Before this patch it returned `CURLE_FAILED_INIT` on init failures, with
the value of 2. Fix it to return `false`.
Seen with clang 18.1.8:
```
../lib/transfer.c(181,12): warning: integer constant not in range of enumerated type 'bool' [-Wassign-enum]
181 | return CURLE_FAILED_INIT;
| ^
../lib/transfer.c(181,12): warning: implicit conversion from enumeration type 'CURLcode' to different enumeration type 'bool' [-Wenum-conversion]
181 | return CURLE_FAILED_INIT;
| ~~~~~~ ^~~~~~~~~~~~~~~~~
../lib/transfer.c(183,12): warning: integer constant not in range of enumerated type 'bool' [-Wassign-enum]
183 | return CURLE_FAILED_INIT;
| ^
../lib/transfer.c(183,12): warning: implicit conversion from enumeration type 'CURLcode' to different enumeration type 'bool' [-Wenum-conversion]
183 | return CURLE_FAILED_INIT;
| ~~~~~~ ^~~~~~~~~~~~~~~~~
```
Follow-up to 35bf766280#14253Closes#16170
As reported in #16166, the STLS hangs with the check for SSL connection
filters, but is working with the old protocol handler way. Revert the
change, although it is unclear why it was no good here.
Fixes#16166
Reported-by: ralfjunker on github
Closes#16172
We don't pursue this, and the necessary `#pragma` got in the way of
compiling curl with gcc 4.2 and older. Drop the logic completely.
Follow-up to 8a266ac488#15939
Reported-by: prpr19xx on Github
Fixes#16152Closes#16157
For ASN.1 tags with indefinite length, curl's own parser for TLS
backends that do not support certificate inspection calls itself
recursively. A malicious server certificate can then lead to high
recursion level exhausting the stack space.
This PR limits the recursion level to 16 which should be safe on all
architectures.
Added unit test 1657 to verify behaviour.
Fixes#16135
Reported-by: z2_
Closes#16137
- replace deprecated `ares_init()` call with `ares_init_options()`.
Follow-up to 0d4fdbf15d#16054
- dedupe `CARES_STATICLIB` initalizations into `curl_setup.h`, to
ensure it's defined before the first (and every) `ares.h` include and
avoid a potential confusion.
- move `CARES_NO_DEPRECATED` from build level to `curl_setup.h`.
To work regardless of build system.
It is necessary because curl calls `ares_getsock()` from two places,
of which one feeds a chain of wrappers: `Curl_ares_getsock()`,
`Curl_resolver_getsock()`, `Curl_resolv_getsock()`.
Closes#16131
Before this patch this macro was used in `vtls/openssl.h` without
setting it first, causing the `keylog_done` member be present in
struct `ossl_ctx` while the code did not use it.
Follow-up to 3210101088#13172Closes#16105
RFC 6455 Section 5.2 notes that for bits RSV1, RSV2, and RSV3 of the
framing header, a non-zero value that is not defined by a negotiated
extension MUST Fail the WebSocket connection.
Test 2310 verifies
Closes#16069
- Use build-time CURL_DEFAULT_SSL_BACKEND as a fallback when environment
variable CURL_SSL_BACKEND contains a backend that is unavailable.
Prior to this change if CURL_SSL_BACKEND was set then
CURL_DEFAULT_SSL_BACKEND was ignored even if the backend of the former
was unavailable. In that case libcurl would instead select the first
available backend in the list of backends.
Bug: https://github.com/curl/curl/issues/16076#issuecomment-2617354254
Reported-by: Jeroen Ooms
Closes https://github.com/curl/curl/pull/16108
- Detach and disconnect an attached connection before performing.
Prior to this change it was not possible to safely reuse an easy handle
with an attached connection in a second call to curl_easy_perform. The
only known case of this is a connect-only type handle where the
connection was detached when curl_easy_perform returned, only to be
reattached by either curl_easy_send/recv.
This commit effectively reverts 2f8ecd5d and be82a360, the latter of
which treated the reuse as an error. Prior to that change undefined
behavior may occur in such a case.
Bug: https://curl.se/mail/lib-2025-01/0044.html
Reported-by: Aleksander Mazur
Closes https://github.com/curl/curl/pull/16008
By better sticking to listing the struct members sorted by size, this
struct is now 48 bytes smaller on my fairly maximized build, without
removing anything.
