Jay Satiro [Sun, 16 Feb 2025 07:49:43 +0000 (02:49 -0500)]
scripts/managen: fix parsing of markdown code sections
- Terminate a code section before parsing a heading line.
Prior to this change when a code line (eg " code") was followed
by a heading line (eg "## heading") the code section in the output
was terminated after converting the header instead of before. That led
to some weird formatting outputs depending on the nroff or roffit etc.
Viktor Szakats [Fri, 7 Feb 2025 12:44:39 +0000 (13:44 +0100)]
cmake: misc tidy-ups
- replace `add_compile_options()`, `add_definitions()` with directory
properties. To harmonize this across all scripts. The new commands are
verbose, but describe better how they work. The syntax is also closer
to setting target properties, helps grepping.
- prefer `CMAKE_INSTALL_PREFIX` over `--prefix` (in tests, CI).
Viktor Szakats [Tue, 11 Feb 2025 01:46:29 +0000 (02:46 +0100)]
build: fix compiler warnings in feature detections
Fix or silence compiler warnings happening in feature detections
to reduce log noise. Warnings may also get promoted to errors in certain
cases, causing missed detections.
It reduces the number of warnings by 4500+ across the linux, linux-old,
macos, non-native and windows GHA workflows (~142 jobs).
Also move picky warning logic for MSVC/Borland to
`CMake/PickyWarnings.cmake. To make them listed in the picky-warnings
log output, and to also apply to feature detections to make them compile
under the same conditions as source code. The hope is to help catching
issues faster. It also improves code quality of feature tests.
Fixed/silenced:
```
warning #177: variable "dummy" was declared but never referenced
warning #177: variable "flag" was declared but never referenced
warning #177: variable "res" was declared but never referenced
warning #592: variable "s" is used before its value is set
warning #1011: missing return statement at end of non-void function "main"
warning #1786: function "SSL_CTX_set_srp_password" (declared at line 1888 of "/usr/include/openssl/ssl.h") was declared deprecated ("Since OpenSSL 3.0")
warning #1786: function "SSL_CTX_set_srp_username" (declared at line 1887 of "/usr/include/openssl/ssl.h") was declared deprecated ("Since OpenSSL 3.0")
warning #2332: a value of type "const char *" cannot be assigned to an entity of type "char *" (dropping qualifiers)
warning: 'SSL_CTX_set_srp_password' is deprecated [-Wdeprecated-declarations]
warning: 'SSL_CTX_set_srp_password' is deprecated: Since OpenSSL 3.0 [-Wdeprecated-declarations]
warning: 'SSL_CTX_set_srp_username' is deprecated [-Wdeprecated-declarations]
warning: 'SSL_CTX_set_srp_username' is deprecated: Since OpenSSL 3.0 [-Wdeprecated-declarations]
warning: 'b' is used uninitialized [-Wuninitialized]
warning: 'gethostname' redeclared without dllimport attribute: previous dllimport ignored [-Wattributes]
warning: Undefined or garbage value returned to caller [core.uninitialized.UndefReturn]
warning: Value stored to 'i' is never read [deadcode.DeadStores]
warning: assigning to 'char *' from 'const char *' discards qualifiers [-Wincompatible-pointer-types-discards-qualifiers]
warning: assignment discards 'const' qualifier from pointer target type [-Wdiscarded-qualifiers]
warning: control reaches end of non-void function [-Wreturn-type]
warning: empty expression statement has no effect; remove unnecessary ';' to silence this warning [-Wextra-semi-stmt]
warning: excess elements in struct initializer
warning: initialization discards 'const' qualifier from pointer target type [-Wdiscarded-qualifiers]
warning: macro "_FILE_OFFSET_BITS" is not used [-Wunused-macros]
warning: macro "_REENTRANT" is not used [-Wunused-macros]
warning: missing braces around initializer [-Wmissing-braces]
