req.no_body was only initialized in Curl_connect, while HTTP/2 server
push adds a duplicated handle via Curl_multi_add_perform and calls
Curl_init_do with conn==NULL, never invoking Curl_connect.
Verify it by amending test 1620
Found by Codex Security
Closes#21194
Up from 1K.
Reduces the risk that someone could flush the list by tricking a user to
do many transfers to new hostnames.
Document the limit.
Follow-up to 03a792b186Closes#21200
To sync names for the same macro logic between lib and src, and to move
it to the curlx namespace, to match `curlx_free()` that it's calling.
Closes#21151
- allow to specify when they are wanted on starting a resolve
- match dns cache entries accordingly. An entry which never
tried to get HTTPS-RRs is no answer for a resolve that wants
it.
- fix late arrivals of resolve answers to match the "async"
records that started them - if it still exists.
- provide for multiple "async" resolves in a transfer at the
same time. We may need to resolve an IP interface while the
main connection resolve has not finished yet.
- allow lookup of HTTPS-RR information as soon as it is
available, even if A/AAAA queries are still ongoing.
For this, the "async" infrastructure is changed:
- Defined bits for DNS queries `CURL_DNSQ_A`, `CURL_DNSQ_AAAA`
and `CURL_DNSQ_HTTPS`. These replace `ip_version` which says
nothing about HTTPS.
Use them in dns cache entries for matching.
- enhance the `async->id` to be a unique `uint32_t` for
resolves inside one multi. This is weak, as the id may
wrap around. However it is combined with the `mid` of
the easy handle, making collisions highly unlikely.
`data->state.async` is only accessed in few places where
the mid/async-id match is performed.
- vtls: for ECH supporting TLS backends (openssl, rustls, wolfssl),
retrieve the HTTPS-RR information from the dns connection filter.
Delay the connect if the HTTPS-RR is needed, but has not
been resolved yet.
The implementation of all this is complete for the threaded
resolver. c-ares resolver and DoH do not take advantage of
all new async features yet. To be done in separate PRs.
Details:
c-ares: cleanup settings and initialisation. Any ares channel
is only being created on starting a resolve and propagating
operations in setopt.c to the channel are not helpful.
Changed threaded+ares pollset handling so that they do not
overwrite each others `ASYNC_NAME` timeouts.
Add trace name 'threads' for tracing thread queue and
pool used by threaded resolver.
Closes#21175
Avoid never-ending growth.
When adding more entries, it now deletes the first entry in the list,
which is the oldest added entry still held in memory. I decided to avoid
a Least Recently Used concept as I suspect with a list with this many
entries most entries have not been used, and we don't save the timestamp
of recent use anyway.
The net effect might (no matter what) be that the removed entry might
feel a bit "random" in the eyes of the user.
Verify with test 1674
Ref #21183Closes#21190
Avoid never-ending growth.
When adding more entries, it now deletes the first entry in the list,
which is the oldest added entry still held in memory. I decided to avoid
a Least Recently Used concept as I suspect with a list with this many
entries most entries have not been used, and we don't save the timestamp
of recent use anyway.
The net effect might (no matter what) be that the removed entry might
feel a bit "random" in the eyes of the user.
Verify with test 1669.
Reported-by: Geeknik Labs
Fixes#21183Closes#21189
New connection filter `cf-dns` that manages DNS queries. If hands
out addresses and HTTPS-RR records to anyone interested. Used by
HTTPS and IP happy eyeballing.
Information may become available *before* the libcurl "dns entry"
is complete, e.g. all queries have been answered. The cf-ip-happy
filter uses this information to start connection attempts as soon
as the first address is available.
The multi MSTATE_RESOLVING was removed. A new connection always
goes to MSTATE_CONNECTING. The connectdata bit `dns_resolved`
indicates when DNS information is complete. This is used for
error reporting and starting the progress meter.
Removed dns entries `data->state.dns[i]`, as the `cf-dns` filter
now keeps the reference now.
Many minor tweaks for making this work and pass address information
around safely.
Closes#21027
- librtmp has no test cases, makes no proper releases and has not had a
single commit within the last year
- librtmp parses the URL itself and requires non-compliant URLs for this
- we have no RTMP tests
- RTMP was used by 2.2% of curl users (self-identified in the 2025
survey)
Closes#20673
- Rename `Curl_resolv_unlink()` to `Curl_dns_entry_unlink()`.
- Change `Curl_dnscache_get()` to return CURLcode result. Returns
now `CURLE_COULDNT_RESOLVE_HOST` for "negative" cache entries.
- Add `Curl_dnscache_add_negative()` to put a "negative" entry
into the cache.
Closes#20864
It had an 'f' too few. Also provide CURLPROTO_WS* unconditionally
internally, so that code can depend on them in all builds.
Follow-up to cd5ca80f00
Spotted by Codex Security
Test case 3219 added to catch this next time.
Closes#21031
When composing the <scheme>_proxy environment variable, we assume the
handler->scheme name is already lowercase.
This makes unit test 1627 verify that is the case.
