When applicable, the flag makes a significant difference in
performacnce. However is easily forgotten in cases where it
is applicable, and cases where it is accidentally used are
easily spotted visually. So arguably, opt-out is a better
default than opt-in.
Change-Id: I88aeb05b42533ba4e23c39573a3ddd016343cd55
Reviewed-by: hjk <hjk@qt.io>
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
These are functional replacements for
QVariant::fromValue(QVariantMap) (or QVariant::fromValue(Store)) and
QVariant::toMap() (or QVariant::toValue<Store>())
We will have a few code paths in the end that need to explicitly
operarate on both QVariantMap and Store (e.g. actual reading/writing
to keep format compatibility etc), so these can't in the end be
simple to/fromValue(OneType) but need an internal 'if' or such.
Change-Id: I954f3cb24fa8fe123162b72bbd25d891dd19b768
Reviewed-by: Marcus Tillmanns <marcus.tillmanns@qt.io>
After this one, is compiles with the opt-in macro. Not changed yet,
though.
Change-Id: I29a66ecb5daa71d0d97566b81fd9f47d92f6368a
Reviewed-by: Marcus Tillmanns <marcus.tillmanns@qt.io>
Also avoid re-instantiation of factories in the build configurations.
Closer to the setups of the other kit aspect(factories).
Change-Id: I1b74a68287b63ee94ff18106d3a00b23624a601b
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
KitAspect -> KitAspectFactory
KitAspectWidget -> KitAspect
KitAspect::createConfigWidget() -> KitAspectFactory::createKitAspect()
Possibly not perfect yet, but closer to what it does and is.
There will be follow-ups to rename derived classes.
Change-Id: I226f6f025f1e6462c348b68663f776f5a2e8c42d
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
This was pretty much ad-hoc so far. Now do the actual value proliferation
inside the aspect first and collect information on what changed. Signal
changes in the end centrally. At that time the aspect internal state is
consistent again.
Additional design desision implemented here: setDefaultValue and
fromMap/readSetting do _not_ signal, user code with unusual needs
has to check on its own.
Change-Id: I1e46282558e014f51933d1044a72bf5c67437ec5
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Marcus Tillmanns <marcus.tillmanns@qt.io>
Reviewed-by: <github-actions-qt-creator@cristianadam.eu>
... by some out-of-band mechanism, keeping previous functionality.
The driving reason is to get the same fromMap signature as in the
AspectContainer base, however, the whole mechanism seems rather useless:
There are only a few places that actually ever could report errors,
in those places "moving on" looks ok, too, and these few places
are not conceptually different than a lot others.
Removal of this (new) mechanism is left for later.
Change-Id: Ibe7a0456bc3b84a84e942231f14446e783372d76
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
To match better with the rest, especially the base AspectContainer.
Change-Id: Ide0966cab3219800aa6b7b9e6012731a3fbe7a6f
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
Not quite harmless, as it was destroying persistent data written by
the normal one.
Amends b9cf021b8e.
Change-Id: Ie2775881535c4d19174177f443a4606acc1576de
Reviewed-by: Eike Ziller <eike.ziller@qt.io>
Generally, aspects should use fromMap/toMap, but some older pages
use {from,to}Settings with always the same ICore::settings().
To make that less intrusive on the user code side, make that globally
implicit.
Task-number: QTCREATORBUG-29167
Change-Id: I29a8e23a31eb8061bb143e93931a07c6741eb7f9
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
`BoolAspect::checkableDecider` was used only in CMakeProjectManager for
askBeforePresetsReload and askBeforeReConfigureInitialParams.
Now the checkbox with "Ask before presets reload" would also match with
the display of the message box.
Change-Id: I45fc7a977dbeb13df051375bd3dac36e7be7bdc4
Reviewed-by: hjk <hjk@qt.io>
This avoids some repetition and could be a step towards
having type storage in (or rather accessible from) the base,
so we can have aspects for more complex data (treemodels...)
that are not easily converted to QVariant.
Change-Id: I9797b3d5646195705212db1830d2b415291ac651
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
This makes sure that one unique argument "-DANDROID_PLATFORM:STRING=..."
is passed to CMake. An entry in a kit's CMake configuration has the
precedence over the ANDROID_PLATFORM defined by Qt's build system.
Fixes: QTCREATORBUG-29112
Change-Id: I1c4d80be9ddfc7a00a5a7eff476497f08d87b741
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Reviewed-by: <github-actions-qt-creator@cristianadam.eu>
This hunk was added as part of 01b0d4f8f5 in 2020.
Change-Id: I3b520005dc2462397ddfb1a2f6671603131ddced
Reviewed-by: Cristian Adam <cristian.adam@qt.io>
Remove function overloads, thes are hard to read, to use and to extend.
I'd even argue this should be a plain default ctor and a few setters +
exec(), pretty much like Process::start() nowadays.
Move "decider" magic into a structure that can be filled ad-hoc outside
checkablemessagebox.cpp paving the ground for:
...removing aspect dependency from CheckableMessageBox, Instead, add a
convenience function to BoolAspect. Arguably, the latter is not needed
and could be done on the user side.
Use pointers instead of mutable references for in-out parameter.
Makes the "specialness" visible on the user side.
Pass ICore::settings() centrally as done elsewhere to reduce line noise
on the user side.
Change-Id: Ibb366353d1ea35401723fd05ce05672617a0a8fd
Reviewed-by: Marcus Tillmanns <marcus.tillmanns@qt.io>
Makes sure that the CMake parameters columns are properly stretched when
Qt Creator window is being resized or maximized / restored from
maximized state.
Fixes: QTCREATORBUG-27257
Change-Id: Ifb4d439fb758dcc5b2593be917ba35e9c79f2840
Reviewed-by: Alessandro Portale <alessandro.portale@qt.io>
QTC_KIT_DEFAULT_CONFIG_HASH was introduced in order to match the
previous CMake presets kits.
This would hash the Qt and the C/C++ compiler paths.
Unfortunately on Linux CMake would pick "cc" and "c++" by default,
whilst Qt Creator uses "gcc" and "g++". The compilers would match from
the Qt Creator point of view, but the hash values would be different.
On mac there is a similar issue with compiler launchers.
This patchset fixes this by removing the hash and fixes also the issue
of allowing broken CMake presets kits (without any compilers), which
were the reason for introducing the hash key in the first place.
Fixes: QTCREATORBUG-29075
Change-Id: Id8dc081febba7bc7023c98055afc6ebfcdc55542
Reviewed-by: <github-actions-qt-creator@cristianadam.eu>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: hjk <hjk@qt.io>
LayoutBuilder is meant to be an implementation detail nowadays.
Change-Id: I777ab934d3d405873e819eeddd27428d8c652f9a
Reviewed-by: Alessandro Portale <alessandro.portale@qt.io>
