| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
Everywhere, except for ARM.
Change-Id: I50e2158aeade4256ad1dfffd17b11ca2d57ad1fb
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The typical examples are views and their corresponding owning
container. Since they are different types, the search is heterogeneous,
hence the name.
We have to do it differently from the Standard Library. There, because
the hasher is a template parameter to std::unordered_{map,set}, it can
be a structure with overloads for different types, all of which the
implementer guarantees produce the same hash for input that also
compares equal. For QHash/QSet, we don't have a template parameter.
One alternative solution would be to detect the existence of
qHashEquals(T1, T2) or qHashHeterogeneousEquals() or something.
Change-Id: I664b9f014ffc48cbb49bfffd17b0318c0775a2b5
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
I didn't do it in a template because the 32-bit code requires a
compatibility hack for Qt 6 and it's using an XOR of the high and low
parts.
Fixes: QTBUG-116080
Change-Id: I5dd50a1a7ca5424d9e7afffd17ae0ba5b9ff52f6
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If we add the typedefs, we also need to add the qHash() overloads. See
code comments for the explanation of how it works. I chose to use an
addition to merge the upper and lower parts because of the comment in
QHashCombineCommutative's operator(), introduced by commit
91b44afdcba3f4cb2ddc7a80c0cff723b408b447.
Found in 6.6 API-review, but didn't make the cut and then was
forgotten.
Drive-by fix long line nearby.
[ChangeLog][QtCore] Added qHash() overloads for quint128 and qint128.
Fixes: QTBUG-116054
Task-number: QTBUG-116080
Change-Id: If484aed08ba476e0eace800b719f435203100f3e
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit c0791ac76ec7cfdc3945efa67a6f72ee3623413c didn't explain why it
was #ifdef'ed out. It's just an alias for double. Maybe compilers at the
time used to complain if you used it, but I can't make Apple's clang
produce a warning now.
Pick-to: 6.7
Change-Id: I664b9f014ffc48cbb49bfffd17b02293403e9571
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
|
|
|
|
|
|
|
|
|
|
| |
It appears to be used only in qlalr, which is, however, not
bootstrapped.
Pick-to: 6.7 6.6 6.5 6.2
Change-Id: Idc16d957bf687238c7b0ee603d8b092e2048ef18
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
| |
Nothing in those files uses QPair; and a local build finished fine without them.
Task-number: QTBUG-115841
Change-Id: I669cfecaa9129bce6b31e464826287f138b159db
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
qhashfunctions.h defines a catch-all 2-arguments qHash(T, seed)
in order to support datatypes that implement a 1-argument overload
of qHash (i.e. qHash(Type)). The catch-all calls the 1-argument
overload and XORs the result with the seed.
The catch-all is constrained on the existence of such a 1-argument
overload. This is done in order to make the catch-all SFINAE-friendly;
otherwise merely instantiating the catch-all would trigger a hard error.
Such an error would make it impossible to build a type trait that
detects if one can call qHash(T, size_t) for a given type T.
The constraint itself is called HasQHashSingleArgOverload and lives in a
private namespace.
It has been observed that HasQHashSingleArgOverload misbehaves for
some datatypes. For instance, HasQHashSingleArgOverload<int> is actually
false, despite qHash(123) being perfectly callable. (The second argument
of qHash(int, size_t) is defaulted, so the call *is* possible.)
--
Why is HasQHashSingleArgOverload<int> false?
This has to do with how HasQHashSingleArgOverload<T> is implemented: as
a detection trait that checks if qHash(declval<T>()) is callable.
The detection itself is not a problem. Consider this code:
template <typename T>
constexpr bool HasQHashSingleArgOverload = /* magic */;
class MyClass {};
size_t qHash(MyClass);
static_assert(HasQHashSingleArgOverload<MyClass>); // OK
Here, the static_assert passes, even if qHash(MyClass) (and MyClass
itself) were not defined at all when HasQHashSingleArgOverload was
defined.
This is nothing but 2-phase lookup at work ([temp.dep.res]): the
detection inside HasQHashSingleArgOverload takes into account the qHash
overloads available when HasQHashSingleArgOverload was declared, as well
as any other overload declared before the "point of instantiation". This
means that qHash(MyClass) will be visible and detected.
