Qt process viewer and manager
Home Page: https://lxqt.github.io
License: GNU General Public License v2.0
Requirements:
Qps is licensed under the terms of the GPLv2 or any later version.
Official binary packages are provided by most major Linux and BSD distributions.
Just use your package manager to search for string qps.
Translations can be done in LXQt-Weblate
Since 1.10.19, qps's build requires lxqt-build-tools. CMakeLists.txt enforces a minimum version of 0.6.0. Tests show that 0.5.0 is sufficient for building.
Build successfully with lxqt-build-tools 0.5.0.
Refuses to build with lxqt-build-tools before 0.6.0.
In CMakeLists.txt, change the setting of LXQTBT_MINIMUM_VERSION from 0.6.0 to 0.5.0. (Sorry, to banal for a PR.)
I couldn't build on Fedora F29, which provides lxqt-build-tools 0.5.0. 0.6.0 was only released a few days ago, and many distributions will be behind.
(Fedora F28 still provides 0.3.2 for some reason. Spoiler: 0.3.2 isn't sufficient for building.)
I haven't actually tested the resulting binary yet, but I assume lxqt-build-tools only affects the build process.
Currently there's no desktop entry file.
Proposal regarding the file itself
[Desktop Entry]
Type=Application
Name=qps
GenericName=Qt process manager
Comment=Qt application to display and manage running processes
Icon=qps
Categories=System;
TryExec=qps
Exec=qps
Terminal=false
StartupNotify=false
Adding this file would obviously involve that an icon is added as well. In my opinion the most suitable choice would be to copy icon/icon.xpm to /usr/share/icons/hicolor/48x48/apps/qps.xpm.
XPM files aren't recommended any longer but still supported by the Specification. In the long rung substitution with a more suitable file type and / or even another icon as a whole is probably something worth considering.
Different story, of course.
The same texts are repeated in it, above the scroll view and inside it. The title (above the scroll view) can just be "Qps VERSION" and, maybe, its icon -- as if there's nothing more important to fix ;)
For example, no process monitor should terminate a process without asking the user, especially when that action has a shortcut.
I'll make a PR after #167
@agaida this is probably a comment for you
This project has two Russian translations on weblate: ru_RU and ru
ru is 100% complete and nice
ru_RU is almost nothing
However, qt loads ru_RU by default so the translation is broken. Could you please delete ru_RU and keep only ru?
At the time Qps was created, they might seem serious. I found the problem only with a special Kvantum theme -- Kvantum is good at revealing incorrect style codes.
Will correct it soon.
you ask for it, you get it :)
DEBUG: version = 31800 QDBusConnection: session D-Bus connection created before QCoreApplication. Application may misbehave. iconTheme "oxygen" Qps BUG: hashstr shortage statlen(385) > hashstr(384), report this message to [email protected]
debian packaging: http://git.siduction.org/summary/qt-desktop!qps.git
add git attribute file
Lines 234 to 262 in c78b406
@paulolieuthier
i'm fine with cmake over qmake - but please add a install part :)
Lines 99 to 100 in 2034ade
Commit 3e80692 is reproducibly crashing and dumping core upon exit on recent Arch Linux.
Problem seems to be related to Qt 5.6 as it got manifest only after compiling the said commit against Qt 5.6 but wasn't when the the same commit was compiled against Qt 5.5 for about a month.
Stack trace / backtrace on https://gist.github.com/pmattern/ccd321bc00789b5e4bdd43c885ac1530.
Due to a regression in 56d9096 the actual binary gets named Project instead of qps.
E. g. running
cmake <path to sources> -DCMAKE_INSTALL_PREFIX=/usr
make
make install
from a suitable build dir yields /usr/bin/Project instead of /usr/bin/qps.
Git master
By default the alignment of PIDs in tree view is frequently faulty in levels ≥ 3 like in this screenshot
The problem doesn't seem to exist right after launching qps but affects the majority of the said PIDs once the view was updated automatically several times.
Maybe it's due to column PID not being wide enough and could be fixed by making the columns' width adjustable.
This could probably best be achieved by making the spacer between the headers draggable, e. g. between PID and TTY, TTY and USER and so on when Basic Fields is chosen from menu Field. A feature like is desirable independent from this issue anyway.
But attempts to avoid an alignment as shown in the screenshot should be made whether or not it's related to the columns' width.
Use qps -reverse to run Qps and you'll see something whose description would need thousands of words.
I have changed some things in the last week that was long overdue:
TODO
That will not change nothing in the functionality nor solve any bugs but might make the code more maintainable.
An unsorted table doesn't make sense. Although the CPU column is sorted on the first run now, It's still possible to make the table unsorted. It should be prevented.
Will fix it after #177
reminder to myself
Although distro maintainer can now do it in their global config files, the sort indicator may show users that they can click other columns and sort them.
