... by making sure we don't count down the "upload left" counter when the
uploaded size is unknown and then it can be allowed to continue forever.
Fixes#996
Since we're using CURLE_FTP_WEIRD_SERVER_REPLY in imap, pop3 and smtp as
more of a generic "failed to parse" introduce an alias without FTP in
the name.
Closes https://github.com/curl/curl/pull/975
This hash is used to verify the original downloaded certificate bundle
and also included in the generated bundle's comment header. Also
rename related internal symbols to algorithm-agnostic names.
CURLINFO_SSL_VERIFYRESULT does not get the certificate verification
result when SSL_connect fails because of a certificate verification
error.
This fix saves the result of SSL_get_verify_result so that it is
returned by CURLINFO_SSL_VERIFYRESULT.
Closes https://github.com/curl/curl/pull/995
While noErr and errSecSuccess are defined as the same value, the API
documentation states that SecPKCS12Import() returns errSecSuccess if
there were no errors in importing. Ensure that a future change of the
defined value doesn't break (however unlikely) and be consistent with
the API docs.
Speed limits (from CURLOPT_MAX_RECV_SPEED_LARGE &
CURLOPT_MAX_SEND_SPEED_LARGE) were applied simply by comparing limits
with the cumulative average speed of the entire transfer; While this
might work at times with good/constant connections, in other cases it
can result to the limits simply being "ignored" for more than "short
bursts" (as told in man page).
Consider a download that goes on much slower than the limit for some
time (because bandwidth is used elsewhere, server is slow, whatever the
reason), then once things get better, curl would simply ignore the limit
up until the average speed (since the beginning of the transfer) reached
the limit. This could prove the limit useless to effectively avoid
using the entire bandwidth (at least for quite some time).
So instead, we now use a "moving starting point" as reference, and every
time at least as much as the limit as been transferred, we can reset
this starting point to the current position. This gets a good limiting
effect that applies to the "current speed" with instant reactivity (in
case of sudden speed burst).
Closes#971
* Added description to Curl_sspi_free_identity()
* Added parameter and return explanations to Curl_sspi_global_init()
* Added parameter explaination to Curl_sspi_global_cleanup()
With HTTP/2 each transfer is made in an indivial logical stream over the
connection, making most previous errors that caused the connection to get
forced-closed now instead just kill the stream and not the connection.
Fixes#941
... instead of if() before the switch(), add a default to the switch so
that the compilers don't warn on "warning: enumeration value
'PLATFORM_DONT_CARE' not handled in switch" anymore.
- Disable ALPN on Wine.
- Don't pass input secbuffer when ALPN is disabled.
When ALPN support was added a change was made to pass an input secbuffer
to initialize the context. When ALPN is enabled the buffer contains the
ALPN information, and when it's disabled the buffer is empty. In either
case this input buffer caused problems with Wine and connections would
not complete.
Bug: https://github.com/curl/curl/issues/983
Reported-by: Christian Fillion
Serialise the call to PK11_FindSlotByName() to avoid spurious errors in
a multi-threaded environment. The underlying cause is a race condition
in nssSlot_IsTokenPresent().
Bug: https://bugzilla.mozilla.org/1297397Closes#985
When we're uploading using FTP and the server issues a tiny pause
between opening the connection to the client's secondary socket, the
client's initial poll() times out, which leads to second poll() which
does not wait for POLLIN on the secondary socket. So that poll() also
has to time out, creating a long (200ms) pause.
This patch adds the correct flag to the secondary socket, making the
second poll() correctly wait for the connection there too.
Signed-off-by: Ales Novak <alnovak@suse.cz>
Closes#978
Instead of displaying the requested hostname the one returned
by the SOCKS5 proxy server is used in case of connection error.
The requested hostname is displayed earlier in the connection sequence.
The upper-value of the port is moved to a temporary variable and
replaced with a 0-byte to make sure the hostname is 0-terminated.
Hooked up the HTTP authentication layer to query the new 'is mechanism
supported' functions when deciding what mechanism to use.
As per commit 00417fd66c existing functionality is maintained for now.
Hooked up the SASL authentication layer to query the new 'is mechanism
supported' functions when deciding what mechanism to use.
