## Summary of the Pull Request
This PR performs a number of miscellaneous clean-up tasks on our schema:
- [X] allow null for some settings
- [X] update some default values to be more useful
- [X] consistently use " instead of '
- [X] remove colorTable
- [X] provide missing description for `backgroundImageOpacity`
- [X] update cursorShape to have options match order displayed in vs code
- [X] if a setting has multiple options, provide a list of those options with a short description on each
- this one I only did a few times
- [X] remove a little bit of ambiguity on some descriptions
## PR Checklist
Closes#5279
## Detailed Description of the Pull Request / Additional comments
Validation was performed for...
- [X] Global Settings
- [X] Profile Settings
- [x] Keybinding Actions
## Validation Steps Performed
I linked VS Code to the my version of the schema. Defined all settings from scratch in settings.json. Anytime I found something off, I updated the schema and saw if it looks right.
This PR fixes a couple of issues with TSFInputControl alignment:
1. The emoji picker IME in particular would show up overlapping the
current buffer row because I stupidly didn't realize that
`TextBlock.ActualHeight` is 0 right after initialization and before
it has any text. So, the emoji picker will show up on the bottom of
the 0 height TextBlock, which overlaps the current buffer row. This
isn't a problem with CJK because inputting text _causes_ the IME to
show up, so by the time the IME shows up, the TextBlock has text and
accordingly has a height.
2. It turns out the emoji picker IME doesn't follow the `TextBlock`
bottom, unlike Chinese and Japanese IME, so if a user were to compose
near the edge of the screen and let the `TextBlock` start line
wrapping, the emoji IME doesn't follow the bottom of the `TextBlock`
as it grows. This means that the `TextBlock` position doesn't update
in the middle of composition, and the `LayoutRequested` event that it
fires at the beginning of composition is the only chance we get to
tell the emoji IME where to place itself. It turns out when we reset
`TextBlock.Text`, the ActualHeight doesn't get immediately reset back
to the min size. So if a user were to bring up the emoji IME before
`ActualHeight` is reset, the IME will show up way below the current
buffer row.
3. We don't currently `TryRedrawCanvas` when the window position changes
(resizing, dragging the window around), so sometimes dragging or
resizing the Terminal doesn't update the position of the IME. Ideally
it should be listening to some "window position changed" event, but
alas we don't have that and it would be much harder than this
incoming fix. We'll just track the window bounds as part of
`TryRedrawCanvas` and redraw if it changes. For the most part, this
will allow the IME to update to where the new window position is, but
it'll only be called if we receive a `LayoutRequested` event.
## PR Checklist
* [x] Closes#5470
* [x] CLA signed.
* [x] Tests added/passed
## Validation Steps Performed
I play with it for quite a bit.
<!-- 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
A tiny performance fix in `renderer.cpp`.
<!-- 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
* [ ] Closes #xxx
* [X] CLA signed. If not, go over [here](https://cla.opensource.microsoft.com/microsoft/Terminal) and sign the CLA
* [ ] Tests added/passed
* [ ] Requires documentation to be 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
The cluster construction code is intensively called during rendering. Even though a single `back()` is fast, but accumulated `back()`s still take a noticiable amount of CPU cycles.
Before:
After:
<!-- Describe how you validated the behavior. Add automated tests wherever possible, but list manual validation steps taken as well -->
## Validation Steps Performed
Manually validated.
This property was deprecated in 0.11. We probably should have also added a warning
message to help the community figure out that this property is gone and won't work
anymore.
This PR adds that warning.
* I'm not going to list the enormous number of duped threads _wait yes I am_
* #5581
* #5547
* #5555
* #5557
* #5573
* #5532
* #5527
* #5535
* #5510
* #5511
* #5512
* #5513
* #5516
* #5515
* #5521
* This literally isn't even all of them
* [x] Also mainly related to #5458
* [x] I work here
* [x] Tests added/passed
Followup to ea61aa3b.
The default foreground in the iTerm2 defaults for the Tango Dark color
scheme is too bright, use the value for ANSI 7 (white) instead.
References #5305
Sorry, I should have really done this in the original PR.
Signed-off-by: Rafael Kitover <rkitover@gmail.com>
<!-- 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
Simple fix to update the example documentation with the recent changes to 0.11.
