constData() was a workaround for the tests
Task-number: QDS-9847
Change-Id: I23fcca794ab50101ae70034079af46303c98c648
Reviewed-by: Marco Bubke <marco.bubke@qt.io>
GCC is recognized by clangso we silence both.
Change-Id: I377caa344a60636e3c73404d9d02a4dfc2eda794
Reviewed-by: Burak Hancerli <burak.hancerli@qt.io>
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
For performance reason we want to get the qml modules directly from the
file system. When the project storage is finished we can get the modules
from there.
Task-number: QDS-9542
Change-Id: I26d4b028fbf5ebc541fcd8e34d285ded1fb14935
Reviewed-by: <github-actions-qt-creator@cristianadam.eu>
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
Since we also license under GPL-3.0 WITH Qt-GPL-exception-1.0,
this applies only to a hypothetical newer version of GPL, that doesn't
exist yet. If such a version emerges, we can still decide to relicense...
While at it, replace (deprecated) GPL-3.0 with more explicit GPL-3.0-only
Change was done by running
find . -type f -exec perl -pi -e "s/LicenseRef-Qt-Commercial OR GPL-3.0\+ OR GPL-3.0 WITH Qt-GPL-exception-1.0/LicenseRef-Qt-Commercial OR GPL-3.0-only WITH Qt-GPL-exception-1.0/g" {} \;
Change-Id: I5097e6ce8d10233993ee30d7e25120e2659eb10b
Reviewed-by: Eike Ziller <eike.ziller@qt.io>
Replace the current license disclaimer in files by
a SPDX-License-Identifier.
Task-number: QTBUG-67283
Change-Id: I708fd1f9f2b73d60f57cc3568646929117825813
Reviewed-by: Eike Ziller <eike.ziller@qt.io>
We expect mostly a string view which has no null termination. Code which
depends on null termination is broken and it is better break early.
Change-Id: I7c2c41fb114e6aaf3a23053b522b37f7af5e1abf
Reviewed-by: <github-actions-qt-creator@cristianadam.eu>
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
There can be quite some copies with the swap method for the default case
that the string is inside the small string area. Instead now we minimize
copies of the smallstring area at the expense of more branches.
The swap approch needs no self tests and it uses the destructor to
remove allocated resources but it can copy three times more often. And
the destructor still has an additional branch in the copy case. Only in
the move case we can use the destructor of the other instance. This can
be surprising for some users which expect that the instance is empty
after a move(actually assuming that a instance is empty after a move is
undefined behavior).
Because SmallString is designed to hold the string in almost all cases
inside the small string area that case should be optimized.
Change-Id: I22cb62aa99b9713a221f103971f149c05e9ff6fa
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
So far the whole class was always copied. For a large SmallString
classes this were often many zeros if it was empty. Now only the non
zero area is copied in that case.
Change-Id: Ic0ea2d6f763513858a993bba3e829c9fc947e1a6
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Reviewed-by: <github-actions-qt-creator@cristianadam.eu>
Using a value after the move is undefined. For example, if you move an
integer, it is not set to zero after the move. So you have to initialize
it again. Adding the take method to SmallString makes this explicit.
So if there is a case that you want to use the string again after
a move, use take or initialize it yourself.
Change-Id: I174116df9639d6a3f63c6f2c2a3bd184852927ce
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Skip the creation auf a QByteArray. Use QStringEncoder instead.
Change-Id: I3d2297b1fef27acb2136a64587c5af9099d92aa6
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
If you write
Utils::SmallStringView view;
Utils::SmallString text{view};
it selects the std::initializer_list contructor. Not the didicated
constructore. It is much to easy to get it wrong so it is better
to make it explicit.
Change-Id: I4240eaf1f39cf71d37df4480fea1ecfa3ea83cb0
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
The small string control block moved to the beginning, so it is more cache
local. The control block is cleanup too, so it should be easier to read.
The alignment is removed because it is creating to big holes.
Change-Id: I401aeb9d55455cbaa5e722dd8192e54b525ddc40
Reviewed-by: Ivan Donchevskii <ivan.donchevskii@qt.io>
Different types are introduced for normalized and native file path. So the
compiler is warning you if you try the wrong format.
Change-Id: I1da0686b142cbf9bb7578468c2b50f90a94cebf9
Reviewed-by: Ivan Donchevskii <ivan.donchevskii@qt.io>
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Before the string was simply swapped with the other string which can lead
to an unexpected behavior for xvalues. Now the destructor of the source is
called and it is default initialized.
foo = std::move(bar);
bar would now hold the value of foo.