Turned 'connect_only' into two bits instead of an unsigned char with two
magic values.
Also put the 'gssapi_delegation' field within ifdef HAVE_GSSAPI.
Closes#16097
The h2 filter mistakenly also checked `sendbuf` when asked
about pending data. The call is only meant to account for
buffered data that still needs to be received.
Also, remove obsolete recvbuf in stream as we write received
headers and data directly.
Fixes#16084Closes#16098
Reported-by: Deniz Sökmen
Works better if we later introduce another way to do the asynch RR
resolves (together with the threaded resolver) that does not use c-ares.
Closes#16090
Rework the way `tool_hugehelp.c` is included in builds.
After this patch, with `./configure` and CMake `tool_hugehelp.c` is only
compiled when building with manuals enabled. With manuals disabled this
source file is not used anymore. The method is similar to how
8a3740bc8e implemented `tool_ca_embed.c`.
`./configure` always generates it as before, otherwise the build fails.
- winbuild: rework to not need `buildconf.bat`, but automatically use
`tool_hugehelp.c` if present (e.g. when building from an official
source tarball) and enable `USE_MANUAL` accordingly.
- `buildconf.bat`: after dropping `tool_hugehelp.c` generation, the only
logic left was `cp Makefile.dist Makefile`. This allowed to launch
winbuild builds via GNU Make in a Git repo. Drop this option together
with the batch file.
- build `libcurltool` without `USE_MANUAL` macro to exclude the manual
and the dependence on the generator commands. Drop relying on
`UNITTESTS` for this purpose.
Follow-up to 96843f4ef7#16068
- `src/mkhelp.pl`: include `tool_hugehelp.h` before using `USE_MANUAL`
to have it set in `config-*.h` builds with source tarballs created
with manual but without zlib.
Closes#16081
Allow building with c-ares and yet use threaded resolver for the main
host A/AAAA resolving:
`--with-ares` provides the c-ares install path and defaults to use
c-ares for name resolving
`--with-threaded-resolver` still uses c-ares in the build (for HTTPS)
but uses the threaded resolver for "normal" resolves.
It works similarly for cmake: ENABLE_ARES enables ares, and if
ENABLE_THREADED_RESOLVER also is set, c-ares is used for HTTPS RR and
the threaded resolver for "normal" resolves.
HTTPSRR and c-ares-rr are new features return by curl_version_info() and
thus shown by curl -V.
The c-ares-rr feature bit is there to make it possible to distinguish
between builds using c-ares for all name resolves and builds that use
the threaded resolves for the regular name resolves and c-ares for
HTTPSRR only. "c-ares-rr" means it does not use c-ares for "plain" name
resolves.
HTTPSRR support is EXPERIMENTAL only.
Closes#16054
- no more malloc/free per chunk
- removes the extra malloc entirely
- make the buffer (much) smaller (10MB => 16KB!)
- rename 'decomp' to 'buffer' to clarify purpose
Closes#16079
Adds a `follow()` callback to protocol handlers, so they may decide how
to act on a `newurl` after a request has been done. This is optional.
This moves the HTTP code for handling redirects from multi.c to http.c
where it should be. If we ever add a protocol with its own logic, it
would install its own follow function.
Closes#16075
The variable `conn->httpversion` was used for several purposes and it
was unclear at which time the value represents what.
- rename `conn->httpversion` to `conn->httpversion_seen`
This makes clear that the variable only records the last
HTTP version seen on the connection - if any. And that it
no longer is an indication of what version to use.
- Change Alt-Svc handling to no longer modify `conn->httpversion`
but set `data->state.httpwant` for influencing the HTTP version
to use on a transfer.
- Add `data->req.httpversion_sent` to have a record of what
HTTP version was sent in a request
- Add connection filter type CF_TYPE_HTTP
- Add filter query `CF_QUERY_HTTP_VERSION` to ask what HTTP
filter version is in place
- Lookup filters HTTP version instead of using `conn->httpversion`
Test test_12_05 now switches to HTTP/1.1 correctly and the
expectations have been fixed.
Removed the connection fitler "is_httpN()" checks and using
the version query instead.