Also known as "Kill HRGN II: Kills Regions Dead (#5485)"
Copying the description from @greg904 in #4778.
--- 8< ---
My understanding is that the XAML framework uses another way of getting
mouse input that doesn't work with `WM_SYSCOMMAND` with `SC_MOVE`. It
looks like it "steals" our mouse messages like `WM_LBUTTONDOWN`.
Before, we were cutting (with `HRGN`s) the drag bar part of the XAML
islands window in order to catch mouse messages and be able to implement
the drag bar that can move the window. However this "cut" doesn't only
apply to input (mouse messages) but also to the graphics so we had to
paint behind with the same color as the drag bar using GDI to hide the
fact that we were cutting the window.
The main issue with this is that we have to replicate exactly the
rendering on the XAML drag bar using GDI and this is bad because:
1. it's hard to keep track of the right color: if a dialog is open, it
will cover the whole window including the drag bar with a transparent
white layer and it's hard to keep track of those things.
2. we can't do acrylic with GDI
So I found another method, which is to instead put a "drag window"
exactly where the drag bar is, but on top of the XAML islands window (in
Z order). I've found that this lets us receive the `WM_LBUTTONDOWN`
messages.
--- >8 ---
Dustin's notes: I've based this on the implementation of the input sink
window in the UWP application frame host.
Tested manually in all configurations (debug, release) with snap,
drag, move, double-click and double-click on the resize handle. Tested
at 200% scale.
Closes#4744Closes#2100Closes#4778 (superseded.)
Takes the lock inside two routines in `TermControl` that were changing
the selection endpoint while a rendering frame was still drawing,
resulting in several variants of graphical glitches from double-struck
selection boxes to duplicated line text.
## References
- Introduced with #5185
## PR Checklist
* [x] Closes#5471
* [x] Already signed life away to company.
* [x] Manual tests passed since it's visual.
* [x] No extra doc besides the comments.
* [x] Am core contributor: Roar.
The renderer base and specific renderer engine do a lot of work to
remember the previous selection and compensate for scrolling regions and
deltas between frames. However, all that work doesn't quite match up
when the endpoints are changed out from under it. Unfortunately,
`TermControl` doesn't have a robust history of locking correctly in step
with the renderer nor does the renderer's `IRenderData` currently
provide any way of 'snapping' state at the beginning of a frame so it
could work without a full lock. So the solution for now is for the
methods that scroll the display in `TermControl` to take the lock that
is shared with the renderer's frame painter so they can't change out of
sync.
## Validation Steps Performed
- Opened terminal with Powershell core.
Did ls a bunch of times.
Clicked to make selection and held mouse button while wheeling around.
- Opened terminal with Powershell core.;
Did ls a bunch of times.
Clicked to make selection and dragged mouse outside the window to make
auto scroll happen.
- Opened terminal with Powershell core.
Did ls a bunch of times.
Clicked to make selection and released. Wheeled around like a crazy
person to make sure I didn't regress that.
Literally just <kbd>ctrl+f</kbd> find-and-replace all the old `profiles.json` that are sitting around in the repo with `settings.json`. I didn't touch the specs, since it seemed better to leave them in the state that they were originally authored in.
* [x] closes#5522
* [x] I work here.
* [x] This is docs.
These were already in the `SettingsSchema.md`, so I just updated the `profiles.schema.json` to have the descriptions as well.
* [x] closes#5520
* [x] I work here.
* [x] This is docs.
It was brought to our attention that shipping a font with ligatures as our default
font could be an accessibility issue for the visually-impaired. Unfortunately, we
don't have a renderer setting to disable ligatures (#759). Fortunately however, we
DO already have a version of Cascadia that doesn't have ligatures.
If we ship that and set it as our default font, we'll at least let people _opt_ to
have ligatures enabled by switching from `Cascadia Mono` to `Cascadia Code`.
## PR Checklist
* [x] Closes internal discussion
* [x] CLA signed
* [ ] Tests added/passed
* [x] Requires documentation to be updated
* [x] I've discussed this with core contributors already.
In order to support a transparent background for the acrylic effect, the
renderer sets the alpha value to zero for the default background color.