warning: no previous extern declaration for non-static variable 'off_t_is_large' [-Wmissing-variable-declarations]
warning: no previous prototype for 'check' [-Wmissing-prototypes]
warning: no previous prototype for function 'check' [-Wmissing-prototypes]
warning: null argument where non-null required (argument 2) [-Wnonnull]
warning: passing 'const char[1]' to parameter of type 'char *' discards qualifiers [-Wincompatible-pointer-types-discards-qualifiers]
warning: passing argument 2 of 'SSL_CTX_set_srp_password' discards 'const' qualifier from pointer target type [-Wdiscarded-qualifiers]
warning: passing argument 2 of 'SSL_CTX_set_srp_username' discards 'const' qualifier from pointer target type [-Wdiscarded-qualifiers]
warning: unused parameter 'c' [-Wunused-parameter]
warning: unused parameter 'f' [-Wunused-parameter]
warning: unused variable 'data' [-Wunused-variable]
warning: unused variable 'dummy' [-Wunused-variable]
warning: unused variable 'flag' [-Wunused-variable]
warning: unused variable 'res' [-Wunused-variable]
warning: unused variable 's' [-Wunused-variable]
warning: variable 's' set but not used [-Wunused-but-set-variable]
warning: variable 'ts' set but not used [-Wunused-but-set-variable]
```
Viktor Szakats [Fri, 27 Dec 2024 23:27:26 +0000 (00:27 +0100)]
cmake: add pre-fill for Unix, enable in GHA/macos, verify pre-fills
TL;DR: Save 10 minutes of CI time for GHA/macos jobs using pre-fills and
add pre-fill verification for Apple and Windows. Also restores Xcode job
and saves 1.5-10 minutes configuring iOS jobs.
Pre-filling feature detection results can bring down the CMake configure
step to ~5 seconds on most GHA runners, ~10 seconds in slow envs like
Cygwin/MSYS2.
The potential savings per job are:
- 5-40 (average 19) seconds on GHA/macos (33 jobs)
- ~10 seconds on GHA for iOS GNU Makefile (1 job)
- 1.5-10 minutes on GHA for iOS Xcode generator (1 job)
- 10 seconds on GHA/linux with native Ubuntu (12 jobs)
- 40 seconds for Cygwin/MSYS2 (2 jobs)
- 5-10 seconds for virtualized BSDs, native CPU (3 jobs)
- ~60 seconds for virtualized BSDs, emulated CPU (1 job)
On native Windows pre-filling has been in place for a long time and
saving 8 minutes (VS2019-VS2015) to 1.5-2 minutes (VS2022), 3 minutes
(VS2022 UWP), and 30-60 seconds (MinGW), per CI job.
The downside is that detection results need to be manually collected and
filtered to those that universally apply to all platforms that they are
enabled on. Another downside is that by using a cache, we're not running
the actual detections, and thus won't catch regressions in them. It
means we must make sure that the cache is solid and matches with actual
detections results. An upside is that it gives a rough overview of which
features are available on which platforms. Another upside is pre-filled
values do work for feature detections skipped for cross-builds, e.g.
`HAVE_WRITABLE_ARGV`.
This PR adds a pre-fill cache that supports all Unixes (except OmniOS)
used in CI, and makes it usable with an internal option. It also enables
it for GHA/macos CI jobs, where the maximum savings are. And also for
the two iOS [1] and two Cygwin/MSYS2 jobs. The latters don't have
pre-fill checks and we can drop them if they turn into a hassle.
Saving:
- 10 minutes of CI time per GHA/macos workflow run. [2]
- ~80 seconds per GHA/windows workflow run with Cygwin/MSYS2.
(offsetting the cost of pre-fill verifications)
- 1.5-10 minutes per GHA/non-native runs with iOS jobs. [3]
You can enable pre-fill locally with `-D_CURL_PREFILL=ON`. It's
experimental, and if you experience a problem, file a PR or an Issue.
This PR also adds a pre-fill checker for macOS and MinGW/MSVC Windows
GHA jobs to catch if the cache diverges from real detections. It also
adds this logic to AppVeyor, but doesn't enable it due to the perf
penalty of 2 minutes mininum.
