vite issueshttps://gitlab.inria.fr/solverstack/vite/-/issues2024-02-28T11:46:52+01:00https://gitlab.inria.fr/solverstack/vite/-/issues/27GUI is broken on Wayland2024-02-28T11:46:52+01:00Philippe SWARTVAGHERGUI is broken on WaylandGUI is broken on Wayland systems. Tested with a fresh install of Debian 12 Gnome in a virtual machine:
![VirtualBox_Debian_Gnome_28_02_2024_11_43_12](/uploads/b71a12cffa8c1bfccf867744a97498d5/VirtualBox_Debian_Gnome_28_02_2024_11_43_12.p...GUI is broken on Wayland systems. Tested with a fresh install of Debian 12 Gnome in a virtual machine:
![VirtualBox_Debian_Gnome_28_02_2024_11_43_12](/uploads/b71a12cffa8c1bfccf867744a97498d5/VirtualBox_Debian_Gnome_28_02_2024_11_43_12.png)
It works fine if I launch Gnome on X11.
Tested with Vite installed from Debian repositories, Vite built from source (commit 9dcc61ffe333510fd98c1295104bb108cf13c7d3), with OpenGL and VBO, with OpenGL without VBO. Not tested with Vulkan.
Maybe it is worth testing !45 before trying to understand this bug.
Reported by @jcollin.https://gitlab.inria.fr/solverstack/vite/-/issues/26files without copyright headers2023-06-29T19:52:32+02:00Johnny Jazeixfiles without copyright headersMaybe some false positives, I did: `grep -rL "CeCILL" src/ plugins/`:
```
src/interface/viteqtreewidget.cpp
src/interface/viteqtreewidget.hpp
src/interface/qxtspanslider.cpp
src/render/vbo/Shader.hpp
src/render/vbo/Shader.cpp
src/render/...Maybe some false positives, I did: `grep -rL "CeCILL" src/ plugins/`:
```
src/interface/viteqtreewidget.cpp
src/interface/viteqtreewidget.hpp
src/interface/qxtspanslider.cpp
src/render/vbo/Shader.hpp
src/render/vbo/Shader.cpp
src/render/texture.cpp
src/render/text2D.hpp
src/render/texture.hpp
src/render/text2D.cpp
src/statistics/Statistics_window.cpp
src/statistics/Diagram.hpp
src/statistics/Diagram.cpp
src/statistics/ChartView.cpp
src/statistics/ChartView.hpp
src/trace/portable_archive_exception.hpp
src/trace/Serializer_container.hpp
src/trace/Serializer_values.hpp
src/trace/Serializer_structs.hpp
src/trace/SerializerDispatcher.cpp
src/trace/Serializer_types.hpp
src/trace/TraceBuilderThread.hpp
src/trace/Serializer.hpp
src/trace/TraceBuilderThread.cpp
src/trace/Serializer.cpp
src/trace/SerializerWriter.cpp
src/common/Palette.hpp
src/common/Palette.cpp
src/common/TraceMemory.hpp
src/parser/ParserSplitted.cpp
src/parser/ParserFactory.hpp
plugins/TraceInfos/TraceInfos.hpp
plugins/TraceInfos/TraceInfos.cpp
plugins/MatrixVisualizer/Helper.cpp
plugins/MatrixVisualizer/Windows/MatrixGLWidget.hpp
plugins/MatrixVisualizer/Windows/MatrixWindow.cpp
plugins/MatrixVisualizer/Windows/MatrixGLWidget.cpp
plugins/MatrixVisualizer/Windows/MatrixWindow.hpp
plugins/MatrixVisualizer/Common/Zoom.cpp
plugins/MatrixVisualizer/Common/Quadtree.cpp
plugins/MatrixVisualizer/Common/Zooming.hpp
plugins/MatrixVisualizer/Common/Quadtree.hpp
plugins/MatrixVisualizer/Common/Zoom.hpp
plugins/MatrixVisualizer/Common/Zooming.cpp
plugins/MatrixVisualizer/Helper.hpp
plugins/MatrixVisualizer/Formats/SymbolMatrix.cpp
plugins/MatrixVisualizer/Formats/SymbolMatrix.hpp
plugins/MatrixVisualizer/MatrixVisualizer.cpp
plugins/MatrixVisualizer/Parsers/SymbolParser.cpp
plugins/MatrixVisualizer/Parsers/SymbolParser.hpp
