For Windows. Based on the code I wrote for QNetworkStatusMonitor.
It also renames the netlistmgr feature, avoiding the abbreviation.
Locally my MinGW fails the networklistmanager feature test so it may
not be supported on MinGW, likely leaving it without a backend at all.
Change-Id: I13bbe4127edc2a9c0bb91602c95f1cb206a85a69
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
While rebasing and resolving conflicts, I've acidentally removed a
declaration of supportsTls13() making a call expression invalid.
Change-Id: If561091db36f0354b61fa050ff9f72b29eddce53
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
which will become parts of TLS plugins in the future.
Task-number: QTBUG-65922
Change-Id: I4ee3c59c435fc34a9f4dacd3ff0e3cfb44251e23
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Remove 'corewlan' feature from the network related configure.json.
The 'bearer' component was removed in the Network module. It was
the only consumer of the 'corewlan' feature. There are no
plans, yet, to use CoreWLAN framework in the future.
Pick-to: 6.0
Fixes: QTBUG-90672
Change-Id: I1afb7b04ea75f8cb09cb570d820d16f54a130cf9
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Member variables have been reordered to reduce the memory
footprint by 8 bytes.
Change-Id: Ib03f916c065861fa317f8802695363fd152e314a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QTlsBackend is a factory itself - it creates TLS/X509 objects. Having
an intermediary between Factory->Backend->TLS primitive does not look
very natural thus let's squash the first two parts. Backend is a factory
creating TLS primitives, but its static functions also provide information
about backends availablei and give access to those backends.
Fixes: QTBUG-90606
Task-number: QTBUG-65922
Change-Id: I8409d81fd11fb46e6ab4465b4937a7680a8c2447
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
By marking undeclared and unused copy/move constructors as deleted.
Change-Id: Ie74b9cb015090dbb425b701a55eadb9d80ab0285
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Read/write/notify properties, 3 out of 5 defined in this class.
Task-number: QTBUG-85520
Change-Id: Ic6c74f90a2fa3c71d71cf9a5d557f1b6fc489d35
Reviewed-by: Andreas Buhr <andreas.buhr@qt.io>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
As part of Qt 6 restructring for the extras modules, this change exposes
the Jni APIs which are very important for Android platform. This patch
adds the APIs QJniObject, QJniEnvironment, QJniExceptionCleaner based
from private QtCore and QtAndroidExtras.
The Jni interface is cross-platform which justifies the name, but
currently, this API is used mainly for Android, and the naming comes
generic without Android keyword to avoid any future limitation on
supporting other platforms.
[ChangeLog][QtCore] Add new QJniObject, QJniEnvironment and
QJniExceptionCleaner APIs.
Task-number: QTBUG-89482
Fixes: QTBUG-89633
Change-Id: I4382dd53a225375759b9d042f6035a4a9810572b
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
Fetched from the authoritative source, verified the content matches
that of the current master revision in the github repository.
Amend one cookie jar test to find the last group in the last chunk
correctly - each group arises from a non-empty hsah-table entry, but
the last few hash-table entries may be empty, in which case the last
group isn't just before the last index, it's earlier by the number of
empty hash table entries. In the process, amend this test and the
related test of the end of the first chunk to iterate all the entries
in the group (in the present version, as it happens, each end-group
has just one entry, but that may vary).
Task-number: QTBUG-90214
Pick-to: 6.0 5.15
Change-Id: I6da365a6ca558124f8275e392735071dc77e04bb
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
This is an abstraction for TLS backend and its factory, preparing to transition
to plugin-based design.
Task-number: QTBUG-65922
Change-Id: Ibe810e77fd1b715a6bea66cd3f44312b015ac274
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
This API gives the names of available backends and provides a basic
information about features/protocols supported by those backends.
Also, it has the 'loadBackend' functions which allow to select
a particular backend (which are becoming plugins).
At the moment, the implementation is still 'hardcoded', the
follow-up patch will allow to select different backends in runtime.
Task-number: QTBUG-65922
Change-Id: I05877de9c02857594e76b24d52e7578bdb01df69
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The plugins are meant to indicate what they do support, meaning users of
QNetworkInformation can choose to not care about which plugin is used
and rather just request what they want.
Task-number: QTBUG-86966
Change-Id: Ie130e1791250ec2a4470e3ba7081d982654af06c
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
OpenSSL adds an '-arm'/'-arm64' suffix to the DLL names when building
for VC-WIN32-ARM or VC-WIN64-ARM. This needs to be accounted for in Qt
to find the OpenSSL DLLs correctly.
Fixes: QTBUG-89647
Pick-to: 5.15
Change-Id: Ibc9b81e06f8e64c0676f335e13024fa7fe3a4fa3
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Those serve no purpose anymore, now that the .pro files are gone.
Task-number: QTBUG-88742
Change-Id: I39943327b8c9871785b58e9973e4e7602371793e
Reviewed-by: Cristian Adam <cristian.adam@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
Remove the qmake project files for most of Qt.
Leave the qmake project files for examples, because we still test those
in the CI to ensure qmake does not regress.
Also leave the qmake project files for utils and other minor parts that
lack CMake project files.
Task-number: QTBUG-88742
Change-Id: I6cdf059e6204816f617f9624f3ea9822703f73cc
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
An invalid (as input data) certificate may have non-zero number of invalid (nullptr)
extensions (if OpenSSL failed to parse them, for example).
Fixes: QTBUG-89547
Pick-to: 6.0 5.15 5.12
Change-Id: I4b93ac9f482f850f02d01b0aea10560dc11b688d
Reviewed-by: Lars Schmertmann <lars.schmertmann@governikus.de>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
This reverts e52239f525, because we do not
have multi-ABI builds currently, and we must link against what the
WrapOpenSSL package provides.
Fixes: QTBUG-89473
Change-Id: I4f38a2c41e90ec8b602e28f8ca77f0ebd009639c
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Another case when an OpenSSL's call can dump some errors into the shared
error queue discovered. An invalid certificate with non-null X509 * may
result in several errors appended to the queue.
