Tweaks around handling of --limit-rate:
* tracing: trace outstanding timeouts by name
* multi: do not mark transfer as dirty that have
an EXPIRE_TOOFAST set
* multi: have one static function to asses speed limits
* multi: when setting EXPIRE_TOOFAST remove the transfers
from the dirty set
* progress: rename vars and comment on how speed limit
timeouts are calculated, for clarity
* transfer: when speed limiting, exit the receive loop
after a quarter of the limit has been received, not
on the first chunk received.
* cf-ip-happy.c: clear EXPIRE_HAPPY_EYEBALLS on connect
* scorecard: add --limit-rate parameter to test with
speed limits in effect
Closes#18454
Only actually needed servers should be listed and none is then implied
if no servers are listed.
Outputs a warning if "none" is still set as a server.
Closes#18466
If a string argument is expected and the first two bytes are 0xe2 ex80
and the third has the 7th bit set, that's enough for curl to warn.
Previously we tried to detect and warn only for the unicode double
quote, but users might use single quotes, other quotes or even lead the
argument with one of the "zero widths" characters. This is an attempt to
detect many of those. Without triggering for "normal" IDN hostnames.
Closes#18459
Split out adding of individual request headers into a switch. Check
the connection http version only on fresh connections, use separate
methods.
Add TE: header directly without allocation. Add bit for indicating
Connection: header has been added and custom headers should not do
that again.
Closes#18444
test_10_08, uploading larger files for a h2 proxy, sporadically fails
with a decrpytion error on received data in AWS-LC. The frequency can
be increased by simulated network receive blocks.
Not setting a 4 * TLS record sized buffer, leaving AWS-LC at its
default buffer size seems to mitigate this problem.
Closes#18434
This allows --remote-time to set dates before 1970.
Due to a minor omission in the API, it will still avoid setting the time
if it is indeed exactly epoch 0 (jan 1 1970).
Verified by test 762
Fixes#18424
Reported-by: Terence Eden
Closes#18443
- using {} with single entries makes little sense
- when using {} sets with two entry lists, there can only be 64 to reach
maximum number of URLs
Verify the max check in test 761
Write out 9-length frames to client's WRITEFUNCTION
Read 0-length frames from READFUNCTION *if* the function
started a new frame via `curl_ws_start_frame()`.
Fixes#18286Closes#18332
Reported-by: Andriy Druk
Every time we first run runtests with -j and then again *without* -j,
this message was shown:
"Warning: $runnerid: cleardir(log) failed"
Not anymore.
Closes#18404
If --trace-config is used to set a level before -v is used, don't reset
the state on first -v (to "-all") as it otherwise does. This way,
--trace-config can be used to set specific trace items before -v on the
command line and it still works.
Previously, the first -v use would otherwise reset and undo the earlier
--trace-config items.
Fixes#18346Closes#18361
- schannel: apply BoringSSL workaround to AWS-LC too.
Affects Schannel + AWS-LC MultiSSL builds. (not tested in CI)
Ref: 274940d743#2643#2634
- curl_ntlm_core: deduplicate macro defines.
- curl_ntlm_core: document version thresholds for an AWS-LC-specific
workaround.
It was necessary between v1.2.0 2022-09-01 and v1.30.1 2024-06-21.
No longer necessary since v1.31.0 2024-07-01:
ba94617d99
Follow-up to 34ef4fab22#10320
- lib758: drop redundant OpenSSL version guards.
`OPENSSL_VERSION_NUMBER > 3` automatically guards against LibreSSL,
BoringSSL and AWS-LC.
Ref: 6ddd8f2c0b
Follow-up to a5f0ab7995#18288
- dllmain, curl_sha512_256: formatting.
Closes#18387
To avoid NULL derferences with many outputs and --next.
Follow-up to 034612cd51
Add test 760 to verify
Reported-by: BobodevMm on github
Fixes#18375Closes#18376
... even if there are tests still pending. To help visualize tests that
might be hanging.
Attempts to set the limit at 10 seconds of silence.
Closes#18349
Changed strategy to start up and terminate resolver thread.
When starting up:
Start the thread with mutex acquired, wait for signal from thread that
it started and has incremented the ref counter. Thread set
pthread_cancel() to disabled before that and only enables cancelling
during resolving itself. This assure that the ref counter is correct and
the unlinking of the resolve context always happens.
When shutting down resolving:
If ref counting shows thread has finished, join it, free everything. If
thread has not finished, try pthread_cancel() (non Windows), but keep
the thread handle around.
When destroying resolving:
Shutdown first, then, if the thread is still there and 'quick_exit' is
not set, join it and free everything. This might occur a delay if
getaddrinfo() hangs and cannot be interrupted by pthread_cancel().
Destroying resolving happens when another resolve is started on an
easy handle or when the easy handle is closed.
Add test795 to check that connect timeout triggers correctly
when resolving is delayed. Add debug env var `CURL_DNS_DELAY_MS`
to simulate delays in resolving.
Fix test1557 to set `quick_exit` and use `xxx.invalid` as domain
instead of `nothing` that was leading to hangers in CI.
