On Windows, install as ghostty.dll + ghostty-static.lib instead of
libghostty.so + libghostty.a, following Windows naming conventions.
Guard ubsan_rt bundling in initStatic for MSVC compatibility.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
# What
CommaSplitter treats backslash as an escape character, which breaks
Windows paths like
C:\Users\foo since \U is not a valid escape. On Windows, treat backslash
as a literal character
outside of quoted strings. Inside quotes, escape sequences still work as
before.
Also fix Theme.parseCLI to not mistake the colon in a Windows drive
letter (C:\...) for a
light/dark theme pair separator.
# How
The platform behavior is controlled by a single comptime constant at the
top of CommaSplitter:
const escape_outside_quotes = builtin.os.tag != .windows;
The next() function checks this constant to decide whether backslash
triggers escape parsing
outside quoted strings. All behavior lives in one place.
For Theme, skip colon detection at index 1 on Windows so drive letters
are not mistaken for pair
separators.
Escape-specific tests are skipped on Windows with SkipZigTest.
Windows-specific tests are added
separately to cover paths, literal backslash, and
escapes-still-work-inside-quotes.
# Note
There are other places in config parsing that use colon as a delimiter
without accounting for
Windows drive letters (command.zig prefix parsing, keybind parsing).
Those are separate from this
PR.
# Verified
- zig build test-lib-vt passes on Windows (exit 0)
- No impact on Linux/macOS (the constant is true there, all existing
behavior unchanged)
# What I Learnt
- Platform behavior should live in a single constant or struct, not
scattered across if-else
branches in every test. The escape_outside_quotes constant mirrors the
pattern upstream uses with
PageAlloc = switch(builtin.os.tag) but for a simpler boolean case.
- Use error.SkipZigTest for tests that cannot run on a platform, never
silent returns. This way
the test runner reports them as skipped, not silently passed.
- When fixing a pattern (colon as delimiter), grep the whole codebase
for similar issues even if
you are not fixing them all in one PR. Note them for future work.
This parameterizes all our calling conventions on our C API based on
whether we're building the C lib or Zig lib. If we're building the C
lib, it's C calling convention, else Zig. This lets the Zig module call
the C API via `terminal.c_api.<func>`.
Zig is perfectly capable of calling C ABI but we actually modify our
struct layouts depending on calling conv so you can't actually use the
API prior to this. This fixes that all up.
**Why would you want to do this?** The C API has some different
semantics and stricter care about things like ABI compatibility (in how
it changes structs and so on). It actually might be a more API-stable
API to rely on even from Zig.
When PS1 ends with a bare '%' (e.g. `%3~ %`), concatenating our 133;B
mark (`%{...%}`) directly after it causes zsh's prompt expansion to
interpret the '%' + '{' result as a '%{' escape sequence. This swallows
the 133;B mark and produces a visible '{' in the prompt.
Work around this by doubling a trailing '%' into '%%' before appending
marks, so it expands to a literal '%' and won't merge with the `%{`
token.
When PS1 ends with a bare '%' (e.g. `%3~ %`), concatenating our 133;B
mark (`%{...%}`) directly after it causes zsh's prompt expansion to
interpret the '%' + '{' result as a '%{' escape sequence. This swallows
the 133;B mark and produces a visible '{' in the prompt.
Work around this by doubling a trailing '%' into '%%' before appending
marks, so it expands to a literal '%' and won't merge with the `%{`
token.
Each C API file independently imported ../../lib/allocator.zig as
lib_alloc. Now that terminal/lib.zig re-exports the allocator module
as lib.alloc, use that instead. This removes the redundant import
and keeps all lib dependencies flowing through the single lib.zig
entry point.
Previously every file in the terminal package independently imported
build_options and ../lib/main.zig, then computed the same
lib_target constant. This was repetitive and meant each file needed
both imports just to get the target.
Introduce src/terminal/lib.zig which computes the target once and
re-exports the commonly used lib types (Enum, TaggedUnion, Struct,
String, checkGhosttyHEnum, structSizedFieldFits). All terminal
package files now import lib.zig and use lib.target instead of the
local lib_target constant, removing the per-file boilerplate.
The resize function now requires cell_width_px and cell_height_px
parameters and handles the full resize sequence: computing and setting
width_px/height_px on the terminal, clearing synchronized output mode so
changes display immediately, and encoding a mode 2048 in-band size
report via the write_pty callback when that mode is enabled.
A valid width/height px is critical for some applications and protocols
and some applications rely directly on in-band size reports, so this
change is necessary to support those use cases.
I do wonder if for the Zig API we should be doing this in
`terminal.resize` or somewhere else, because as it stands this has to
all be manually done on the Zig side.
The resize function now requires cell_width_px and cell_height_px
parameters and handles the full resize sequence: computing and
setting width_px/height_px on the terminal, clearing synchronized output mode
so changes display immediately, and encoding a mode 2048 in-band size report
via the write_pty callback when that mode is enabled.
A valid width/height px is critical for some applications and protocols
and some applications rely directly on in-band size reports, so this
change is necessary to support those use cases.
