of the Pull Request / Additional comments
- I found four settings that weren't hot reloadable with the 3927 comment above them:
1. Experimental retro terminal effect
2. Experimental software rendering
3. Experimental full repaint rendering
4. Antialiasing settings for text
I made them all hot reloadable by telling the `TermControl` to propagate them on settings change to the `DxEngine`.
Then I set up the `DxEngine` inside the setters to only set them if they changed. And if they do change, to trigger a full repaint and/or a complete drop and recreate of the entire DX device chain (as would happen if it were lost for another reason like a user-mode graphics failure, disconnected display, etc.)
I made the boolean an atomic because the settings can be coming in off of another thread (the XAML eventing one) and the renderer is picking the status up on its thread at the top of the BeginPaint frame.
## Validation Steps Performed
- [x] Opened it up and toggled all the settings while staring at PowerShell
- [x] Opened it up and toggled all the settings while staring at something intensive like a `cacafire` fire
This pull request reduces input lag, especially with selection, by using
`IDXGISwapChain2::GetFrameLatencyWaitableObject`.
This is based on the [DXGI 1.3 documentation].
Excerpt from the [DXGI 1.3 improvement list]:
> The following functionality has been added in Microsoft DirectX
Graphics Infrastructure (DXGI) 1.3, which is included starting in
Windows 8.1.
Before, during rendering:
1. render frame
2. call `Present` on swap chain:
2.a. blocks until it can present
2.b. meanwhile, selection/text in terminal might have changed, but
we're still using the frame that we rendered before blocking
2.c. presents
After, during rendering:
1. block until we can present
2. render frame with latest data
3. call `Present` on swap chain:
3.a. present without blocking
[DXGI 1.3 documentation]: https://docs.microsoft.com/en-us/windows/uwp/gaming/reduce-latency-with-dxgi-1-3-swap-chains
[DXGI 1.3 improvement list]: https://docs.microsoft.com/en-us/windows/win32/direct3ddxgi/dxgi-1-3-improvements:
## Summary of the Pull Request
This is the plan that @miniksa suggested to me. Instead of trying to do lots of work in all the renderers to do backgrounds as one pass, and foregrounds as another, we can localize this change to basically just the DX renderer.
1. First, we give the DX engine a "heads up" on where the cursor is going to be drawn during the frame, in `PrepareRenderInfo`.
- This function is left unimplemented in the other render engines.
2. While printing runs of text, the DX renderer will try to paint the cursor in `CustomTextRenderer::DrawGlyphRun` INSTEAD of `DxEngine::PaintCursor`. This lets us weave the cursor background between the text background and the text.
## References
* #6151 was a spec in this general area. I should probably go back and update it, and we should probably approve that first.
* #6193 is also right up in this mess
## PR Checklist
* [x] Closes#1203
* [x] I work here
* [ ] Tests added/passed
* [n/a] Requires documentation to be updated
## Detailed Description of the Pull Request / Additional comments
* This is essentially `"cursorTextColor": "textForeground"` from #6151.
* A follow up work item is needed to add support for the current behavior, (`"cursorTextColor": null`), and hooking up that setting to the renderer.
## Summary of the Pull Request
When resizing the window title, a GDI object would be leaked. This has to do with our island message handler using `wil` to track these objects and `wil` having a bug.
## References
microsoft/wil#100
## PR Checklist
* [x] Closes#5949
* [x] I work here.
* [x] Tested manually
* [x] Doc not required.
* [x] Am core contributor.
## Validation Steps Performed
* [x] Added the GDI Objects column to Task Manager, set the Terminal to use the `titleWidth` size tabs, then changed the title a bunch with PowerShell. Confirmed repro before (increasing GDI count). Confirmed it's gone after (no change to object count).