However, when the _reversed video_ attribute is set, the background is
actually filled with the foreground color, and will not be displayed
correctly if it is made transparent. This PR addresses that issue by
making sure the rendered background color is opaque if the reversed
video attribute is set.
## References
* This is not a major issue at the moment, since the _reverse video_
attribute is not typically forwarded though conpty, but that will
change once #2661 is fixed.
## PR Checklist
* [x] Closes#5498
* [x] CLA signed.
* [ ] Tests added/passed
* [ ] Requires documentation to be 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: #5498
## Detailed Description of the Pull Request / Additional comments
This simply adds an additional check in `Terminal::GetBackgroundColor`
to make sure the returned color is opaque if the _reverse video_
attribute is set. At some point in the future this check may need to be
extended to support the `DECSCNM` reverse screen mode, but for now
that's not an issue.
## Validation Steps Performed
I've run the test case from issue #5498, and confirmed that it now works
as expected. I've also got an experimental fix for #2661 that I've
tested with this patch, and that now displays _reverse video_ attributes
correctly too.
Closes#5498
If Terminal is spawned by a shortcut that requests that it run in a new process group
while attached to a console session, that request is nonsense. That request will, however,
cause WT to start with Ctrl-C disabled. This wouldn't matter, because it's a Windows-subsystem
application. Unfortunately, that state is heritable. In short, if you start WT using cmd in
a weird way, ^C stops working inside the terminal. Mad.
Fixes#5460.
## Summary of the Pull Request
- Adjusts scaling practices in `DxEngine` (and related scaling practices in `TerminalControl`) for pixel-perfect row baselines and spacing at High DPI such that differential row-by-row rendering can be applied at High DPI.
## References
- #5185
## PR Checklist
* [x] Closes#5320, closes#3515, closes#1064
* [x] I work here.
* [x] Manually tested.
* [x] No doc.
* [x] Am core contributor. Also discussed with some of them already via Teams.
## Detailed Description of the Pull Request / Additional comments
**WAS:**
- We were using implicit DPI scaling on the `ID2D1RenderTarget` and running all of our processing in DIPs (Device-Independent Pixels). That's all well and good for getting things bootstrapped quickly, but it leaves the actual scaling of the draw commands up to the discretion of the rendering target.
- When we don't get to explicitly choose exactly how many pixels tall/wide and our X/Y placement perfectly, the nature of floating point multiplication and division required to do the presentation can cause us to drift off slightly out of our control depending on what the final display resolution actually is.
- Differential drawing cannot work unless we can know the exact integer pixels that need to be copied/moved/preserved/replaced between frames to give to the `IDXGISwapChain1::Present1` method. If things spill into fractional pixels or the sizes of rows/columns vary as they are rounded up and down implicitly, then we cannot do the differential rendering.
**NOW:**
- When deciding on a font, the `DxEngine` will take the scale factor into account and adjust the proposed height of the requested font. Then the remainder of the existing code that adjusts the baseline and integer-ifies each character cell will run naturally from there. That code already works correctly to align the height at normal DPI and scale out the font heights and advances to take an exact integer of pixels.
- `TermControl` has to use the scale now, in some places, and stop scaling in other places. This has to do with how the target's nature used to be implicit and is now explicit. For instance, determining where the cursor click hits must be scaled now. And determining the pixel size of the display canvas must no longer be scaled.
- `DxEngine` will no longer attempt to scale the invalid regions per my attempts in #5185 because the cell size is scaled. So it should work the same as at 96 DPI.
- The block is removed from the `DxEngine` that was causing a full invalidate on every frame at High DPI.
- A TODO was removed from `TermControl` that was invalidating everything when the DPI changed because the underlying renderer will already do that.
## Validation Steps Performed
* [x] Check at 150% DPI. Print text, scroll text down and up, do selection.
* [x] Check at 100% DPI. Print text, scroll text down and up, do selection.
* [x] Span two different DPI monitors and drag between them.
* [x] Giant pile of tests in https://github.com/microsoft/terminal/pull/5345#issuecomment-614127648
Co-authored-by: Dustin Howett <duhowett@microsoft.com>
Co-authored-by: Mike Griese <migrie@microsoft.com>
## Summary of the Pull Request
Users were not able to intercept Ctrl-C input using `$Host.UI.RawUI.ReadKey("IncludeKeyUp")`, because we weren't sending a Ctrl-C KeyUp event. This PR simply adds a KeyUp event alongside the existing KeyDown.
