cirrus: Add FreeBSD 13.0 job and disable sanitizer build
As alluded to the in the now removed comment, a 13.0 image became
available and is now ready to be used.
The sanitizer builds were running on the 12.1 image which since has
been removed from the config, leaving the builds not running at all.
When enabled it turns out that they don't actually work due to very
long timeouts in executing the tests, so keep the disabled for now
but a bit more controlled.
MAntoniak [Tue, 17 Aug 2021 16:40:25 +0000 (18:40 +0200)]
mbedtls: avoid using a large buffer on the stack
Use dynamic memory allocation for the buffer used in checking "pinned
public key". The PUB_DER_MAX_BYTES parameter with default settings is
set to a value greater than 2kB.
Daniel Stenberg [Sat, 4 Sep 2021 21:41:07 +0000 (23:41 +0200)]
configure: make --disable-hsts work
The AC_ARG_ENABLE() macro itself uses a variable called
'enable_[option]', so when our script also used a variable with that
name for the purpose of storing what the user wants, it also
accidentally made it impossible to switch off the feature with
--disable-hsts. Fix this by renaming our variable.
Reported-by: Michał Antoniak
Fixes #7669
Closes #7672
Daniel Stenberg [Wed, 1 Sep 2021 08:26:36 +0000 (10:26 +0200)]
opt-docs: verify man page sections + order
In every libcurl option man page there are now 8 mandatory sections that
must use the right name in the correct order and test 1173 verifies
this. Only 14 man pages needed adjustments.
The sections and the order is as follows:
- NAME
- SYNOPSIS
- DESCRIPTION
- PROTOCOLS
- EXAMPLE
- AVAILABILITY
- RETURN VALUE
- SEE ALSO
Daniel Stenberg [Wed, 1 Sep 2021 07:23:37 +0000 (09:23 +0200)]
opt-docs: make sure all man pages have examples
Extended manpage-syntax.pl (run by test 1173) to check that every man
page for a libcurl option has an EXAMPLE section that is more than two
lines. Then fixed all errors it found and added examples.
Daniel Stenberg [Thu, 2 Sep 2021 13:57:25 +0000 (15:57 +0200)]
connect: get local port + ip also when reusing connections
Regression. In d6a37c23a3c (7.75.0) we removed the duplicated storage
(connection + easy handle), so this info needs be extracted again even
for re-used connections.
Daniel Stenberg [Tue, 31 Aug 2021 14:37:14 +0000 (16:37 +0200)]
curl.1: provide examples for each option
The file format for each option now features a "Example:" header that
can provide one or more examples that get rendered appropriately in the
output. All options MUST have at least one example or gen.pl complains
at build-time.
This fix also does a few other minor format and consistency cleanups.
Daniel Stenberg [Fri, 27 Aug 2021 15:42:40 +0000 (17:42 +0200)]
http_proxy: only wait for writable socket while sending request
Otherwise it would wait socket writability even after the entire CONNECT
request has sent and make curl basically busy-loop while waiting for a
response to come back.
The previous fix attempt in #7484 (c27a70a591a4) was inadequate.
Daniel Stenberg [Thu, 26 Aug 2021 05:43:34 +0000 (07:43 +0200)]
curl: better error message when -O fails to get a good name
Due to how this currently works internally, it needs a working initial
file name to store contents in, so it may still fail even with -J is
used (and thus accepting a name from content-disposition:) if the file
name part of the URL isn't "good enough".
Don [Tue, 24 Aug 2021 18:27:39 +0000 (11:27 -0700)]
cmake: sync CURL_DISABLE options
Adds the full listing of CURL_DISABLE options to the CMake build. Moves
all option code, except for CURL_DISABLE_OPENSSL_AUTO_LOA_CONFIG which
resides near OpenSSL configuration, to the same block of code. Also
sorts the options here and in the cmake config header.
Additionally sorted the CURL-DISABLE listing and fixed the
CURL_DISABLE_POP3 option.
