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)
4.9 KiB
Vendored
Generated
c, SPDX-License-Identifier, Title, Section, Source, See-also, Protocol, Added-in
| c | SPDX-License-Identifier | Title | Section | Source | See-also | Protocol | Added-in | ||
|---|---|---|---|---|---|---|---|---|---|
| Copyright (C) Daniel Stenberg, <daniel@haxx.se>, et al. | curl | libcurl-thread | 3 | libcurl |
|
|
n/a |
NAME
libcurl-thread - libcurl thread-safety
Multi-threading with libcurl
libcurl is thread-safe but has no internal thread synchronization. You may have to provide your own locking should you meet any of the thread-safety exceptions below.
Handles
You must never share the same handle in multiple threads. You can pass the handles around among threads, but you must never use a single handle from more than one thread at any given time.
Shared objects
You can share certain data between multiple handles by using the share interface but you must provide your own locking and set curl_share_setopt(3) CURLSHOPT_LOCKFUNC and CURLSHOPT_UNLOCKFUNC.
Note that some items are specifically documented as not thread-safe in the share API (the connection pool and HSTS cache for example).
TLS
All current TLS libraries libcurl supports are thread-safe.
OpenSSL
OpenSSL 1.1.0+ can be safely used in multi-threaded applications provided that support for the underlying OS threading API is built-in. For older versions of OpenSSL, the user must set mutex callbacks.
libcurl may not be able to fully clean up after multi-threaded OpenSSL depending on how OpenSSL was built and loaded as a library. It is possible in some rare circumstances a memory leak could occur unless you implement your own OpenSSL thread cleanup.
For example, on Windows if both libcurl and OpenSSL are linked statically to a DLL or application then OpenSSL may leak memory unless the DLL or application calls OPENSSL_thread_stop() before each thread terminates. If OpenSSL is built as a DLL then it does this cleanup automatically and there is no leak. If libcurl is built as a DLL and OpenSSL is linked statically to it then libcurl does this cleanup automatically and there is no leak (added in libcurl 8.8.0).
Please review the OpenSSL documentation for a full list of circumstances: https://docs.openssl.org/3.0/man3/OPENSSL_init_crypto/#notes
mbedTLS
mbedTLS can be used safely in a multi-threaded environment provided that mbedTLS is compiled with MBEDTLS_THREADING_C enabled.
https://mbed-tls.readthedocs.io/en/latest/kb/development/thread-safety-and-multi-threading
Signals
Signals are used for timing out name resolves (during DNS lookup) - when built without using either the c-ares or threaded resolver backends. On systems that have a signal concept.
When using multiple threads you should set the CURLOPT_NOSIGNAL(3) option to 1L for all handles. Everything works fine except that timeouts cannot be honored during DNS lookups - which you can work around by building libcurl with c-ares or threaded-resolver support. c-ares is a library that provides asynchronous name resolves. On some platforms, libcurl cannot function properly multi-threaded unless the CURLOPT_NOSIGNAL(3) option is set.
When CURLOPT_NOSIGNAL(3) is set to 1L, your application needs to deal with the risk of a SIGPIPE (that at least the OpenSSL backend can trigger). Note that setting CURLOPT_NOSIGNAL(3) to 0L does not work in a threaded situation as there is a race condition where libcurl risks restoring the former signal handler while another thread should still ignore it.
Name resolving
The gethostbyname or getaddrinfo and other name resolving system calls used by libcurl are provided by your operating system and must be thread-safe. It is important that libcurl can find and use thread-safe versions of these and other system calls, as otherwise it cannot function fully thread-safe. Some operating systems are known to have faulty thread implementations. We have previously received problem reports on *BSD (at least in the past, they may be working fine these days). Some operating systems that are known to have solid and working thread support are Linux, Solaris and Windows.
curl_global_* functions
These functions are thread-safe since libcurl 7.84.0 if curl_version_info(3) has the CURL_VERSION_THREADSAFE feature bit set (most platforms).
If these functions are not thread-safe and you are using libcurl with multiple threads it is especially important that before use you call curl_global_init(3) or curl_global_init_mem(3) to explicitly initialize the library and its dependents, rather than rely on the "lazy" fail-safe initialization that takes place the first time curl_easy_init(3) is called. For an in-depth explanation refer to libcurl(3) section GLOBAL CONSTANTS.
Memory functions
These functions, provided either by your operating system or your own replacements, must be thread-safe. You can use curl_global_init_mem(3) to set your own replacement memory functions.
Non-safe functions
CURLOPT_DNS_USE_GLOBAL_CACHE(3) is not thread-safe.
curl_version_info(3) is not thread-safe before libcurl initialization.