Currently the ownership of the processor is unknown to the codeassist
when running an async operation. Move the ownership of the processor to
the assist and delete the processor after the proposal was completed.
Change-Id: I6a2e023c47cbc876669dba866bee12b481447cb7
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
In particular, clang with version number 10 is not really clang 10 on
macOS.
Change-Id: I8be489fa9cffd63c1ea8f13d181686ab1a575ec6
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
You can now change the preview size with
rootModelNode.setAuxiliaryData("previewSize@Internal", size);
If size is null it will use the bounding box size.
Change-Id: Icbe747ccc5a2e26e79783825d2ed94ecc640012b
Reviewed-by: Michael Winkelmann <michael.winkelmann@qt.io>
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
When running benchmarks display them on the visual results
pane as well.
So far, they had been ignored and where visible only in
the text output.
Omitting some additional information present inside the xml
output to be closer on the default text display.
Change-Id: I9dbcad917ab522c12aaaa81704add59efedf7905
Reviewed-by: David Schulz <david.schulz@qt.io>
Uses a timestamp file in the build directory to detect when the input
changes using DEPENDS from add_custom_command.
Total copied files size sum up to more than 260 MB. This was copied at every
incremental build, even if nothing changed.
This leads to faster build, especially when the copied files are not
cached in RAM or when not using a SSD.
Change-Id: I3a0b691b961e1d8591068bacf40ae2467c8637cf
Reviewed-by: Eike Ziller <eike.ziller@qt.io>
Reviewed-by: Cristian Adam <cristian.adam@qt.io>
When the completion request is triggered by an activation character,
set the triggerCharacter accordingly.
This can help the language server to propose more sensible items.
Detected while trying cmake language server. It was throwing an error
when triggerKind is TriggerCharacter and triggerCharacter is not present.
Change-Id: Ifde578e39f2eebfc5c6551f08985f4408f6c7933
Reviewed-by: David Schulz <david.schulz@qt.io>
When the language server is initialized, the Initialize request can contain
user provided data in initializationOptions field.
Allow the user to set data inside. This can be required to let the
language server have some context.
Change-Id: Ib057fdb940c21b3fd032853fb84253d42ad1e321
Reviewed-by: David Schulz <david.schulz@qt.io>
When the default assist provider is not DocumentContentCompletionProvider
because of some plugin that also provide an assist provider, this would
fail.
If the user has configured a language server, assume is prefer that and
always propose the server's assist providers when available.
This the case for CMake files but probably C/C++ files too if the CppEditor
plugin is configured to handle them.
Change-Id: I641d73466c8b520beb494dae8e6b274debd2881f
Reviewed-by: David Schulz <david.schulz@qt.io>
The proposal of SemanticHighlighting uses a versioned text document
identifier, but there is a suggestion to be more coherent with other
existing requests and use text document identifier instead (non versionned).
See here: https://github.com/microsoft/vscode-languageserver-node/pull/367/files#r225879268
The clangd language server implements the non versioned version.
Existing code was returning errors with clangd because
SemanticHighlightingParams::isValid would return false because of the
missing version field.
Change-Id: I1c36151342437adad2405118ca6b303db0936e41
Reviewed-by: David Schulz <david.schulz@qt.io>
They can now be created independently of any toolchains, and we expose
them in the build and run configurations, so that users can easily get
tasks for output that comes from custom tools or is otherwise specific
to the user's environment.
Fixes: QTCREATORBUG-23993
Change-Id: I405753b9b68508ffe5deb4fcac08d6b213c7554d
Reviewed-by: André Hartmann <aha_1980@gmx.de>
Reviewed-by: hjk <hjk@qt.io>
So far, the user had to press "Apply" to generate a kit for the
currently selected target. The kit got silently added, there was no
immediate UI feedback that something happened.
Also, there was no possibility to remove a specific Qt for MCUs kit once
it was added, becoause it is under the "Auto-dected" kit category.