plugins/MatrixVisualizer/Parsers/ValuesParser.hpp
plugins/MatrixVisualizer/Parsers/ValuesParser.cpp
plugins/MatrixVisualizer/Parsers/OrderParser.cpp
plugins/MatrixVisualizer/Parsers/OrderParser.hpp
plugins/MatrixVisualizer/Parsers/Readers/Pastix.hpp
plugins/MatrixVisualizer/Parsers/Readers/Pastix.cpp
plugins/MatrixVisualizer/Parsers/Parser.hpp
plugins/MatrixVisualizer/MatrixVisualizer.hpp
plugins/Distribution/Distribution.cpp
plugins/CriticalPath/ParserStats.hpp
plugins/CriticalPath/CriticalPath.hpp
plugins/CriticalPath/ParserStats.cpp
plugins/CriticalPath/DrawStats.cpp
plugins/CriticalPath/ParseTasks.hpp
plugins/CriticalPath/ParseDAG.cpp
plugins/CriticalPath/DrawStats.hpp
plugins/CriticalPath/ParseDAG.hpp
plugins/CriticalPath/ParseTasks.cpp
```https://gitlab.inria.fr/solverstack/vite/-/issues/25clang-format2023-07-09T00:21:35+02:00Johnny Jazeixclang-formathttps://gitlab.inria.fr/solverstack/vite/-/issues/24OpenGL migration2023-07-08T19:43:19+02:00Johnny JazeixOpenGL migrationJust to track the migration. QGLWidget is deprecated and removed in Qt6 so it needs to be updatedJust to track the migration. QGLWidget is deprecated and removed in Qt6 so it needs to be updatedhttps://gitlab.inria.fr/solverstack/vite/-/issues/23Qt6 migration2023-07-08T19:43:19+02:00Johnny JazeixQt6 migrationJust to track the migrationJust to track the migrationhttps://gitlab.inria.fr/solverstack/vite/-/issues/22Installing with spack and OTF2 support: issue with default OTF2 version (3.0)2023-06-28T16:26:38+02:00Matthieu KuhnInstalling with spack and OTF2 support: issue with default OTF2 version (3.0)Hi,
It seems a more or less recent move of otf2 from version 2.3 to 3.0 broke the compilation of vite (branch master).
The interface has certainly changed, involving small updates into the OTF2 parser of vite (into src/parser/OTF2Parser...Hi,
It seems a more or less recent move of otf2 from version 2.3 to 3.0 broke the compilation of vite (branch master).
The interface has certainly changed, involving small updates into the OTF2 parser of vite (into src/parser/OTF2Parser/ParserDefinitionOTF2.hpp and/or src/parser/OTF2Parser/ParserDefinitionOTF2.cpp).
Please find spack error messages when trying to install vite with the default version of OTF2 which is now otf2@3.0 below.
Forcing installation with otf version 2.3 allows to install vite with spack (`spack install vite+otf2 ^otf2@2.3`).https://gitlab.inria.fr/solverstack/vite/-/issues/21Preferences -> Settings takes a very long time to open2023-04-03T14:57:11+02:00THIBAULT Samuelsamuel.thibault@inria.frPreferences -> Settings takes a very long time to openHello,
It seems like the time that the Preferences->Settings menu takes to open depends on the size of the trace currently open. For instance, with this trace: [paje.trace](/uploads/b4b7ecd824f608c777c79c7d51e57410/paje.trace) it takes ...Hello,
It seems like the time that the Preferences->Settings menu takes to open depends on the size of the trace currently open. For instance, with this trace: [paje.trace](/uploads/b4b7ecd824f608c777c79c7d51e57410/paje.trace) it takes a few seconds, while for this trace: https://dept-info.labri.fr/~thibault/tmp/gpu.trace it takes minutes.