Closes#16073
Make transfer attach/detach to/from connections chepaer.
- the "attach" event was no longer implemented by any filter
- the "detach" did the same as the "done" event for the filters
who still implemented it. It should be superfluous as the "done"
must always happen.
Closes#16067
- Prefer Sleep(1) over sched_yield() for pre-Vista thread yield.
On Windows sched_yield is often implemented as Sleep(0) which only
yields to threads of highest priority to current priority. However,
during libcurl initialization if there is thread contention then it's
possible that there is a wait for a different library or OS thread of
a lesser priority and then the yield is not effective during that time.
On the other hand Sleep(1) will wait the minimum time slice which is
usually like 15ms or more.
Prior to this change 2c4bfef removed sched_yield detection on Windows,
which effectively removed the yield in the spin lock, and therefore this
change restores the yield but in a different way.
For Windows Vista and later we use SRW locks and do not have this issue.
Ref: https://github.com/curl/curl/pull/16037#issuecomment-2600161764
Ref: https://devblogs.microsoft.com/oldnewthing/20051004-09/?p=33923
Closes https://github.com/curl/curl/pull/16048
They were more or less the same, but each missed some things the other
had. Windows CE is a subset of Win32, make the headers reflect that and
avoid duplications.
Ref: #15975Closes#16038
Protocol handler option PROTOPT_SSL is used to setup a connection
filters. Once that is done, used `Curl_conn_is_ssl()` to check if
a connection uses SSL.
There may be other reasons to add SSL to a connection, e.g. starttls.
Closes#16034
- setup-win32: drop duplicate `WIN32_LEAN_AND_MEAN` and `NOGDI` macros
(set earlier in `curl_setup.h`)
- config-win32ce: detect ARM for `CURL_OS` macro.
Based on: https://cegcc.sourceforge.net/docs/details.html
It should also apply to MSVC. (untested)
- indentation.
Closes#16029
- Fix a bug in EAGAIN handling when sending frames that led to a
corrupted last byte of the frame sent.
- Restore sanity to curl_ws_send() behaviour:
- Partial writes are reported as OK with the actual number of
payload bytes sent.
- CURLE_AGAIN is only returned when none of the payload bytes
(or for 0-length frames, not all of the frame header bytes)
could be sent.
- curl_ws_send() now behaves like a common send() call.
- Change 'ws-data' test client to allow concurrent send/recv
operations and vary frame sizes and repeat count.
- Add DEBUG env var CURL_WS_CHUNK_EAGAIN to simulate blocking
after a chunk of an encoded websocket frame has been sent.
- Add tests.
Prior to this change data corruption may occur when sending websocket
messages due to two bugs:
1) 3e64569a (precedes 8.10.0) caused a data corruption bug in the last
byte of frame of large messages.
2) curl_ws_send had non-traditional send behavior and could return
CURLE_AGAIN with bytes sent and expect the caller to adjust buffer
and buflen in a subsequent call. That behavior was not documented.
Reported-by: na-trium-144@users.noreply.github.com
Fixes https://github.com/curl/curl/issues/15865
Fixes https://github.com/curl/curl/issues/15865#issuecomment-2569870144
Closes https://github.com/curl/curl/pull/15901
Since c-ares has no function to restore this value to default, it
instead closes and re-inits the c-ares channel.
Reported-by: Deniz Sökmen
Fixes#16015Closes#16016
Improved the filter implementation to be flexible in which order h3 and
h2/h1 are attempted. When HTTPSRR is enabled, look at the ALPNs it found
and use the order given for connecting in default setups.
Closes#16012
When enabled in the build.
Update test2100: verify with HTTPS RR included
Adjust runtests and server/disabled.c to include "HTTPSRR" as a feature
in the test suite.
Also, decode the ALPN list in HTTPS records straight into IDs. There's
no point in storing everything in string format. Skip ALPNs we do not
support.
Closes#16007
This typically happens if CURL_CONNECT_ONLY is used and a second
curl_easy_perform() is attempted.
A connection "taken over" with CURL_CONNECT_ONLY cannot be ended any
other way than a curl_easy_cleanup() on the easy handle that holds it.
Add test 696 to verify.
Closes#16003
