Speed ups the typical use cases that can profit from this:
* Change a header file and switch then to source file
* Open documents one after the other (Follow Symbol)
* Change documents visible in splits (e.g. by refactoring action)
* Restore a session with multiple splits
Fixes the test ClangIpcServer.GetCodeCompletionDependingOnArgumets.
Change-Id: Ia575bd59780df14146dfc091a4d48794e4a0543d
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
...in preparation for concurrent processing of documents.
Parsing and reparsing is handled by TranslationUnit. Since we will do
this in a different thread, extract the core logic into the new class
TranslationUnitUpdater, so that we can prepare the necessary data for
the run and then later incorporate the results of the parse/reparse.
Change-Id: Ic9d936d193ee6795a755f0cfc38c0b2a7bd402cc
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
We want to share more functionality of the IPC mechanism and for what we
need more interface classes. But we use this names already for the
ClangCodeModel implementation. So we rename the them to ClangCodeModel*.
Change-Id: Ie320e0d3b993586a9bcc6a5aa0d32427af41202e
Reviewed-by: Tim Jenssen <tim.jenssen@theqtcompany.com>
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@qt.io>
There are some compiler problems and before I change the code I want to
investigate them deeper.
Change-Id: I497aa594fe0ea67be5927cc75bf6d64158952663
Reviewed-by: Orgad Shaneh <orgads@gmail.com>
The position of the dot character was determined on an outdated
translation unit.
We queried the translation unit for the source location of the dot
character, but apparently clang_codeCompleteAt() does not update the
source locations for the translation unit. And we do not want to reparse
since this is expensive. Thus, determine the byte position manually by
scanning over the document until the right line/column is reached.
Change-Id: I25e256bb81a83bb71c7e46a0fb3e927bf4031b16
Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>
Reviewed-by: Eike Ziller <eike.ziller@qt.io>
When using stdlib (default on Linux) and compiling with "-std=c++14", errors
like the following occurr. Avoid them by not requiring c++14.
$ clang --version
Ubuntu clang version 3.6.0-2ubuntu1~trusty1 (tags/RELEASE_360/final) (based on LLVM 3.6.0)
Target: x86_64-pc-linux-gnu
Thread model: posix
$ cat clangtest.cpp
#include <random>
$ clang -std=c++11 -fsyntax-only clangtest.cpp # Fine with c++11/stdlibc++
$ clang -std=c++14 -fsyntax-only clangtest.cpp # Ops with c++14/stdlibc++
In file included from clangtest.cpp:1:
In file included from /usr/bin/../lib/gcc/x86_64-linux-gnu/4.8/../../../../include/c++/4.8/algorithm:62:
In file included from /usr/bin/../lib/gcc/x86_64-linux-gnu/4.8/../../../../include/c++/4.8/bits/stl_algo.h:65:
In file included from /usr/bin/../lib/gcc/x86_64-linux-gnu/4.8/../../../../include/c++/4.8/random:39:
/usr/bin/../lib/gcc/x86_64-linux-gnu/4.8/../../../../include/c++/4.8/cstdio:120:11: error: no member named 'gets' in the global namespace
using ::gets;
~~^
1 error generated.
$ clang -std=c++14 -stdlib=libc++ -fsyntax-only clangtest.cpp # Fine with c++14/libc++
See also https://llvm.org/bugs/show_bug.cgi?id=18402.
Change-Id: I2ab3ea837b80196b51d8816ee8aa26a4a531e939
Reviewed-by: Christian Stenger <christian.stenger@theqtcompany.com>
Sometimes you need a very small array which has a size and max size.
QVarLengthArray is simular but has some size and run time overhead and it
has no max size. It will instead malloc. So this array is for very small
collections under 256 values which never allocate and have only a size
overhead of a byte.
Change-Id: Ia392c750d566c4accc6077c3dc4d9d4ae501e599
Reviewed-by: Eike Ziller <eike.ziller@theqtcompany.com>
Perf is not producing reliable stack traces without.
Change-Id: I5f594d45b06a66d91b0c687937dd948d6c724c62
Reviewed-by: Tim Jenssen <tim.jenssen@theqtcompany.com>
SmallString is a very simple utf8 string class. It's purpose is performance.