I have tried with git 38510db9a5a06a4fe7a154ac04fd9c738c5e8c2c (Mon Aug 8 10:11:39 2022)
Samuelhttps://gitlab.inria.fr/solverstack/vite/-/issues/19Wrong values in ruler when max displayed above 100002020-08-18T20:18:09+02:00Johnny JazeixWrong values in ruler when max displayed above 10000See attached files.
![export](/uploads/585072ffd7c7784a4dd183ae5e54915d/export.png)[per-color-state.trace](/uploads/84731c36bd4bad39234f6390e1717239/per-color-state.trace)
The ruler displays 4000 but the value is 5000.
If you zoom betwe...See attached files.
![export](/uploads/585072ffd7c7784a4dd183ae5e54915d/export.png)[per-color-state.trace](/uploads/84731c36bd4bad39234f6390e1717239/per-color-state.trace)
The ruler displays 4000 but the value is 5000.
If you zoom between 6000 and 10000+, all will be rounded to thousands (so we can have 6000, 6000, 7000, 7000 in the graduations instead of the good values).https://gitlab.inria.fr/solverstack/vite/-/issues/18svg export does not display states2020-10-19T21:32:17+02:00Johnny Jazeixsvg export does not display statesWhether it's command-line or via the menu, the svg export does not display states.Whether it's command-line or via the menu, the svg export does not display states.https://gitlab.inria.fr/solverstack/vite/-/issues/17command-line svg export segfaults2020-10-19T22:23:26+02:00Antoine Jegocommand-line svg export segfaultsHello!
I have pajé traces I want to export to svg.
When opening my trace (`vite paje.trace`) then `Ctrl+E > "export.svg" > Save`, the export is fine.
However when I try to use the command-line interface (`vite paje.trace -e export.svg`...Hello!
I have pajé traces I want to export to svg.
When opening my trace (`vite paje.trace`) then `Ctrl+E > "export.svg" > Save`, the export is fine.
However when I try to use the command-line interface (`vite paje.trace -e export.svg`) it crashes with the following message
> export of paje.trace to export.svg between the beginning of the trace to the end of the trace.
>
> Loading of the trace : 100% ~ Parsing... Remaining: 0 s
>
> double free or corruption (out)
>
> Aborted (core dumped)
My version of vite is the following
> the Visual Trace Explorer - version 1.2 - July 2010
(I got vite through `sudo apt-get install vite`)
Antoinehttps://gitlab.inria.fr/solverstack/vite/-/issues/16Vite displays type of link in the field Value2020-10-19T23:41:20+02:00Philippe SWARTVAGHERVite displays type of link in the field ValueHello,
If I display informations about a link representing a communcation, the field **Value** contains `MPI communication` (so the same as the field **Type**). According to the definition of the Paje file, the field **Value** should co...Hello,
If I display informations about a link representing a communcation, the field **Value** contains `MPI communication` (so the same as the field **Type**). According to the definition of the Paje file, the field **Value** should contains the size of transfered data.
The same occurs if the link represents a work stealing: StarPU puts 0 as value of the link, but Vite displays `Work stealing`.
I can try to fix it, but before, confirm it's a bug, please.https://gitlab.inria.fr/solverstack/vite/-/issues/15build failure with designated initializers2018-12-06T13:10:05+01:00Ghost Userbuild failure with designated initializersMerci pour la vitesse avec Qt4!