<!-- Please review the items on the PR checklist before submitting-->
## PR Checklist
* [x] Closes#1894
* [x] CLA signed.
* [x] Tests added/passed
<!-- Describe how you validated the behavior. Add automated tests wherever possible, but list manual validation steps taken as well -->
## Validation Steps Performed
The repro script in #1894 now works, both options for `ReadKey`: `IncludeKeyUp` and `IncludeKeyDown` work fine.
This PR adds a test for #5428. Mysteriously, after #5398 merged, 5428 went away. However, I already wrote this test for it, so we might as well add it to our collection.
* [x] Closes#5428
* [x] I work here
* [x] Is a test
- build: move oss required to build conhost out of dep/
This change is necessary as the dep/ folder is not synced into the
Windows source tree.
I've also added a build rule producing a lib for {fmt}.
This will be required for our next OS ingestion.
Related work items: #26069643
This change is necessary as the dep/ folder is not synced into the
Windows source tree.
I've also added a build rule producing a lib for {fmt}.
This will be required for our next OS ingestion.
* Cleaning up the whitelist a bit.
* The magic to exclude repeated characters worked 👍
* Every successful run on master now logs its suggested cleanup, e.g. for 5740e197c2 has https://github.com/microsoft/terminal/runs/596271627#step:4:37
* ⚠️ This check-spelling 0.0.15a+ tolerates Windows line endings in the `whitelist.txt` file (another project I touched had some `.gitconfig` magic which required supporting them).
This means that if someone edits the file w/ something that likes Windows line endings, the file will successfully convert (instead of it being ignored and check-spelling complaining about everything). Most likely anyone else who then edits the file will use something that will maintain the line endings.
Improve wide glyph support in UIA (GH-4946)
Add enhanced key support for ConPty (GH-5021)
Set DxRenderer non-text alias mode (GH-5149)
Reduce CursorChanged Events for Accessibility (GH-5196)
Add more object ID tracing for Accessibility (GH-5215)
Add SS3 cursor key encoding to ConPty (GH-5383)
UIA: Prevent crash from invalid UTR endpoint comparison (GH-5399)
Make CodepointWidthDetector::GetWidth faster (CC-3727)
add til::math, use it for float conversions to point, size (GH-5150)
Add support for renderer backoff, don't FAIL_FAST on 3x failures, add UI (GH-5353)
Fix a deadlock and a bounding rects issue in UIA (GH-5385)
Don't duplicate spaces from potentially-wrapped EOL-deferred lines (GH-5398)
Reimplement the VT tab stop functionality (CC-5173)
Clamp parameter values to a maximum of 32767. (CC-5200)
Prevent the cursor type being reset when changing the visibility (CC-5251)
Make RIS switch back to the main buffer (CC-5248)
Add support for the DSR-OS operating status report (CC-5300)
Update the virtual bottom location if the cursor moves below it (CC-5317)
ci: run spell check in CI, fix remaining issues (CC-4799) (CC-5352)
Set Cascadia Code as default font (GH-5121)
Show a double width cursor for double width characters (GH-5319)
Delegate all character input to the character event handler (CC-4192)
Update til::bitmap to use dynamic_bitset<> + libpopcnt (GH-5092)
Merged PR 4465022: [Git2Git] Merged PR 4464559: Console: Ingest OSS changes up to e0550798
Correct scrolling invalidation region for tmux in pty w/ bitmap (GH-5122)
Render row-by-row instead of invalidating entire screen (GH-5185)
Make conechokey use ReadConsoleInputW by default (GH-5148)
Manually pass mouse wheel messages to TermControls (GH-5131)
This fixes C-M-space for WSL but not for Win32, but I'm not sure there's a problem in Win32 quite yet. (GH-5208)
Fix copying wrapped lines by implementing better scrolling (GH-5181)
Emit lines wrapped due to spaces at the end correctly (GH-5294)
Remove unneeded whitespace (CC-5162)
[Git2Git] Git Train: Merge of building/rs_onecore_dep_uxp/200414-1630 into official/rs_onecore_dep_uxp Retrieved from https://microsoft.visualstudio.com os OS official/rs_onecore_dep_uxp 90031afa8114b7d95e3993573fc8449a1524a7fd
Related work items: #25439646
The "Campbell Powershell" color scheme does not have a high enough color contrast
ratio. Campbell does, so we're changing it there.