Pick-to: 6.0 5.15
Change-Id: I1278b371bd1edf2d656760c371bfb6da5dcab6ec
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Something defined is always greater than undefined
Pick-to: 6.0
Fixes: QTBUG-89530
Change-Id: I7633dd4233f7b13c3c5be17352841ec873531ef9
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
If the error occurs during the call to
QHttpNetworkConnectionPrivate::queueRequest coming from the http thread
delegate then we will not yet have connected to the signal! But the http
thread delegate checks if the error code is not NoError, and handles
those situations. To let that work we must update the replies.
Pick-to: 6.0 5.15
Change-Id: I47188e9439920694aaad1765ab28add1e86ccdff
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Compiler (gcc) keeps nagging and complaining about an address of function
to be never nullptr in any particular instantiation, when argument is an
address of a function.
Change-Id: If67e80f2ff4d408608429d53814083777cc8441c
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
At the same time I'll add a generic protection against being called in
unprotected mode in the schannel backend (openssl already has it in a
different form).
Pick-to: 5.15 6.0
Change-Id: I97c1be6239c27e306de0af7ad568fbcfde09da71
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The URL for the PAC proxy that is passed needs to be preserved for the
main URL part and not entirely percent encoded, only the query part
typically needs to be encoded. So use toEncoded instead for a URL to
ensure they are not percent encoded. This amends
c163ec1dbf
Pick-to: 6.0 5.15
Change-Id: Ie41ab55f71be8e25c18775e61ce7b4d110c2ddbf
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Following a41c61fb2d QIODevice may try to
copy the QByteArray itself (rather than the data it points to). This can
lead referencing dangling data when the QByteArray is initialized with
raw data.
Pick-to: 6.0
Change-Id: I481695b33f251f750ef482d72b81636f0d4bf462
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Previously we only checked if there was something in the h2 queue or
if we had not yet switched to http2. These means in some cases that
following an error it would enter this branch and, because it had not
yet switched to http2, it would call 'ensureConnection' which would try
start a reconnection. This could make the proxy socket engines try to
emit 'auth required' with no replies, meaning nothing happens.
Fixes: QTBUG-88417
Pick-to: 5.15 6.0
Change-Id: If37b8c71ac36597adc305fb1b1ea4fa7b48b0d28
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The requests will remove themselves once they get deleted
but since the deletion is done through a _queued_ invokeMethod
to 'deleteLater' we will call
QHttpNetworkConnection::_q_startNextRequest first which may
end up starting a reconnect of the TCP socket which we had the error on.
In this specific instance it manifested as a race condition where we
either don't get a proxyAuthorizationRequired signal at all (it was
emitted while we didn't have any valid replies), or we get the signal
emitted too late and it gets emitted on whatever the next reply was.
Task-number: QTBUG-88417
Pick-to: 5.15 6.0
Change-Id: If3f8ececc5550f1868c90124559cb8e3029646d8
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
One patch from many years ago already did most of it, but the key of
the hash was still int.
Change-Id: I557fa19d2573addbef7557d1a16fdd6554a117cc
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The correct entry for this lib is 'openssl/nolink', not
'openssl_headers'.
In configurejson2cmake's helper.py we already have
this information encoded such that the no_link_so_name field of
'openssl' points to 'openssl_headers'. Extend configurejson2cmake to
take this into account, and re-generate src/network/configure.cmake.
Pick-to: 6.0
Change-Id: I9817ea7722503a373d309b7e6fa201448d403e8d
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
QSslCertificate::verify() has an undocumented and not very desirable property -
on some platorms it updates the default configuration, which can be surprising.
For example, we deprecated QSslSocket::setDefaultCaCertificates() and recommend
using QSslConfiguration::defaultConfiguration(), QSslConfiguration::setDefaultConfiguration(),
and QSslConfiguration::setCaCertificates(). If an application does this to select
CA roots it trusts explicitly, and then for some reason is calling verify, the
application can have its QSslSockets successfully connecting to a host, whose
root was not trusted by the application. Also, on Windows, defaultCaCertificates()
include system roots already, no need to have them twice.
[ChangeLog][QtCore][QtNetwork] QSslSocket::verify - do not change the default configuration
Pick-to: 5.15
Pick-to: 6.0
Pick-to: 6.0.0
Fixes: QTBUG-88639
Change-Id: I1cd40b259d0a6dcd15c78d1e7c027ff10859595c
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Q_MOVABLE_TYPE was conceived before C++ had move semantics. Now, with
move semantics, its name is misleading. Q_RELOCATABLE_TYPE was
introduced as a synonym to Q_MOVABLE_TYPE. Usage of Q_MOVABLE_TYPE
is discouraged now. This patch replaces all usages of Q_MOVABLE_TYPE
by Q_RELOCATABLE_TYPE in QtBase. As the two are synonymous, this
patch should have no impact on users.
Pick-to: 6.0
Change-Id: Ie653984363198c1aeb1f70f8e0fa189aae38eb5c
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
The Professional CMake book suggests always using explicit linkage with
target_link_libraries, so let's use that.
Whether to use PUBLIC or PRIVATE depends on the context. But let's
be conservative and advise on using PRIVATE by default.
Task-number: QTBUG-88935
Pick-to: 6.0
Change-Id: I12b80ee85be9f6916f1e4dea6b1c9cb29e03c20f
Reviewed-by: Craig Scott <craig.scott@qt.io>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
As noted by LCOV, the part with q_BIO_free(bio) was never executed
since we were taking the result from QScopedPointer before returning.
While it's a what RAII idiom is for, there is quite a low probability
that SSL_set_bio() one day will start throwing exceptions.
Pick-to: 6.0
Pick-to: 5.15
Change-Id: Id24e480dac34166c627b71bb2972de558c644339
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
This should've been caught a long time ago, but long story short: it wasn't.
This has been deprecated since 5.14 and is now asserting.
Pick-to: 5.15
Change-Id: Iddee22b1a68e5d1a03006831f795db8ec1c472be
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
In general the protocolHandler isn't deleted unless
the channel is being destructed. So instead of reset()ing
the pointer we keep it around.
Also update the http2protocolhandler to mimic the http1
handler a little closer: shutting down the channel in
receiveReply if there's no reply/activeStreams, and not
calling receiveReply at all if there's no activeStreams.