Ehren Bendler [Thu, 19 Aug 2021 17:45:55 +0000 (13:45 -0400)]
wolfssl: clean up wolfcrypt error queue
If wolfSSL is built in certain ways (OPENSSL_EXTRA or Debug), the error
queue gets added on to for each session and never freed. Fix it by
calling ERR_clear_error() like in vtls/openssl when needed. This func is
a no-op in wolfcrypt if the error queue is not enabled.
Don [Fri, 20 Aug 2021 16:29:10 +0000 (09:29 -0700)]
curl_setup.h: sync values for HTTP_ONLY
The values for HTTP_ONLY differed between CMakeLists.txt and
curl_setup.h. Sync them and sort the values in curl_setup.h to make it
easier to spot differences.
Jay Satiro [Tue, 17 Aug 2021 06:24:48 +0000 (02:24 -0400)]
configure: set classic mingw minimum OS version to XP
- If the user has not specified a minimum OS version (via WINVER or
_WIN32_WINNT macros) then set it to Windows XP.
Prior to this change classic MinGW defaulted the minimum OS version
to Windows NT 4.0 which is way too old. At least Windows XP is needed
for getaddrinfo (which resolves hostnames to IPv6 addresses).
Jay Satiro [Tue, 17 Aug 2021 06:17:21 +0000 (02:17 -0400)]
schannel: Work around typo in classic mingw macro
- Define ALG_CLASS_DHASH (the typo from the include) to ALG_CLASS_HASH.
Prior to this change there was an incomplete fix to ignore the
CALG_TLS1PRF macro on those versions of MinGW where it uses the
ALG_CLASS_DHASH typoed macro.
Marc Hoersken [Tue, 17 Aug 2021 19:16:41 +0000 (21:16 +0200)]
tests/*server.pl: flush output before executing subprocess
Also avoid shell processes staying around by using exec.
This is necessary to avoid output data being buffering
inside the process chain of Perl, Bash/Shell and our
test server binaries. On non-Windows systems the exec
will also make the subprocess replace the intermediate
shell, but on Windows it will at least bind the processes
together since there is no real fork or exec available.
Daniel Stenberg [Tue, 17 Aug 2021 14:52:55 +0000 (16:52 +0200)]
openssl: when creating a new context, there cannot be an old one
Remove the previous handling that would call SSL_CTX_free(), and instead
add an assert that halts a debug build if there ever is a context
already set at this point.
Marc Hoersken [Sun, 15 Aug 2021 10:40:54 +0000 (12:40 +0200)]
tests/server/*.c: align handling of portfile argument and file
1. Call the internal variable portname (like pidname) everywhere.
2. Have a variable wroteportfile (like wrotepidfile) everywhere.
3. Make sure the file is cleaned up on exit (like pidfile).
4. Add parameter --portfile to usage outputs everywhere.
Reviewed-by: Daniel Stenberg
Replaces #7523
Closes #7574
Jay Satiro [Thu, 29 Apr 2021 21:06:49 +0000 (17:06 -0400)]
tool_operate: Fix --fail-early with parallel transfers
- Abort via progress callback to fail early during parallel transfers.
When a critical error occurs during a transfer (eg --fail-early
constraint) then other running transfers will be aborted via progress
callback and finish with error CURLE_ABORTED_BY_CALLBACK (42). In this
case, the callback error does not become the most recent error and a
custom error message is used for those transfers:
Patrick Monnerat [Mon, 16 Aug 2021 06:35:22 +0000 (08:35 +0200)]
auth: do not append zero-terminator to authorisation id in kerberos
RFC4752 Section 3.1 states "The authorization identity is not terminated
with a zero-valued (%x00) octet". Although a comment in code said it may
be needed anyway, nothing confirms it. In addition, servers may consider
it as part of the identity, causing a failure.
The signal handler registered for SIGINT is only handling SIGINT
so there isn't much need for inspecting the signo. While there,
rename the handler to be more specific.
g_should_exit should really be of sig_atomic_t type, but relying
on autoconf in the examples seems like a bad idea so keep that
for now.
projects / thirdparty / curl.git / log
