Values "GDB", "LLDB", "CDB", the "Engine" prefix is a bit too
technical.
Change-Id: I81d97a435c796125015e8b7425e832420f338b9a
Reviewed-by: Eike Ziller <eike.ziller@theqtcompany.com>
If we fail to parse a declaration, we rewind, eat the token and look for the
next token that might be a good candidate for a declaration start (e.g. an
identifier). This becomes cpu and memory expensive with super long and invalid
expressions like
typedef b::m::if_< b::m::bool_<
(sizeof(fun((Dummy *) 0, (ThisT *) 0, (b::m::int_<70> *) 0)) ==
sizeof(defined_)) >, b::m::if_< b::m::bool_<
(sizeof(fun((Dummy *) 0, (ThisT *) 0, (b::m::int_<71> *) 0)) ==
sizeof(defined_)) >, b::m::if_< b::m::bool_<
(sizeof(fun((Dummy *) 0, (ThisT *) 0, (b::m::int_<72> *) 0)) ==
sizeof(defined_)) >, b::m::if_< b::m::bool_<
(sizeof(fun((Dummy *) 0, (ThisT *) 0, (b::m::int_<73> *) 0)) ==
sizeof(defined_)) >, b::m::if_< b::m::bool_<
(sizeof(fun((Dummy *) 0, (ThisT *) 0, (b::m::int_<74> *) 0)) ==
sizeof(defined_)) >, b::m::if_< b::m::bool_<
(sizeof(fun((Dummy *) 0, (ThisT *) 0, (b::m::int_<75> *) 0)) ==
sizeof(defined_)) >, b::m::if_< b::m::bool_<
// ...some more crazy lines like this
Therefore, stop trying after two failures by looking for the next semicolon or
closing curly brace.
Task-number: QTCREATORBUG-12890
Change-Id: I6637daeb840dd549d669080775228fa91fc932eb
Reviewed-by: Erik Verbruggen <erik.verbruggen@theqtcompany.com>
This works around missing support for standalone installations
of signed packages in androiddeployqt. The tool will use its
arguments to decide what the package to install is named, so
the arguments need to match the ones used to build the package.
In particular, we need to tell it that we are creating a signed
package.
This is intended as a temporary solution. The tool should be fixed
to support this use case properly, but the changes needed there
were deemed too risky at this point in the release cycle, so we
apply the work-around for now.
Task-number: QTCREATORBUG-13431
Change-Id: I03163582979715e7daa389c49bf10a88b5879db7
Reviewed-by: Daniel Teske <daniel.teske@theqtcompany.com>
Reviewed-by: BogDan Vatra <bogdan@kde.org>
We cannot just set it unconditionally on setRecording() because the
client recording state is supposed to reflect the user's intention.
Instead, if the recording button's state differs from the intention
and then spontaneously changes to match it (through a change in the
application), we toggle the client recording state twice to make sure
everything is properly synchronized.
Change-Id: I070d38364e63ab9cf0597fabed3d326ac3685589
Task-number: QTCREATORBUG-13322, QTCREATORBUG-13430
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
The --release argument for androiddeployqt is really not very
useful, since it will create an unsigned package which cannot
be installed on a device or published in any way. This is a very
special use case, which I don't think we need to support in Creator,
and it should at least not be connected to the build flags for the
C++ code, since it's quite possible that you want to build your
application code in release, but still sign the apk with a debug key
to test it.
Task-number: QTCREATORBUG-13431
Change-Id: I752bfdf396eab596637ab111c7844128962db1fd
Reviewed-by: Daniel Teske <daniel.teske@theqtcompany.com>
Reviewed-by: BogDan Vatra <bogdan@kde.org>
Nested event loops rearing its ugly head. Avoid all the trouble by
opening the message box on the next iteration of the event loop.
Task-number: QTCREATORBUG-13428
Change-Id: I3c06d649464c34875da9601fb2e13227a0397836
Reviewed-by: Robert Loehning <robert.loehning@theqtcompany.com>
Reviewed-by: Tobias Hunger <tobias.hunger@theqtcompany.com>
Debug Deployment broke with Android 5. Also show Android 5 devices
as incompatible in the device dialog if we are using debug deployment.
Task-number: QTCREATORBUG-13419
Change-Id: Ic321cfa46eb724f87f338af9c4b50face06c7c06
Reviewed-by: Eike Ziller <eike.ziller@theqtcompany.com>
Reviewed-by: BogDan Vatra <bogdan@kde.org>
Reviewed-by: Daniel Teske <daniel.teske@theqtcompany.com>
We want to only navigate the text editor when browsing through profile
data.
Change-Id: I2156b9853a6a1547a52c491ea021a68ca4f6dde9
Task-number: QTCREATORBUG-13433
Reviewed-by: Eike Ziller <eike.ziller@theqtcompany.com>
If the qmljsmodel has no background task, then isIdle returns true
Change-Id: I31420343c560c5ed118f371f7be347eadb6622f4
Reviewed-by: Thomas Hartmann <Thomas.Hartmann@digia.com>
In situation:
void f() {
// Cursor is HERE.
}
the code after ">i{" command is indented as (with shiftwidth=4):
void f() {
// Cursor is HERE.