Let's try something slightly different:
template <typename T>
constexpr bool HasQHashSingleArgOverload = /* magic */;
size_t qHash(int);
static_assert(HasQHashSingleArgOverload<int>); // ERROR
This one *does not work*. How is it possible? The answer is that 2-phase
name lookup combines the names found at definition time with the names
_found at instantiation time using argument-dependent lookup only_.
`int` is a fundamental type and does not participate in ADL. In the
example, HasQHashSingleArgOverload has actually no qHash overloads to
even consider, and therefore its detection fails.
You can restore detection by moving the declaration of the qHash(int)
overload *before* the definition of HasQHashSingleArgOverload, so it's
captured at definition time:
size_t qHash(int);
template <typename T>
constexpr bool HasQHashSingleArgOverload = /* magic */;
static_assert(HasQHashSingleArgOverload<int>); // OK!
This is why HasQHashSingleArgOverload<int> is currently returning
`false`: because HasQHashSingleArgOverload is defined *before* all the
qHash(fundamental_type) overloads in qhashfunctions.h.
--
Now consider this variation of the above, where we keep the qHash(int)
overload after the detector (so, it's not found), but also prepend an
Evil class implicitly convertible from int:
struct Evil { Evil(int); };
size_t qHash(Evil);
template <typename T> constexpr bool HasQHashSingleArgOverload = /* magic */;
size_t qHash(int);
static_assert(HasQHashSingleArgOverload<int>); // OK
Now the static_assert passes. HasQHashSingleArgOverload is still not
considering qHash(int) (it's declared after), but it's considering
qHash(Evil). Can you call *that* one with an int? Yes, after a
conversion to Evil.
This is extremely fragile and likely an ODR violation (if not ODR, then
likely falls into [temp.dep.candidate/1]).
--
Does this "really matter" for a type like `int`? The answer is no. If
HasQHashSingleArgOverload<int> is true, then a call like
qHash(42, 123uz);
will have two overloads in its overloads set:
1) qHash(int, size_t)
2) qHash(T, size_t), i.e. the catch-all template. To be pedantic,
qHash<int>(const int &, size_t), that is, the instantiation of the
catch-all after template type deduction for T (= int)
([over.match.funcs.general/8]).
Although it may look like this is ambiguous as both calls have perfect
matches for the arguments, 1) is actually a better match than 2) because
it is not a template specialization ([over.match.best/2.4]).
In other words: qHash(int, size_t) is *always* called when the argument
is `int`, no matter the value of HasQHashSingleArgOverload<int>. The
catch-all template may be added or not to the overload set, but it's
a worse match anyways.
--
Now, let's consider this code:
enum MyEnum { E1, E2, E3 };
qHash(E1, 42uz);
This code compiles, although we do not define any qHash overload
specifically for enumeration types (nor one is defined by MyEnum's
author).
Which qHash overload gets called? Again there are two possible
overloads available:
1) qHash(int, size_t). E1 can be converted to `int` ([conv.prom/3]),
and this overload selected.
2) qHash(T, size_t), which after instantiation, is qHash<MyEnum>(const
MyEnum &, size_t).
In this case, 2) is a better match than 1), because it does not require
any conversion for the arguments.
Is 2) a viable overload? Unfortunately the answer here is "it depends",
because it's subject to what we've learned before: since the catch-all
is constrained by the HasQHashSingleArgOverload trait, names introduced
before the trait may exclude or include the overload.
This code:
#include <qhashfunctions.h>
enum MyEnum { E1, E2, E3 };
qHash(E1, 42uz);
static_assert(HasQHashSingleArgOverload<MyEnum>); // ERROR
will fail the static_assert. This means that only qHash(int, size_t) is
in the overload set.
However, this code:
struct Evil { Evil(int); };
size_t qHash(Evil);
#include <qhashfunctions.h>
enum MyEnum { E1, E2, E3 };
qHash(E1, 42uz);
static_assert(HasQHashSingleArgOverload<MyEnum>); // OK
will pass the static_assert. qHash(Evil) can be called with an object of
type MyEnum after an user-defined conversion sequence
([over.best.ics.general], [over.ics.user]: a standard conversion
sequence, made of a lvalue-to-rvalue conversion + a integral promotion,
followed by a conversion by constructor [class.conv.ctor]).