Pick-to: 5.15
Change-Id: I702547f594deb6b0c1384068f7e93e560527e8e2
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
With TLS 1.3 the client goes through renegotiation
when using Schannel. The status returned is OK and
we can immediately continue, so do that.
Change-Id: I831eaae318df9d94b5fb7672db7e407d94f9da56
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The libproxy issue was fixed 4 years ago and the related
Qt and libproxy bug reports were also closed long ago.
Change-Id: I5af18f1eaaffb8d33556ab94ca1759959a2c272d
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
The content in the porting guides are closer to a changelog than
a porting guide. At this point, it is easier for maintainers and
contributors to write in a changelog than a guide. This change
should help with readability and is closer to the usage of
"Changes" in documentation.
Part of a rename in other submodules.
Task-number: QTBUG-87156
Change-Id: Ib59de7976fae19e7b39962e80161df6628e4070b
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
The former *.platform.sh entry has been replaced by more specific
entries, replacing the wildcard with only selected second levels.
Task-number: QTBUG-87925
Change-Id: Ie4ba7c66ba9aa40eafe23f02faa12f19d791cff2
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The header is now in src/network/kernel/ rather than src/corelib/io/,
but the qt_attribution.json got left behind and the update program was
still in a sub-dir of util/corelib/. Renamed the latter to
util/publicSuffix/ (second-layer sub-directory was overkill, util/
isn't crowded and it was the only thing in util/corelib/; and there
was no util/network/). This is a follow-up to
commit 4f076db3d2
Change-Id: I51c2c7892752ddc47390966044eb5650dfdfa9c2
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
When -DFEATURE_schannel=ON, -DFEATURE_ssl=ON, and -DFEATURE_openssl=OFF
cmake configuration would fail.
Fixes: QTBUG-88126
Change-Id: I7325a6c6bc4b9ff448893c7f05b7f46c34eefc1e
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Details described in a comment.
Task-number: QTBUG-88063
Pick-to: 5.15
Change-Id: I763ecfedf518de97615e04a8eaae0fe1fd784f52
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Gets rid of moving strings back and forth, which was extra expensive
without QList's prepend optimization (which is returning, but getting
rid of the moves is even better).
Change-Id: I520fc140310a0e1f096dda97f2e999d0e7993e77
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QSslPreSharedKeyAuthenticator::isEqual() as accessed
by inline operators is not exported.
Amends 57e57d9bcd.
Change-Id: Ia0f0de81a371a31422d07af5b11f0db03265504e
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The non-versioned one's do miss some properties. It's therefore best
to not advocate using them.
Change-Id: I53645e65ed4de4e0100e59905c024cdfe40be0c5
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
In preparation for some further regeneration.
Also modify pro2cmake to add forgotten mapping for the
Qt::EglFsKmsGbmSupportPrivate module.
Change-Id: I92425c566c2b275b40eec8c652496290754ac385
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Make them hidden friends, add a private isEqual helper where needed.
Adjust and add documentation.
Fixes: QTBUG-87976
Change-Id: If7c19eeab5be7452364eb76193981100f5516d6b
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Make global operators hidden friends. No change to the member-operators.
Task-number: QTBUG-87976
Change-Id: If7b08a30700d4e2f1a304d4b6cc4b5d02ee5e251
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
It is an important new details in cookies, as a minimum pass it through,
before we add API for it.
Pick-to: 5.15
Change-Id: I5222a24e0f50f3822a94cce126b5055fed1a8008
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Following 77d812683f QMutexLocker is now a
templated class which should in most cases be deduced automatically.
Doesn't work here because of the global static wrapper.
Change-Id: I62fda0651eac12f3234396cb7e41a26df8cc8b46
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Since we're going to split QMutex and QRecursiveMutex into
separate classes, make sure QMutexLocker is prepared for that.
Change-Id: Id5e9a955d1db7c8ee663dd3811ad6448dad0aeae
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Avoid detaching where possible
Change-Id: I438d3e66689aeef05951af86a48af2a6910da7c2
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
When parsing subjectAltName extension to make it more like OpenSSL
counterpart.
Fixes: QTBUG-86830
Change-Id: If1a4e72ee0b19f2cf40aa53632f9ec1468178c3b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
When a peer closes the connection, the device remains opened for reading
purposes. However, we should disable writing on disconnected socket.
Otherwise, if the user issues a write() call, a new pipe writer object
will be created and the write call occurs with invalid handle value.
Pick-to: 5.15
Change-Id: Id136798c7663df1fce7ed0aa4e3c6f5c65218a11
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
It's not possible to connect to microsoft.com with Schannel TLS 1.3 for
some reason (also tested with Internet Explorer), but other sites work
fine. Must be something they have to iron out for later.
In my experience this needs a preview release of Windows. One of my
machines is opted into the dev channel of Windows where they enabled TLS
1.3 by default, and it works well in my tests except for the part above.
On my other machine, after enabling TLS 1.3 through the registry, I fail
to complete the handshake with any site. So around March/April next year
is when this code would activate for most people.
MinGW apparently defines NTDDI_VERSION as the one for Windows Server
2003, so it currently doesn't build the new TLS 1.3 code. In Qt (as a
project) we could consider setting this higher, but that's out of scope
for this patch!
Fixes: QTBUG-81294
Change-Id: If329959c3a30ecbfbb8c0d335cc39ccb6d012890
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
It was there for binary compatibility reasons, but it's not needed anymore
Change-Id: I659dadc710a5bcdbcba74f13fd4db6044a497cd2
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
This time based on grepping to also include documentation, tests and
examples previously missed by the automatic tool.
Change-Id: Ied1703f4bcc470fbc275f759ed5b7c588a5c4e9f
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Rename Android package name org.qtproject.qt5.android to
org.qtproject.qt.android to avoid inconsistency with Qt 6 name.
Also, we include the major version number in the jar target.
Task-number: QTBUG-86969
Change-Id: Ibb68947289be1079911b34ea157bf089cc52c47f
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
These value gaps were there for binary compatibility, but we don't need
those in Qt 6.
Change-Id: I72a40d601b61f90e14562578e88493b0d54ff398
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
It was already used many places directly making the code inconsistent.