Closes (only upon validation) #5393.
## Summary of the Pull Request
This is a quick-and-easy solution to #5309. If the ITextRangeProvider API allows us to take in two UiaTextRanges, we need to verify that they are both valid.
With this PR, we make sure they both fit in the current TextBuffer. If not, we return `E_FAIL`. Though this doesn't prove that both UiaTextRanges are from the same TextBuffer, at the very least we don't crash and in cases where we can't make a valid comparison, we return an HRESULT failure.
## References
#5406 - This should be the proper solution to this problem. Each UiaTextRange needs to be aware of which TextBuffer it came from.
## PR Checklist
* [X] Closes#5309
## Validation Steps Performed
1. generate enough output to cause the terminal to scroll
2. execute `nano` to make us go into the alternate buffer
This previously crashed, now NVDA seems to detect that there was an error and keeps moving along.
The logic here, regarding deleting the spaces and just instantly adding
them bad, is incredibly suspect. Given that we're close to 0.11, I don't
think I can change it.
I've added a TODO with an issue number to figure out the right logic
here.
Fixes#5386.
## PR Checklist
* [x] Closes#5386
* [x] CLA signed
* [x] Tests added/passed
* [ ] Requires documentation to be updated
* [x] I'm horrified.
## Validation Steps Performed
Tests, manual validation of the scenario in 5386 and a repro program.
This commit adds a specific error message to the build that tells people
to restore git submodules if they forgot to read the README.
#5416 was the straw that broke the camel's back.
Selection would act up when you were using shift to ignore VT mouse
mode: we would get hundreds of WM_KEYDOWN for VK_SHIFT and dismiss the
selection every time.
I took the opportunity to move the actual responsibility for key event
dispatch into HwndTerminal. In the future, I'd like to make more of the
TerminalXxx calls just call impl methods on HwndTerminal.
This commit adds a `WT_PROFILE_ID` environment variable, which contains
the guid of the active profile.
It also teaches ConptyConnection to take an environment map on creation.
We had to do a little manual jiggery with the WSLENV environment
variable as passed by the creator.
* [x] CLA signed
* [ ] Tests added/passed
* [ ] Requires documentation to be updated
* [x] I've discussed this with core contributors already.
Ran terminal, validated vars and translated paths under windows and WSL.
References #4566 (this PR originally introduced WT_SETTINGS/DEFAULTS)
Closes#3589
The scroll locking rework that landed with the DxRenderer's partial
invalidation change introduced a deadlock. UIA locks the buffer for
reading before asking it to scroll (which now requires a write lock.)
Scrolling is probably _okay_ to have a little bit of torn state that
might arise from us unlocking the read lock early before triggering the
write lock down the line.
While investigating this, I also noticed that our bounding rects stopped
being viewport-relative (and were instead buffer-relative.) That
regressed with the switch to `GetTextRects` in #4991
## PR Checklist
* [ ] Closes (issues noticed in investigating the DPI changes)
* [x] CLA signed.
* [ ] Tests added/passed
* [ ] Requires documentation to be updated
* [x] Core contributor badge
## Validation Steps Performed
Manual pass with inspect.exe
## Summary of the Pull Request
Adds SS3 cursor encoding for cursor keys and home/end button. Reverts a portion of #4913 that checks for VT Input Mode.
## PR Checklist
* [X] Closes#4873
## Validation Steps Performed
1. Open pwsh
2. run `wsl`
3. execute `printf "\e[?1h"`
4. verify keys work
5. exit back to pwsh
6. verify keys work still (didn't previously)
Also verified that those keys work in vim when connected to my Raspberry Pi over SSH.
## Summary of the Pull Request
This pull request ports the VT mouse code from TermControl to WpfTerminalControl. Our WPF control is a lot closer to Win32 than to Xaml, so our mouse event handler looks _nothing_ like the one that we got from Xaml. We can pass events through almost directly, because the window message handling in the mouse input code actually came from _conhost_. It's awesome.