Mais avec gcc7:
```
[ 76%] Building CXX object src/CMakeFiles/vite.dir/parser/OTF2Parser/ParserDefinitionOTF2.cpp.o
cd /builddir/build/BUILD/vite-297d71d28a55198a67faec1701532bdd48fb5e43/src && /opt/rh/de...Merci pour la vitesse avec Qt4!
Mais avec gcc7:
```
[ 76%] Building CXX object src/CMakeFiles/vite.dir/parser/OTF2Parser/ParserDefinitionOTF2.cpp.o
cd /builddir/build/BUILD/vite-297d71d28a55198a67faec1701532bdd48fb5e43/src && /opt/rh/devtoolset-7/root/usr/bin/c++ -DQT_CORE_LIB -DQT_GUI_LIB -DQT_NO_DEBUG -DQT_OPENGL_LIB -DQT_UITOOLS_LIB -DQT_XML_LIB -DWITH_OTF2 -I/opt/rh/loveshack-profiling-backports/root/usr/include -I/builddir/build/BUILD/vite-297d71d28a55198a67faec1701532bdd48fb5e43/src/common -I/builddir/build/BUILD/vite-297d71d28a55198a67faec1701532bdd48fb5e43/src -I/builddir/build/BUILD/vite-297d71d28a55198a67faec1701532bdd48fb5e43/externals/qtcolorpicker/src -isystem /usr/include/QtOpenGL -isystem /usr/include/QtUiTools -isystem /usr/include/QtGui -isystem /usr/include/QtXml -isystem /usr/include/QtCore -std=gnu++14 -O2 -g -DNDEBUG -o CMakeFiles/vite.dir/parser/OTF2Parser/ParserDefinitionOTF2.cpp.o -c /builddir/build/BUILD/vite-297d71d28a55198a67faec1701532bdd48fb5e43/src/parser/OTF2Parser/ParserDefinitionOTF2.cpp
/builddir/build/BUILD/vite-297d71d28a55198a67faec1701532bdd48fb5e43/src/parser/OTF2Parser/ParserDefinitionOTF2.cpp: In static member function 'static OTF2_CallbackCode ParserDefinitionOTF2::handler_DefMetricClass(void*, OTF2_MetricRef, uint8_t, const OTF2_MetricMemberRef*, OTF2_MetricOccurrence, OTF2_RecorderKind)':
/builddir/build/BUILD/vite-297d71d28a55198a67faec1701532bdd48fb5e43/src/parser/OTF2Parser/ParserDefinitionOTF2.cpp:359:5: sorry, unimplemented: non-trivial designated initializers not supported
};
^
```
[J'ais seulement un petit peu de Français :-(]https://gitlab.inria.fr/solverstack/vite/-/issues/14Qt4 build failure2018-12-03T19:18:47+01:00Ghost UserQt4 build failureI'm trying to build the current development ViTE (f96665f) on centos7.
I can't use Qt5 as centos/fedora is missing QtUiTools, and the Qt4 build
fails with
```
/builddir/build/BUILD/vite-f96665f224ddb90657bc1f85da87e6ed14e2e3eb/src/com...I'm trying to build the current development ViTE (f96665f) on centos7.
I can't use Qt5 as centos/fedora is missing QtUiTools, and the Qt4 build
fails with
```
/builddir/build/BUILD/vite-f96665f224ddb90657bc1f85da87e6ed14e2e3eb/src/common/Session.cpp: In static member function 'static Palette* Session::get_palette(const string&, const string&)':
/builddir/build/BUILD/vite-f96665f224ddb90657bc1f85da87e6ed14e2e3eb/src/common/Session.cpp:310:63: error: 'class QMap<QString, QVariant>' has no member named 'cbegin'; did you mean 'begin'?
for(QMap<QString, QVariant>::const_iterator it = qmap.cbegin() ;
^~~~~~
/builddir/build/BUILD/vite-f96665f224ddb90657bc1f85da87e6ed14e2e3eb/src/common/Session.cpp:311:24: error: 'class QMap<QString, QVariant>' has no member named 'cend'; did you mean 'end'?