Follow-up to c294f9cb56
Spotted by Codex Security
Closes#21033
- move defines to header file
- make bit2str require < 8 unused bits
- make bool strings stricter
- make UTime2str show + or - for custom time zones
- removed unused 'type' argument to ASN1tostr() function
- fix int2str for negative values. All values below 10000 are now shown
in decimal properly, also possibly negative values.
Add unit test 1667 to verify ASN1tostr
Closes#21013
- make extract-unit-protos handle multi-line prototypes - but they need
to be above the implementation
- Prototypes for static functions we use in unit tests should not be in
header files. We generate lib/unitprotos.h for this purpose
- Removed some function wrappers written for unit tests and make them
use UNITTEST function directly.
- Renamed time2str() in the tool to timebuf() since we have the same
name in lib/ and in unit tests they can both be used non-static in a
build.
This reverts commit f95fadd116.
Follow-up to #21010Closes#21014
- return error on zero length input
- return error on OOM or doing too large output
- fix full 32-bit number support
- fix the broken handling of the first and second numbers
- support up to 32-bit minus 80 for the second number
- a field with a leading 0x80 is now considered an error, since it only
works as padding and is then no longer the shortest possible version
Add unit tests in 1666
Bonus: removed the last argument to OID2str() as it was always set TRUE.
Closes#21003
Extend `#pragma diagnostic push`/`pop` guards to the whole codebase
(from tests and examples only) to disable it for GCC <4.6. Rename guard
to `CURL_HAVE_DIAG` and make it include llvm/clang to be interchangeable
with `__GNUC__ || __clang__` in this context.
The above means no longer disabling certain warnings locally, so pair
this with disabling all picky warnings for GCC <4.6.
Also:
- drop global workarounds for misbehaving GCC <4.6 compiler warnings.
Not needed with picky warnings disabled.
Reported-by: fds242 on github
Reported-by: Sergey Fedorov
Thanks-to: Orgad Shaneh
Follow-up to f07a98ae11#20366Fixes#20892Fixes#20924Closes#20902Closes#20907
Paths starting with one or two leading dots but without a following
slash were not handled correctly.
Follow-up to c31dd6631f
Extended test 1395 accordingly with a set of new test string.
Reported by Codex Security
Closes#20974
- tool_getparam: revert an unnecessary/no-op C89 warning silencer.
Follow-up to 09c9afdd71#20363
- tool_writeout: add comment saying silencing is a no-op for llvm/clang.
For `strftime()` it is a GCC-specific, as of llvm/clang v22.1.0.
Follow-up to f07a98ae11#20366
- unit1652: drop always-false `!defined(__clang__)` guard.
Pointed-out-by: Orgad Shaneh
Ref: #20902
Follow-up to 7e814c8717#16062
- unit1652: document that `-Wformat` is necessary for GCC v5 to v8.
Follow-up to 71cf0d1fca#14772Closes#20908
It's mostly a filler word. I've read through each use of it in the code
base and did minor rephrasings when "simply" carried some meaning. The
overwhelming majority of cases, removing it improved the text
significantly. Inspired by #20793.
Closes#20822
To allow building test binaries without test certs, replacing
`_CURL_SKIP_BUILD_CERTS` internal option with a build target that is
similar to `testdeps`, but without building the test certificates.
To make building test binaries a little bit faster, with less noisy
output, and without having to reconfigure the build.
Closes#20708
Tests are build in "unity"-style, by including sources into an umbrella
C files (similar to how CMake unity works). This does not play well with
clang-tidy, which seems to unconditionally ignore C sources included
like this. To fix it, curl's CMake implements a manual clang-tidy
support for tests, which compiles sources one-by-one, while also making
sure sources compile cleanly standalone (e.g. all sources need to
include `first.h`). The manual clang-tidy implementation is fragile, and
performance, in particular when targeting Windows, is abysmal.
This patch introduces an alternate solution, enabled by the
`_CURL_TESTS_CONCAT=ON` option. In this mode, umbrella sources include
the actual sources instead of `#including` them. Allowing to use CMake's
built-in clang-tidy support to compile them, with clang-tidy actually
checking the sources. Making the manual clang-tidy support unnecessary.
In the Windows CI job it results in a 4x performance improvement (4m ->
1m), making it practical to run clang-tidy on tests on Windows, in CI.
The main downside is that clang-tidy doesn't understand the `#line`
directive. Meaning issues found show the wrong filename and line number
next to them. It's not impossible to locate errors this way, but also
not convenient.
Minor/potential downside is that the concatenated source needs to be
reassembled each time an original source is updated. This may result in
more copying on the disk when used in local development. The largest
source is 1.4MB, so probably not a show-stopper on most machines.
Another is the complexity of maintaining two methods in parallel, which
may be necessary till clang-tidy understands `#line`:
https://github.com/llvm/llvm-project/issues/62405
This solution may in theory also enable adding clang-tidy support for
tests in autotools, though I haven't tried.
Targeted for curl CI for now, and used in a GHA/windows job. 100%
experimental, not recommended outside these.