For now existing functionality is maintained.
As Windows SSPI authentication calls fail when a particular mechanism
isn't available, introduced these functions for DIGEST, NTLM, Kerberos 5
and Negotiate to allow both HTTP and SASL authentication the opportunity
to query support for a supported mechanism before selecting it.
For now each function returns TRUE to maintain compatability with the
existing code when called.
I discovered some people have been using "https://example.com" style
strings as proxy and it "works" (curl doesn't complain) because curl
ignores unknown schemes and then assumes plain HTTP instead.
I think this misleads users into believing curl uses HTTPS to proxies
when it doesn't. Now curl rejects proxy strings using unsupported
schemes instead of just ignoring and defaulting to HTTP.
Undo change introduced in d4643d6 which caused iPAddress match to be
ignored if dNSName was present but did not match.
Also, if iPAddress is present but does not match, and dNSName is not
present, fail as no-match. Prior to this change in such a case the CN
would be checked for a match.
Bug: https://github.com/curl/curl/issues/959
Reported-by: wmsch@users.noreply.github.com
Mark's new document about HTTP Retries
(https://mnot.github.io/I-D/httpbis-retry/) made me check our code and I
spotted that we don't retry failed HEAD requests which seems totally
inconsistent and I can't see any reason for that separate treatment.
So, no separate treatment for HEAD starting now. A HTTP request sent
over a reused connection that gets cut off before a single byte is
received will be retried on a fresh connection.
Made-aware-by: Mark Nottingham
Makes libcurl work in communication with gstreamer-based RTSP
servers. The original code validates the session id to be in accordance
with the RFC. I think it is better not to do that:
- For curl the actual content is a don't care.
- The clarity of the RFC is debatable, is $ allowed or only as \$, that
is imho not clear
- Gstreamer seems to url-encode the session id but % is not allowed by
the RFC
- less code
With this patch curl will correctly handle real-life lines like:
Session: biTN4Kc.8%2B1w-AF.; timeout=60
Bug: https://curl.haxx.se/mail/lib-2016-08/0076.html
All compilers used by cmake in Windows should support large files.
- Add test SIZEOF_OFF_T
- Remove outdated test SIZEOF_CURL_OFF_T
- Turn on USE_WIN32_LARGE_FILES in Windows
- Check for 'Largefile' during the features output
Since the server can at any time send a HTTP/2 frame to us, we need to
wait for the socket to be readable during all transfers so that we can
act on incoming frames even when uploading etc.
Reminded-by: Tatsuhiro Tsujikawa
In order to make MBEDTLS_DEBUG work, the debug threshold must be unequal
to 0. This patch also adds a comment how mbedtls must be compiled in
order to make debugging work, and explains the possible debug levels.
After a few wasted hours hunting down the reason for slowness during a
TLS handshake that turned out to be because of TCP_NODELAY not being
set, I think we have enough motivation to toggle the default for this
option. We now enable TCP_NODELAY by default and allow applications to
switch it off.
This also makes --tcp-nodelay unnecessary, but --no-tcp-nodelay can be
used to disable it.
Thanks-to: Tim Rühsen
Bug: https://curl.haxx.se/mail/lib-2016-06/0143.html
When input stream for curl is stdin and input stream is not a file but
generated by a script then curl can truncate data transfer to arbitrary
size since a partial packet is treated as end of transfer by TFTP.
Fixes#857
Previously, passing a timeout of zero to Curl_expire() was a magic code
for clearing all timeouts for the handle. That is now instead made with
the new Curl_expire_clear() function and thus a 0 timeout is fine to set
and will trigger a timeout ASAP.
This will help removing short delays, in particular notable when doing
HTTP/2.
Regression added in 790d6de485. The was then added to avoid one
particular transfer to starve out others. But when aborting due to
reading the maxcount, the connection must be marked to be read from
again without first doing a select as for some protocols (like SFTP/SCP)
the data may already have been read off the socket.
Reported-by: Dan Donahue
Bug: https://curl.haxx.se/mail/lib-2016-07/0057.html
If a call to GetSystemDirectory fails, the `path` pointer that was
previously allocated would be leaked. This makes sure that `path` is
always freed.