it != qmap.cend() ; ++ it) {
^~~~
make[2]: *** [src/CMakeFiles/vite.dir/common/Session.cpp.o] Error 1
```
I'm afraid I'm not sufficiently confident to try to fix it in a reasonable time, but it seems to have been broken by the OTF2 merge (which I want!).https://gitlab.inria.fr/solverstack/vite/-/issues/13clang, clazy, cppcheck application to use c++11 improvements2018-12-03T08:45:28+01:00Johnny Jazeixclang, clazy, cppcheck application to use c++11 improvementsHi,
as I can't do pull request, I open a bug to list a few patches I've done applying clang-tidy, clazy and cppcheck.
I didn't fix all the existing warnings/errors. I've separated in different patches to not have a too long diff.
[1-mod...Hi,
as I can't do pull request, I open a bug to list a few patches I've done applying clang-tidy, clazy and cppcheck.
I didn't fix all the existing warnings/errors. I've separated in different patches to not have a too long diff.
[1-modernize-nullptr.diff](/uploads/9505aaf698914dbc0648e71bc36136ed/1-modernize-nullptr.diff)
[2-modernize-use-equals-default.diff](/uploads/0972e72a67f27b5d61e2e067fb7013a8/2-modernize-use-equals-default.diff)
[3-modernize-use-default-member-init.diff](/uploads/a0840e418b90e0f7b25eca520d2aee67/3-modernize-use-default-member-init.diff)
[4-clazy-qvariant-template-instantiation.diff](/uploads/06467c1e4bf718e4a90398b3794463f5/4-clazy-qvariant-template-instantiation.diff)
[5-clazy-incorrect-emit.diff](/uploads/39ad084596ddae73bc3729bd9627dc35/5-clazy-incorrect-emit.diff)
[6-clang-performance.diff](/uploads/359d48b567ca716e2461529e2687c198/6-clang-performance.diff)
[7-gcc-partial-Wall.diff](/uploads/adc7d6caea449d936ff19bbdf1f03a38/7-gcc-partial-Wall.diff)
[8-modernize-use-override.diff](/uploads/1822ef8359e30d7295268c003d5c6907/8-modernize-use-override.diff)
[9-modernize-loop-convert.diff](/uploads/23e248354b1334df710c4c47bff99b9e/9-modernize-loop-convert.diff)
[99-fix-Distribution-plugin.diff](/uploads/219ff524f9a61bf90b3cb9d8b3f6510d/99-fix-Distribution-plugin.diff)
Johnnyhttps://gitlab.inria.fr/solverstack/vite/-/issues/12Compilation issue on Arch Linux2018-11-21T09:56:42+01:00François TrahayCompilation issue on Arch LinuxHi,
On Arch Linux, Vite fails to compile because some opengl features it use are now flags as "experimental":
> [ 81%] Building CXX object src/CMakeFiles/vite.dir/core/Core.cpp.o
> In file included from /var/lib/runner/build/vite-svn/s...Hi,
On Arch Linux, Vite fails to compile because some opengl features it use are now flags as "experimental":
> [ 81%] Building CXX object src/CMakeFiles/vite.dir/core/Core.cpp.o
> In file included from /var/lib/runner/build/vite-svn/src/vite-svn/src/render/Render_alternate.hpp:60,
> from /var/lib/runner/build/vite-svn/src/vite-svn/src/core/Core.cpp:121:
> /usr/include/glm/gtx/transform.hpp:23:3: error: #error "GLM: GLM_GTX_transform is an experimental extension and may change in the future. Use #define GLM_ENABLE_EXPERIMENTAL before including it, if you really want to use it."
> # error "GLM: GLM_GTX_transform is an experimental extension and may change in the future. Use #define GLM_ENABLE_EXPERIMENTAL before including it, if you really want to use it."