Closes#20667
- vms/curlmsg_vms.h: delete unused/commented code.
- vtls/schannel_verify: sort includes.
- typecheck-gcc.h: fix indent and alignment.
- lib/config-win32.h: drop idle `#undef`.
- spacecheck: check for stray empty lines before after curly braces.
- make literals more readable: 1048576 -> 1024 * 1024
- scope variables.
- use ISO date in a comment.
- drop redundant parentheses.
- drop empty comments.
- unfold lines.
- duplicate/stray spaces in comments.
- fix indent, whitespace, minor typos.
Closes#20690
For consistency.
Also:
- one remaining in `src/tool_writeout.c`.
- replace casting an `int` to `CURLcode`.
- lib758: rename `CURLMcode` `result` to `mresult`.
- move literals to the right side of if expressions.
Follow-up to d0dc6e2ec0#20426
Follow-up to 56f600ec23Closes#20432
And a few variables around.
There remain cases where the accepted pointer is const, yet the returned
pointer is written to.
Partly addressing (glibc 2.43):
```
* For ISO C23, the functions bsearch, memchr, strchr, strpbrk, strrchr,
strstr, wcschr, wcspbrk, wcsrchr, wcsstr and wmemchr that return
pointers into their input arrays now have definitions as macros that
return a pointer to a const-qualified type when the input argument is
a pointer to a const-qualified type.
```
Ref: https://lists.gnu.org/archive/html/info-gnu/2026-01/msg00005.html
Reported-by: Rudi Heitbaum
Ref: #20420Closes#20421
When the compiler supports C99.
- map logging functions to macro stubs when verbose logging is disabled
and the compiler is C99. Make sure these stubs silence unused variable
warnings for non-variadic arguments.
Before this patch they mapped to function stubs, the same codepath
used for C89 compiler in this configuration.
- introduce new macros to tell the compiler which code to include
when verbose code is active, or inactive:
- `CURLVERBOSE`: defined when verbose code is active.
To enclose blocks of code only used for verbose logging.
- `VERBOSE(statement);`:
compile statement when verbose code is active.
To mark code lines only used for verbose logging.
- `NOVERBOSE(statement);`:
compile statement when verbose code is inactive.
To suppress warnings for arguments passed to logging functions via
printf masks, e.g. `NOVERBOSE((void)ipaddress);`, yet keeping
the warning in verbose builds.
Note these macros are not the same as `CURL_DISABLE_VERBOSE_STRINGS`.
Verbose code is always active in C89 mode (without variadic macro
support).
- drop existing uses of `CURL_DISABLE_VERBOSE_STRINGS` where redundant,
or replace with the above macros. Ending up reducing the number of
`#ifdef`s, and also the number of lines.
Assisted-by: Daniel Stenberg
Assisted-by: Jay Satiro
Reported-by: Dan Fandrich
Fixes#20341
Refs: #12105#12167Closes#20353
When checking a transfer for being expired via `Curl_timeleft_ms()`,
eleminate the `bool connecting` parameter and have the function check
the `mstate` of the transfer instead.
Advantages:
* eleminate the caller needing awareness if the transfer is
connecting or in a later state
* fix pingpong timeout handling to check the correct timeout
during "proto_connect" phases
* avoid using "connecting" timeouts during establishing a secondary
connection (e.g. FTP) since this would use the timestamp from
the original, primary connect and thus be wrong
Reported-by: Wyuer on github
Fixes#20347Closes#20354
Drop separate `TrackMemory` (aka `CURLDEBUG`) debug feature.
After recent changes (thread-safety,
193cb00ce9, and updates leading up to
it), `TrackMemory` is unlikely to cause build or runtime issues.
To simplify builds and debug options, enable `TrackMemory`
unconditionally for debug-enabled (aka `DEBUGBUILD`) builds. Before
this patch, this was already the default, with an option to disable
it, or enable it in non-debug-enabled builds.
Note, in practice these two debug options already went hand in hand. It
was not possible to toggle them separately for a long time due to bugs,
before 59dc9f7e69 (2024-05-28) fixed it.
This patch also removes/deprecates separate knobs and feature flags for
`TrackMemory`:
- autotools: `--enable-curldebug`/`--disable-curldebug`
- cmake: `-DENABLE_CURLDEBUG=ON`/`OFF`
- C macro: `CURLDEBUG`
- libcurl: `CURL_VERSION_CURLDEBUG` symbol deprecated in favor
of `CURL_VERSION_DEBUG`. They always return the same value after this
patch.
Also:
- drop `TrackMemory` from `curl -V` output.
- rename internal `CURLDEBUG` macro to `CURL_MEMDEBUG` internally.
To avoid confusion with `DEBUGBUILD`, but to keep guarding
`TrackMemory`-related internals for readability.
- runtests: bind `TrackMemory` to debug feature. Keep it a separate
test feature requirement, for clarity.
- CI: drop test builds for combinations of the two options.
- GHA/linux: no longer disable TrackMemory in the TSAN job.
Ref: https://github.com/curl/curl/pull/20328#issuecomment-3754528407Closes#20331