Therefore, HasQHashSingleArgOverload<MyEnum> is true here; the catch-all
template is added to the overload set; and it's a best match for the
qHash(E1, 42uz) call.
--
Is this a problem? **Yes**, and a huge one: the catch-all template does
not yield the same value as the qHash(int, size_t) overload. This means
that calculating hash values (e.g. QHash, QSet) will have different
results depending on include ordering!
A translation unit TU1 may have
#include <QSet>
#include <Evil>
QSet<MyEnum> calculateSet { /* ... */ }
And another translation unit TU2 may have
#include <Evil>
#include <QSet> // different order
void use() {
QSet<MyEnum> set = calculateSet();
}
And now the two TUs cannot exchange QHash/QSet objects as they would
hash the contents differently.
--
`Evil` actually exists in Qt. The bug report specifies QKeySequence,
which has an implicit constructor from int, but one can concoct infinite
other examples.
--
Congratulations if you've read so far.
=========================
=== PROPOSED SOLUTION ===
=========================
1) Move the HasQHashSingleArgOverload detection after declaring the
overloads for all the fundamental types (which we already do anyways).
This means that HasQHashSingleArgOverload<fundamental_type> will now
be true. It also means that the catch-all becomes available for all
fundamental types, but as discussed before, for all of them we have
better matches anyways.
2) For unscoped enumeration types, this means however an ABI break: the
catch-all template becomes always the best match. Code compiled before
this change would call qHash(int, size_t), and code compiled after this
change would call the catch-all qHash<Enum>(Enum, size_t); as discussed
before, the two don't yield the same results, so mixing old code and new
code will break.
In order to restore the old behavior, add a qHash overload for
enumeration types that forwards the implementation to the integer
overloads (using qToUnderlying¹).
(Here I'm considering the "old", correct behavior the one that one gets
by simply including QHash/QSet, declaring an enumeration and calling
qHash on it. In other words, without having Evil around before including
QHash.)
This avoids an ABI break for most enumeration types, for which one
does not explicitly define a qHash overload. It however *introduces*
an ABI break for enumeration types for which there is a single-argument
qHash(E) overload. This is because
- before this change, the catch-all template was called, and that
in turn called qHash(E) and XOR'ed the result with the seed;
- after this change, the newly introduced qHash overload for
enumerations gets called. It's very likely that it would not give
the same result as before.
I don't have a solution for this, so we'll have to accept the ABI
break.
Note that if one defines a two-arguments overload for an enum type,
then nothing changes there (the overload is still the best match).
3) Make plans to kill the catch-all template, for Qt 7.0 at the latest.
We've asked users to provide a two-args qHash overload for a very long
time, it's time to stop working around that.
4) Make plans to switch from overloading qHash to specializing std::hash
(or equivalent). Specializations don't overload, and we'd get rid of
all these troubles with implicit conversions.
--
¹ To nitpick, qToUnderlying may select a *different* overload than
the one selected by an implicit conversion.
That's because an unscoped enumeration without a fixed underlying type
is allowed to have an underlying type U, and implicitly convert to V,
with U and V being two different types (!).
U is "an integral type that can represent all the enumerator values"
([dcl.enum/7]). V is selected in a specific list in a specific order
([conv.prom]/3). This means that in theory a compiler can take enum E {
E1, E2 }, give it `unsigned long long` as underlying type, and still
allow for a conversion to `int`.
As far as I know, no compiler we use does something as crazy as that,
but if it's a concern, it needs to be fixed.
[ChangeLog][Deprecation Notice] Support for overloads of qHash with only
one argument is going to be removed in Qt 7. Users are encouraged to
upgrade to the two-arguments overload. Please refer to the QHash
documentation for more information.