Neither TermControl nor conhost pass hover events through when the control isn't focused, so I wired up focus events to make sure we acted the same.
Just like Terminal and conhost, mouse events are suppressed when <kbd>Shift</kbd> is held.
## Validation Steps Performed
Tested with MC, and tested by manually engaging SGR events in an Echo terminal.
## Summary of the Pull Request
This pull request ports #5096 to WpfTerminalControl, bringing it in line with the selection mechanics in Terminal. It also introduces double- and triple-click selection and makes sure we clear the selection when we resize.
Please read #5096 for more details.
## Detailed Description of the Pull Request / Additional comments
This code is, largely, copy-and-pasted from TermControl with some updates to use `std::chrono` and `til::point`. I love `til::point`. A lot.
## Validation Steps Performed
Lots of manual selection.
## Summary of the Pull Request
This pull request fixes a crash on scrolling down (overflow exception cramming the signed short upper WORD of the wParam into an actual signed short on x64) and a crash on key input caused by improper use of `Marshal.ReadByte` on an integer (instead of a memory address).
## Detailed Description of the Pull Request / Additional comments
## Validation Steps Performed
Manually did those things on x64.
This commit fixes a number of problems and code quality/health issues
with the AzureConnection.
This is a general tidying-up of the azure connection. It improves error
logging (like: it actually emits error logs...) and retry logic and the
state machine and it audits the exit points of the state machine for
exceptions and removes the HRESULT returns (so they either succeed and
transition to a new state or throw an exception or are going down
anyway).
There's also a change in here that changes how we display tenants. It
adds the "default domain" to the name, so that instead of seeing this:
Conhost (aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa)
Default Directory (bbbbbbbb-bbbb-bbbb-bbbb-bbbbbbbbbbbb)
you see this
Conhost (conhost.onmicrosoft.com)
Default Directory (dustinhowett.onmicrosoft.com)
Changes:
* rework tenant/tenant storage and fix display names
Switch to the 2020 tenant API.
Instead of passing around four loose variables, create a Tenant class
and use that for packing/unpacking into/out of json (and the windows
credential store, where we "cleverly" used json for the tenant info
there too).
When displaying a tenant, use its display name if there is one, the
unknown resource string if there isn't, and the default domain if
there is one and the ID if there isn't.
Fixes#5325.
* use {fmt} for formatting request bodies
* remove dead strings
* rework/rename Request/HeaderHelper to
Send(Authenticated)ReqReturningJson
* rewrite polling to use std::chrono
* remove HR returns from state machine
* rename state handlers from _XHelper to _RunXState
* cleanup namespaces, prefix user input with >, remove namespaces
* Rework error handling
- _RequestHelper no longer eats exceptions.
- Delete the "no internet" error message.
- Wrap exceptions coming out of Azure API in a well-known type.
- Catch by type.
- Extract error codes for known failures (keep polling, invalid
grant).
- When we get an Invalid Grant, dispose of the cached refresh token
and force the user to log in again.
- Catch all printable exceptions and print them.
- Remove the NoConnect state completely -- just bail out when an
exception hits the toplevel of the output thread.
- Move 3x logic into _RefreshTokens and pop exceptions out of it.
- Begin abstracting into AzureClient
Fixes#5325 (by addressing its chief complaint).
Fixes#4803 (by triggering auth flow again if the token expires).
Improves diagnosability for #4575.
# Summary of the Pull Request
This PR will allow the cursor to be double width when on top of a double width character. This required changing `IsCursorDoubleWidth` to check whether the glyph the cursor's on top of is double width. This code is exactly the same as the original PR that addressed this issue in #2932. That one got reverted at some point due to the crashes related to it, but due to a combination of Terminal having come further since that PR and other changes to address use-after-frees, some of the crashes may/may not be relevant now. The ones that seemed to be relevant/repro-able, I attempt to address in this PR.
The `IsCursorDoubleWidth` check would fail during the `TextBuffer::Reflow` call inside of `Terminal::UserResize` occasionally, particularly when `newCursor.EndDeferDrawing()` is called. This is because when we tell the newCursor to `EndDefer`, the renderer will attempt to redraw the cursor. As part of this redraw, it'll ask if `IsCursorDoubleWidth`, and if the renderer managed to ask this before `UserResize` swapped out the old buffer with the new one from `Reflow`, the renderer will be asking the old buffer if its out-of-bounds cursor is double width. This was pretty easily repro'd using `cmatrix -u0` and resizing the window like a madman.