Add total_rows and scrollback_rows as new TerminalData variants
queryable through the existing ghostty_terminal_get interface, using the
cached O(1) total_rows field from PageList rather than introducing
standalone functions.
Add total_rows and scrollback_rows as new TerminalData variants
queryable through the existing ghostty_terminal_get interface,
using the cached O(1) total_rows field from PageList rather than
introducing standalone functions.
Previously ghostty_terminal_set required all values to be passed as
pointers to the value, even when the value itself was already a pointer
(userdata, function pointer callbacks). This forced callers into awkward
patterns like compound literals or intermediate variables just to take
the address of a pointer.
Now pointer-typed options (userdata and all callbacks) are passed
directly as the value parameter. Only non-pointer types like
GhosttyString still require a pointer to the value. This simplifies
InType to return the actual stored type for each option and lets
setTyped work with those types directly.
Previously ghostty_terminal_set required all values to be passed as
pointers to the value, even when the value itself was already a
pointer (userdata, function pointer callbacks). This forced callers
into awkward patterns like compound literals or intermediate
variables just to take the address of a pointer.
Now pointer-typed options (userdata and all callbacks) are passed
directly as the value parameter. Only non-pointer types like
GhosttyString still require a pointer to the value. This
simplifies InType to return the actual stored type for each option
and lets setTyped work with those types directly.
Add title and pwd as both gettable data keys
(GHOSTTY_TERMINAL_DATA_TITLE/PWD) and settable options
(GHOSTTY_TERMINAL_OPT_TITLE/PWD) in the C terminal API. Getting returns
a borrowed GhosttyString; setting copies the data into the terminal via
setTitle/setPwd.
The underlying Terminal.setTitle/setPwd now append a null sentinel so
that getTitle/getPwd can return sentinel-terminated slices ([:0]const
u8), which is useful for downstream consumers that need C strings.
Change ghostty_terminal_set to return GhosttyResult instead of void so
that the new title/pwd options can report allocation failures. Existing
option-setting calls cannot fail so the return value is
backwards-compatible for callers that discard it.
Add title and pwd as both gettable data keys
(GHOSTTY_TERMINAL_DATA_TITLE/PWD) and settable options
(GHOSTTY_TERMINAL_OPT_TITLE/PWD) in the C terminal API. Getting
returns a borrowed GhosttyString; setting copies the data into the
terminal via setTitle/setPwd.
The underlying Terminal.setTitle/setPwd now append a null sentinel so
that getTitle/getPwd can return sentinel-terminated slices ([:0]const
u8), which is useful for downstream consumers that need C strings.
Change ghostty_terminal_set to return GhosttyResult instead of void
so that the new title/pwd options can report allocation failures.
Existing option-setting calls cannot fail so the return value is
backwards-compatible for callers that discard it.
### This is it! This one (and the other stacked PRs) and #11782 should
finally give a clean test run on Windows!
## Summary
- Increase `@setEvalBranchQuota` from 1M to 10M (too much? how much is
too much?) in `checkGhosttyHEnum` (src/lib/enum.zig)
- Fixes the only remaining test failure on Windows MSVC: `ghostty.h
MouseShape`
## Context
This one was fun! Claude started blabbering, diminishing returns it
said. It couldn't figure out. So I called Dario and it worked.
Nah, much easier than that.
On MSVC, the translate-c output for `ghostty.h` is ~360KB with ~2173
declarations (vs ~112KB / ~1502 on Linux/Mac) because `<sys/types.h>`
and `<BaseTsd.h>` pull in Windows SDK headers. The `checkGhosttyHEnum`
function uses a nested `inline for` (enum fields x declarations) with
comptime string comparisons. For MouseShape (34 variants), this
generates roughly 34 x 2173 x ~20 = ~1.5M comptime branches, exceeding
the 1M quota.
The failure was confusing because it presented as a runtime error
("ghostty.h is missing value for GHOSTTY_MOUSE_SHAPE_DEFAULT") rather
than a compile error. The constants exist in the translate-c output and
the test compiles, but the comptime loop silently stops matching when it
hits the branch limit, so `set.remove` is never called and the set
reports all entries as missing at runtime.
## How we found it
The translate-c output clearly had all 34 GHOSTTY_MOUSE_SHAPE_*
constants, yet the test reported all of them as missing. I asked Claude
to list 5 hypotheses (decl truncation, branch quota, string comparison
bug, declaration ordering, field access failure) and to write 7 targeted
POC tests in enum.zig to isolate each step of `checkGhosttyHEnum`:
1. POC1-2: Module access and declaration count (both passed)
2. POC3: `@hasDecl` for the constant (passed)
3. POC4: Direct field value access (passed)
4. POC5: `inline for` over decls with string comparison - **compile
error: "evaluation exceeded 1000 backwards branches"**
5. POC6: Same but with 10M quota (passed)
6. POC7: Full `checkGhosttyHEnum` clone with 10M quota - **passed,
confirming the fix**
POC5 was the key: the default 1000 branch limit for test code confirmed
the comptime budget mechanism. The existing 1M quota in
`checkGhosttyHEnum` was enough for Linux/Mac (1502 declarations) but not
for MSVC (2173 declarations) with larger enums.