[ChangeLog][Potentially Binary-Incompatible Changes] If an enumeration
type for which a single-argument qHash overload has been declared is
being used as a key type in QHash, QMultiHash or QSet, then objects of
these types are no longer binary compatible with code compiled against
an earlier version of Qt. It is very unlikely that such qHash overloads
exist, because enumeration types work out of the box as keys Qt
unordered associative containers; users do not need to define qHash
overloads for their custom enumerations. Note that there is no binary
incompatibity if a *two* arguments qHash overload has been declared
instead.
Fixes: QTBUG-108032
Fixes: QTBUG-107033
Pick-to: 6.2 6.4
Change-Id: I2ebffb2820c553e5fdc3a341019433793a58e3ab
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
To allow the user to customize the C++ code that QDoc sees, so as to be
able to work-around some limitations on QDoc itself, QDoc defines two
symbols: Q_QDOC and Q_CLANG_QDOC, both of which are "true" during an
entire execution of QDoc.
At a certain point in time, QDoc allowed the user the choice between a
custom C++ parser and a Clang based one.
The Q_QDOC symbol would always be defined while the Q_CLANG_QDOC symbol
would be defined only when the Clang based parser was chosen.
In more recent times, QDoc always uses a Clang based parser, such that
both Q_CLANG_QDOC and Q_QDOC are always defined, making them equivalent.
To avoid using different symbols, and the possible confusion and
fragmentation that derives from it, all usages of Q_CLANG_QDOC are now
replaced by the equivalent usages of Q_QDOC.
Change-Id: I5810abb9ad1016a4c5bbea99acd03381b8514b3f
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
| |
It's not used and not useful.
The macro itself has to stay, for now, because Qt5Compat uses it, too.
Task-number: QTBUG-100861
Pick-to: 6.4 6.3 6.2
Change-Id: I5d0557a6c959d6facf6e47f26786a9d365339e95
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We have in the past found ourselves in need of disambiguating new
overloads from older, QT_REMOVED_SINCE'ed ones, which we usually did
by adding a defaulted int argument, cf. e.g. QMetaType::id().
The problem with a defaulted int argument is twofold: First, an int is
a valid argument type, and users may get confused as to its meaning
when presented with the signature in the IDEs auto-completion popup,
and a lot of things implicitly convert to an int, so any errornous
parameter passing may stay unnoticed until we remove these fake
arguments come Qt 7.
Second, this way of doing things requires a lot of ifdef'ery to keep
the additional argument out of both the docs and future Qt 7 builds.
The solution presented in this patch is to create a tag type,
QDisambiguated_t, which a) more clearly communicates its purpose and
b) doesn't implicitly convert from anything.
To help with the invariably ugly ifdef'ery, provide a set of macros
that hide this stuff from qdoc and Qt 7 builds.
Use the macros to replace the fake int arguments that were added for
6.4.
Pick-to: 6.4
Change-Id: I6916f38c8eb9793ad6dea5f61e7e5fff7e75e273
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The QBAV one should pass the parameter by value, like QStringView. And
now that we have it, the non-View one should call the View one in an
inline function, like we already do for QString.
The extra, defaulted parameter is there only so we get a different
signature in the new inline function compared to the removed one.
Pick-to: 6.4
Change-Id: If05aeeb7176e4f13af9afffd16e7f08062b1dc86
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Replace the current license disclaimer in files by
a SPDX-License-Identifier.
Files that have to be modified by hand are modified.
License files are organized under LICENSES directory.
Task-number: QTBUG-67283
Change-Id: Id880c92784c40f3bbde861c0d93f58151c18b9f1
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
|
|
|
|
|
|
|
|
| |
[ChangeLog][QtCore][QByteArrayView] Added std::hash specialization.
Change-Id: If05aeeb7176e4f13af9afffd16e7f03dfa413a66
Reviewed-by: Rui Oliveira
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
|
|
|
|
|
|
|
| |
Task-number: QTBUG-98434
Change-Id: Ib7c5fc0aaca6ef33b93c7486e99502c555bf20bc
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
|
|
|
|
|
|
| |
Task-number: QTBUG-98434
Change-Id: I37b70a1b7bc164fa7dc9cb702827297ca66a6fdc
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
While 'pure' in GCC is weaker than Stepanov's Regular Procedure from
Elements of Programming (equal result for equal inputs), it does not
allow accesses to volatile memory:
> functions declared with the pure attribute can safely read any
> non-volatile objects
The globalSeed() function reads from an atomic variable that can be
changed at any time from another thread.