Closes#938
This is a follow up to the parent commit dcdd4be which fixes one leak
but creates another by failing to free the credentials handle if out of
memory. Also there's a second location a few lines down where we fail to
do same. This commit fixes both of those issues.
- the expression of an 'if' was always true
- a 'while' contained a condition that was always true
- use 'if(k->exp100 > EXP100_SEND_DATA)' instead of 'if(k->exp100)'
- fixed a typo
Closes#889
... as otherwise we could get a 0 which would count as no error and we'd
wrongly continue and could end up segfaulting.
Bug: https://curl.haxx.se/mail/lib-2016-06/0052.html
Reported-by: 暖和的和暖
Prior to this change we called Curl_ssl_getsessionid and
Curl_ssl_addsessionid regardless of whether session ID reusing was
enabled. According to comments that is in case session ID reuse was
disabled but then later enabled.
The old way was not intuitive and probably not something users expected.
When a user disables session ID caching I'd guess they don't expect the
session ID to be cached anyway in case the caching is later enabled.
- Enable protocol family logic for IPv6 resolves even when support
for synthesized addresses is enabled.
This is a follow up to the parent commit that added support for
synthesized IPv6 addresses from IPv4 on iOS/OS X. The protocol family
logic needed for IPv6 was inadvertently excluded if support for
synthesized addresses was enabled.
Bug: https://github.com/curl/curl/issues/863
Ref: https://github.com/curl/curl/pull/866
Ref: https://github.com/curl/curl/pull/867
Use getaddrinfo() to resolve the IPv4 address literal on iOS/Mac OS X.
If the current network interface doesn’t support IPv4, but supports
IPv6, NAT64, and DNS64.
Closes#866Fixes#863
Calling QueryContextAttributes with SECPKG_ATTR_APPLICATION_PROTOCOL
fails on Windows < 8.1 so we need to disable ALPN on these OS versions.
Inspiration provide by: Daniel Seither
Closes#848Fixes#840
- Change the parser to not require a minor version for HTTP/2.
HTTP/2 connection reuse broke when we changed from HTTP/2.0 to HTTP/2
in 8243a95 because the parser still expected a minor version.
Bug: https://github.com/curl/curl/issues/855
Reported-by: Andrew Robbins, Frank Gevaerts
Sessionid cache management is inseparable from managing individual
session lifetimes. E.g. for reference-counted sessions (like those in
SChannel and OpenSSL engines) every session addition and removal
should be accompanied with refcount increment and decrement
respectively. Failing to do so synchronously leads to a race condition
that causes symptoms like use-after-free and memory corruption.
This commit:
- makes existing session cache locking explicit, thus allowing
individual engines to manage lock's scope.
- fixes OpenSSL and SChannel engines by putting refcount management
inside this lock's scope in relevant places.
- adds these explicit locking calls to other engines that use
sessionid cache to accommodate for this change. Note, however,
that it is unknown whether any of these engines could also have
this race.
Bug: https://github.com/curl/curl/issues/815Fixes#815Closes#847
Mostly in order to support broken web sites that redirect to broken URLs
that are accepted by browsers.
Browsers are typically even more leniant than this as the WHATWG URL
spec they should allow an _infinite_ amount. I tested 8000 slashes with
Firefox and it just worked.
Added test case 1141, 1142 and 1143 to verify the new parser.
Closes#791
Regression from the previous *printf() rearrangements, this file missed to
include the correct header to make sure snprintf() works universally.
Reported-by: Moti Avrahami
Bug: https://curl.haxx.se/mail/lib-2016-05/0196.html
While compiling lib/curl_multibyte.c with '-DUSE_WIN32_IDN' etc. I was
getting:
f:\mingw32\src\inet\curl\lib\memdebug.h(38): error C2054: expected '('
to follow 'CURL_EXTERN'
f:\mingw32\src\inet\curl\lib\memdebug.h(38): error C2085:
'curl_domalloc': not in formal parameter list
...as otherwise the TLS libs will skip the CN/SAN check and just allow
connection to any server. curl previously skipped this function when SNI
wasn't used or when connecting to an IP address specified host.