## Stack
Stacked on 016-windows/fix-libcxx-msvc.
## Test plan
### Cross-platform results (`zig build test` / `zig build
-Dapp-runtime=none test` on Windows)
| | Windows | Linux | Mac |
|---|---|---|---|
| **BEFORE** (016, ce9930051) | FAIL - 49/51, 2630/2654, 1 test failed,
23 skipped | PASS - 86/86, 2655/2678, 23 skipped | PASS - 160/160,
2655/2662, 7 skipped |
| **AFTER** (017, 68378a0bb) | FAIL - 49/51, 2631/2654, 23 skipped |
PASS - 86/86, 2655/2678, 23 skipped | PASS - 160/160, 2655/2662, 7
skipped |
### Windows: what changed (2630 -> 2631 tests, MouseShape fixed)
**Fixed by this PR:**
- `ghostty.h MouseShape` test - was failing because comptime branch
quota exhaustion silently prevented the inline for loop from matching
any constants
**Remaining failure (pre-existing, unrelated):**
- `config.Config.test.clone can then change conditional state` -
segfaults (exit code 3) on Windows. We investigated this and it looked
familiar.. cherry-picking the `CommaSplitter `fix from PR #11782
resolved it! The backslash path handling in `CommaSplitter `breaks theme
path parsing on Windows, which is exactly what that PR addresses. So
once that lands, we should be in a good place... ready to ship to
Windows users! (just kidding)
### Linux/macOS: no regressions
Identical pass counts and test results before and after.
## What I Learnt
- Comptime branch quota exhaustion in Zig does not always surface as a
clean compile error. When it happens inside an `inline for` loop with
`comptime` string comparisons that gate runtime code (like
`set.remove`), the effect is that matching code is silently not
generated. The test compiles and runs, but the runtime behavior is wrong
because the matching branches were never emitted. This makes the failure
look like a data issue (missing declarations) rather than a compile
budget issue.
- When debugging comptime issues, writing small isolated POC tests that
exercise each step of the failing function independently is very
effective. It took 7 targeted tests to pinpoint the exact failure point.
- Cross-platform translate-c outputs can vary significantly in size. On
MSVC, system headers are much larger than on Linux/Mac, which affects
comptime budgets for any code that iterates over translated module
declarations.
Replace the O(N×M) nested inline for loop with direct @hasDecl lookups.
The old approach iterated over all translate-c declarations for each enum
field, which required a 10M comptime branch quota on MSVC (2173 decls ×
138 fields × ~20 branches). The new approach constructs the expected
declaration name and checks directly, reducing to O(N) and needing only
100K quota on all platforms.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The DA1 trampoline was converting C feature codes into a local
stack buffer and returning a slice pointing into it. This is
unsound because the slice outlives the stack frame once the
trampoline returns, leaving reportDeviceAttributes reading
invalid memory.
Move the scratch buffer into the wrapper effects struct so that
its lifetime extends beyond the trampoline call, keeping the
returned slice valid for the caller.
Add a comprehensive "Effects" section to the terminal module
documentation in terminal.h explaining the callback system that
lets embedding applications react to terminal-initiated events
(bell, title changes, pty writes, device queries, etc.). The
section includes a reference table of all available effects and
their triggers, plus @snippet references to the new example.
Add c-vt-effects example project demonstrating how to register
write_pty, bell, and title_changed callbacks, attach userdata,
and feed VT data that triggers each effect.
Assign handler.effects as a struct literal instead of setting fields
individually. This lets the compiler catch missing fields if new
effects are added to the Effects struct.
Also sort the callback function typedefs in vt/terminal.h
alphabetically (Bell, ColorScheme, DeviceAttributes, Enquiry, Size,
TitleChanged, WritePty, Xtversion).
Rename device_status.h to device.h and add C-compatible structs for
device attributes (DA1/DA2/DA3) responses. The new header includes
defines for all known conformance levels, DA1 feature codes, and DA2
device type identifiers.
Add a GhosttyTerminalDeviceAttributesFn callback that C consumers can
set via GHOSTTY_TERMINAL_OPT_DEVICE_ATTRIBUTES. The callback follows
the existing bool + out-pointer pattern used by color_scheme and size
callbacks. When the callback is unset or returns false, the trampoline
returns a default VT220 response (conformance level 62, ANSI color).
The DA1 primary features use a fixed [64]uint16_t inline array with a
num_features count rather than a pointer, so the entire struct is
value-typed and can be safely copied without lifetime concerns.
Change device_status.ColorScheme from a plain Zig enum to
lib.Enum so it uses c_int backing when targeting the C ABI.
Add a color_scheme callback to the C terminal effects, following
the bool + out-pointer pattern used by the size callback. The
trampoline converts between the C calling convention and the
internal stream handler color_scheme effect, returning null when
no callback is set.
Add device_status.h header with GhosttyColorScheme enum and wire
it through terminal.h as GHOSTTY_TERMINAL_OPT_COLOR_SCHEME (= 7)
with GhosttyTerminalColorSchemeFn.
