Closes#7877.
Small disclaimer: First Ghostty PR and Zig PR, but looking forward to
contributing to the project.
This PR supports shell-integration-features `ssh-env` and `ssh-terminfo`
as per other shells, but not the rest as this is what the issue states.
That being said, with this PR, then you would see this:
- `warning(io_exec): shell could not be detected, no automatic shell
integration will be injected`, but given that the default mode is
`detect` it will pick up the executable and if ssh features are enabled
it will integrate it. This might be confusing for users.
- I decided to not add `nu` to `pub const Shell` because if we do so,
then from what I understand from the code, then the code flow would
imply that "shell integration will be injected" but it will only do so
if those `ssh-*` features are enabled, which may be misleading. But on
the other hand, providing `ssh` shell integration but returning `null`
for `?!ShellIntegration` does not seem very correct either.
- I dont like that I added `features` argument to `setupshell`, just to
check them if `nu` was used. The reasoning is because the way Nushell
works, if we autoload the `nushell` directory (by `setupXdgDatadirs()`)
even if no `ssh` features were present, it will wrap the `ssh` function
and I think that is not desirable, even if we end up just forwarding the
arguments.
Sorry for the long wall of text, but I think it was worth to add some of
the doubts I have had myself, plus the ones that you folks may add. I am
very happy to iterate on this, even if its a very "Easy" one, so I much
welcome the feedback.
> [!NOTE]
>
> Used `GPT` for helping with nushell variable naming
verification/improvement
> Used `Gemini` for helping with understanding the `Zsh` ssh integration
so that I could replicate the logic with nushell. Just because I find
`zsh` language very difficult to understand in detail.
This changes our OSC133 parser to parse options lazily. We do this for
multiple reasons:
1. Parsing all our options ahead of time balloons our required
osc.Command tagged union type which has C ABI implications. Adding
all supported options (including Kitty extensions) today already
breaks our C ABI.
2. Invalid options are allowed by the specification and should be
explicitly ignored, so we don't need to validate options at all
during parse time.
3. Semantic prompt markers don't need to be high throughput, so we
can afford to do some extra work at processing time to gather
the options. They're also rather short usually.
Related to #5932
This updates our OSC parser to parse the full OSC 133 specification:
https://gitlab.freedesktop.org/Per_Bothner/specifications/blob/master/proposals/semantic-prompts.md
The logic for handling these events was _unchanged_ from our prior
implementation. This is just a parser-only update. As such, we ignore a
bunch of semantic prompt command we should definitely handle, and
incorrectly handle others. This is the crux of #5932 that I want to head
towards fixing. This PR just contains the parser updates.
I also retained all the Kitty parser extensions.
**AI disclosure:** AI helped a lot of the rote tasks once I manually did
a few. I'm still reviewing this manually but will do so shortly.
Refer to discussion #10000
When a tab contains only a single split, resize_split and
toggle_split_zoom actions now return false (not performed). This allows
keybindings marked with `performable: true` to pass the event through to
the terminal program.
The performable flag causes unperformed actions to be treated as if the
binding didn't exist, so the key event is sent to the terminal instead
of being consumed.
- Add isSplit() helper to SplitTree to detect single-pane vs split state
- Update GTK resizeSplit/toggleSplitZoom to return false when single
pane
- Update macOS resizeSplit/toggleSplitZoom to return Bool and check
isSplit
- Add unit test for isSplit method
This adds a new single-file library called "Tripwire" in
`src/tripwire.zig`. This library helps inject failures around `try`
cases for the purpose of testing `errdefer`. It is fully optimized away
in non-test builds (even debug), turning into zero space and zero
assembly.
From this, I've verified (via unit tests w/ tripwire) and fixed a number
of errdefer issues:
* PageList init with non-standard pages that requires more than 1 page
can leak on allocation error on the 2nd+ loop
* Tabstop allocation failure on resize corrupts the internal state
(invalid cols)
* `Screen.selectionString` would leak memory on late allocation failures
* Screen search could leak memory on late allocation failures
* `SharedGrid.renderGlyph` in our font subsystem would corrupt the glyph
cache if failure occurred
* `SharedGrid.init` could leak memory if loading font metrics failed
In addition to the bugs found, there is now tripwire coverage around
more of our core and we should continue to add more. I've also added
significantly more explicit error sets as I found them.
**AI disclosure:** AI wrote some of the tests, but tripwire itself is
all handwritten and everything was reviewed.
Add errdefer cleanup for codepoints and glyphs hash maps in init().
Previously, if ensureTotalCapacity or reloadMetrics() failed after
allocating these maps, they would leak.
Add tripwire test to verify all failure points in init().
Add errdefer to remove cache entry after getOrPut if subsequent
operations fail (getPresentation, atlas.grow, renderGlyph). Without
this, failed renders would leave uninitialized/garbage entries in
the glyph cache, potentially causing crashes or incorrect rendering.
Add tripwire test to verify the rollback behavior.