Change-Id: I3b14bc6c333640fb3ba33c71eba97e78c973e44b
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Modify special case locations to use the new API as well.
Clean up some stale .prev files that are not needed anymore.
Clean up some project files that are not used anymore.
Task-number: QTBUG-86815
Change-Id: I9947da921f98686023c6bb053dfcc101851276b5
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
To avoid potential decompression bombs. This is implemented with just
a simple check that the ratio doesn't pass some hardcoded preset.
Change-Id: I17246f0f43e73280cdb35a8f03d65885f5678ad6
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Clean up the state of the projects,
before changing the internal CMake API function names.
Task-number: QTBUG-86815
Change-Id: I90f1b21b8ae4439a4a293872c3bb728dab44a50d
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
For HTTP connections, QNAM defaults to opening its TCP socket in
unbuffered mode. This means that Qt will send the data written
into the socket right to the kernel, queueing only if the kernel
says it doesn't want more data for the moment being.
QNAM itself then uses separate write() calls to write the HTTP
headers and the body of the request (like POST or PUT). These 2+
writes result in headers and body being sent over different TCP
segments -- even if, in principle, a POST with a few bytes of data
(e.g. a HTML form, or a REST or SOAP request) could fit in the same
segment as the request.
Multiple writes like this interact extremely poorly with other
TCP features, e.g. delayed ACKs, Nagle's algorithm and the like.
In a typical scenario, the kernel will send a segment containing just
the headers, wait for the ACK (which may be delayed), and only then
send the body (it wasn't sent before because Nagle was blocking it).
The reply at this point is immediate (because the server can process
the request and starts replying), but the delayed ACK is typically
40-50ms, and documented up to 500ms (!). If one uses QNAM to access a
service, this introduces unacceptable latency.
These multiple writes to the OS should be avoided.
The first thing that comes into mind is to use buffered sockets.
Now, there are good reasons to keep the socket unbuffered, so we
don't want to change that. But the deal breaker is that even buffered
sockets won't help in general: for instance, on Windows, a buffered
write will immediately detect that the socket is ready for write and
flush the buffer right away (not 100% sure of why this is necessary;
basically, after populating the QTcpSocket write buffer, Qt enables
a write socket notifier on the socket -- notifier that fires
synchronously and immediately, without even returning to the event
loop, and that causes the write buffer flush).
Linux of course offers the perfect solution: corking the socket via
TCP_CORK, which tells the kernel not to send the data right away but
to buffer it up to a timeout (or when the option gets disabled
again, whichever comes first). It's explicitly designed to support
the case of sending headers followed by something like a
sendfile(2). Setting this socket option moves the problem to
the kernel and we could happily keep issuing multiple writes.
Ça va sans dire, no other OS supports that option or any other
similar option.
We have therefore to deal with this in userspace: don't write in the
socket multiple times, but try and coalesce the write of the headers
with the writing of the data. This patch implements that, by storing
the headers and sending them together with the very first chunk of
data. If the data is small enough, this sends the entire request
in one TCP segment.
Interestingly enough, QNAM has a call setting TCP_NODELAY
currently commented out because Qt doesn't combine "HTTP requests"
(whatever that means). The call comes all the way back
from pre-public history (before 2011) (!). This patch doesn't
touch it.
Fixes: QTBUG-41907
Change-Id: Id555d14e0702c9f75c3134b18277692eb3659afe
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Also add existing such docs to the new document group
Task-number: QTBUG-84051
Change-Id: I76f033f0846e09943f249d2beeb1606869eef382
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
Certain servers, like smtp.live.com, will send a request for a
certificate even though they don't require one. In Schannel this
manifests as a warning/info status (SEC_I_INCOMPLETE_CREDENTIALS).
In the cases where it's not needed we should suppress the warning and
try to connect anyway, which is done by calling
InitializeSecurityContext again when we get the status.
Pick-to: 5.15
Change-Id: I3c48140f2949d8557251a49a2b66946da9395736
Reviewed-by: Joshua GPBeta <studiocghibli@gmail.com>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Fix our API, so that QStringList and QList<QString> are the
same thing.
This required a bit of refactoring in QList and moving the
indexOf(), lastIndexOf() and contains() method into
QListSpecialMethods. In addition, we need to ensure that
the QStringList(const QString&) constructor is still available
for compatibility with Qt 5.
Once those two are done, all methods in QStringList can be moved
into QListSpecialMethods<QString>.
Change-Id: Ib8afbf5b6d9df4d0d47051252233506f62335fa3
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
CentOS it seems not only backported some OpenSSL 3 functions,
but also raised the default security level to 2, making some of
our keys (and MDs?) 'too weak' and failing auto-tests here and
there as a result. For our auto-test we lower the level to 1,
as it is expected to be.
Fixes: QTBUG-86336
Pick-to: 5.15
Change-Id: I7062a1b292e8b60eb9c2b2e82bd002f09f9da603
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
AFAICT with GSSAPI the normal workflow is to run kinit or similar and
authenticate before running programs relying on it. Therefore
we can try to get the credentials before we choose whether or not
to use Negotiate.
Pick-to: 5.15
Task-number: QTBUG-85123
Change-Id: If0478fdd45389b2939ad87c2f582776fe56959bb
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The GSSAPI thing is a bit noisy and not useful unless you're
debugging it specifically.
Pick-to: 5.15
Change-Id: I4a8c14159ec889776d06e0970ddf66083d788b63
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
In QAuthenticator "detach()" does not do what you expect.
First off it doesn't detach at all, and secondly it will reset the phase
that the authentication is in. This last part is intended, but it has
one issue: if setUser/setPassword is called with the same arguments
every time we ask for credentials then we never reach a fail-state since
it thinks we will have a new chance to authenticate.
Pick-to: 5.15
Change-Id: I02e2e42242220f3fced3572323e6492429cf173e
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
It's problematic if we have already used these credentials and it
failed. Since then we emit authenticationRequired only for QNAM to
intercept it and "helpfully" suggest we use the same credentials again.
By moving on we can check the cache or ask the user about the necessary
credentials.