Atomics, while not volatile objects in the sense of the keyword, must
fall under the pure attribute doc's exclusion criterion:
The difference between a volatile and an atomic access, while
important for the implementation of the function, is indistinguishable
to the caller of the function: both volatile and atomic objects can
change value without the current thread of execution changing them,
with no way for the caller of the function to distinguish which one
occurred.
Therefore, globalSeed() should not be pure.
5.15 is not affected, as qGlobalQHashSeed() is not marked as pure.
Task-number: QTBUG-62185
Pick-to: 6.3 6.2
Change-Id: I6fc52e2bd41ef4aa3f8039072b47c7a1314b98fa
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
| |
... instead of QT_PREPEND_NAMESPACE(qHash), which is qualified (prepends at least '::'), and therefore disables ADL.
This is not a problem as long as we wrote our qHash() overloads as free functions (incl. non-hidden friends), but it should™ fail for hidden friends, so use the old using-std::swap() trick to bring QT_PREPEND_NAMESPACE(qHash) into scope, proceeding with an unqualified lookup.
Pick-to: 6.2
Change-Id: I00860b2313699849f86bfe3dd9f41db4ce993cd3
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
There were two problems here: first, qHash(std::pair) must be declared
before qHashMulti that might call back to qHash(std::pair) (i.e., a pair
with one element that is also a pair). But moving the declaration above
causes the second problem: the noexcept expression can't refer to qHash
functions that aren't declared yet. So we forward-declare a constexpr
function for that result, but implement it far below.
Fixes: QTBUG-92910
Change-Id: Ia8e48103a54446509e3bfffd16767ed2e29b026c
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
| |
That's two years from when the replacements were added (6.2).
Change-Id: Id2983978ad544ff79911fffd1671f7dd38fede02
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It is noexcept, except when initializing. When initializing, let's just
use qEnvironmentVariableIntValue (which we should have used anyway),
which avoids the memory allocation and is noexcept. The QRandomGenerator
functions are not marked noexcept, but are mostly so: they can't throw
regular exceptions, but some implementations do call POSIX Thread
Cancellation Points, which may cause forced stack unwinding. That's
unlikely to happen at the moment of the QHash initialization.
This is also mitigated in the next commit.
Change-Id: Id2983978ad544ff79911fffd1671fd16f8d6378d
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
|
|
| |
Commit 37e0953613ef9a3db137bc8d3076441d9ae317d9 added a to-do, but we
can actually change the type, since we've documented since Qt 5.10 that
setting a non-zero value (aside from -1) with qSetGlobalQHashSeed was
not allowed. Storing a value to be reset later is simply not supported.
Change-Id: Id2983978ad544ff79911fffd1671f7b5de284bab
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
|
|
|
|
|
|
|
|
|
| |
Stop going through the implicit int conversion.
[ChangeLog][QtCore][QFlags] QFlags now has a qHash() overload.
Change-Id: Id380ed252695f24af2e8c239b650dcb6f44e2893
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
| |
The hashing seed's type has been changed from int to size_t in Qt 6.
However the functions setting/getting the seed, and the seed itself,
are still simply int, meaning that we've crippled our seeding.
Add a TODO to amend it.
Change-Id: Ie9dd177149ec299ccf16d4e31f9f4b065804cfed
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In some cases, the default equality operator for a class is not suitable
for using in hashing (for example because it uses fuzzy comparisons).
Add a qHashEquals() method that by default uses the equality operator, but
allows to tailor the operations that should be used when using the class
as a key in QHash.
Task-number: QTBUG-88966
Change-Id: I346cf0e6e923277a8b42a79e50342a1c2511fd80
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
(cherry picked from commit 5d8b586e73e37070b0303bee24372550854637eb)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
To support qHash overloads without a seed we have a qHash(T, size_t)
catch-all that calls qHash(T) and XORs the seed. The problem is
that this catch-all is not SFINAE friendly. For a type Foo which
does not have any qHash overload, we can't ask if qHash(Foo, size_t)
is callable because it would instantiate the catch-all and fail
to compile.