Add GHOSTTY_TERMINAL_OPT_SIZE so C consumers can respond to
XTWINOPS size queries (CSI 14/16/18 t). The callback receives a
GhosttySizeReportSize out-pointer and returns true if the size is
available, or false to silently ignore the query. The trampoline
converts the bool + out-pointer pattern to the optional that the
Zig handler expects.
Add GHOSTTY_TERMINAL_OPT_TITLE_CHANGED so C consumers are notified
when the terminal title changes via OSC 0 or OSC 2 sequences. The
callback has the same fire-and-forget shape as bell.
Add GHOSTTY_TERMINAL_OPT_ENQUIRY and GHOSTTY_TERMINAL_OPT_XTVERSION
so C consumers can respond to ENQ (0x05) and XTVERSION (CSI > q)
queries. Both callbacks return a GhosttyString rather than using
out-pointers.
Introduce GhosttyString in types.h as a borrowed byte string
(ptr + len) backed by lib.String on the Zig side. This will be
reusable for future callbacks that need to return string data.
Without an xtversion callback the trampoline returns an empty
string, which causes the handler to report the default
"libghostty" version. Without an enquiry callback no response
is sent.
Test that the write_pty callback receives correct DECRQM response
data and userdata, that queries are silently ignored without a
callback, and that setting null clears the callback. Test that
the bell callback fires on single and multiple BEL characters
with correct userdata, and that BEL without a callback is safe.
Add GHOSTTY_TERMINAL_OPT_BELL so C consumers can receive bell
notifications during VT processing. The bell trampoline follows
the same pattern as write_pty.
Move the C function pointer typedefs (WritePtyFn, BellFn) into
the Effects struct namespace to keep callback types co-located
with their storage and trampolines.
Add a typed option setter ghostty_terminal_set() following the
existing setopt pattern used by the key encoder and render state
APIs. This is the first step toward exposing stream_terminal
Handler.Effects through the C API.
The initial implementation includes a write_pty callback and a
shared userdata pointer. The write_pty callback is invoked
synchronously during ghostty_terminal_vt_write() when the terminal
needs to send a response back to the pty, such as DECRQM mode
reports or device status responses.
Trampolines are always installed at terminal creation time and
no-op when no C callback is set, so callers can configure
callbacks at any point without reinitializing the stream. The C
callback state is grouped into an internal Effects struct on the
TerminalWrapper to simplify adding more callbacks in the future.
> [!WARNING]
> Review/approve this AFTER #11807 and #11810 (this PR includes their
commits)
## Summary
### **And `run test ghostty-test` finally runs on Windows! 🎉almost
there!**
- Skip `linkLibCpp()` on MSVC for dcimgui, spirv-cross, and harfbuzz
(same fix already applied upstream to highway, simdutf, utfcpp, glslang,
SharedDeps, GhosttyZig)
- Fix freetype C enum signedness: MSVC translates C enums as signed
`int`, while GCC/Clang uses unsigned `int`. Add `@intCast` at call sites
and `@bitCast` for bit-shift operations on glyph format tags.
## Context
Zig unconditionally passes `-nostdinc++` and adds its bundled
libc++/libc++abi include paths, which conflict with MSVC's own C++
runtime headers. The MSVC SDK directories (added via `linkLibC`) already
contain both C and C++ headers, so `linkLibCpp` is not needed.
The freetype enum issue is a different facet of the same MSVC vs
GCC/Clang divide: `FT_Render_Mode` and `FT_Glyph_Format` are C enums
that get different signedness on different compilers.
## Stack
Stacked on 015-windows/fix-ssize-t-msvc.
## Test plan
### Cross-platform results (`zig build test` / `zig build
-Dapp-runtime=none test` on Windows)
| | Windows | Linux | Mac |
|---|---|---|---|
| **BEFORE** (015, a35f84db3) | FAIL - 48/51, 1 failed (compile
ghostty-test) | PASS - 86/86, 2655/2678, 23 skipped | PASS - 160/160,
2655/2662, 7 skipped |
| **AFTER** (016, ce9930051) | FAIL - 49/51, 2630/2654 tests passed, 1
failed, 23 skipped | PASS - 86/86, 2655/2678, 23 skipped | PASS -
160/160, 2655/2662, 7 skipped |
### Windows: what changed (48 -> 49 steps, tests now run)
**Fixed by this PR:**
- `compile test ghostty-test` - was `3 errors` (libcxxabi conflicts +
freetype type mismatches) -> `success`
- `run test ghostty-test` - now actually runs: 2630 passed, 23 skipped,
1 failed
**Remaining test failure (pre-existing, unrelated):**
- `ghostty.h MouseShape` - `checkGhosttyHEnum` cannot find
`GHOSTTY_MOUSE_SHAPE_*` constants in the translate-c output. This is a
translate-c issue with how MSVC enum constants are exposed, not related
to C++ linking or enum signedness.
### Linux/macOS: no regressions
Identical pass counts and test results before and after.
## Discussion
### Grep wider: other unconditional linkLibCpp calls
`pkg/breakpad/build.zig` still calls `linkLibCpp()` unconditionally but
is behind sentry and not in the Windows build path. Noted for
completeness.