Pick-to: 5.15
Change-Id: Idaac5ae71167462369b66194ab6b1f77113d636a
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Although this is more of a QNAM thing.
In some cases what we load from the cache might be credentials we have
already tried (and failed with, thus leading us to emit the
authenticationRequired signal). With this patch we will fall through
more often and ask the user for credentials.
Pick-to: 5.15
Change-Id: If2a556883c3ea5b0b225f4df273d38353b552b54
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Windows only so far, need a similar patch for *nix, or an alternate
approach when there is no server set up.
Pick-to: 5.15
Task-number: QTBUG-85123
Change-Id: Iff7a6b1540a2f1984153a237eea07c7bb1970064
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
If we compiled without support for it then we shouldn't consider it an
option either.
Pick-to: 5.15
Change-Id: If6e0a6afa738f375e360bf3d439196b39e47bee8
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
It has been in Qt for some years now and 6.0 marks a good point to
enable it by default. The exception is connectToHostEncrypted where we
still require the users to enable it explicitly since there's no logical
way to disable it.
[ChangeLog][QtNetwork][QNetworkAccessManager] HTTP/2 is now enabled by
default.
Fixes: QTBUG-85902
Change-Id: Ia029a045727cc593d77df9eb3a5888522ad19199
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
It really could only manifest itself if you started a request and then
immediately cancelled it and then started another one to the same
site. But only if in a certain race outcome - the connection that the
backend was establishing had to finish connecting after aborting but
before a new request had been queued!
Change-Id: I7cad2cf4ac1f64cc838498cefa076cd2c6d26701
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Mostly related to qstrlen().
Change-Id: I69e2052c83766e4fc466ed398d0d0eac011a77ec
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
While it's possible to instantiate safe_delete using a nullptr, the
check in if-statement is 99.9(9) % of time redundant and equal
to if (true && object). Some compilers will issue a compilation
error (if warnings are treated as errors for example).
Change-Id: Ib593dc53deb6d2e4b77ea5c896610dc536c61b7c
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Android 11's SELinux rules block the RTM_GETLINK requests that
qnetworkinterface_linux.cpp needs to make. So disable this completely
and fall back on the older rimplementation. Bionic has some workarounds
in their getifaddrs() implementation to make this work.
Fixes: QTBUG-86394
Pick-to: 5.15 5.15.1
Change-Id: I87447a0ecdee4dc7b506fffd163180c2b1db7835
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Remove compile warning for -Wdeprecated-declarations.
Change-Id: I747c8a390f26cfc4626e40c4684034c394bbf3c9
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Lars Schmertmann <lars.schmertmann@governikus.de>
It triggers when connecting to localhost (or when using socks proxy) and
is nothing you can act on most of the time, so change it to qCDebug.
Change-Id: I669fbde4fa0ed194703ea6a4dab13790aa771852
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Makes the QNetworkAccessBackendFactory a real interface to be used in
plugins. Requires exporting some classes but they're not made public
yet.
Removes unused features and functions.
Some things are likely still unused due to being specific for HTTP
but the HTTP network replies don't use this backend system.
Changes QNetworkAccessBackend to use a more traditional
read(char*, qint64) function for the "downloaded" data.
And an optional readPointer if supported. So far no backends have it
so it's somewhat useless, but it may be useful going forward.
If not it shall be deleted
Converts all current backends to the new setup
Easy enough, also gets rid of some unused functions.
Task-number: QTBUG-80340
Change-Id: I9339e6c6eb394c471c921f5cafd3af6175936399
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
While we initially wanted to just disable the functionality the way it
is currently built forces people to do a full build of Qt just to enable
it. Instead of doing this half-measure let's just remove the code
completely from QtBase and rather prepare QtNetwork to handle being a
plugin that can be compiled at any time.
Task-number: QTBUG-80340
Change-Id: I19155c8c167cf932088f01b2a9706d0e7ab792d1
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
And put a note about it in the documentation
Change-Id: I29126e4a80f83c256190e03b8fe01f3c869fd46d
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
And not the ssl configuration we have on the reply since it's missing
e.g. the newly received session ticket.
Change-Id: Idfeb09012a847605a76d1fe4fb881c663d019b4a
Reviewed-by: Peter Hartmann <peter@edelhirsch.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
and not just the active one(s). When starting the connection with
HTTP/2 and calling QNetworkReply::ignoreSslErrors() it would previously
only ignore the errors in the single active channel that was used, but
if we then fall back to HTTP/1.1 then we'll use the 5 other channels as
well, and those would then fail due to the ignored ssl errors.
Change-Id: I7aeb5b59897dd3a53579f0d38bd255bc2d97c2bb
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
And don't use int based type mapping anymore.
Change-Id: I456e76d1933ef646a7bd39ce565886b89e938a44
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
And remove one of the type id to name mapping that still
existed in QMetaType. QMetaTypeInterface can provide that,
so there's no need to have a second copy of the data.
qMetaTypeTypeInternal() can still map all the names of all
builtin types to ids. That functionality is for now still
required by moc and can't be removed yet.
Change-Id: Ib4f8e9c71e1e7d99d52da9e44477c9a1f1805e57
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
It works as follows:
- user calls write(const QByteArray &);
- this function keeps a pointer to the chunk and calls a regular
write(data, len);
- write(data, len) calls a virtual writeData();
- subclass calls a new QIODevicePrivate::write();
- QIODevicePrivate::write() makes a shallow copy of
the byte array.
Proposed solution is fully compatible with existing subclasses.
By replacing a call to d->writeBuffer.append() with d->write(),
subclasses can improve their performance.
Bump the TypeInformationVersion field in qtHookData, to notify the
Qt Creator developers that the offset of QFilePrivate::fileName was
changed and dumpers should be adapted.
Change-Id: I24713386cc74a9f37e5223c617e4b1ba97f968dc
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
- Remove obsolete functions and enumeration values
- Remove QObject * parameter from QMetaProperty accessors
- Fix renamed enumerations in QSsl
- Fix list items to be \li
- Fix function signatures and variable names
Change-Id: I37c7e6bf2c8ff92bc7b82620bae0a27796f866ab
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
QAbstractSocket::abort() is not a virtual function and
QSslSocket::abort() does not override it. Having two alternatives
requires a dynamic typecasting and violates the principles of object-
oriented programming.