The pre-fill checker works by configuring out-of-tree with and without
pre-fill, then diffing their `lib/curl_config.h` outputs.
Exceptions are 3 detection results exposed indirectly [4], and missing
to expose 2, of which one is the C89 header `stddef.h`. While we assume
the C99 `stdint.h` available outside iOS. We can expose them in the
future, if necessary.
The pre-fill checks cost in total:
- ~20 seconds for macOS
- ~40 seconds for MinGW on GHA
- ~80 seconds for MSVC on GHA (UWP would be 2x this)
An extra time saving potential is caching type sizes. They are
well-known, and seldom change, esp. in CI. GHA/Windows jobs spend 8-17
seconds per job on these ~12 feature checks. ~5s on Cygwin/MSYS2. Couple
of seconds on other platforms. (This PR doesn't make this optimization.)
Another opportunity is doing the same for autotools, which typically
spends more time in the configuration step than cmake.
[4] detection results exposed indirectly in `curl_config.h`:
- `HAVE_FILE_OFFSET_BITS` via `_FILE_OFFSET_BITS`
- `HAVE_GETHOSTBYNAME_R_*_REENTRANT` via `NEED_REENTRANT`
- `HAVE_SOCKADDR_IN6_SIN6_ADDR` via `USE_IPV6`
Jay Satiro [Sat, 15 Feb 2025 18:32:34 +0000 (13:32 -0500)]
curl_msh3: remove verify bypass from DEBUGBUILDs
- Remove the workaround that disabled peer verification in DEBUGBUILDs
when CA certs were provided.
The workaround was part of a TODO that disabled verification in
DEBUGBUILDs with a CAfile/path because apparently there's no way to set
those options in msh3 and that caused some tests to fail. Instead the
tests should fail and this problem should not be covered up.
Daniel Stenberg [Fri, 14 Feb 2025 10:29:08 +0000 (11:29 +0100)]
strparse: switch to curl_off_t as base data type
- add hex and octal parsers to the Curl_str_* family
- make curlx_strtoofft use these parsers
- remove all use of strtol() and strtoul() in library code
- generally use Curl_str_* more than strtoofft, for stricter parsing
- supports 64-bit universally, instead of 'long' which differs in size
between platforms
Extended the unit test 1664 to verify hex and octal parsing.
Viktor Szakats [Fri, 14 Feb 2025 21:38:15 +0000 (22:38 +0100)]
cmake: drop two stray TLS feature checks for wolfSSL
Drop check for `SSL_set0_wbio`, `SSL_CTX_set_srp_username`.
The wolfSSL backend doesn't implement these features. The checks were
wrong, and also missing from `./configure`.
If they get implemented, the feature checks should use distinct macros
from OpenSSL; they should check for the `wolfSSL_`-prefixed APIs via
wolfSSL headers; and matching checks should be added to `./configure`.
Viktor Szakats [Sun, 9 Feb 2025 17:12:14 +0000 (18:12 +0100)]
tidy-up: delete, comment or scope C macros reported unused
To reduce the number `-Wunused-macro` compiler warnings:
- delete unused macros.
- comment out unused macro that are part of a set.
- move macros into the scope they are used.
This may be useful to enable by default, but there are tricky cases that
I didn't manage to fix and paused the effort. E.g. internal features
checks in `openssl.c`. There is more, once those are fixed.
By dropping the unused enum wrappers for `AF_*` macros.
Also fix `./configure` to apply `--enable-werror` options to
`tests/servers`, to catch this next time.
Seen with Intel C compiler:
```
socksd.c(184): warning #188: enumerated type mixed with another type
socksd.c(881): warning #188: enumerated type mixed with another type
[...]
sws.c(76): warning #188: enumerated type mixed with another type
sws.c(229): warning #188: enumerated type mixed with another type
[...]
```
Ref: https://github.com/curl/curl/actions/runs/13296520425/job/37129676921#step:40:338
Viktor Szakats [Thu, 13 Feb 2025 01:56:11 +0000 (02:56 +0100)]
cmake: fix `HAVE_ATOMIC`/`HAVE_STDATOMIC` pre-fill for clang-cl
`HAVE_ATOMIC` and `HAVE_STDATOMIC` is available in clang-cl builds.