As a solution, I've moved the Start/End DeferDrawing calls out of `Reflow` and into `UserResize`. This way, I can "clamp" the portion of the code where the newBuffer is getting created and reflowed and swapped into the Terminal buffer, and only allow the renderer to draw once the swap is done. This also means that ConHost's `ResizeWithReflow` needed to change slightly.
In addition, I've added a WriteLock to `SetCursorOn`. It was mentioned as a fix for a crash in #2965 (although I can't repro), and I also figured it would be good to try to emulate where ConHost locks with regards to Cursor operations, and this seemed to be one that we were missing.
## PR Checklist
* [x] Closes#2713
* [x] CLA signed
* [x] Tests added/passed
## Validation Steps Performed
Manual validation that the cursor is indeed chonky, added a test case to check that we are correctly saying that the cursor is double width (not too sure if I put it in the right place). Also open to other test case ideas and thoughts on what else I should be careful for since I am quite nervous about what other crashes might occur.
## Summary of the Pull Request
When WSL vim prints the initial empty buffer (the one that's just a bunch of '\~'s), it prints this by doing the following:
* Print '\~' followed by enough spaces to clear the line
* Use CUP (`^[[H`) to move the cursor to the start of the next line
* repeat until the buffer is full
When we'd get the line of "\~ "... in conhost, we'd mark that line as wrapped.
Logically, it doesn't really make any sense that when we follow that up by moving the cursor, the line is wrapped. However, this is just how conhost is right now.
This wasn't ever a problem in just conhost before, because we really didn't care if lines in the alt buffer were "wrapped" or not. Plus, when vim would get resized, it would just reprint it's own buffer anyways. Nor was this a problem in conpty before this year (2020). We've only just recently added logic to conpty to try and preserve wrapped lines.
Initially, I tried fixing this by breaking the line manually when the cursor was moved. This seemed to work great, except for the win32 vim.exe. Vim.exe doesn't emit a newline or a CUP to get to the next line. It just _goes for it_ and keeps printing. So there's _no way_ for us to know the line broke, because they're essentially just printing one long line, assuming we'll automatically move the cursor.
So instead, I'm making sure to emit the proper number of spaces at the end of a line when the line is wrapped. We won't do any funny business in that scenario and try to optimize for them, we'll _just print the spaces_.
## References
* #5181 - This change regressed this
* #4415 - Actually implemented wrapped lines in conpty
## PR Checklist
* [x] Closes#5291
* [x] I work here
* [x] Tests added/passed
* [n/a] Requires documentation to be updated
## Validation Steps Performed
* Wrote a unittest first and foremost
* Checked vtpipeterm to make sure vim still works
* checked Terminal to make sure vim still works
Loc issues are given to us through the internal bug tracker.
* Lock some strings, or parts of strings, that should not be localized.
* Switch to positional format parameters
* Remove the forced newlines in the warning resources; insert them at
runtime
Fixes MSFT:25936156.
If an application writes to the screen while not in VT mode, and the
user has scrolled forward in the screen buffer, the _virtual bottom_
location is not updated to take that new content into account. As a
result, the viewport can later jump back to the previous _virtual
bottom_, making the content disappear off screen. This PR attempts to
fix that issue by updating the _virtual bottom_ location whenever the
cursor moves below that point.
## PR Checklist
* [x] CLA signed.
* [x] Tests added/passed
## Detailed Description of the Pull Request / Additional comments
This simply adds a condition in the
`SCREEN_INFORMATION::SetCursorPosition` to check if the new _Y_
coordinate is below the current _virtual bottom_, and if so, updates the
_virtual bottom_ to that new value.
I considered trying to make it only update when something is actually
written to the screen, but this seemed like a cleaner solution, and is
less likely to miss out on a needed update.
## Validation Steps Performed
I've manually tested the case described in issue #5302, and confirmed
that it now works as expected. I've also added a unit test that checks
the virtual bottom is updated correctly under similar conditions.
Closes#5302