Due to the BC, we were unable to fix that in Qt5. Now, we can modify
QSslSocket::close() to handle QAbstractSocket::abort() requests and
remove the duplicate.
Change-Id: I49d6f32a571ae6e35b08cb366816f917e580dae8
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
1. Remove a useless forward declaration of a non-existing class.
2. Simplify the cipher filtering.
3. A missing private key (when local cert is present) found
by the Qt, not OpenSSL, so no need in asking OpenSSL for errors
in queue.
3. Fix a potential double-free (for opaque keys).
4. read/write BIOs normally owned by SSL object, but if
we fail to allocate any of them, we return early,
potentially failing to free the one that was allocated.
Change-Id: Ifb52fbc9fd1a38f101bd7ff02e79b82d6eb7e5b0
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
After calling this function, the user expects that the connection is
terminated, I/O device is closed and the socket is ready for a new
attempt. But, if the socket was disconnected before the call, close()
is not called and I/O device remains opened.
Because QAbstractSocket::close() and QIODevice::close() can handle
reentering, we can call close() unconditionally, which makes
the behavior obvious.
Change-Id: I90a9cbb1a1fe8f866b55ef0bd68d286b34e853f5
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QtNetwork requires the socket library as a public dependency. Similar
to Windows's ws2_32 library.
Task-number: QTBUG-83202
Change-Id: I92bb48fddc2cbea07700cde65b4737500b9820f7
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Now that I need this stuff both in OpenSSL and woldSSL code, makes sense
to introduce such helpers as a separate change (instead of duplicating
the code in wolfSSL patch).
Change-Id: I9ce600ebe709d103209372ea4c2bdb6fa6b6ce3a
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
One warning about unused variable (dataLeftover), which was at some
point replaced by the class variable 'decoderHasData'.
The second warning was a fault of logic: checking that the unsigned
value retValue was greater than or equal to zero. Which only came about
because they initially did different things but the branches got merged
and the logic became flawed.
Change-Id: Ia3a04516c1b7b5f962226998bf3f4d101dd38148
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
We extend configurejson2cmake to read the "commandline"
information from configure.json. This data is then translated to CMake function
calls and written it into commandline.cmake files.
We extend QtProcessConfigureArgs.cmake to pick up those commandline.cmake
files to feed our command line handling code, which is a
re-implementation of the command line handling in qt_configure.prf.
The command line handler sets INPUT_xxx variables, similar to
configure/qmake's config.input.xxx variables. The INPUT_xxx values are
translated
- to -DFEATURE_xxx=ON/OFF arguments if the input represents a feature,
- to corresponding CMake variables if such a variable is known,
- or to -DINPUT_xxx=yyy CMake arguments.
Configure arguments that have an entry in
cmake/configure-cmake-mapping.md are actually implemented. Other
arguments are likely to need more work.
Task-number: QTBUG-85373
Change-Id: Ia96baa673fc1fb88e73ba05a1afb473aa074b37d
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Move the QIODevice::OpenMode enum into a base class, so that
we can remove the full QIODevice (and thus QObject) dependency
from qdatastream.h and qtextstream.h.
This is required so that we can include QDataStream in qmetatype.h
without getting circular dependencies.
As a nice side effect, QDataStream and QTextStream can now inherit
QIODeviceBase and provide the OpenMode enum directly in their
class scope.
Change-Id: Ifa68b7b1d8d95687ed032f6c9206f92e63bfacdf
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@qt.io>
Both normal and relaxed constexpr are required by our new minimum of
C++17.
Change-Id: Ic028b88a2e7a6cb7d5925f3133b9d54859a81744
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Also take this opportunity to reshuffle the content-encodings in the
intended ordering since the ordering is used to signify priority.
Task-number: QTBUG-83269
Change-Id: I022eecf1ba03b54dbd9c98a9d63d05fb05fd2124
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Not following redirects is not a feature, but just a hastle for everyone.
The main issue with switching this default is that applications that
actually do manual redirect handling will break in various ways.
FollowRedirectsAttribute was removed as it no longer serves any
purpose beyond duplicating the default value.
[ChangeLog][Network] QNetworkAccessManager now follows redirects by
default with the NoLessSafeRedirectPolicy.
[ChangeLog][Potentially Source-Incompatible Changes]
QNetworkRequest::FollowRedirectsAttribute was removed and has been
superseded by QNetworkRequest::RedirectsPolicyAttribute
Fixes: QTBUG-85901
Change-Id: Ic5b776180a4b84ac4fc895158bb5a66a3c91a042
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
If we ask for HTTP/2 or 1.1 and the server doesn't list either then we
should still try to connect using HTTP/1(.1) just in case, to keep
compatibility.
Task-number: QTBUG-85902
Change-Id: I6ff2e38ac9d767e482a19ee4c81d101be37d3fab
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
This seems to only be happening when we have a single channel because
otherwise it will try IPv4 in one channel and IPv6 in the second.
Change-Id: I0d513e25fefffeabfc733e895827aa12da335ef9
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The user-agent should be propagated to the proxy as well or else we
get our default one.
Change-Id: Id2283a8f2ade1a32f7fcf3d691be8d380d334b50
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
In some cases (i.e. with SOCKS) we don't have a HTTP/2 request ready
yet so it would just error out.
Change-Id: I6449de5fb52f5208d5641062c5906c3baff18b77
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
In that case we haven't created the h2 handler yet, so it will
crash when trying to access various members.
Task-number: QTBUG-85902
Change-Id: Id0699ff06ef67748a16622703f731db0b0867771
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Before the introduction of QDecompressHelper gzip was preferred over
deflate. The change seems to be accidental. Amends commit
7b76379a89.
Change-Id: I70f33d551912465d63f49ea3db1ac3575d19a92d
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
QIODevice::skip() called a virtual QIODevicePrivate::skip() to
implement an efficient skipping on I/O devices for the internal
subclasses. The user subclasses cannot inherit QIODevicePrivate, so
this functionality was not externally accessible.