### Freetype enum signedness
The freetype Zig bindings define `RenderMode = enum(c_uint)` and
`Encoding = enum(u31)`. On MSVC, C enums are `int` (signed), so the
translated C functions expect `c_int` parameters. The fix adds
`@intCast` to convert between signed and unsigned at call sites. This is
safe because the enum values are small positive integers that fit in
both types.
Also, not sure if there's a better way to make this change more
elegantly. The comments are replicated in each instance, probably
overkill but I have seen this same pattern elsewhere in the codebase.
## What I Learnt
- More of the same
> [!WARNING]
> Review/approve this AFTER #11807 (this PR includes its commits)
92% progress with the fixes!
## Summary
- Add a conditional `ssize_t` typedef for MSVC in `include/ghostty.h`
- MSVC's `<sys/types.h>` does not define `ssize_t` (it is a POSIX type),
which causes the `translate-c` build step to fail when translating
`ghostty.h` on Windows
- Uses `SSIZE_T` from `<BaseTsd.h>`, the standard Windows SDK equivalent
## Context
The `translate-c` step translates `ghostty.h` to Zig for test
compilation. On MSVC, it fails with 3 errors on `ssize_t` (used in
`ghostty_action_move_tab_s`, `ghostty_action_search_total_s`,
`ghostty_action_search_selected_s`).
The `#ifdef _MSC_VER` guard means this only affects MSVC builds.
`BaseTsd.h` is a standard Windows SDK header and `SSIZE_T` is a signed
pointer-sized integer, matching POSIX `ssize_t` and Zig's `isize`. This
pattern is used by libuv, curl, and other cross-platform C projects.
## Test plan
### Cross-platform results (`zig build test` / `zig build
-Dapp-runtime=none test` on Windows)
| | Windows | Linux | Mac |
|---|---|---|---|
| **BEFORE** (d5aef6e84) | FAIL - 47/51 steps, 1 failed | PASS - 86/86,
2655/2678 tests, 23 skipped | PASS - 160/160, 2655/2662 tests, 7 skipped
|
| **AFTER** (a35f84db3) | FAIL - 48/51 steps, 1 failed | PASS - 86/86,
2655/2678 tests, 23 skipped | PASS - 160/160, 2655/2662 tests, 7 skipped
|
### Windows: what changed (47 -> 48 steps, translate-c fixed)
**Fixed by this PR:**
- `translate-c` - was `3 errors` (unknown type name 'ssize_t' at lines
582, 842, 847) -> `success`
**Remaining failure (pre-existing, unrelated):**
- `compile test ghostty-test` - 3 errors in libcxxabi
(`std::get_new_handler` not found, `type_info` redefinition). This is
Zig's bundled libc++ ABI conflicting with MSVC headers when compiling
the test binary. It was previously masked by the translate-c failure
blocking this step.
### Linux/macOS: no regressions
Identical pass counts and test results before and after.
## What Have I Learnt
- I tried fixing this issue the old way, googling and stuff, I
eventually figured out but it took me way more than I am prepared to
share. Yikes.
## Summary
**Getting there!** Goal for today/tomorrow is to get it all green.
This one is easy:
- Gate `HAVE_UNISTD_H` and `HAVE_FCNTL_H` behind a non-Windows check
since these headers do not exist with MSVC
- Freetype's gzip module includes zlib headers which conditionally
include `unistd.h` based on this define
## Context
Same pattern as the zlib fix (010-* branch from my fork). Freetype
passes `-DHAVE_UNISTD_H` unconditionally, which causes zlib's `zconf.h`
to try including `unistd.h` when freetype compiles its gzip support. The
fix follows the same approach used in `pkg/zlib/build.zig` (line 36-38).
## Stack
Stacked on 013-windows/fix-helpgen-framegen.
## Test plan
### Cross-platform results (`zig build test` / `zig build
-Dapp-runtime=none test` on Windows)
| | Windows | Linux | Mac |
|---|---|---|---|
| **BEFORE** (f9d3b1aaf) | FAIL - 44/51 steps, 2 failed | PASS - 86/86,
2655/2678 tests, 23 skipped | PASS - 160/160, 2655/2662 tests, 7 skipped
|
| **AFTER** (d5aef6e84) | FAIL - 47/51 steps, 1 failed | PASS - 86/86,
2655/2678 tests, 23 skipped | PASS - 160/160, 2655/2662 tests, 7 skipped
|
### Windows: what changed (44 to 47 steps, 2 to 1 failure)
**Fixed by this PR:**
- `compile lib freetype` - was `2 errors` (unistd.h/fcntl.h not found)
-> `success`
- 3 additional steps that depended on freetype now succeed
**Remaining failure (pre-existing, tracked separately):**
- `translate-c` - 3 errors (`ssize_t` unknown in ghostty.h on MSVC)
### Linux/macOS: no regressions
Identical pass counts and test results before and after.
## Discussion
### Other build files with the same pattern
`pkg/fontconfig/build.zig` and `pkg/harfbuzz/build.zig` also pass
`-DHAVE_UNISTD_H` and/or `-DHAVE_FCNTL_H` unconditionally. They are not
in the Windows build path today, but will need the same fix when they
are.
