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)
5.0 KiB
Vendored
Generated
c, SPDX-License-Identifier, Long, Short, Arg, Help, Protocols, Mutexed, Category, Added, Multi, See-also, Example
| c | SPDX-License-Identifier | Long | Short | Arg | Help | Protocols | Mutexed | Category | Added | Multi | See-also | Example | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Copyright (C) Daniel Stenberg, <daniel@haxx.se>, et al. | curl | form | F | <name=content> | Specify multipart MIME data | HTTP SMTP IMAP | data head upload-file | http upload post imap smtp | 5.0 | append |
|
|
--form
For the HTTP protocol family, emulate a filled-in form in which a user has pressed the submit button. This makes curl POST data using the Content-Type multipart/form-data according to RFC 2388.
For SMTP and IMAP protocols, this composes a multipart mail message to transmit.
This enables uploading of binary files etc. To force the 'content' part to be a file, prefix the filename with an @ sign. To get the content part from a file, prefix the filename with the symbol <. The difference between @ and < is then that @ makes a file get attached in the post as a file upload, while the < makes a text field and gets the contents for that text field from a file.
Read content from stdin instead of a file by using a single "-" as filename. This goes for both @ and < constructs. When stdin is used, the contents is buffered in memory first by curl to determine its size and allow a possible resend. Defining a part's data from a named non-regular file (such as a named pipe or similar) is not subject to buffering and is instead read at transmission time; since the full size is unknown before the transfer starts, such data is sent as chunks by HTTP and rejected by IMAP.
Example: send an image to an HTTP server, where 'profile' is the name of the form-field to which the file portrait.jpg is the input:
curl -F profile=@portrait.jpg https://example.com/upload.cgi
Example: send your name and shoe size in two text fields to the server:
curl -F name=John -F shoesize=11 https://example.com/
Example: send your essay in a text field to the server. Send it as a plain text field, but get the contents for it from a local file:
curl -F "story=<hugefile.txt" https://example.com/
You can also instruct curl what Content-Type to use by using type=, in a
manner similar to:
curl -F "web=@index.html;type=text/html" example.com
or
curl -F "name=daniel;type=text/foo" example.com
You can also explicitly change the name field of a file upload part by setting filename=, like this:
curl -F "file=@localfile;filename=nameinpost" example.com
If filename/path contains ',' or ';', it must be quoted by double-quotes like:
curl -F "file=@\"local,file\";filename=\"name;in;post\"" \
https://example.com
or
curl -F 'file=@"local,file";filename="name;in;post"' \
https://example.com
Note that if a filename/path is quoted by double-quotes, any double-quote or backslash within the filename must be escaped by backslash.
Quoting must also be applied to non-file data if it contains semicolons, leading/trailing spaces or leading double quotes:
curl -F 'colors="red; green; blue";type=text/x-myapp' \
https://example.com
You can add custom headers to the field by setting headers=, like
curl -F "submit=OK;headers=\"X-submit-type: OK\"" example.com
or
curl -F "submit=OK;headers=@headerfile" example.com
The headers= keyword may appear more than once and above notes about quoting apply. When headers are read from a file, empty lines and lines starting with '#' are ignored; each header can be folded by splitting between two words and starting the continuation line with a space; embedded carriage-returns and trailing spaces are stripped. Here is an example of a header file contents:
# This file contains two headers.
X-header-1: this is a header
# The following header is folded.
X-header-2: this is
another header
To support sending multipart mail messages, the syntax is extended as follows:
-
name can be omitted: the equal sign is the first character of the argument,
-
if data starts with '(', this signals to start a new multipart: it can be followed by a content type specification.
-
a multipart can be terminated with a '=)' argument.
Example: the following command sends an SMTP mime email consisting in an inline part in two alternative formats: plain text and HTML. It attaches a text file:
curl -F '=(;type=multipart/alternative' \
-F '=plain text message' \
-F '= <body>HTML message</body>;type=text/html' \
-F '=)' -F '=@textfile.txt' ... smtp://example.com
Data can be encoded for transfer using encoder=. Available encodings are binary and 8bit that do nothing else than adding the corresponding Content-Transfer-Encoding header, 7bit that only rejects 8-bit characters with a transfer error, quoted-printable and base64 that encodes data according to the corresponding schemes, limiting lines length to 76 characters.
Example: send multipart mail with a quoted-printable text message and a base64 attached file:
curl -F '=text message;encoder=quoted-printable' \
-F '=@localfile;encoder=base64' ... smtp://example.com