Printing a wide character at the right edge of the screen with an active
hyperlink triggered a page integrity violation (UnwrappedSpacerHead).
printCell wrote the spacer_head to the cell and then called
cursorSetHyperlink, whose internal integrity check observed the
spacer_head before printWrap had a chance to set the row wrap flag.
Fix by setting row.wrap = true before calling printCell for the
spacer_head case, so all integrity checks see a consistent state.
printWrap sets wrap again afterward, which is harmless. Found by AFL++
stream fuzzer.
CSI @ (ICH) with an explicit parameter of 0 should be clamped to 1,
matching xterm behavior. Previously, a zero count reached
Terminal.insertBlanks which called clearCells with an empty slice,
triggering an out-of-bounds panic.
Fix the stream dispatch to clamp 0 to 1 via @max, and add a defensive
guard in insertBlanks for count == 0. Found by AFL++ stream fuzzer.
This works around Fish (at least v4.2) having a non-compliant OSC133
implementation paired with not having the hooks to fix this via shell
integration. We have to instead resort to heuristics in the terminal
emulator. Womp, womp.
The issue is that Fish does not emit OSC133 secondary prompt (`k=s`)
markers at the beginning of continuation lines. And, since Fish doesn't
provide a PS2-equivalent, we can't do this via shell integration.
We fix this by assuming on newline (`\n`) that a cursor that is already
painting prompt cells is continuing a prior prompt line, and
pre-emptively mark it as a prompt line. But this has two further issues
we have to work around:
1. Newline/index (`\n`) is one of the _hottest path_ functions in
terminal emulation. It sucks to add any new conditional logic here.
We do our best to gate this on unlikely conditions that the branch
predictor can easily optimize away.
2. Fish also emits these for auto-complete hints that may be deleted
later. So, we also have to handle the scenario where a prompt is
continued, then replaced by command output, and fix up the prompt
continuation flag to go back to output mode.
Point 2 is ALMOST automatically handled, because Fish does emit a `CSI J`
(erase display below) to erase the auto-complete hint. This resets all
our rows back to output rows. **Unfortunately**, Fish emits `\n` before
triggering the preexec hooks which set OSC133C. So we get the newline
logic FIRST (sets the prompt line), THEN sets the output cursor. If they
switched ordering here everything would just work (with the one
heuristic). But now, we need two!
To address this, I put some extra heuristic logic in the OSC133C
(output starting) handler: if our row is marked as a prompt AND our
cursor is at x=0, we assume that the prompt continuation was deleted
and we unmark it.
I put the heuristic logic dependent on OSC133C because that's way colder
of a path than putting something in `printCell` (which is the actual
hottest path in Ghostty).
We could get more rigorous here by also checking if every cell is empty
but that doesn't seem to be necessary at this time for any Fish version
I've tested. I hope thats correct.
I'd really love for Fish to improve their OSC133 implementation to
conform more closely to the terminal-wg spec, but we're going to need
these workarounds indefinitely to handle older Fish versions anyway.
This PR builds on https://github.com/ghostty-org/ghostty/pull/9678 ~so
the diff from there is included here (it's not possible to stack PRs
unless it's a PR against my own fork)--review that one first!~
This PR updates the `graphemeBreak` calculation to use `uucode`'s
`computeGraphemeBreakNoControl`, which has [tests in
uucode](215ff09730/src/x/grapheme.zig (L753))
that confirm it passes the `GraphemeBreakTest.txt` (minus some
exceptions).
Note that the `grapheme_break` (and `grapheme_break_no_control`)
property in `uucode` incorporates `emoji_modifier` and
`emoji_modifier_base`, diverging from UAX #29 but matching UTS #51. See
[this comment in
uucode](215ff09730/src/grapheme.zig (L420-L434))
for details.
The `grapheme_break_no_control` property and
`computeGraphemeBreakNoControl` both assume `control`, `cr`, and `lf`
have been filtered out, matching the current grapheme break logic in
Ghostty.