It uses a short string opimization which provides 31 bytes of heap free
memory to save a 30 bytes long string plus null terminator. If the string
gets larger heap is allocated. The grow strategy is 1.5 to improve reuse
of allocated memory.
It uses optionally constexpr to provide string literals.
Change-Id: I7757fb51abfeca200d074cbfce2f1d99edc0ecb0
Reviewed-by: Tobias Hunger <tobias.hunger@theqtcompany.com>
We used to call QTextCharFormat::setToolTip from the ExtraSelection to
install the diagnostic tooltip. Since this allows to set only text
tooltips and we would like to introduce a custom tooltip widget for
diagnostics, make use of CppHoverHandler, which is more flexible.
Change-Id: Ia1b2c3c50810596ce4a3a025002e6e4efd8789db
Reviewed-by: Alessandro Portale <alessandro.portale@theqtcompany.com>
This simplifies UnsavedFiles and makes TemporaryModifiedUnsavedFiles
useless.
Change-Id: I1896f971215ed22ce7aa7bf21b16381862b7469d
Reviewed-by: Marco Bubke <marco.bubke@theqtcompany.com>
1 struct Foo { int member; };
2 void f(Foo *foo)
3 {
4 foo.<REQUEST COMPLETION> // correct '.' to '->' and provide results
5 }
The preferred approach would be to check if "foo" in line 4 is of
pointer type, but there is no suitable cursor (only CompoundStmt) at
that position since the code is usually not yet parsed and thus invalid.
Thus, just run the completion as is. If there are not any results for a
dot completion, re-run the completion with "." exchanged by "->". This
approach is inherently slower than the preferred approach implemented in
the built-in code model.
The following rare cases are not handled:
1) Requesting completion after white space:
Works: foo.<COMPLETE HERE>
Fails: foo. <COMPLETE HERE>
2) Opening a file and requesting completion (ctrl+space) without prior
editing. No editing before triggering completion means that no
unsaved file is generated on the backend side, which is a
requirement for the correction.
Task-number: QTCREATORBUG-11581
Change-Id: I6bc8e8594778774ab342755fdb01a8a3e5c52ba0
Reviewed-by: Marco Bubke <marco.bubke@theqtcompany.com>
"Editor Updates" are not to the point since what we generate is for the
document.
Change-Id: I94ceeca5c85c7cf01cc659ca55320d07010c8617
Reviewed-by: Marco Bubke <marco.bubke@theqtcompany.com>
We send first the current editor, next the visible editors and the end
everything else.
Change-Id: I4a7b5924ffe563d6a74251739ddedcd005ce046c
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@theqtcompany.com>
Prepare the move of the semantic highlighting to the clang back end. We
have it under tests too so it should be quite easy to make changes or
corrections.
Change-Id: I5706a8a06fde5a9ba2eba3a8ba62782102ac0bd3
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@theqtcompany.com>
They are invokable by the usual means (Alt+Enter, editor's context menu
> Refactor) plus by the context menu of the editor's left margin for the
related line.
The fixit text comes directly from libclang and is thus not translated.
We modify the text slighty by stripping the diagnostic category prefix
("note:", "error:", ...) and capitalizing the first letter.
A follow-up change should properly indicate available refactorings with
a refactoring icon in the editor's left margin.
Task-number: QTCREATORBUG-14868
Change-Id: I86157c9f824d2a9dedf19087476d02ad1e6cc854
Reviewed-by: Marco Bubke <marco.bubke@theqtcompany.com>
...by introducing ClangDiagnosticManager and the helper class
ClangDiagnosticFilter.
ClangDiagnosticManager will get more state in a follow-up change.
Change-Id: Id2c312bc897ea41ed67292b56b24dcfb7975ff4a
Reviewed-by: Marco Bubke <marco.bubke@theqtcompany.com>
Reparsing a document is expensive so we should avoid it by all means. In
this patch we prevent that the same document is send again. It isn't send
too in advance of a code completion if there was no changes before the
the completion position.
Change-Id: I0bb786ba1d4e7ce08611a518cb32f8cf8f4d0037
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@theqtcompany.com>
Diagnostics are now moved to the clang backend process. Fixits are
supported too.
Change-Id: I20faacf466bbf78dec479220c3d7b336a47bc453
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@theqtcompany.com>
