FilePath::absoluteFilePath() would consider an empty path to be
relative, and resolved it wrt the current working directory. That is
unexpected in the sense that QFileInfo::absoluteFilePath behaves
differently, and lead to a crash down the line when e.g. a diff editor
was opened when the current working directory is '/' (the default on
macOS when started with "open" or from Finder, Dock or Spotlight).
Make FilePath::absoluteFilePath() return and empty path if the input was
empty.
Change-Id: Ie0d4da50afa24134bd56505b2f4abcf382eee982
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>
Unfortunately, QDir::cleanPath() only cleans according to the rules
of the host system, which can be wrong in remote setups.
As the implementation is not accessible/tweakable from the outside,
copy the relevant code and remove the platform #ifdef's.
Change-Id: Ife9a925412a12d3cef21ed3721a387c61c152ddf
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
It's too late, and apparently "close-without-open" is triggered:
QDEBUG: tst_fsengine::testRootPathContainsFakeDir() SOFT ASSERT: "m_tempStorage" in /data/dev/creator/src/libs/utils/fsengine/fsengine_impl.cpp:64
QDEBUG: tst_fsengine::testListDir() SOFT ASSERT: "m_tempStorage" in /data/dev/creator/src/libs/utils/fsengine/fsengine_impl.cpp:64
Change-Id: Ic7100d7abddfffca909b412ea7ea19e295c86729
Reviewed-by: Marcus Tillmanns <marcus.tillmanns@qt.io>
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
We apply not very often the parent to views. So it can lead to dangling
pointer if the parent is used by other objects and does not handle the
null pointer case. It can lead to double deletion if the parent is
deleted before the object when it is on the stack or handled by smart
pointer.
If you really want to use it there is still setParent.
Change-Id: I1fc6b145a50f037a0e9d415fb36e7970ea7296ed
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Intended as helpers for parentDir() to stop when cutting components.
Change-Id: I7615803743351a733d6f3cc9813e9de85973dc61
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
Reviewed-by: Marcus Tillmanns <marcus.tillmanns@qt.io>
"Fixes" one Q_SKIP and makes the implementation OS-agnostic.
FilePath::osType() actually depends on the host system for not fully
qualified file paths, and that doesn't seem to be the right thing to do
in all cases. The caller may actually have more context to decide
whether this is needed (e.g. in pathchooser.cpp in the context of
"return value from a GUI element, i.e. 'host' implied).
Change-Id: Iff80c17094d2deec79d34d0527223fbcf8294935
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
Changing the toString() to prepend system-specific prefixes
to cater for the remote filesystem model handling had unintended
sideeffects.
This here is the first step to split both paths.
Change-Id: I66cdd9a87802408e2ba95c2e29b9d7cce7fe2553
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
Don't hardcode a single type of OS (Windows/Non-Windows) via
QDir::rootPath(), actually check both, independent of the host OS.
Most of the new tests are actually failing, as the parsing code uses
the host OS. Mark those tests as "Expexted fail" for now.
Change-Id: Ic907582ad3959ef95b9618826d987d4eab463b49
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
Recent changes in FilePath made it incompatible with windows targets
(by design). Therefore we disable the tests here.
Change-Id: I2182c64f36268f5ab6282d149930653a79cc9aa1
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
DeviceShell does not support piping commands at the moment.
To continue to support ::bytesAvailable the parsing of the
output from "df" is moved to FileUtils::bytesAvailableFromDFOutput.
Change-Id: Ia229208748aa6c572b99899e6ae042bdd5654f4a
Reviewed-by: Eike Ziller <eike.ziller@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
Origanally my idea was that there can be no invalid model nodes created.
Because it was practically to return invalid nodes we added a default
constructor. So now we have invalid nodes because we explicitly created
them and invalid nodes because we destroyed them(except the root node
which has to be always valid). To remove all the isValid test two new
approches are introduced.
Instead of throwing an exception a reasonable return value is provided
or the function is simply doing nothing.
For example if you try to remove an non existing property nothing is
happen. If you try to get an non existing property you get an invalid
property. If you then try to get an value from that property you get an
null(false) value.
All object can now be tested for true or false. So you can write:
if (auto property = getnode().property("foo"))
// do something with if node and property are positive
Task-number: QDS-7454
Change-Id: I9ebb0bae45354cedc0b6cfe610d1a71435ee1f38
Reviewed-by: Aleksei German <aleksei.german@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
This saves about 2 seconds out of 5 seconds spent in
displayResult() for converting from QString into
FilePath in main thread (use case: gitgrep with
2 million hits).
Change-Id: I6334e8559986501c34c90a5c7c6b4acc9a6953e0
Reviewed-by: <github-actions-qt-creator@cristianadam.eu>
Reviewed-by: hjk <hjk@qt.io>
If no base64 is installed on the target, the shell script
cannot work. Previously this would lead to the shell functions
being unavailable.
This change adds a fallback path. In case no base64 is found,
the shell will create a process for each run request instead.
This is much slower than the shell script, but acceptable as
a fallback.
Change-Id: I70591d7e610c4e1c3c258a8e4bef354221d05cb9
Reviewed-by: hjk <hjk@qt.io>
Since we are now requiring macOS 10.14 we can remove our local
implementation of optional and use std::optional for macOS too.
Change-Id: I2bd018261b68da64f7f031a812045dd7784697e1
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: <github-actions-qt-creator@cristianadam.eu>
Reviewed-by: Marco Bubke <marco.bubke@qt.io>
FilePath::setFromString failed previously for URLs
without a path, e.g. "http://qt.io" as it would add
the scheme and host to its path.
Also added a test for various URLs
Change-Id: I6c22489c0af7154734390aa1e5e12693864210b1
Reviewed-by: hjk <hjk@qt.io>
IsSubClass is renamed to isBasedOn and takes now NodeMetaInfo as a
parameter. But for most cases there are is... functions as short cut.
The model is providing shortcut NodeMetaInfos too. This is done in the
sake of caching.
Task-number: QDS-7445
Change-Id: Iff2dea66e253b412105427134bd49cb16ed76193
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: <github-actions-qt-creator@cristianadam.eu>
Reviewed-by: Thomas Hartmann <thomas.hartmann@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>
It return true if it is the same or a derived type. You can use multiple
base types.
Task-number: QDS-7424
Change-Id: I965a7c5b87ef0eb0183be1c705ebc18dee1943c3
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