Add a suitable trait and enable_if.
Pick-to: 6.0
Change-Id: Idffd48a537eebaf77cee7030b8d91a302643ffde
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
|
|
|
|
|
| |
Change-Id: Ibe796c398a8e5488b7203abb07aa54740744f1ab
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
C++20 via P1120 is deprecating arithmetic operations between
unrelated enumeration types, and GCC 10 is already complaining.
Hence, these operations might become illegal in C++23 or C++26 at
the latest.
A case of this that affects Qt is in key combinations: a
QKeySequence can be constructed by summing / ORing modifiers and a
key, for instance:
Qt::CTRL + Qt::Key_A
Qt::SHIFT | Qt::CTRL | Qt::Key_G (recommended, see below)
The problem is that the modifiers and the key belong to different
enumerations (and there's 2 enumerations for the modifier, and one
for the key).
To solve this: add a dedicated class to represent a combination of
keys, and operators between those enumerations to build instances
of this class.
I would've simply defined operator|, but again docs and pre-existing
code use operator+ as well, so added both to at least tackle simple
cases (modifier + key).
Multiple modifiers create a problem: operator+ between them yields
int, not the corresponding flags type (because operator+ is not
overloaded for this use case):
Qt::CTRL + Qt::SHIFT + Qt::Key_A
\__________________/ /
int /
\______________/
int
Not only this loses track of the datatypes involved, but it would
also then "add" the key (with NO warnings, now its int + enum, so
it's not mixing enums!) and yielding int again.
I don't want to special-case this; the point of the class is
that int is the wrong datatype. Everything works just fine when
using operator| instead:
Qt::CTRL | Qt::SHIFT | Qt::Key_A
\__________________/ /
Qt::Modifiers /
\______________/
QKeyCombination
So I'm defining operator+ so that the simple cases still work,
but also deprecating it.
Port some code around Qt to the new class. In certain cases,
it's a huge win for clarity. In some others, I've just added
the necessary casts to make it still compile without warnings,
without attempting refactorings.
[ChangeLog][QtCore][QKeyCombination] New class to represent
a combination of a key and zero or more modifiers, to be used
when defining shortcuts or similar.
[ChangeLog][Potentially Source-Incompatible Changes] A keyboard
modifier (such as Qt::CTRL, Qt::AltModifier, etc.) should be
combined with a key (such as Qt::Key_A, Qt::Key_F1, etc.) by using
operator|, not operator+. The result is now an object of type
QKeyCombination, that stores the key and the modifiers.
Change-Id: I657a3a328232f059023fff69c5031ee31cc91dd6
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
|
|
|
|
|
|
|
|
|
| |
Export some private functions from QUtf8 to resolve
undefined symbols in Qt5Compat after moving QStringRef.
Task-number: QTBUG-84437
Change-Id: I9046dcb14ed520d8868a511d79da6e721e26f72b
Reviewed-by: Lars Knoll <lars.knoll@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>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Created a QByteArrayView in symmetry with QStringView.
Added the basic tests symmetrical to QStringView tests.
Moved the implementations of non-modifying methods of QByteArray to
namespace QtPrivate, to be reused inline from both QByteArray and
QByteArrayView. Changed QByteArray's counterparts of those methods to
take QByteArrayView as argument instead of QByteArray. Removed
QByteArray's operator QNoImplicitBoolCast(), because it was causing
ambiguity when calling those methods with QByteArray argument (it was
there to perevnt if(!ba)/if(ba) from compiling, but currently that would
be ambiguous and won't compile anyway).
[ChangeLog][QtCore][QByteArrayView] New class.
Task-number: QTBUG-84321
Change-Id: I05f92e654cf65c95f2bb31b9c9018746ac110426
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
| |
Task-number: QTBUG-84319
Change-Id: If77bc94c18e8d522b4577050091cd7d7aa941311
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Make QPair an alias for std::pair, and qMakePair just a forwarder
towards std::make_pair.
Why? Fundamentally to ditch a bunch of NIH code; gain for free
structured bindings, std::tuple and std::reference_wrapper
compatibility, and so on.
Breakages:
* Some that code manually forward declares QPair.