Adjust the pre-filled values accordingly.
Detected by a temporary job comparing pre-filled and actual values
on AppVeyor CI:
https://ci.appveyor.com/project/curlorg/curl/builds/51506692/job/2v8qrytgdnlah348#L416
Viktor Szakats [Sun, 9 Feb 2025 04:12:24 +0000 (05:12 +0100)]
addrinfo: add curl macro to avoid redefining foreign symbols
Before this patch curl code was redefining `getaddrinfo` and
`freeaddrinfo` system symbols to plug in its debug wrappers. This was
causing pains to avoid applying the redefinitions to system headers
defining these functions, and to the local debug wrappers. Especially
in unity builds. It also required workarounds for systems where these
symbols are already macros.
Introduce curl-namespaced macros for these functions and use them.
This allows to drop all workarounds and makes it work in all envs,
local targets and unity/bundle combinations.
Viktor Szakats [Wed, 12 Feb 2025 09:35:39 +0000 (10:35 +0100)]
GHA/linux: drop Linux arm job for runner image flakiness with stunnel4
Since last week the Ubuntu arm runner became flaky while installing `stunnel`.
```
08:07:26 Setting up stunnel4 (3:5.72-1build2) ...
08:07:26 Failed to check if group stunnel4 already exists: Connection refused
08:07:26 Group stunnel4 not found.
08:07:28 Reload daemon failed: Failed to activate service 'org.freedesktop.systemd1': timed out (service_start_timeout=25000ms)
08:07:28 Created symlink /etc/systemd/system/multi-user.target.wants/stunnel.target -> /usr/lib/systemd/system/stunnel.target.
08:08:18 Failed to get unit file state for stunnel.target: Connection timed out
08:08:43 Failed to retrieve unit state: Connection timed out
08:08:43 stunnel.target is a disabled or a static unit, not starting it.
08:08:43 /bin/chown: invalid user: ‘stunnel4:stunnel4’
08:08:43 dpkg: error processing package stunnel4 (--configure):
08:08:43 installed stunnel4 package post-installation script subprocess returned error exit status 1
08:08:43 [...]
08:08:47 Errors were encountered while processing:
08:08:47 stunnel4
08:08:54 Error: Timeout was reached
08:08:55 E: Sub-process /usr/bin/dpkg returned an error code (1)
08:08:55 Error: Process completed with exit code 100.
```
Ref: https://github.com/curl/curl/actions/runs/13280736653/job/37078440398?pr=16300#step:2:94
Viktor Szakats [Wed, 12 Feb 2025 09:06:09 +0000 (10:06 +0100)]
GHA/non-native: drop cmake Xcode generator iOS job for poor performance
The Xcode configure phase is slow. In most cases it's >10x slower than
the GNU Makefile configuration. This is after days of attempts to make
it faster. With GNU Makefile it takes 20-30 seconds, while Xcode can
take 3 to 10 minutes, which sometimes makes the job run out of its time
limit and fail.
CMake recommends this generator, but with such poor, and inconsistent
performance, it isn't practical. It helped detecting an Xcode-specific
CMake script issue, but aside from that it's not adding much value.
Viktor Szakats [Sat, 8 Feb 2025 02:20:02 +0000 (03:20 +0100)]
cmake: fix/add missing feature detections for Windows/MS-DOS
Almost all feature detection results are pre-filled on Windows
for performance, so none of the issues fixed here affected builds.
For good measure, this patch add missing detections and fixes others
to make sure they work even when omitting the pre-fill.
It also fixes detecting IPv6 for MS-DOS.
- fix `HAVE_STRUCT_TIMEVAL` detection for MSVC.
Follow-up to c1bc090d65b8d7d14e811dd36f5e8674be43dff3 #12495
- add `HAVE_SOCKADDR_IN6_SIN6_SCOPE_ID` detection for Windows.
- fix `HAVE_STRDUP` detection for MSVC.