This PR keeps the `Precompute.data` logic mostly equivalent, since the
`uucode` `precomputedGraphemeBreak` lacks benchmarks in the `uucode`
repository (it was benchmarked in [the original PR adding `uucode` to
Ghostty](https://github.com/ghostty-org/ghostty/pull/8757)). Note
however, that due to `grapheme_break` being one bit larger than
`grapheme_boundary_class` and the new `BreakState` also being one bit
larger, the state jumps up by a factor of 8 (u10 -> u13), to 8KB.
## Benchmarks
~I benchmarked the old `main` version versus this PR for
`+grapheme-break` and surprisingly this PR is 2% faster (?). Looking at
the assembly though, I'm thinking something else might be causing that.
Once I get to the bottom of that I'll remove the below TODO and include
the benchmark results here.~
When seeing the speedup with `data.txt` and maybe a tiny speedup on
English wiki, I was surprised given the 1KB -> 8KB tables. Here's what
AI said when I asked it to inspect the assembly:
https://ampcode.com/threads/T-979b1743-19e7-47c9-8074-9778b4b2a61e, and
here's what it said when I asked it to predict the faster version:
https://ampcode.com/threads/T-3291dcd3-7a21-4d24-a192-7b3f6e18cd31
It looks like two loads got reordered and that put the load that
depended on stage1 -> stage2 -> stage3 second, "hiding memory latency".
So that makes the new one faster when looking up the `grapheme_break`
property. These gains go away with the Japanese and Arabic benchmarks,
which spend more time processing utf8, and may even have more grapheme
clusters too.
### with data.txt (200 MB ghostty-gen random utf8)
<img width="1822" height="464" alt="CleanShot 2025-11-26 at 08 42 03@2x"
src="https://github.com/user-attachments/assets/56d4ee98-21db-4eab-93ab-a0463a653883"
/>
### with English wiki dump
<img width="2012" height="506" alt="CleanShot 2025-11-26 at 08 43 15@2x"
src="https://github.com/user-attachments/assets/230fbfb7-272d-4a2a-93e7-7268962a9814"
/>
### with Japanese wiki dump
<img width="2008" height="518" alt="CleanShot 2025-11-26 at 08 43 49@2x"
src="https://github.com/user-attachments/assets/edb408c8-a604-4a8f-bd5b-80f19e3d65ee"
/>
### with Arabic wiki dump
<img width="2010" height="512" alt="CleanShot 2025-11-26 at 08 44 25@2x"
src="https://github.com/user-attachments/assets/81a29ac8-0586-4e82-8276-d7fa90c31c90"
/>
TODO:
* [x] Take a closer look at the assembly and understand why this PR (8
KB vs 1 KB table) is faster on my machine.
* [x] _(**edit**: checking this off because it seems unnecessary)_ If
this turns out to actually be unacceptably slower, one possibility is to
switch to `uucode`'s `precomputedGraphemeBreak` which uses a 1445 byte
table since it uses a dense table (indexed using multiplication instead
of bitCast, though, which did show up in the initial benchmarks from
https://github.com/ghostty-org/ghostty/pull/8757 a small amount.)
AI was used in some of the uucode changes in
https://github.com/ghostty-org/ghostty/pull/9678 (Amp--primarily for
tests), but everything was carefully vetted and much of it done by hand.
This PR was made without AI with the exception of consulting AI about
whether the "Prepend + ASCII" scenario is common (hopefully it's right
about that being uncommon).
We need to have sane behavior in error handling because the running
program that sends the restore cursor command has no way to realize it
failed. So if our style fails to add (our only fail case) then we revert
to no style.
https://ampcode.com/threads/T-019bd7dc-cf0b-7439-ad2f-218b3406277a
Fixes#9905
This fixes a major compatibility issues with the CSI S sequence:
When our top margin is at the top (row 0) without left/right
margins, we should be creating scrollback. Previously, we were
only deleting.
