GitLab

It turns out that one "Connection" per RunControl doesn't map
well to the uses we have. Instead, RunWorkers need to know
individually how to connect to the place where they can work,
but they are already specific enough to be able to use a
standard class (like QUrl) as their way to specify the needed
entry point.

In theory one could see a RunControl's connection as an
aggregation of its workers connection bits, but that does
not really seem to be needed in code.

As consequence, replace UrlConnection by a plain QUrl, and also
the HostName connection by a QUrl with hostName set.

Change-Id: I40c97e37779314ac0a77041e864a18eadb78f987
Reviewed-by: Vikas Pachdha <vikas.pachdha@qt.io>

Having the overall runworker setup closer to the general pattern
allows to re-use SimpleTargetRunner.

Change-Id: Iff151cbebaa6ae6615b933f4277b0581a43d7f7f
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

Orthogonal concepts, that only happen to coincide.

Also, make the server directly settable instead of relying
on the runControl's connection().

Change-Id: I2472acafcc50aede2cb6f99421901f0e67531b91
Reviewed-by: Christian Stenger <christian.stenger@qt.io>

... and QmlProfilerRunner::Configuaration by PE::UrlConnection,
and call it 'serverUrl' on the user side.

That's the only variant we ever had and avoids "translations"
between three structures that are essential the same.

Change-Id: I33386b2b8d2a7985ff934f6f8f840de0831bf9c1
Reviewed-by: Christian Stenger <christian.stenger@qt.io>

Also streamline code paths in the result.

Change-Id: Id7d96343a8f778ba8f415b1a850cc78576afa475
Reviewed-by: Christian Stenger <christian.stenger@qt.io>

This also re-enables the remote linux case and enables
the recording of a single run of events.

Change-Id: I9ea55017c8e5f2f17e6f32c5453df48093e41931
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
Reviewed-by: Christian Stenger <christian.stenger@qt.io>

This increases re-usability of activities like 'port gathering',
and makes their use less dependent on actual device implementations.

Change-Id: I017cb74874f2b38c487ba2d03906a675d5618647
Reviewed-by: Christian Stenger <christian.stenger@qt.io>

Apparently macOS will block the event loop until a real mouse click
happens if you open and close context menus in quick succession. When
it does that, the timer won't hit anymore and the mouse event to trigger
the next test step won't be generated.

Change-Id: Ie0c05d8a5a2020fe46381358133cd7cdbbf42299
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

The context menu event can be generated multiple times within one loop
of QTRY_VERIFY. The result is that multiple showFullRange() signals can
be generated before we check again for spy.count() == 1. Thus, the check
never succeeds and the number of signals keeps growing.

We connect the showFullRange() signal to the model manager in order to
get a more realistic test setup. This way the action that generates the
showFullRange() signal is disabled in any further context menus, just as
it is supposed to be. In addition we can now check for the manager to
actually show the full range.

Change-Id: I5e13c2666ce1a15c7a5fad57affd4274d9656656
Reviewed-by: Christian Stenger <christian.stenger@qt.io>

Q*Application classes have unusually many static functions. In many
cases in our code, these functions are unnecessarily called as instance
functions, using the qApp helper.

This patch replaces many occurencies of qApp with the according
Q*Application classname.

Change-Id: I6099a419fa7bf969891269c37ed7a9e817ef5124
Reviewed-by: hjk <hjk@qt.io>

The change is "conceptually wrong", the AnalyzerRunControl derived
classes' functionality should be provided by ToolRunners based classes
encapsulating/"being" the current Analyzer*Runner classes.

However, the AnalyzerRunControl is only three (empty even) virtual
functions, but a big obstacle in merging attempt due to a lot of
mechanical followup changes in downstream users.

The current construction mechanism of analyzer run controls is actually
two different mechanisms (locally direct RunControlFactories, and a
"generic" createAnalyzerRunControl wrapper for remote cases). The generic
createAnalyzerRunControl makes it difficult to migrated them one-by-one,
due to the various downstream users.

So instead of merging the per-analyzer two uses directly reduce
the "indirection" distance by removing the AnalyzerRunControl
intermediate layer. After that the createAnalyzerRunControl mechanism
can be dissolved by using normal RunControlFactories also for
the remote cases. After that, porting to ToolRunner, and combining
with ther local equivalent can be done one by one.

Change-Id: I0ddace33fcce210cf3a547ac5bb23b3d85013934
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Tobias Hunger <tobias.hunger@qt.io>

If the context menu disappears for whatever reason, just open a new one.
Making sure that the context menu stays around when we don't do anything
is not Qt Creator's job, so we don't have to test for it. Also, when
trying to remove the context menu, click a spot that's certainly inside
the screen.

Change-Id: If6e574d84bedf2ab552c8c29d756501e9d2282bd
Reviewed-by: Christian Stenger <christian.stenger@qt.io>

This introduces a mini-state-"machine" to handle RunControl states
Intialized->[Starting->Running->Stopping->Stopped->]*->Finished.

Needing time between trying to start and getting feedback is nowadays
the normal setup for all remote targets as well as for most local tools.
Making that the default for all runs simplifies the code and provides an
opportunity to (a) fix some currently wrong reports of "stopped
immediately" and (b) to remove target-specific (WinRT) or tool-specific
(Valgrind, GammaRay) state members doing essentially the same.