This change modifies the UI interaction so, that there is a specific
"Create Kit" button, and in case that the kit already exists, there is a
"Remove Kit" button.
Task-number: QTCREATORBUG-23929
Change-Id: I334ed75f86905ac762e0293bd3aa70b3433acf28
Reviewed-by: hjk <hjk@qt.io>
The docs are still generated under doc/qtdesignstudio/doc/html.
Change-Id: I9916ccf8cdfa2541b632a680fb6b4f7103f6a1a5
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
This prints a QDoc warning when building with QDoc 5.15.
Change-Id: Id6f4646738f3e09956a243dc0da639f5567470f3
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Apparently, glibc uses an overly aggressive caching strategy, causing
Creator to accumulate an unreasonable amount of resident memory over
time.
We deal with this by manually forcing glibc to free unused memory once a
minute, provided the application is actually in use.
Fixes: QTCREATORBUG-22832
Change-Id: I300eafd8714eb7e49774129758998db32ca1a1de
Reviewed-by: hjk <hjk@qt.io>
The button has been removed.
Task-number: QDS-2056
Change-Id: I7495107e68920f02effef0618461f5ed6c88bfa8
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
...when resizing items
Fixes: QDS-2034
Change-Id: I4b4589a19c3c0ca96d6f8a80c7e4a3cc03731690
Reviewed-by: Aleksei German <aleksei.german@qt.io>
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
FTS5 is proving it. So for full text search we should use FTS5.
Change-Id: Iac5abdfa9bc6fd9f9c2515c1f15ef04cf445baf4
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
It is still only support references in columns but so far it is enough.
Change-Id: Iebb4866cf738d651270e54357b5e4a2837f05417
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Until, we have a mechanism to define a default/fallback in a read-
only setting, it is better to stick to old file names rather than to
create redundant entries.
Change-Id: Ifbc78160e9c57025c66a266efab2b43646ee477a
Reviewed-by: Brook Cronin <brook.cronin@qt.io>
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
As the addresses were only matched against upper case letters
all registers with lower case characters were not correctly parsed
and had zero values
Change-Id: Idc3da0a33b9f075d7a209cad74768e81cced2d67
Reviewed-by: André Hartmann <aha_1980@gmx.de>
Reviewed-by: hjk <hjk@qt.io>
This can appear as (non-existing) preferred frame index in backtraces
consisting only of frames without debug info. Fall back to top most
frame in that case.
Change-Id: Id7bb941d5fc3080aff3853cebb7b92afdd0ba4ea
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
Presumably, this has changed in newer llvm.
Fixes: QTCREATORBUG-23992
Change-Id: I82aaae893e3cf159be7c43fb4d7762875d2be8a8
Reviewed-by: André Hartmann <aha_1980@gmx.de>
No need to start the debugger event loop if the engine setup fails
Change-Id: I6bdeda0d2609c79aa5fafff7bc4b373a0bf14fbd
Reviewed-by: David Schulz <david.schulz@qt.io>
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
The original reason for this was gdb's inability to set certain
break point properties in one go (commands and so parts of the setup are async
anyway, and using only async paths made the overall setup simpler.
The need to tweak some properties asyncronously still remains, but
for the caller side it's better when at least the initial round is
synchronous.
Change-Id: Ibe542533a6026c7f0f4e49f4244b7e9d2aea7062
Reviewed-by: David Schulz <david.schulz@qt.io>
Breakpoints inserting is async, so they did end up when the process
was running already.
This adds an extra roundtrip. Better, but more intrusive solution
might be to set the initial breakpoints synchronously, but that
would touch all engines.
Change-Id: Ia728a9c5ae8f1d6d29d3cc02b9e2d04952091fe9
Reviewed-by: David Schulz <david.schulz@qt.io>
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
This avoids appearing as there is still something happening.
Change-Id: I34c34e054a5454d1030234f482630e3d3fdd5a72
Reviewed-by: David Schulz <david.schulz@qt.io>