Change-Id: Ibea3f18333a168634b7faf2fdaf9b5b52c82d5cc
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
The database is using file path integer ids to handle file paths because
otherwise we would save many redundant data. This patch is improving it
further with the introduction of a database based file path cache. The
entries are now divided in a directory path and file name. This is quite
handy for directory based file watching.
Change-Id: I03f2e388e43f3d521d6bf8e39dfb95eb2309dc73
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
It is required by many template function because it is used in the STL.
Change-Id: I5a2b9266d0c9d79bdc37cbbc4e7b62fac7a82b8b
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
The reserve function is quite complicated because it is implementing a
grow pattern. Something we seldom need. So we now align to the next cache
line size.
Change-Id: I14bb88c12bd740a7afa7cd08969a4e07fb0f9add
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
It is more efficient to use compare instead of less and equal together.
Change-Id: I772b03ba02509c0ebc5d5d770be14cf7d5bd431e
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
We called reserve for even if we do not found anything to replace.
Change-Id: I01dfbb9e8152605c01fa007eb93586f9a30cf56e
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
You need to allocate the memory before you replace a smaller text with
a larger one. It has some overhead because you go two times over the text
but it is anyway not designed for a large text.
Change-Id: I2f56f1300a5daf9e26632b5b0667023a09ff4bd2
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
The short string capacity is the data area size minus the size of \0.
Change-Id: Iffb4428bb81fc78bdc4c00218cf9f2c9ac8c25e5
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
If you call reserve on a read only reference we always allocated on the
heap which is not that smart.
Change-Id: Ib9653c6fc87bc65716a966545c13f7ecb3712039
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
The bug was that we returned the new end() but not the old reallocated
after a reallocation of the memory. The return value is used to move the
string after the found sub string to the next found sub string to the
new position of after the replaced sub string. Because the new end moved
more to the right we moved a much bigger chunk of memory. So we
accessed memory outside of the allocated memory.
Change-Id: I9ac276e201eabee896190da0cf0ee1d86860777d
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
It should not be used for chained concatenation. Use initializer list
instead.
auto text = in + 'x';
auto text = Utils::SmallString{in, "x", other, ", "};
Change-Id: I453d986913eae89fd5e1f525b9f4e0c4bd089467
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
We use std::to_string except for integer where we want a low overhead
solution.
Change-Id: I16ce7d575d83ff56e61a5038aa7f9a1febfaf34f
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Reviewed-by: Orgad Shaneh <orgads@gmail.com>
It is much more performant to have a overload of char instead of providing
a string all the time.
Change-Id: I1a4ed82bf056f6af0c1f91c236b3fc30afa7f5d8
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
We used only 6 bit to save the short size but for SmallString with a size
over 64 it is not enough. So we have now to use a uint16 instead of a
uint8 if the size if over 64.
Change-Id: I53558e492b6cb40b739b23a8af83d192a2e11bd2
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
You can now write
SmallString text = {"Oh ", women[4], " how much I miss you"};
if it can be casted to SmallStringView.
Change-Id: I86b69ee8d735017cac4391e7c4e68355eb5f227b
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Compiling every header file again and again is quite time comsuming. There
are technics to improve this like preambles(a kind of automated
precompiled header) but they don't share their data between translation
units. This approach provides an automatically generated precompiled
header for every project and subproject to improve the loading time.
Change-Id: I34f5bd4db21951175920e2a9bbf6b97b1d705969
Reviewed-by: Eike Ziller <eike.ziller@qt.io>
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Thomas provided a smaller fix.
This reverts commit 78fb7f44bf.
Change-Id: I1edff6477526fe4ce29df38852f47fe49117561e
Reviewed-by: Alessandro Portale <alessandro.portale@qt.io>
We now support constexpr, noexcept and alignas on all compilers. Next will
be C++ 14 constexpr support.
Change-Id: I254049c60dcd82fc373ec907cbb5702a4b344978
Reviewed-by: Tobias Hunger <tobias.hunger@qt.io>
The capacity of a real only reference is zero but the size is larger than
zero. So if we reserve memory the new capacity has to be bigger than the
size and the capacity.
Change-Id: I8b423da7e7c1cf7cee081d1b3f3464b6fb3f67cf
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Reviewed-by: Tobias Hunger <tobias.hunger@qt.io>
We have to copy the content of SmallString before we instantiate a new
constructor in the same memory. So the content to which data() is pointing
can be already invalid.
Change-Id: I3a0ab4f9ac0c1219c2bd75fc4412eaf56209ca64
Reviewed-by: Tobias Hunger <tobias.hunger@qt.io>
Sometimes you want to compare unix and windows texts in tests. This is a
convenient function to remove the carriage returns.
Change-Id: I164298b70d3d775dd45903ea3753ac0e68ed2fdc
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