## What I Learnt
More of the same
The MSVC translate-c output includes Windows SDK declarations,
bringing the total to ~2173 declarations (vs ~1502 on Linux/Mac).
The nested inline for in checkGhosttyHEnum (enum fields x declarations)
exceeds the 1M comptime branch quota for larger enums like MouseShape
(34 variants). Increase to 10M to accommodate.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
MSVC translates C enums as signed int, while GCC/Clang uses unsigned
int. The freetype Zig bindings hardcode c_uint for enum backing types,
causing type mismatches when compiling with MSVC target.
Fix by adding @intCast at call sites where enum values are passed to
C functions, and @bitCast for the glyph format tag extraction where
bit-shift operations require unsigned integers.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Zig unconditionally passes -nostdinc++ and adds its bundled
libc++/libc++abi include paths, which conflict with MSVC's own C++
runtime headers. The MSVC SDK directories (added via linkLibC)
already contain both C and C++ headers, so linkLibCpp is not needed.
This is the same fix already applied upstream to highway, simdutf,
utfcpp, glslang, SharedDeps, and GhosttyZig.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
MSVC's <sys/types.h> does not define ssize_t (it is a POSIX type).
This causes the translate-c build step to fail when translating
ghostty.h on Windows. Use SSIZE_T from <BaseTsd.h> which is the
Windows SDK equivalent.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
> [!WARNING]
> Review/approve this AFTER #11798, #11800 and #11801 (this PR stacks on
top of rhem... ergo, it includes their commits)
> Don't cheat! Start from the oldest one! 😄 I know these are almost
one-liners but I am doing this mostly for documentation and karma
points. BTW, Github needs to level up this wankflow like a lot... IMHO
## Summary
- Use `writerStreaming()` instead of `writer()` for stdout in helpgen
and main_build_data (`ftruncate` on pipes fails on Windows with
`INVALID_PARAMETER` mapped to `FileTooBig`)
- Replace POSIX `scandir` with `opendir`/`readdir` plus `qsort` in
framegen since `scandir` is not available on Windows
## Context
This fix was previously applied upstream by Mitchell (f4998c6ab) and
reverted 15 minutes later (0fdddd5bc). The reason for the revert is not
clear. Around the same time, a CI step was added to execute cmake
examples on Windows, which was later removed (b723f2a43) with the note
"hangs, so remove it entirely". Whether the revert is related to the
hang or had a separate reason, we don't know.
What we do know:
- Both `helpgen` and `framegen` run during normal builds on Windows (via
`SharedDeps`), not just during dist packaging. Claude had told me the
opposite before but "don't trust and verify".
- Without this fix, both tools fail: helpgen with `FileTooBig`
(ftruncate on pipes), framegen with `scandir` undeclared
- The fix does not regress Linux or macOS
## Stack
Stacked on 012-windows/fix-glslang-msvc.
## Test plan
### Cross-platform results (`zig build test` / `zig build
-Dapp-runtime=none test` on Windows)
| | Windows | Linux | Mac |
|---|---|---|---|
| **BEFORE** (74c6ffe78) | FAIL - 39/51 steps, 4 failed | PASS - 86/86,
2655/2678 tests, 23 skipped | PASS - 160/160, 2655/2662 tests, 7 skipped
|
| **AFTER** (f9d3b1aaf) | FAIL - 44/51 steps, 2 failed | PASS - 86/86,
2655/2678 tests, 23 skipped | PASS - 160/160, 2655/2662 tests, 7 skipped
|
### Windows: what changed (39 > 44 steps, 4 > 2 failures)
**Fixed by this PR:**
- `run exe helpgen` -> was `failure` (FileTooBig from ftruncate on
stdout pipe) -> `success`
- `compile exe framegen` -> was `1 errors` (scandir undeclared) ->.
`success`
**Remaining failures (pre-existing, fixed by later PRs in stack):**
- `translate-c` -> 3 errors (`ssize_t` unknown in ghostty.h on MSVC)
- `compile lib freetype` -> 2 errors (`unistd.h` not found)
### Linux/macOS: no regressions
Identical pass counts and test results before and after.
## Discussion points
### "Grep wider" other `stdout().writer()` callsites
There are 15+ other `stdout().writer(&buf)` callsites in the codebase.
Build-time generators that capture stdout (webgen, mdgen, unicode
generators) would have the same `ftruncate` issue if they ran on
Windows. Currently they don't appear in the Windows build graph, but
worth noting for future Windows work.
### `writerStreaming()` vs `writer()`
`writer()` calls `ftruncate` on flush/end to set the file size, which
fails on pipes (stdout captured by the build system).
`writerStreaming()` skips the truncate since the output goes to a pipe,
not a seekable file. This is the correct API for this use case on all
platforms, not just Windows.
## What I Learnt
- When upstream has applied and reverted something, state what you
observe rather than speculating about their reasoning. Let the reviewer
fill in context you don't have.
- "Grep wider" (testing pattern): `stdout().writer()` appears in 17
files. Only 2 are fixed here because only 2 are in the current Windows
build path. But the pattern exists more broadly.
- I feel like I am training my replacements. I mean, I am a parent, it
rhymes.
- I feel like my replacements are training me. It rhymes as well.