- fix `HAVE_SNPRINTF` detection for Windows.
Regression from 8e345057761a8f796403923a96f2c8fd3edca647 #15164
- fix `HAVE_IOCTLSOCKET` detection for non-UWP MSVC.
- exclude `if_nametoindex` detection for Windows.
Although it exists on Windows, detection, usage and availability is
complicated, and curl doesn't use it on this platform.
Regression from 8e345057761a8f796403923a96f2c8fd3edca647 #15164
- move IPv6 detections so that pre-filling and MS-DOS Watt-32
configuration applies to them. This fixes
`HAVE_SOCKADDR_IN6_SIN6_SCOPE_ID` detection with MS-DOS.
Ref: https://github.com/curl/curl/actions/runs/13260511764/job/37015877585#step:7:306
Follow-up to a3585c9576abccddbd27200058912cef900c3c0f #15543
Also:
- add debug option to test without pre-filling.
- replace `NOT LESS` with `GREATER_EQUAL`
Viktor Szakats [Mon, 10 Feb 2025 22:09:55 +0000 (23:09 +0100)]
libssh2: fix to use non-deprecated `libssh2_scp_send64()`
Seen in curl-for-win daily, building against libssh2 1.11.2_DEV:
```
curl-for-win/curl/lib/vssh/libssh2.c:2644:9: warning: 'libssh2_scp_send_ex' is deprecated:
since libssh2 1.2.6. Use libssh2_scp_send64() [-Wdeprecated-declarations]
2644 | SCP_SEND(sshc->ssh_session, sshp->path, data->set.new_file_perms,
```
Ref: https://github.com/curl/curl-for-win/actions/runs/13229370277/job/36924363438#step:3:5805
Viktor Szakats [Mon, 10 Feb 2025 18:39:41 +0000 (19:39 +0100)]
tests: fix test 558, 1330 for MSVC, allow TrackMemory with MSVC in cmake
Extend output filter to pick up backslashes. This makes them pass in CI
when run in the vcpkg MSVC job, for example.
Also:
- cmake: allow TrackMemory, aka `ENABLE_CURLDEBUG` again. Drop FIXME.
- cmake: drop stale TODO.
- runtests: include the word 'TrackMemory' in the message disabling it.
Viktor Szakats [Mon, 10 Feb 2025 04:57:49 +0000 (05:57 +0100)]
cmake: fix `HAVE_GETHOSTBYNAME_R_*` detections with `CURL_WERROR=ON`
Fix `HAVE_GETHOSTBYNAME_R_*` feature detections always failing with
`CURL_WERROR=ON` due to stripping a const.
Also fix the GHA/cmake-vs-configure to enable `CURL_WERROR=ON` to sync
this setting with `./configure` which enables it by default. With that,
CI detects this issue.
Viktor Szakats [Sat, 8 Feb 2025 23:50:07 +0000 (00:50 +0100)]
GHA/windows: improve build perf with cmake unity batches
Default curl unity builds make a single unit for each target. It means
all target sources are batched together and built in a single compiler
invocation. With multi-core CPUs this doesn't always result in the best
possible performance. This patch enables smaller batches for jobs where
this resulted in shorter build times. These jobs are Cygwin, MSYS2,
MinGW, running on the Windows runners.
Use batch of 30 (meaning 30 sources batched into units), and 32 for
Cygwin/MSYS2 to avoid a unity fallout that's subject to a different PR.
(CMake allows to set the number of sources per unit, not the number
of units, though the latter may be more practical to max out CPU cores.)
Also override to not batch the `curlu` target because batching lost
a little bit of time there, due to the already existing parallelism when
building the `testdeps` targets.
For jobs on the macOS and Linux runners jobs were already mostly single
digit or below teen seconds, and batching didn't improve on them
noticeably. On VM jobs, the virtual CPUs are limited, so I didn't
make a try. In AppVeyor and GHA vcpkg jobs (using msbuild), batching
didn't result in conclusive or any gains.
