This fixes the VS16 issues found in this test:
https://ucs-detect.readthedocs.io/sw_results/ghostty.html#ghostty
This is also a more robust way to handle VS15/16 in general.
This commit also changes our propeties to be a packed struct which
reduces its size from 4 bytes to 1 and likewise drops our unicode table
size 4x.
This updates uucode. As part of this, the wcwidth implementation was
updated (in uucode) to make emoji modifiers width ZERO. But if they're
standalone, we want them as width 2.
So this also contains a change to force them as width 2 for our width
calculation. This only matters for standalone emoji modifiers, because
when they form a valid grapheme we don't use this width calculation.
This adds a new stream handler implementation that updates terminal
state in reaction to VT sequences, but doesn't perform any of the
actions that would require responses (e.g. queries).
This is exposed in two ways: first, as a standalone `ReadonlyStream` and
`ReadonlyHandler` type that contains all the implementation. Second, as
a convenience func on `Terminal` as `vtStream` and `vtHandler` which
return their respective types preconfigured to update the calling
terminal state.
This dramatically simplifies libghostty-vt usage from Zig (and will
eventually be exposed to C, too) since a Terminal on its own is ready to
go as a full VT parser and state machine without needing to build any
custom types!
There's a second big bonus here which is that our `stream_readonly.zig`
tests are true end-to-end tests for raw bytes to terminal state. This
will let us test a wider variety of situations more broadly. To start,
there are only a handful of tests implemented here.
**AI disclosure:** Amp wrote basically this whole thing, but I reviewed
it. https://ampcode.com/threads/T-3490efd2-1137-4112-96f6-4bf8a0141ff5
This makes it cleaner to add new sources of table generation and also
avoids inadvertently depending on different modules (despite Zig's lazy
analysis).
This also fixes up terminal to only use our look up tables which avoids
bringing ziglyph in for the terminal module.
Without this change, a phantom space appears after any character with
default emoji presentation that is converted to text with VS15. The only
other terminal I know of that respects variation selectors is Kitty, and
it walks the cursor back, which feels like the best choice, since that
way the behavior is observable (no way to know if the terminal supports
variation selectors otherwise without hard-coding that info per term)
and "dumb" programs like `cat` will output things correctly, and not
gain a phantom space after any VS15'd emoji.
This brings the behavior of mode 47, 1047, and 1049 much closer to
xterm's behavior. I found that our prior implementation had many
deficiencies.
For example, we weren't properly copying the cursor state back to the
primary screen from the alternate screen for modes 47 and 1047. And we
weren't saving/restoring cursor state unconditionally for mode 1049 even
if we were already in the alternate screen.
These are weird, edgy behaviors that I don't think anyone expected
(evidence by there being no bug reports about them), but they are bugs
nontheless.
Many tests added.
As of Zig 0.14.0, `@splat` can be used for array types, which eliminates
a lot of redundant syntax and makes things generally cleaner.
I've explicitly avoided applying this change in the renderer files for
now since it would just create rebasing conflicts in my renderer rework
branch which I'll be PR-ing pretty soon.
This commit changes a LOT of areas of the code to use decl literals
instead of redundantly referring to the type.
These changes were mostly driven by some regex searches and then manual
adjustment on a case-by-case basis.
I almost certainly missed quite a few places where decl literals could
be used, but this is a good first step in converting things, and other
instances can be addressed when they're discovered.
I tested GLFW+Metal and building the framework on macOS and tested a GTK
build on Linux, so I'm 99% sure I didn't introduce any syntax errors or
other problems with this. (fingers crossed)
Fixes#5718
When a terminal is resized with text reflow (i.e. soft-wrapped text), the cursor
is generally reflowed with it.
For example, imagine a terminal window 5-columns wide and you type the
following without pressing enter. The cursor is on the X.
```
OOOOO
OOX
```
If you resize the window now to 8 or more columns, this happens, as expected:
```
OOOOOOOX
```
As expected, the cursor remains on the "X". This behaves like any other text
input...
Terminals also provide an escape sequence to
[save the cursor (ESC 7 aka DECSC)](https://ghostty.org/docs/vt/esc/decsc).
This includes, amongst other things, the cursor position. The cursor can be
restored with [DECRC](https://ghostty.org/docs/vt/esc/decrc).