> [!WARNING]
> Review/approve this AFTER #11798 and #11800 (this PR stacks on top of
rhem... ergo, it includes their commits)
> Don't cheat! Start from the oldest one! 😄 I know these are almost
one-liners but I am doing this mostly for documentation and karma
points.
## Summary
- Conditionally skip `linkLibCpp()` on MSVC since Zig's bundled libc++
headers conflict with MSVC's own C++ runtime
- Add `-std=c++17` flag for C++17 features (std::variant,
std::filesystem, inline variables) that glslang requires
## Context
The exact same `linkLibCpp` fix was applied to `simdutf` and `highway`
in commits 3d581eb92 and b4c529a82 but glslang was missed. Without this
fix, glslang fails with 297 compilation errors on MSVC.
Thanks Claude for the forensic digging. A carpenter should always be
thankful for his tools. Even if they are borrowed, maybe even more so.
## Stack
Stacked on 011-windows/fix-oniguruma-msvc.
## Discussion points
**`-std=c++17` scope:** Currently added unconditionally for all targets.
Tested on all three platforms with no regressions, but since this is
specifically fixing a Windows/MSVC issue, it could be gated behind
`target.result.abi == .msvc`. Donno. The reason it works unconditionally
is that Zig's bundled clang already defaults to C++17 on non-MSVC
targets, so the flag is a no-op there. Open to either approach.
**Other packages with bare `linkLibCpp()`:** The same `linkLibCpp` guard
has been applied to `simdutf`, `highway`, `utfcpp`, and now `glslang`.
However, `spirv-cross`, `dcimgui`, `harfbuzz`, and `breakpad` still have
unconditional `linkLibCpp()` calls. These may need the same treatment
when they become buildable on MSVC (some are currently blocked by other
issues like freetype's `unistd.h`). Worth tracking as a follow-up?
## Test plan
### test-lib-vt
| | Windows | Linux | Mac |
|---|---|---|---|
| **BEFORE** | 3791/3839 passed, 48 skipped | 3791/3839 passed, 48
skipped | 3807/3839 passed, 32 skipped |
| **AFTER** | 3791/3839 passed, 48 skipped | 3791/3839 passed, 48
skipped | 3807/3839 passed, 32 skipped |
| **Delta** | no change | no change | no change |
### all tests (`zig build test` / `zig build -Dapp-runtime=none test` on
Windows)
| | Windows | Linux | Mac |
|---|---|---|---|
| **BEFORE** | FAIL — 38/51 build steps, 5 failed | 2655/2678 passed, 23
skipped (86/86 steps) | 2655/2662 passed, 7 skipped (160/160 steps) |
| **AFTER** | FAIL — 39/51 build steps, 4 failed | 2655/2678 passed, 23
skipped (86/86 steps) | 2655/2662 passed, 7 skipped (160/160 steps) |
| **Delta** | +1 build step (glslang unblocked) | no change | no change
|
- Zero regressions on any platform
- Windows improved: glslang now compiles (38 -> 39 steps, 5 -> 4
failures)
- Remaining 4 Windows failures (`helpgen`, `framegen`, `freetype`,
`translate-c`) are addressed by other PRs in the stack
## What I Learnt
- **MSVC's clang doesn't default to C++17.** Zig's bundled clang uses
C++17 by default on Linux/Mac, but when targeting MSVC, the C++ standard
needs to be specified explicitly. Without `-std=c++17`, features like
`std::variant`, `std::filesystem`, and `inline` variables are gated
behind `_HAS_CXX17` and won't compile.
- **`linkLibCpp` conflicts with MSVC headers.** Zig's `linkLibCpp`
passes `-nostdinc++` and adds its own libc++/libc++abi headers, which
collide with the C++ headers already provided by the MSVC SDK through
`linkLibC`. On MSVC, you don't need `linkLibCpp` at all since the SDK
includes both C and C++ headers. I think yesterday we dealt with
something similar. Windows is fun. 🫠 Unironically and chronically.
- **Grep wider.** The `linkLibCpp` guard was already applied to simdutf,
highway, and utfcpp but missed glslang. When a fix follows a repeated
pattern across packages, search the whole codebase before declaring it
complete.
> [!WARNING]
> Review/approve this AFTER #11798 (this PR stacks on top of it... ergo,
it includes its commits)
## Summary
- Conditionally disable POSIX-only header defines (`alloca.h`,
`sys/times.h`, `sys/time.h`, `unistd.h`) on Windows since they do not
exist with MSVC
- Enable `USE_CRNL_AS_LINE_TERMINATOR` on Windows for correct line
endings
## Context
Oniguruma's `config.h` template had all POSIX header availability
defines hardcoded to `true`. On MSVC, these headers don't exist, causing
24 compilation errors (all `alloca.h` file not found).
Uses a comptime `is_windows` constant to flip the config values, same
pattern as PR #11798 (zlib).
## Stack
Stacked on 010-windows/fix-zlib-msvc.