Closes#18263
`runtests.pl` defaults to `-` if a command is not set, since
c43ad0f972.
Also:
- drop a stray `</file>`.
- replace a `nothing` with a guaranteed invalid hostname.
Ref: 4334033b43
- replace unused URLs with `-`.
Closes#18319
`Win32::Process::List` and `Win32::Process`.
To replace external calls to `tasklist.exe` and `taskkill.exe`.
The perl modules are wrappers/binding to Win32 API calls. They avoid
launching external processes with a command shell (including MSYS2),
the external tool, and creating command-lines and parsing tool output.
According to local tests and the CI, one test session calls
`tasklist.exe` 350-400 times. `taskkill.exe` is rarely called:
https://github.com/curl/curl/actions/runs/17012376726?pr=18296
It's hard to predict any possible side-effect of dynamically loading
the two necessary, module DLLs into the Perl process. The MSYS2 runtime
if prone to fail when doing this in fork operations, as seen earlier
with the Win32.DLL module. But, is looks like a symptom, not the root
cause for these failures, because the failures are present with or
without perl.exe loading the Win32.DLL.
Ref: be01b60ce5#18287
Cherry-picked from #18296Closes#18308
`CMAKE_C_FLAGS` is not set by curl, but may contain custom options
required for a successful compiler run, when invoked by these tests.
One such case is when configuring Visual Studio or clang-cl via compiler
options, instead of envs.
Cherry-picked from #18301Closes#18307
When an application install its own OpenSSL verify callback and that
callback invokes `SSL_set_retry_verify()`, the transfer is automatically
paused and does not progress the connect attempt any further until
unpaused via `curl_easy_pause().
Added test758 to verify.
Ref: #18284
Original PR by @Natris
Bug: https://curl.se/mail/lib-2025-08/0012.htmlCloses#18288
Easily triggered when doing an FTP ranged download with the end range
larger than the file size.
A regression since 30de937bda (shipped in curl 8.9.0)
Add test 2307 to verify.
Reported-by: kupavcevdenis on github
Fixes#18283Closes#18295
`Time::HiRes` was already used unconditionally before this patch in
`servers.pm`. This package, and functions used by runtests (`sleep` and
`gettimeofday`) are supported by the minimum Perl version required for
curl:
https://perldoc.perl.org/5.8.0/Time::HiRes
- Drop the `portable_sleep()` wrapper in favor of `Time::HiRes::sleep()`.
- Use `Time::HiRes` unconditionally in `serverhelp.pm`.
- Stop using the `Win32` package where available. It was included
to provide a Windows fallback for `Time::HiRes::sleep()`. It was never
actually called, but the dependency may have loaded `Win32.dll`, which
often appears in failed fork operations in GHA logs.
Ref: a6fed41f6f#5054#5034
Ref: https://github.com/curl/curl/discussions/14854Closes#18287
To not create a file named `nul` in the `tests` directory when running
tests with MSYS2/Cygwin Perl (the only supported Perl for Windows).
I imagine this may cause issues when being written in parallel. Also
`nul` is a reserved filename on Windows.
Closes#18282
Add support for CURLOPT_READFUNCTION with WebSocket urls when *not* in
connect-only mode, e.g. when using curl_multi_perform.
Install the callback function and set CURLOPT_UPLOAD. Return
CURL_READFUNC_PAUSE when having nothing more to send and unpause the
transfer when more data is ready.
This will send the read bytes in a WebSocket BINARY frame.
Add support for this mode in the pytest "ws_data" client and have all
tests run in 'curl_ws_send/recv' and 'peform' mode as well.
Add `curl_ws_start_frame()`. Document, cover in libcurl-ws.md and
explain the READFUNCTION mode for websockets.
Add example `websocket-updown` for this.
Closes#17683
in the user home directory.
Before this patch, the curl tool found the system curlrc first, ignoring
the custom one set by the test via `CURL_HOME`.
Closes#18242
Instead of doing malloc/free every time and "manual" buffer population.
This reuses the dynbuf for this purpose for the duration of parsing the
file.
Closes#18230
The GlobalConfig only exists in a single instance and it has worked like
this since the dawn of time. It is about time we stop passing around
pointers to what was already essentially a global object and instead
just use a... global.
It simplifies things.
Closes#18213
By micromanaging the project dependency and its inclusion into the test
project. It feels like an awkward construct, but perhaps better than
nothing.
It's also fragile because it's a static build with no assistance from
the external project (curl in this case). Mitigated in test by disabling
all dependencies and some features.
Since there is no special core cmake logic to be tested here, in CI
the test is tested really. To keep CI jobs at minimum, only add 3 of
them, taking 42s in total. (All 6 would take 270s.)
Follow-up to e2a23d5d0d#17203Closes#18208
curl_easy_reset() did not reset the `rewind_read` flag. This caused any
handles that previously had a CURLE_SEND_FAIL_REWIND error to get stuck
with that error, failing any subsequent requests, even if they didn't
have any body at all.
Verified in test 3034
Fixes#18206Closes#18207