We don't care about it (<QContainerFwd> is the proper way).
* Some code that overloads on std::pair and QPair. Luckily
it's mostly centralized: debug, metatypes, testing macros.
Just remove the QPair overload.
* Usages of qMakePair forcing the template type parameters.
There are a handful of these in qtbase, but only one was actually
broken.
* std::pair is NOT (and will never likely be) trivially copiable.
This is agreed to be a mistake done by practically all implementations
in C++11, can can't be fixed without breaking ABI.
Some code using QPair assuming it's trivially copiable may break;
exactly one occurrence was in qtbase.
* QMetaType logic extracts the type names in two different ways,
one by looking at the source code string (e.g. extracted by moc)
and one via some ad-hoc reflection in C++. We need to make
"QPair" (as spelled in the source code) be the same as "std::pair"
(gathered via reflection, which will see through the alias)
when compared. The way it's already done e.g. for QList is
by actually replacing the moc-extracted name with the name
of the actual type used in C++; do the same here.
On libc++, std::pair is actually in an inline namespace --
i.e. std::__1::pair; the reflection will extract and store
"std::__1::pair" so we need an ad-hoc fix to QMetaType.
[ChangeLog][QtCore][QPair] QPair is now an alias to std::pair,
and does not exist as a class in Qt any more. This may break
code such as functions overloaded for both QPair and std::pair.
Usually, the overload taking a QPair can be safely discarded,
leaving only the one taking a std::pair. QPair API has not changed,
and qMakePair is still available for compatibility (although
new code is encouraged to use std::pair and std::make_pair
directly instead).
Change-Id: I7725c751bf23946cde577b1406e86a336c0a3dcf
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add a helper function so that we have a shortcut.
Instead of writing:
QHashCombine hash;
seed = hash(seed, fieldA);
seed = hash(seed, fieldB);
// etc.
return seed;
one can now simply write:
return qHashMulti(seed, fieldA, fieldB, fieldC);
Port a few usages inside qtbase as a demonstration.
[ChangeLog][QtCore][QHash] Added the qHashMulti and
qHashMultiCommutative functions as convenience helpers
to calculate a hash from multiple variables (typically,
data members of a class).
Change-Id: I881a9ad41168df20ceecc6588a94abe7ddc6a532
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
|
|
|
|
|
|
|
|
|
|
| |
Except for (void and) bool, which we may not want to have
to avoid accidental implicit conversions.
Drive-by, rearrange qHash overloads to C++ types first,
and then Qt ones.
Change-Id: I9c4ecef5f28568d35ca52e339583851ce53b3bae
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
| |
Condition is a compile-time one.
Change-Id: I6e60f12cc51e96b2528c375017357c0631e2fc0b
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
|
|
|
|
|
|
| |
Change-Id: Id5e091b135c006b10987f229f45319228edb8675
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
|
|
|
|
|
|
| |
Change-Id: I8728ba339161e210772e73c633cb2309dfb01b8e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
|
|
|
|
|
|
|
|
|
|
|
| |
This is required, so that QHash and QSet can hold more
than 2^32 items on 64 bit platforms.
The actual hashing functions for strings are still 32bit, this will
be changed in a follow-up commit.
Change-Id: I4372125252486075ff3a0b45ecfa818359fe103b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The hash function provides rather good mixing of the input bits.
It spreads numbers out evenly through the uint range, a change of
one bit in the input changes around half the output bits, and it is
pretty fast.
Using this as a hash function over the simple hash(int) == int has the
advantage that it reduces the amount of collisions for badly distributed
keys. In addition, it allows us to always use power of two sizes for the
hash table, leading to better performance for inserts and lookups.
the 32 and 64 bit hash functions where chosen from
https://nullprogram.com/blog/2018/07/31/. I selected the ones that give
a very good distribution of the hash values while using the integer for
both multiplication steps. This should be slighty faster than using two
different numbers.
While the result is still being cast to a uint, the method is prepared
so it can handle 64 bit keys and seeds.
Fixes: QTBUG-29009
Change-Id: Id7a1b97b3c0d219e65de2e6e1fe6faf092f8ce16
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We have a problem. Our types don't play well with the std unordered
containers, because they do not specialize std::hash. We therefore
force our users to come up with an implementation, hindering
interoperability, since any two developers are unlikely to come up
with compatible implementations. So combining libraries written by
different developers will result in ODR violations.