}
Change-Id: I48283c91c32fc407bbdb24349f2491461e401ee9
Reviewed-by: hjk <hjk121@nokiamail.com>
Remove the "Paste from clipboard" action and move the functionality
into "Paste Snippet". Introduce a new overload for post() taking
a flag mask specifying the sources and use this for the
CodePasterService class (used by the diff editor).
Task-number: QTCREATORBUG-13401
Change-Id: Iadc3560a8a3bdaa817bc4a86007c2682feeb4b77
Reviewed-by: Eike Ziller <eike.ziller@theqtcompany.com>
Only show the "Do you want to recompile now?" modal dialog if the
"Enable QML debugging" or "Enable Qt Quick Compiler" checkboxes have
been changed directly by the user.
The previous change let the QML debugging checkbox update also when the
.pro file has changed. Showing a modal dialog in such cases interrupts
the user's flow, and is unexpected.
Change-Id: I9304cbecee7e201694ff72ac8c0f38cedf5fb416
Reviewed-by: Eike Ziller <eike.ziller@theqtcompany.com>
A change of the project languages (for example when the qml code
model is updated, which happens quite late) can add or remove
the QMLJS language, and thus change the value of
linkQmlDebuggingLibrary. This should be shown in the UI.
Change-Id: I55b6a866be6d0e356ed516568124bf3e80045078
Reviewed-by: Eike Ziller <eike.ziller@theqtcompany.com>
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
Pending breakpoints have zero resolved locations.
Change-Id: I18cecfabe8dcfebb7733030a404f75603464d157
Reviewed-by: Christian Stenger <christian.stenger@theqtcompany.com>
Return when the check fails, so there won't be a nullptr deref 2 lines
down.
Change-Id: I4dd1ba13798a57700c486e4b1f24b45afb2837ad
Reviewed-by: Lorenz Haas <lykurg@gmail.com>
Reviewed-by: Erik Verbruggen <erik.verbruggen@theqtcompany.com>
The usual case is probably to want normal editors stay open,
but memory/diassembler removed, but all other combinations
may be useful at times. So use two options instead of one.
Change-Id: I58c3951cc507c7e66a8d7a65b459a036e6a5d801
Reviewed-by: Christian Stenger <christian.stenger@theqtcompany.com>
Reviewed-by: Eike Ziller <eike.ziller@theqtcompany.com>
Exporting DO_NOT_BUILD_QMLDESIGNER disables the build of
the QmlDesigner plugin.
Change-Id: Ia3809268816cd84ef23b872f8df4709e381b3618
Reviewed-by: Eike Ziller <eike.ziller@theqtcompany.com>
Unneeded effort and also re-sets the marker to the top frame
somewhat spoiling part of the result.
Change-Id: Iebaa52dc8c0703f3245cac2ca15b942e3006993a
Reviewed-by: Christian Stenger <christian.stenger@theqtcompany.com>
Provide Project:Name there by default, and make it accessible
from the target.
Change-Id: I2faa52dc8c0703f3245cac2ca15b942e3006993a
Reviewed-by: Tobias Hunger <tobias.hunger@theqtcompany.com>
Make sure to refresh the runMode if it is changed by Qbs. Make any
user change stick though.
Task-number: QTCREATORBUG-12947
Change-Id: Ie03fad764f1a4d927213ecbb226c0aecfcf791a5
Reviewed-by: Tobias Hunger <tobias.hunger@theqtcompany.com>
There's some bad interaction between the delayed hide,
the deleteLater() from CloseOnDelete, and the async
evaluation. Work around by start from scratch for each
tooltip. On the downside we have more flicker.
Change-Id: I4801368c07895b652e10fe1d173b21cef7869ed0
Reviewed-by: hjk <hjk121@nokiamail.com>
This avoids the "StayOnTop" situation described in QTCREATORBUG-13397
for the case of pinned tooltips.
Task-number: QTCREATORBUG-13397
Change-Id: I3801368c07895b652e10fe1d173b21cef7869ed0
Reviewed-by: hjk <hjk121@nokiamail.com>
True expansion was setting another WidgetContent to
an exististing Tooltip. Arguably wrong in the base
Tooltip handling, but avoidable on the debugger side,
too. This was visible only on Mac.
Change-Id: I2e5a3273b571658b4dd4200c9b3a0e9542a16015
Reviewed-by: hjk <hjk121@nokiamail.com>
... if source file does not exist. This is the same
logic as for GDB now.
Change-Id: I3e5a3273b571658b4dd4200c9b3a0e9542a16015
Reviewed-by: hjk <hjk121@nokiamail.com>
More robust to variations in tool chains than function names.
(e.g. GDB reports 'foo' and LLDB 'foo()')
Change-Id: I1e5a3273b571658b4dd4200c9b3a0e9542a16015
Reviewed-by: hjk <hjk121@nokiamail.com>
Previously it "released" on step as the scope paths from the code
model and the debugger report "differed" ("a/b/../c" vs "a/c")
Change-Id: Ida094581cfaa53b00f9186495870854ca431f041
Reviewed-by: Christian Stenger <christian.stenger@theqtcompany.com>