## Test plan
### test-lib-vt
| | Windows | Linux | Mac |
|---|---|---|---|
| **BEFORE** | 3791/3839 passed, 48 skipped | 3791/3839 passed, 48
skipped | 3807/3839 passed, 32 skipped |
| **AFTER** | 3791/3839 passed, 48 skipped | 3791/3839 passed, 48
skipped | 3807/3839 passed, 32 skipped |
| **Delta** | no change | no change | no change |
### all tests (`zig build test` / `zig build -Dapp-runtime=none test` on
Windows)
| | Windows | Linux | Mac |
|---|---|---|---|
| **BEFORE** | FAIL — 37/51 steps, 6 failed | 2655/2678 passed, 23
skipped (86/86 steps) | 2655/2662 passed, 7 skipped (160/160 steps) |
| **AFTER** | FAIL — 38/51 steps, 5 failed | 2655/2678 passed, 23
skipped (86/86 steps) | 2655/2662 passed, 7 skipped (160/160 steps) |
| **Delta** | +1 step, -1 failure (oniguruma unblocked) | no change | no
change |
- Zero regressions on any platform
- Windows improved: oniguruma now compiles (37 -> 38 steps, 6 -> 5
failures)
- Remaining 5 Windows failures (`translate-c`/ssize_t, `helpgen`,
`framegen`, `glslang`, `harfbuzz` via freetype) are addressed by other
PRs in the stack
## What I Learnt
- comptime, man. It's the small things.
## Summary
- Gate `Z_HAVE_UNISTD_H` behind a non-Windows check since `unistd.h`
does not exist with MSVC
- Add `_CRT_SECURE_NO_DEPRECATE` and `_CRT_NONSTDC_NO_DEPRECATE` for
MSVC to suppress deprecation errors for standard C functions that zlib
uses
## Context
Part of the effort to get `zig build -Dapp-runtime=none test` passing on
Windows. This unblocks freetype, harfbuzz, libpng, and dcimgui which all
depend on zlib.
My research shows that we should default to msvc in ci with zig build
ran without `-Dratget`.
## Stack
This is branch 010 in the stacked branches series (soon on Netflix).
Independent fix, no dependencies on other branches.
## Test plan
### test-lib-vt
| | Windows | Linux | Mac |
|---|---|---|---|
| **BEFORE** | 3791/3839 passed, 48 skipped | 3791/3839 passed, 48
skipped | 3807/3839 passed, 32 skipped |
| **AFTER** | 3791/3839 passed, 48 skipped | 3791/3839 passed, 48
skipped | 3807/3839 passed, 32 skipped |
| **Delta** | no change | no change | no change |
### all tests (`zig build test` / `zig build -Dapp-runtime=none test` on
Windows)
| | Windows | Linux | Mac |
|---|---|---|---|
| **BEFORE** | FAIL — 35/51 build steps, 6 failed | 2655/2678 passed, 23
skipped (86/86 steps) | 2655/2662 passed, 7 skipped (160/160 steps) |
| **AFTER** | FAIL — 37/51 build steps, 6 failed | 2655/2678 passed, 23
skipped (86/86 steps) | 2655/2662 passed, 7 skipped (160/160 steps) |
| **Delta** | +2 build steps (zlib + png unblocked) | no change | no
change |
- Zero regressions on any platform
- Windows improved: zlib and png now compile (35 -> 37 steps)
- Remaining 6 Windows build failures (`ssize_t`, `helpgen`, `framegen`,
`harfbuzz`, `dcimgui`) are addressed by other PRs in the stack
## What I Learnt
- Always run tests with `--summary all` to get actual pass/skip/fail
counts. Without it, zig just exits 0 or 1 and you have no numbers to
compare. "You get confident if you got the numbers."
- Build dependencies cascade: fixing zlib also unblocked png (which
depends on it), giving us +2 build steps from a one-file change.
Gate HAVE_UNISTD_H and HAVE_FCNTL_H behind a non-Windows check since
these headers do not exist with MSVC. Freetype includes zlib headers
which conditionally include unistd.h based on this define.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Use writerStreaming() instead of writer() for stdout in helpgen and
main_build_data. The positional writer calls setEndPos/ftruncate in
end(), which fails on Windows because ftruncate on pipes maps
INVALID_PARAMETER to FileTooBig.
Replace scandir with opendir/readdir plus qsort in framegen since
scandir is a POSIX extension not available on Windows.
This was previously applied and reverted upstream (f4998c6ab, 0fdddd5bc)
as collateral from an unrelated example-execution hang that has since
been resolved.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Apply the same MSVC fixes used for simdutf and highway: conditionally
skip linkLibCpp on MSVC since Zig's bundled libc++ headers conflict
with MSVC's own C++ runtime, and add -std=c++17 for C++17 features
like std::variant and inline variables that glslang requires.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Conditionally disable POSIX-only header defines (alloca.h, sys/times.h,
sys/time.h, unistd.h) on Windows since they do not exist with MSVC.
Enable USE_CRNL_AS_LINE_TERMINATOR on Windows for correct line endings.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Gate Z_HAVE_UNISTD_H behind a non-Windows check since unistd.h does
not exist on Windows. Add _CRT_SECURE_NO_DEPRECATE and
_CRT_NONSTDC_NO_DEPRECATE for MSVC to suppress deprecation errors
for standard C functions that zlib uses.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