Change-Id: I7f52fee41144188ee8389e922fdc265f8c0a6459
Reviewed-by: David Schulz <david.schulz@qt.io>
Reviewed-by: Tobias Hunger <tobias.hunger@qt.io>

We keep it private to QmlProfilerModelManager and proxy the last few
methods that were directly called on the model. This enables us to
remove the QmlProfilerDataModel class by integrating what is left of it
into QmlProfilerModelManagerPrivate in a next step.

Change-Id: Ie9b4e03fb286e5a0040374d00b7b26f810426278
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

Format initializer lists code style like.

Change-Id: Ib82c235e4ba7dc75ee96a7abc0c47eff7b0a9013
Reviewed-by: hjk <hjk@qt.io>

This is the place where we need it most. In addition, this allows us to
use the DetailsRewriter's cache also on gotoSourceLocation.

Change-Id: I14e0f11ba7c8a2a3888b2e8439a375068b36e29a
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>

This is the logical place to do it. Adding the event first to the QML
model and then passing it back to the manager in order to have it
dispatched to the other models is somewhat backwards.

Change-Id: I64b1cb38f97331b62d83fa5ae49b9b2690810d40
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>

Since the cache size is not given for individual file events anymore,
the test should only check it for the events in row 1.

Change-Id: Idfc7ec90ab8da214f4eedce36dd3b84fb00ed553
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

... instead of guessing it. This is more robust.

Change-Id: I4855b78da23cb1caefa3a7a95bfa60c5733b32c5
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

This tests that the file name is correctly stored in the model and it
removes the detour via the model manager.

Change-Id: Icc592b6383edcb18bfe31a81eb7ae0736a4f5508
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

We should use it for all time printing instead of duplicating the code
everywhere.

Change-Id: I530baa31fd7044aefce6201fec0ab27c99a61a1d
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

The test should also run if there is no kit.

Change-Id: Ia2b5d34f2fc792235fb9da9c77f9235c186addde
Reviewed-by: Christian Stenger <christian.stenger@qt.io>

Change-Id: Ia3d1ab6be249fb7193ce084b7312e2b6192edad6
Reviewed-by: Orgad Shaneh <orgads@gmail.com>

The flame graph keeps separate "flames" for compile and other QML/JS
events now. The test has to be adapted to check for this.

Change-Id: Ifdbc3331240d30a3acbb9df62a6a44654466273e
Reviewed-by: Christian Stenger <christian.stenger@qt.io>

When given an empty string as executable and some non-empty arguments
QProcess might emit error() but not errorOccurred(). This was fixed by
2f0ffba638bdb868a810bc0349c240b6b6e172cd in qtbase, but we cannot rely
on that being present, yet. Thus, avoid passing empty strings as
executable.

Change-Id: Ia1bb5ebceb6edeca1e54489d12b838f3cea8851e
Reviewed-by: Orgad Shaneh <orgads@gmail.com>
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

Change-Id: I468ed3c7f75b79d16231b57246e494e29a89b920
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

The client manager should not be bothered with details of QML events,
but rather just connect the client, the model manager, and the state
manager.

Change-Id: Iec4499f8441a06d4ef5cbcf7bfe23da6f5e7f239
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
Reviewed-by: Alessandro Portale <alessandro.portale@qt.io>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>

Change-Id: I0d6deea53302022c18c4aacef395393d60308b84
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

If the application outputs the "QML debugging is disabled" line (and
Qt Creator can read it), the QmlProfilerRunControl will kill the
application. In the test this happens if Qt Creator was compiled
without QML debugging enabled.

Change-Id: I958a991d304ec9af8643e88dd9709f580a28753f
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

We can actually get several TCP connection attempts from the client
manager. That is the whole point of retrying. The QQmlDebugServer
explicitly allows sequential connections, so we should allow them, too.

Change-Id: Id38290dbfd86fadfb0c648189f89827330ea1174
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

On macOS funnny things happen if two popups are open at the same time.

Change-Id: Ic83d9e93184fd3f6a5f7e1a8d497da805eeeb6c6
Reviewed-by: Christian Stenger <christian.stenger@qt.io>

Depending on the build configuration of Qt and Qt Creator there can
be various kinds of messages and we cannot reliably determine their
number in advance.

Change-Id: I162cc02a72200026d3cbe7498fa705911f4cda39
Reviewed-by: Christian Stenger <christian.stenger@qt.io>

Instead of starting Qt Creator to get a debug server we can just
simulate the hello message on a locally created connection. That is
enough to make the client manager signal "connectionOpened()".

This way we don't get excessive timeouts on this test.

Change-Id: I89fa12b1bdda41a62dec47de042224f26fee79a9
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

We want to know why it's failing.

Change-Id: I794f183d8aae7f0f7c6444b432d9a2e80db16f22
Reviewed-by: Christian Stenger <christian.stenger@qt.io>

We want to use it for other profilers, too.

Change-Id: Ice4bd7fdfce6e0153d62a7c9a83dc7de6d5cba30
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

Change-Id: I9d758857dc65daa564ab6a0394250c02ea006832
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

Change-Id: I64fbc8fae00e859ce6433ff85ceb5706675b88cc
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

Change-Id: I040ef23f780e361037386a92444fcf45aded7a3f
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

Change-Id: I0227d9c97b2b61d68ae5ed9e3f2f7e3776075338
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>

Change-Id: I3735e70429e67dca9b8f02ef13174bef8530abf6
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>