CVE-2016-3739
Bug: https://curl.haxx.se/docs/adv_20160518A.html
Reported-by: Moti Avrahami
CID 1024412: Memory - illegal accesses (OVERRUN). Claimed to happen when
we run over 'workend' but the condition says <= workend and for all I
can see it should be safe. Compensating for the warning by adding a byte
margin in the buffer.
Also, removed the extra brace level indentation in the code and made it
so that 'workend' is only assigned once within the function.
The proper FTP wildcard init is now more properly done in Curl_pretransfer()
and the corresponding cleanup in Curl_close().
The previous place of init/cleanup code made the internal pointer to be NULL
when this feature was used with the multi_socket() API, as it was made within
the curl_multi_perform() function.
Reported-by: Jonathan Cardoso Machado
Fixes#800
Prior to this change a width arg could be erroneously output, and also
width and precision args could not be used together without crashing.
"%0*d%s", 2, 9, "foo"
Before: "092"
After: "09foo"
"%*.*s", 5, 2, "foo"
Before: crash
After: " fo"
Test 557 is updated to verify this and more
The new way of disabling certificate verification doesn't work on
Mountain Lion (OS X 10.8) so we need to use the old way in that version
too. I've tested this solution on versions 10.7.5, 10.8, 10.9, 10.10.2
and 10.11.
Closes#802
curl's representation of HTTP/2 responses involves transforming the
response to a format that is similar to HTTP/1.1. Prior to this change,
curl would do this by separating header names and values with only a
colon, without introducing a space after the colon.
While this is technically a valid way to represent a HTTP/1.1 header
block, it is much more common to see a space following the colon. This
change introduces that space, to ensure that incautious tools are safely
able to parse the header block.
This also ensures that the difference between the HTTP/1.1 and HTTP/2
response layout is as minimal as possible.
Bug: https://github.com/curl/curl/issues/797Closes#798Fixes#797
... introduced in curl-7_48_0-293-g2968c83:
Error: COMPILER_WARNING:
lib/vtls/openssl.c: scope_hint: In function ‘Curl_ossl_check_cxn’
lib/vtls/openssl.c:767:15: warning: conversion to ‘int’ from ‘ssize_t’
may alter its value [-Wconversion]
- In the case of recv error, limit returning 'connection still in place'
to EINPROGRESS, EAGAIN and EWOULDBLOCK.
This is an improvement on the parent commit which changed the openssl
connection check to use recv MSG_PEEK instead of SSL_peek.
Ref: https://github.com/curl/curl/commit/856baf5#comments
Calling SSL_peek can cause bytes to be read from the raw socket which in
turn can upset the select machinery that determines whether there's data
available on the socket.
Since Curl_ossl_check_cxn only tries to determine whether the socket is
alive and doesn't actually need to see the bytes SSL_peek seems like
the wrong function to call.
We're able to occasionally reproduce a connect timeout due to this
bug. What happens is that Curl doesn't know to call SSL_connect again
after the peek happens since data is buffered in the SSL buffer and thus
select won't fire for this socket.
Closes#795
Only protocols that actually have a protocol registered for ALPN and NPN
should try to get that negotiated in the TLS handshake. That is only
HTTPS (well, http/1.1 and http/2) right now. Previously ALPN and NPN
would wrongly be used in all handshakes if libcurl was built with it
enabled.
Reported-by: Jay Satiro
Fixes#789
Sometimes, in systems with both ipv4 and ipv6 addresses but where the
network doesn't support ipv6, Curl_is_connected returns an error
(intermittently) even if the ipv4 socket connects successfully.
This happens because there's a for-loop that iterates on the sockets but
the error variable is not resetted when the ipv4 is checked and is ok.
This patch fixes this problem by setting error to 0 when checking the
second socket and not having a result yet.
Fixes#794
curl_printf.h defines printf to curl_mprintf, etc. This can cause
problems with external headers which may use
__attribute__((format(printf, ...))) markers etc.
To avoid that they cause problems with system includes, we include
curl_printf.h after any system headers. That makes the three last
headers to always be, and we keep them in this order:
curl_printf.h
curl_memory.h
memdebug.h
None of them include system headers, they all do funny #defines.