> ^~~~~
This can be fixed by adding a #define:
[glx_experimental.patch](/uploads/a005d931d6c395910a6970dd8af91e75/glx_experimental.patch)https://gitlab.inria.fr/solverstack/vite/-/issues/11Bad variable drawing2020-10-20T00:16:17+02:00THIBAULT Samuelsamuel.thibault@inria.frBad variable drawingHello,
Here is an example [paje.trace.zip](/uploads/a327b33662591889844d33d2ced68fb5/paje.trace.zip) of trace with bad-looking variable drawing.
If I really set to 0.0 the timing of all the first events until 0_w0's gf variable, I avoi...Hello,
Here is an example [paje.trace.zip](/uploads/a327b33662591889844d33d2ced68fb5/paje.trace.zip) of trace with bad-looking variable drawing.
If I really set to 0.0 the timing of all the first events until 0_w0's gf variable, I avoid the issue, but variables could show up on the fly during execution, so we would not have to do it, this drawing is really odd :)
Samuelhttps://gitlab.inria.fr/solverstack/vite/-/issues/10Please support per-state Color2020-10-20T00:14:55+02:00THIBAULT Samuelsamuel.thibault@inria.frPlease support per-state ColorHello,
In some cases, I'd like to be able to set different colors for different states, but without having to define a state type for each color (because there can e.g. be a *lot* of different colors). The attached [trace.zip](/uploads/...Hello,
In some cases, I'd like to be able to set different colors for different states, but without having to define a state type for each color (because there can e.g. be a *lot* of different colors). The attached [trace.zip](/uploads/d3b0f560d0e72c5ec0335c00c84c6c92/trace.zip) would be an example: PajeSetState 22 has an additional Color field, which is used by 22 at line 537 & 538 (time 1407.695097000), which should thus make the first task show up as almost black instead of green.. Note that just like in https://gitlab.inria.fr/solverstack/vite/issues/9 , vite currently crashes like every other opening of the file, so https://gitlab.inria.fr/solverstack/vite/issues/9 probably needs to be fixed first :)
Samuelhttps://gitlab.inria.fr/solverstack/vite/-/issues/9Crashes on extended Event type definition2018-12-03T19:14:17+01:00THIBAULT Samuelsamuel.thibault@inria.frCrashes on extended Event type definitionHello,
It is allowed to set different types of events with different sets of values, for instance in the attached [paje.trace.zip](/uploads/af668b6927e60ea42b8629a493c1cb12/paje.trace.zip) , there is both a PajeNewEvent 9 with the stand...Hello,
It is allowed to set different types of events with different sets of values, for instance in the attached [paje.trace.zip](/uploads/af668b6927e60ea42b8629a493c1cb12/paje.trace.zip) , there is both a PajeNewEvent 9 with the standard values, and a PajeNewEvent 20 with more values. Currently vite sometimes crashes on opening such file (I mean, if I keep opening the file, vite crashes like one time every two).
SamuelMathieu FavergeMathieu Favergehttps://gitlab.inria.fr/solverstack/vite/-/issues/8fails to build with recent version of glm2018-03-02T17:37:28+01:00THIBAULT Samuelsamuel.thibault@inria.frfails to build with recent version of glmAs reported on http://bugs.debian.org/888795 , the new version of glm requires to #define GLM_ENABLE_EXPERIMENTAL to be able to include glm/gtx/transform.hpp, because its interface may change.As reported on http://bugs.debian.org/888795 , the new version of glm requires to #define GLM_ENABLE_EXPERIMENTAL to be able to include glm/gtx/transform.hpp, because its interface may change.https://gitlab.inria.fr/solverstack/vite/-/issues/7Increase the windows size2018-01-11T16:32:45+01:00Mathieu FavergeIncrease the windows sizeWe should be able to define the matrix size to display the symbol matrix in order to increase/decrease its size.We should be able to define the matrix size to display the symbol matrix in order to increase/decrease its size.CHEVALIER ArthurCHEVALIER Arthur