This patch replaces QIODevicePrivate::skip() with a virtual protected
QIODevice::skipData(). While the basic implementation simply discards
the data by reading into a dummy buffer, users can reimplement this
function to improve the performance in their subclasses.
[ChangeLog][QtCore][QIODevice] Added virtual protected skipData().
Now, subclasses can implement device-specific skipping of data.
Change-Id: I9522f7f7ab9d03ac06e972a525f8ec2fa909a617
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Currently it is possible to compare a QSslError with a
QSslError::SslError because QSslError has an implicit
constructor. But the comparison often fails because
a QSslError received from the system contains a
certificate.
[ChangeLog][QtNetwork][QSslError] The constructor
QSslError(QSslError::SslError error) is now explicit.
Change-Id: I36cc5895245d3b43ab4b8d65a9635893d6b0e6a4
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Changes are not too big for now. Just replaces use of the previous
calls to the zlib decompression function. And initialize
QDecompressHelper when we know the content-encoding.
Task-number: QTBUG-83269
Change-Id: I41358feaef2e7ac5f48f14e3f95ec094e0c110b7
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Some servers have an upper limit on how many times a nonce can be
reused. Because the nc was not being reset the server would get a
high nc and would close the connection.
Change-Id: I71d5a316f79777a45c7323d93e28b3845aefcfad
Original-patch-by: Benjamin Reikowski
Pick-to: 5.15
Fixes: QTBUG-85729
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
To support streaming decompression in QNAM.
Will also be used to refactor existing decompression code in QNAM.
Task-number: QTBUG-83269
Change-Id: Iecf3e359734163f15686c949f75d41fa4794a00e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
We have several classes inheriting QObject (directly or not) and disabling
copy (thus non-movable implicitly). Let's make them non-movable explicitly
now.
Change-Id: I509a066e9977480b30842afef26319b8afec6001
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This amends 0ef46b41d8 to ensure the
new configuration is styled correctly for qmake.
Task-number: QTBUG-80938
Change-Id: Iebfff3ff0570d2e92d01d4eb1178c0d6261468f9
Pick-to: 5.15
Reviewed-by: Lars Schmertmann <lars.schmertmann@governikus.de>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
WebAssembly does not need the http feature, but these functions were
hidden behind that feature.
Fixes: QTBUG-83867
Pick-to: 5.15
Change-Id: I7cd9aa73003f45ccbc2613c67c21e4067d6edddd
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
The CA fetcher on Windows works on a separate thread, it can take
quite some time to finish its job and if a connection was meanwhile
closed (via 'abort', 'close' or 'disconnectFromHost') but the socket
is still alive/re-used - we don't want to be fooled by the previous
fetch 'finished' signal, only if it's fetching for the same certificate.
Change-Id: Ibd0a70000ad10cff10207d37d7b47c38e615d0f1
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Also, remove the deprecated function and its auto-test.
Change-Id: If04a54c4b91e0f76523960c7b1a5bcb8fe883ac6
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
When a first call to ::WSASend() consumes less data than requested, we
are trying to make a second call, leaving 'bytesToSend' unmodified.
Instead of setting a new value for 'bytesToSend', it seems reasonable
to break the loop, if the system buffer overflow detected.
Pick-to: 5.15
Change-Id: I8038e95c20221518955efe786f2d0dadb98360fa
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QAbstractSocket::flush stopped being virtual and, instead, QAbstractSocketPrivate::flush
is now virtual. So the code in QSslSocket is essentially a duplicate.
Change-Id: Iebb33c8490268270b90a662ab64b1ae47c4cbac7
Reviewed-by: Alex Trotsenko <alex1973tr@gmail.com>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
If QSslConfiguration::setCaCertificates is set to empty list openssl
cannot verify the OCSP-Response.
Qt will provide it as QSslError::OcspResponseCannotBeTrusted that can be
ignored. But the openssl error is still in the error queue and prevents
a successful reply in QNetworkReply::finished.
So let's clear the queue after OCSP checking to avoid side affects.
Change-Id: I44a7f45a2eebd20ea86a235a0534f80986c40a26
Fixes: QTBUG-85638
Pick-to: 5.15
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
OpenSSL 1.1.1 has a convenient function for parsing ASN.1 times;
use that instead of the hand-rolled code.
Change-Id: Ic0aecc915f362c02b220819305f3f3c347a75297
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Certain resources and compiled jar files have to be copied into the
non-prefix build dir location. They were previously only installed,
which doesn't do anything in a non-prefix build.
Change pro2cmake to generate code that places the compiled java
jars into QT_BUILD_DIR, so that non-prefix builds work.
Place the module dependencies xml files into lib folder in non-prefix
builds.
Don't special case the output and install location of the Android QPA
plugin.
Task-number: QTBUG-85399
Change-Id: I4ac9d3929ea8ecc95ec99a77e621ad2121b68832
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
This is required to remove the ; from the macro with Qt 6.
Task-number: QTBUG-82978
Change-Id: I3f0b6717956ca8fa486bed9817b89dfa19f5e0e1
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
The change creates a slight source incompatibility. The main
things to take care of are
* code using printf statements on list.size(). Using qsizetype in
printf statements will always require a cast to work on both 32
and 64 bit.
* A few places where overloads now get ambiguous. One example is
QRandomGenerator::bounded() that has overloads for int, uint and
double, but not int64.
* Streaming list.size() to a QDataStream will change the format
depending on the architecture.
[ChangeLog][QtCore][QList] QList now uses qsizetype to index into
elements.
Change-Id: Iaff562a4d072b97f458417b670f95971bd47cbc6
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
CFURLStringByAddingXXX was deprecated, so we use QUrl combined
with existing QString-based converters (from/to CFStringRef).
Pick-to: 5.15
Task-number: QTBUG-83992
Change-Id: I6f1a3fdafa9532bb7816a1262169df3421169cb7
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Also add a ; where it is missing.
Task-number: QTBUG-82978
Change-Id: Ic5d2a07363c25ab641d234baca89bc62238458cb
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
No client of QDateTimeParser actually uses it unless datestring was
enabled, nor is it any use without datestring. Various methods
conditioned on datestring are broken unless datetimeparser is enabled.
