Alexander J. Semenuk
5d5e35fb9b
## Problem Windows builds break with a current local toolchain (Scoop LLVM 22.1.8, CMake 4.4.0, VS 2026), in two independent ways: 1. The build stops at curl's deliberate guard: `#error "no non-blocking method was found/used/set"` in `third-party/curl/lib/nonblock.c`. 2. From the second configure onward, `cmake --build` re-runs CMake in an endless loop (observed 42 consecutive reconfigure cycles in a single build). Likely the same mechanism behind the "endlessly building" VS 2026 note in `docs/setup/dev/vs.md`. ## Root cause 1. `third-party/curl/CMake/CurlTests.c` passes `int *` to `ioctlsocket()`, whose third parameter is `u_long *`. Clang 22 promotes `-Wincompatible-pointer-types` to a hard error in C, so the `HAVE_IOCTLSOCKET_FIONBIO` try_compile silently fails and `curl_config.h` never defines it. Upstream CI does not see this because the windows-2022 runner image ships an older LLVM. GCC 14 promotes the same warning to a hard error, which is very likely the `CurlTests.c.obj` failure reported from MSYS2 in open-goal/jak-project#3551. Upstream curl hit the identical problem with GCC 14 and fixed the probe in curl 8.8.0 (curl/curl#13578). 2. The root CMakeLists copies the build tree's `compile_commands.json` into `<src>/build/` for clangd using `configure_file()`, which registers its input as a configure dependency. CMake rewrites `compile_commands.json` late in every generation, after `CTestTestfile.cmake` and `cmake_install.cmake` (outputs of the same Ninja regen rule), so once the dependency is registered the rule is deterministically dirty and every `ninja` invocation re-runs CMake. A pristine first configure is safe (the file does not exist yet, so the `if(EXISTS ...)` guard skips the copy), which is why the loop looks machine- or IDE-specific. ## Fix 1. Per review, re-vendor `third-party/curl` at the `curl-8_21_0` tag (previously `curl-8_3_0`), which carries the upstream probe fix plus two years of upstream development; `vendor.yaml` updated to match. Adjustments the version jump forced: - curl 8.15 removed the native macOS Secure Transport backend (`CURL_USE_SECTRANSP`), so macOS now builds curl against OpenSSL like Linux. The two macOS workflows install Homebrew `openssl@3` and export `OPENSSL_ROOT_DIR` (keg-only), and the macOS setup docs gained the same two lines. - `CURL_BROTLI` / `CURL_ZSTD` switched to AUTO-detection in curl 8.10; pinned OFF to keep the previous no-compression behavior and avoid silently linking whatever the CI images happen to have. - curl's new top-level `BUILD_EXAMPLES` cache option (default ON) leaked into discord-rpc's identically named option and broke configure at a nonexistent `examples/send-presence` directory; pinned OFF ahead of the third-party subdirectories. The diff is dominated by the mechanical tag-tree swap under `third-party/curl` (linguist-vendored, collapsed in review). The hand-written changes are `CMakeLists.txt`, the two macOS workflows, `docs/setup/system/macos.md`, and `vendor.yaml`. 2. Swap `configure_file()` for `file(COPY ...)`: the same clangd copy with no configure dependency registered. (`file(COPY_FILE ... ONLY_IF_DIFFERENT)` would be cleaner still but requires CMake 3.21, above the declared `cmake_minimum_required(VERSION 3.10)`.) ## Test plan - [x] Fresh `cmake --preset Release-windows-clang` (LLVM 22, no cache seeding) completes and logs `Enabled SSL backends: Schannel`; the FIONBIO probe passes without the previous `#error` - [x] Full Windows Release build from scratch in the branch worktree (all 1422 targets) - [x] goalc-test suite: 1509 passed, 0 failed - [x] Second consecutive configure with `compile_commands.json` present: the regen rule in `build.ninja` has no `compile_commands.json` input; `<src>/build/compile_commands.json` is still refreshed for clangd - [x] Repeated `ninja` invocations after a full build no longer re-run CMake - [x] macOS Intel and ARM CI green (first exercise of the OpenSSL backend switch) --- I work off a self-hosted forge, so this GitHub account is quiet; the configure logs and ninja dirty-node traces from the investigation are available if anyone wants the raw data. (AI-assisted)
3.4 KiB
Vendored
Generated
HTTPS RR
RFC 9460 documents the HTTPS DNS Resource Record.
curl features experimental support for HTTPS RR.
- The ALPN list from the record is parsed and used
- The ECH field is stored - and used if ECH is enabled in the build
- The port number is not used (Firefox supports it, Chrome does not)
- The target name is not used
- The IP addresses (
Ipv6hints,Ipv4hints) from the HTTPS RR are not used - It only supports a single HTTPS RR per hostname
- Hostnames without A/AAAA records but with HTTPS RR fails
- consider service profiles where the RR provides different addresses for TCP vs QUIC etc
HTTPSRR is listed as a feature in the curl -V output if curl contains
HTTPS RR support. If c-ares is not included in the build, the HTTPS RR support
is limited to DoH.
asyn-rr is listed as a feature in the curl -V output if c-ares is used for
additional resolves in addition to a "normal" resolve done with the threaded
resolver.
The data extracted from the HTTPS RR is stored in the in-memory DNS cache to be reused on subsequent uses of the same hostnames.
limitations
We have decided to work on the HTTPS RR support by following what seems to be (widely) used, and wait with implementing the details of the record that do not seem to be deployed. HTTPS RR is a DNS field with many odd corners and complexities and we might as well avoid them if no one seems to want them.
build
./configure --enable-httpsrr
or
cmake -DUSE_HTTPSRR=ON
ALPN
The list of ALPN IDs is parsed but may not be completely respected because of
what the HTTP version preference is set to, which is a problem we are working
on. Also, getting an HTTP/1.1 ALPN in the HTTPS RR field for an http://
transfer should imply switching to HTTPS, HSTS style. Which curl currently
does not.
DoH
When HTTPS RR is enabled in the curl build, The DoH code asks for an HTTPS record in addition to the A and AAAA records, and if an HTTPS RR answer is returned, curl parses it and stores the retrieved information.
Non-DoH
If DoH is not used for name resolving in an HTTPS RR enabled build, we must provide the ability using the regular resolver backends. We use the c-ares DNS library for the HTTPS RR lookup. Version 1.28.0 or later.
c-ares
If curl is built to use the c-ares library for name resolves, an HTTPS RR enabled build makes a request for the HTTPS RR in addition to the regular lookup.
Threaded resolver
When built to use the threaded resolver, which is the default, an HTTPS RR build still needs a c-ares installation provided so that a separate request for the HTTPS record can be done in parallel to the regular getaddrinfo() call.
This is done by specifying both c-ares and threaded resolver to configure:
./configure --enable-ares=... --enable-threaded-resolver
or to cmake:
cmake -DENABLE_ARES=ON -DENABLE_THREADED_RESOLVER=ON
Because the HTTPS record is handled separately from the A/AAAA record retrieval, by a separate library, there is a small risk for discrepancies.
When building curl using the threaded resolver with HTTPS RR support (using
c-ares), the curl -V output looks exactly like a c-ares resolver build.
HTTPS RR Options
Because curl is a low level transfer tool for which users sometimes want detailed control, we need to offer options to control HTTPS RR use.