The behavior of the position of the _saved cursor_ in the context of text
reflow is unspecified and varies wildly between terminals Ghostty does this
right now (as do many other terminals):
```
OOOOOOOO
X
```
This commit changes the behavior so that we reflow the saved cursor.
In multiple tests we create 1 or more pages by growing them 1 row at a
time, which results in an integrity check of the page for each row grown
which is just... horrible. By simply pausing integrity checks while
growing these pages (since growing them is not the point of the test) we
MASSIVELY speed up all of these tests.
Also reduced grapheme bytes during testing and made the Terminal "glitch
text" test actually assert what it intends to achieve, rather than just
blindly assuming 100 copies of the text will be the right amount -- this
lets us stop a lot earlier, making it take practically no time.
In multiple tests we create 1 or more pages by growing them 1 row at a
time, which results in an integrity check of the page for each row grown
which is just... horrible. By simply pausing integrity checks while
growing these pages (since growing them is not the point of the test) we
MASSIVELY speed up all of these tests.
Also reduced grapheme bytes during testing and made the Terminal "glitch
text" test actually assert what it intends to achieve, rather than just
blindly assuming 100 copies of the text will be the right amount -- this
lets us stop a lot earlier, making it take practically no time.
Fixes#2941
This fixes the rendering of the text below. For those that can't see it,
it is the following in UTF-32: `0x22 0x1F3FF 0x22`.
```
"🏿"
```
`0x1F3FF` is the Fitzpatrick modifier for dark skin tone. It has the
Unicode property `Emoji_Modifier`. Emoji modifiers are defined in UTS
#51 and are only valid based on ED-13:
```
emoji_modifier_sequence := emoji_modifier_base emoji_modifier
emoji_modifier_base := \p{Emoji_Modifier_Base}
emoji_modifier := \p{Emoji_Modifier}
```
Additional quote from UTS #51:
> To have an effect on an emoji, an emoji modifier must immediately follow
> that base emoji character. Emoji presentation selectors are neither needed
> nor recommended for emoji characters when they are followed by emoji
> modifiers, and should not be used in newly generated emoji modifier
> sequences; the emoji modifier automatically implies the emoji presentation
> style.
Our precomputed grapheme break table was mistakingly not following this
rule. This commit fixes that by adding a check for that every
`Emoji_Modifier` character must be preceded by an `Emoji_Modifier_Base`.
This only has a cost during compilation (table generation). The runtime
cost is identical; the table size didn't increase since we had leftover
bits we could use.
Fixes#2857
Some terminal modes always reset, but there are others that should be
conditional based on how the terminal's default state is configured.
Primarily from #2857 is the grapheme clustering mode (mode 2027) which
was always resetting to false but should be conditional based on the
the `grapheme-width-method` configuration.
Fixes#2817
The test is pretty explanatory. I also renamed `end` to `count` since I
think this poor naming was the reason for the bug. In `eraseChars`, the
`count` (nee `end`) is the number of cells to erase, not the index of
the last cell to erase.
This is more correct: a pagelist is a linked list of nodes, not pages.
The nodes themselves contain pages but we were previously calling the
nodes "pages" which was confusing, especially as I plan some future
changes to the way pages are stored.
When resetting the terminal screen, the memory buffer allocated for the
scrollback is now cleared by reinitializing the screen and falling back to the
current method if any of the attempts to reinitialize fail.
Closes#2464
When the focus reporting mode (1004) is enabled, send the current focus
state. This allows applications to track their own focus state without
first having to wait for a focus event (or query
it by sending a DECSET followed by a DECRST).
Ghostty's focus state is stored only in the renderer, where the termio
thread cannot access it. We duplicate the focus state tracking in the
Terminal struct with the addition of a new (1-bit) flag. We duplicate
the state because the renderer uses the focus state for its own purposes
(in particular, the Metal renderer uses the focus state to manage
its DisplayLink), and synchronizing access to the shared terminal state
is more cumbersome than simply tracking the focus state in the renderer
in addition to the terminal.
Handle `clonePartialRowFrom` errors in `insertLines` and `deleteLines`
by adjusting page capacity. To do this, I've rewritten both functions
with a new way of iterating rows by moving a tracked pin up/down.
Benchmarks seem to indicate that this has no effect on performance.
Mitchell Hashimoto