Reported-by: David Benjamin
Fixes#743
Mostly because they're not needed, because memdebug.h is always included
last of all headers so the others already included the correct ones.
But also, starting now we don't want this to accidentally include any
system headers, as the header included _before_ this header may add
defines and other fun stuff that we won't want used in system includes.
Previously, connections were closed immediately before the user had a
chance to extract the socket when the proxy required Negotiate
authentication.
This regression was brought in with the security fix in commit
79b9d5f1a4Closes#655
WinSock destroys recv() buffer if send() is failed. As result - server
response may be lost if server sent it while curl is still sending
request. This behavior noticeable on HTTP server short replies if
libcurl use several send() for request (usually for POST request).
To workaround this problem, libcurl use recv() before every send() and
keeps received data in intermediate buffer for further processing.
Fixes: #657Closes: #668
This commit fixes a Clang warning introduced in curl-7_48_0-190-g8f72b13:
Error: CLANG_WARNING:
lib/connect.c:1120:11: warning: The right operand of '==' is a garbage value
1118| }
1119|
1120|-> if(-1 == rc)
1121| error = SOCKERRNO;
1122| }
... but output non-stripped version of the line, even if that then can
make the script identify the wrong position in the line at
times. Showing the line stripped (ie without comments) is just too
surprising.
- Error if a header line is larger than supported.
- Warn if cumulative header line length may be larger than supported.
- Allow spaces when parsing the path component.
- Make sure each header line ends in \r\n. This fixes an out of bounds.
- Disallow header continuation lines until we decide what to do.
Ref: https://github.com/curl/curl/issues/659
Ref: https://github.com/curl/curl/pull/663
This commit ensures that streams which was closed in on_stream_close
callback gets passed to http2_handle_stream_close. Previously, this
might not happen. To achieve this, we increment drain property to
forcibly call recv function for that stream.
To more accurately check that we have no pending event before shutting
down HTTP/2 session, we sum up drain property into
http_conn.drain_total. We only shutdown session if that value is 0.
With this commit, when stream was closed before reading response
header fields, error code CURLE_HTTP2_STREAM is returned even if
HTTP/2 level error is NO_ERROR. This signals the upper layer that
stream was closed by error just like TCP connection close in HTTP/1.
Ref: https://github.com/curl/curl/issues/659
Ref: https://github.com/curl/curl/pull/663
This commit ensures that data from network are processed before HTTP/2
session is terminated. This is achieved by pausing nghttp2 whenever
different stream than current easy handle receives data.
This commit also fixes the bug that sometimes processing hangs when
multiple HTTP/2 streams are multiplexed.
Ref: https://github.com/curl/curl/issues/659
Ref: https://github.com/curl/curl/pull/663
Previously, when a stream was closed with other than NGHTTP2_NO_ERROR
by RST_STREAM, underlying TCP connection was dropped. This is
undesirable since there may be other streams multiplexed and they are
very much fine. This change introduce new error code
CURLE_HTTP2_STREAM, which indicates stream error that only affects the
relevant stream, and connection should be kept open. The existing
CURLE_HTTP2 means connection error in general.
Ref: https://github.com/curl/curl/issues/659
Ref: https://github.com/curl/curl/pull/663
... but ignore EAGAIN if the stream has ended so that we don't end up in
a loop. This is a follow-up to c8ab613 in order to avoid the problem
d261652 was made to fix.
Reported-by: Jay Satiro
Clues-provided-by: Tatsuhiro Tsujikawa
Discussed in #750
As these two options provide identical functionality, the former for
SOCK5 proxies and the latter for HTTP proxies, merged the two options
together.
As such CURLOPT_SOCKS5_GSSAPI_SERVICE is marked as deprecated as of
7.49.0.
Calculate the service name and proxy service names locally, rather than
in url.c which will allow for us to support overriding the service name
for other protocols such as FTP, IMAP, POP3 and SMTP.
It turns out the google GFE HTTP/2 servers send a PING frame immediately
after a stream ends and its last DATA has been received by curl. So if
we don't drain that from the socket, it makes the socket readable in
subsequent checks and libcurl then (wrongly) assumes the connection is
dead when trying to reuse the connection.
Reported-by: Joonas Kuorilehto
Discussed in #750