Now that we depend on C++11, and thus the presence of std::hash, we
still face the problem that the standard does not provide us with a
means to compose new hash functions out of old ones. In particular, we
cannot, yet, depend on C++17's std::hash<std::string_view> to
implement std::hash<QByteArray>, say. There's also no std::hash for
std::tuple, which would allow easy composition by using std::tie().
So piggy-back on the work we have done over the years on qHash()
functions, and implement the std::hash specializations for Qt types
using the existing qHash() functions, with a twist: The standard
allows implementations to provide means against predictable hash
values. Qt has this, too, but the seed is managed by the container and
passed to the qHash() function as a separate argument. The standard
does not have this explicit seed, so any protection must be implicit
in the normal use of std::hash.
To reap whatever protection that std library has on offer, if any, we
calculate a seed value by hashing int(0). This will be subject to
constant folding if there's no actual seed, but will produce a value
dependent on the seed if there is one.
Add some tests.
A question that remains is how to document the specialization. Can we
have a \stdhashable QDoc macro that does everything for us?
Task-number: QTBUG-33428
Change-Id: Idfe775f1661f8489587353c4b148d76611ac76f3
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
|
|\
| |
| |
| | |
refs/staging/dev
|
| |
| |
| |
| |
| |
| |
| | |
In preparation of Qt6 move away from pre-C++11 macros.
Change-Id: I44126693c20c18eca5620caab4f7e746218e0ce3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|/
|
|
|
| |
Change-Id: I91ac9e714a465cab226b211812aa46e8fe5ff2ab
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
| |
This way we can easily use them as keys in QHash and QSet.
Change-Id: Ie744c3b5ad1176ba2ab035c7e650af483757a0c9
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
| |
Two of the three functions were for functions that
should not be documented. The third was a function
protected by #ifndef Q_OS_DARWIN, which required a
test of Q_CLANG_QDOC in the header and cpp files.
Change-Id: Id2ab3e4f2ea896dc628a622de2e80a19c18eb9fe
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
|
|
|
|
|
|
|
|
|
| |
There are some callers of qt_hash that first build a string just to hash it.
By allowing to pass an initial value for 'h', we can chain qt_hash() calls
to avoid having to allocate memory just to hash a two-part string.
Change-Id: Ifaca82d47b2fb8c707912342c3ddd84f91e70267
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
| |
[ChangeLog][QtCore] Added qHash(QStringView).
Change-Id: I69c9203cf301fe586e924168381884aab2e19e5c
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Anton Kudryavtsev <antkudr@mail.ru>
|
|
|
|
|
|
| |
Change-Id: I49d07689e642d26b4bceda5cace738aadd828ce0
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Anton Kudryavtsev <antkudr@mail.ru>
|
|\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Conflicts:
configure
src/3rdparty/angle/src/libANGLE/renderer/d3d/d3d11/Renderer11.cpp
src/network/access/qnetworkaccessmanager.cpp
src/plugins/platforms/cocoa/qcocoacolordialoghelper.mm
src/plugins/platforms/eglfs/deviceintegration/eglfs_kms/qeglfskmsgbmcursor.cpp
src/plugins/platforms/eglfs/deviceintegration/eglfs_kms/qeglfskmsgbmcursor.h
src/widgets/widgets/qlineedit_p.cpp
src/widgets/widgets/qlineedit_p.h
src/winmain/winmain.pro
tests/auto/corelib/io/qstorageinfo/tst_qstorageinfo.cpp
tests/auto/dbus/qdbusconnection/tst_qdbusconnection.cpp
tests/auto/dbus/qdbusconnection/tst_qdbusconnection.h
tests/auto/testlib/selftests/expected_cmptest.teamcity
tests/auto/testlib/selftests/expected_cmptest.txt
tests/auto/widgets/itemviews/qtableview/tst_qtableview.cpp
tools/configure/configureapp.cpp
Change-Id: Ib9997b0d0f91946e4081d36c0c6b696c5c983b2a
|