Daniel Stenberg [Mon, 10 Feb 2025 09:15:44 +0000 (10:15 +0100)]
tool_progress: fix percent output of large parallel transfers
When the total transfered amount (upload or download) for parallel
transfers was larger than 2^63/100 bytes (81 petabytes) the progress
percent counter would show wrong.
Jay Satiro [Sat, 8 Feb 2025 08:19:27 +0000 (03:19 -0500)]
curl_global_sslset.md: Add SSL backend names
- Add the names that can be used to select an SSL backend by name.
This change syncs the names in 3 places: curl_global_sslset
documentation, the curl tool environment variable documentation for
CURL_SSL_BACKEND and the libcurl environment variable documentation for
CURL_SSL_BACKEND.
Viktor Szakats [Sun, 9 Feb 2025 01:35:07 +0000 (02:35 +0100)]
lib: silence LibreSSL collision warning on non-MSVC Windows
LibreSSL headers emit this warning because we included `wincrypt.h`
before them. We have to include `wincrypt.h` before OpenSSL headers
to avoid symbol collisions when using other forks. LibreSSL 3.8.2+
offers a macro to silence its warnings to avoid this issue. This patch
sets it.
This allows to stop setting this macro in curl-for-win builds.
Warnings seen with MinGW with cmake non-unity (also unity batch=30):
```
[156/219] Building C object lib/CMakeFiles/libcurl_object.dir/vtls/openssl.c.obj
In file included from lib/vtls/openssl.h:35,
from lib/vtls/openssl.c:53:
dep/libressl-win-x64/include/openssl/ossl_typ.h:90:2: warning: #warning overriding WinCrypt defines [-Wcpp]
90 | #warning overriding WinCrypt defines
| ^~~~~~~
In file included from dep/libressl-win-x64/include/openssl/pem.h:71,
from dep/libressl-win-x64/include/openssl/ssl.h:151,
from lib/vtls/openssl.h:36:
dep/libressl-win-x64/include/openssl/x509.h:108:2: warning: #warning overriding WinCrypt defines [-Wcpp]
108 | #warning overriding WinCrypt defines
| ^~~~~~~
In file included from dep/libressl-win-x64/include/openssl/x509.h:319:
dep/libressl-win-x64/include/openssl/pkcs7.h:77:2: warning: #warning overriding WinCrypt defines [-Wcpp]
77 | #warning overriding WinCrypt defines
| ^~~~~~~
```
Viktor Szakats [Sat, 8 Feb 2025 12:51:59 +0000 (13:51 +0100)]
GHA/windows: restore runtests perf with last known good Git for Windows
Use the last known good release of Git for Windows by installing it
manually. It restores `runtests.pl` performance to the levels before
the October 2024 and this week's fallouts which gradually deployed
MSYS2 runtimes with pipe/signal/concurrency issues.
Also:
- restore vcpkg job's test parallelism to `-j8` (from `-j4`).
- keep using the default shell for jobs not running tests.
To avoid the unnecessary Git for Windows install overhead.
Upsides:
- good performance again.
- easy to experiment with any version.
Downsides:
- installing the Git for Windows package takes 15-30 seconds.
- we're pinned to an old package version.
- no canary to tell when the issue is fixed on the runner images.
Unknown:
- stability. (no MSYS2 runtimes were ever stable and it's difficult
to quantify if a version improves or worsens stability/flakiness, and
intermittent env failures.
Stefan Eissing [Sat, 8 Feb 2025 11:49:34 +0000 (12:49 +0100)]
vtsl: eliminate 'data->state.ssl_scache'
Keeping the relevant 'ssl_scache' in 'data->state' leads to problems
when the owner of the cache is cleaned up and this reference is left
dangling.
Remove the ref entirely and always find the ssl_scache at the current
share or multi.
Folded in #16260 (test 3208) to verify this fixes the bug with a
dangling reference when an easy handle is used with easy_perform first
and in a multi_perform after.
CURLSHOPT_SHARE.md: adjust for the new SSL session cache
curl 8.12.0 introduced an improved SSL session cache. All easy handles
that are added to the same multi handle automatically use the multi
handle's SSL session cache.