We can't condition public API on datetimeparser, as it's a private
feature, but client code can condition use of it on the private
feature. All string-to-date/time conversions that use a string format
(this includes all locale-specific formats) depend on feature
datetimeparser.
Change #if-ery (or add it) in all client (including test) code to test
the right feature.
Tidied up some code in the process. Killed some already-redundant
textdate #if-ery. Renamed a test whose name claimed it involved
locale, which it doesn't, in the course of #if-ing it.
This simplifies the condition for feature datetimeedit (which overtly
depended on textdate, redundantly since it depends on datestring which
depends on textdate; its dependence on datetimeparser now makes its
dependency on datestring also redundant).
It also removes the need for assorted datestring checks in
QDateTimeParser itself.
Change-Id: I5dfe3a977042134b2cfb16cbcc795070634e7adf
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
We had such an overloaded version in QSslSocket, it was deprecated without
providing any alternative. Now this function has some use and may be
introduced in Qt6, as QSslConfiguration::setCiphers(const QString &).
Last but not the least - a useless and strange auto-test was removed
(it was creating a list of 5 QSslCiphers each with isNull() == true).
That's becasue '!MD5' or 'ALL' (for example) is not a cipher to be found
in supportedCiphers.
Change-Id: I47eb4c0faa9b52885e883751dd992cd9cb3d26fe
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Not to pollute the global namespace with rather generic names
(especially in case QT_NAMESPACE is none); also drop the
(now)redundant 'Q' prefix in the names.
Change-Id: I57ea7e3996cced705f7ddbdbc1e0231191b31c43
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Closing it early means the user may get this message when reading data
without checking for error:
'QIODevice::read (QNetworkReplyWasmImpl): device not open'
Change-Id: I377fc3da68e9d9f14f8504ace5addbb4f8e3c539
Fixes: QTBUG-85176
Pick-to: 5.15
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Remove quote character, fixing:
Performing C++ SOURCE FILE Test HAVE_openssl_headers succeeded with the following output:
Run Build Command(s)
src.cxx 8 88: warning: missing terminating ' character
8 | # error OpenSSL was reported as >= 1.1.1 but is missing required features, possibly it's libressl which is unsupported
which makes one think the test failed.
Change-Id: I0ac02deee4282ef7c953186b820a31309e82b5e1
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
LibResolv uses a QLibrary which is a QObject that must be deleted
if the QCoreApplication is being destroyed to release the underlying library.
A Q_GLOBAL_STATIC won't release any memory and is not able to
manually release it.
Pick-to: 5.15
Task-number: QTBUG-84234
Change-Id: I97fe5faca309e9c1e85435f602ad7f8c3f633b48
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QHostInfoLookupManager has a QThreadPool as member.
QThreadPool is a QObject and must be deleted if the QCoreApplication
is being destroyed to release the underlying ThreadData.
A Q_GLOBAL_STATIC won't release any memory is not able to
manually release it.
Pick-to: 5.15
Task-number: QTBUG-84234
Change-Id: I96be4601c3af38fa7c753a6f7acb8273ee277a27
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Requires moving construction and destruction of the IConnectionPoint
(contained in QNetworkListManagerEvents contained in
QNetworkStatusMonitorPrivate) and the calls to CoInitialize and
CoUninitialize to start()/stop() so it can be taken down
and brought up in a reasonable way.
Partial squashed cherry pick of bde6b27882
and the following fixup patch 84be436ef0.
Task-number: QTBUG-84031
Change-Id: I9d644261c6c730ea511e28d75da91bd34504bd03
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
(cherry picked from commit bde6b27882)
(cherry picked from commit 84be436ef0)
They were deprecated with replacements in QSslConfiguration proposed (and
some without alternative, which we'll provide if there is any demand
in such an API). Special thanks to M.N. for a nice hint on how to
amend the test without introducing a new API.
Change-Id: I7841a5b3f30469d8204b61cb65921c34275e0650
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
And name the main class QList. That's also the one we document.
This gives less porting pain for our users, and a lot less churn
in our API, as we use QList in Qt 5 in 95% of our API.
In addition, it gives more consistent naming with QStringList and
QByteArrayList and disambiguates QList vs QVector(2|3|4)D.
Fixes: QTBUG-84468
Change-Id: I3cba9d1d3179969d8bf9320b31be2230d021d1a9
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
There is no reason for keep using our macro now that we have C++17.
The macro itself is left in for the moment being, as well as its
detection logic, because it's needed for C code (not everything
supports C11 yet). A few more cleanups will arrive in the next few
patches.
Note that this is a mere search/replace; some places were using
double braces to work around the presence of commas in a macro, no
attempt has been done to fix those.
tst_qglobal had just some minor changes to keep testing the macro.
Change-Id: I1c1c397d9f3e63db3338842bf350c9069ea57639
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The addition of CMakeLists.txt in other directories made the reference
ambiguous. Also fix the Qt version.
Change-Id: I679d86a302ae6c41b84ded66e3f21c5320e4b325
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
I removed the line locally, and then added it back for testing and
forgot to remove it again.
Amends a0bfa4e1f8.
Task-number: QTBUG-84907
Change-Id: I7d6d0237742f4c3fc3ca0a2532d776325f396aab
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The connectivity pre-check has been around for a long time, and it has
caused various issues in that time. Certain scenarios, like using
certain VPN configurations, might confuse the OS into thinking you don't
have and network connectivity at all and abort the connection.
Especially noticeable/frustrating when the connection was going to a
host inside the local network.
The negative impact of this change would at worst be that we might try
to connect and it will wait some amount of time before the OS tells us
the connection failed in situations where it would previously have been
aborted before it started. But the false-negatives are not really an OK
sacrifice in that case.
Fixes: QTBUG-84907
Change-Id: I37fc69051e39df3c1a1fecb56ef54521a4d3d0c3
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
CL = Content-Length
The uploadByteDevice was kept after a redirect which caused the
internals to assume that we had to upload the data. Even if this was
not the case we still transmitted the Content-Length header from the
first request which was now stored in two places.
Fixes: QTBUG-84162
Pick-to: 5.15
Change-Id: Ic86b1ef0766ffcc50beeed96c1c915b721d40209
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>