Sorting the CPU column by default can also show the update interval on the fist run.
qps is worth it.
qps.x86_64: I: binary-or-shlib-calls-gethostbyname /usr/bin/qps
The binary calls gethostbyname(). Please port the code to use getaddrinfo().
As in src/lookup.cpp#L190.
See Ulrich Drepper's explanation.
Hello, I made a patch for display the time in the qps miniHistory (very useful for timing the duration of the "peaks" of processor/IO that a program can generate), I've been trying for a few hours and so far it seems work well. Here I leave the content of this:
--- qps-1.10.16/infobar-original.cpp 2011-07-17 08:20:06.000000000 -0300
+++ qps-1.10.16/infobar.cpp 2015-12-18 03:13:46.378411055 -0300
@@ -7,6 +7,7 @@
#include <stdio.h>
#include <sys/types.h>
#include <time.h>
+#include <string>
#include "infobar.h"
#include "proc.h"
@@ -44,6 +45,16 @@
};
VCursor vcursor;
+// Convert epoch time to HH:mm:ss
+const std::string formatTime(time_t epoch_seconds) {
+ struct tm tstruct;
+ char buf[9];
+ tstruct = *localtime(&epoch_seconds);
+ strftime(buf, sizeof(buf), "%T", &tstruct);
+
+ return buf;
+}
+
//DEL
void Infobar::hideEvent( QHideEvent * event )
{
@@ -759,7 +770,7 @@
} */
-// TODO: 1.sort 2. time(?)
+// TODO: 1.sort
QString doHistory(SysHistory *sysh)
{
QString str;
@@ -768,7 +779,8 @@
char buf[128];
//sprintf(buf,"miniHistory /* %.02f%%",sysh->load_cpu*100);
- sprintf(buf,"miniHistory CPU");
+ const std::string saved_time = formatTime(sysh->time);
+ sprintf(buf,"miniHistory CPU [%s]", saved_time.c_str());
str+=QString::fromAscii(buf);
foreach(p,sysh->procs)
@@ -849,7 +861,7 @@
update(p.x()-5 -rel_x,0,p.x()+5 -rel_x,height()); // only Vcursor update
}
-// TODO: 1.sort 2. time
+// TODO: 1.sort
QString IO_Graph::doHistoryTXT(SysHistory *sysh)
{
QString str;
@@ -858,7 +870,8 @@
char buf[64],mem_str[64];
//sprintf(buf,"miniHistory /* %.02f%%",sysh->load_cpu*100);
- sprintf(buf,"miniHistory IO");
+ const std::string saved_time = formatTime(sysh->time);
+ sprintf(buf,"miniHistory IO [%s]", saved_time.c_str());
str+=QString::fromAscii(buf);
foreach(p,sysh->procs)
The patch is for QT4 version because I do not have QT5 due to it being not quite stable in my distribution but it should not be difficult to adapt the code to your fork of the program.
By the way, my knowledge of C and C++ are quite limited so do not rule out that there are more "clean" ways to do the same.
PS: If you need to clarify, the patch is released under public domain. ;-)
In less than 19 minutes it can take more memory than gnome shell
#149 helped a little, but there is still a huge leak left somewhere else.
My proposal is to track everything regarding the leaks here.
Why was "da" changed to "da_DK" in fe1f847#diff-ccd42e734cc98b6cd957b075d3448605 ?
I would like this not to happen in the future.
Can i change it on Weblate?
Line 1989 in df5cb58
Hello, please make a release, so distros can pack qps with a version number.
The code lacks reasonable information regarding licensing.
An informational file in the root directory doesn't exist, the headers of the actual source files only refer to a file like that.
I want to watch Swap size.
SWAP size is very large.
SWAP of Qps and SWAP of Top are different.
Screenshot:
https://github.com/lxqt/qps/blob/master/src/proc_linux.cpp#L804
https://github.com/lxqt/qps/blob/master/src/proc_common.cpp#L216
&size: 669587×4096/1024=2678348 kB (VmSize)
&resident: 102843×4096/1024=411372 kB (VmRSS)
sizeK: 2678348-411372=2266976 kB # SWAP field
I think that "&size" is not total memory but virtual memory(VmSize).
So I don't think that value of "sizeK" is Swap size.
I checked it and found it here.
Screenshot:
My PC have 2GB memory. Therefore, my PC requires SWAP. I want to watch Swap size.
However, many recent PCs have a lot of memory. SWAP may not be required.
This issue is not important.
You may encounter it in special circumstances. For example, when you type a filter that can't be matched (like "blahblah"), the horizontal scrollbar doesn't disappear, although it isn't needed. There are worse scenarios too. In short, the table width and its horizontal scrollbar position may not be updated correctly.
The reason is a few lines of bad code in qttableview.cpp and htable.cpp. I'll fix them after #174 is merged (because of conflicts).
The memory should not leak
Memory probably leaks
Rework
https://github.com/lxqt/qps/blob/master/src/command.cpp#L256-L271
https://github.com/lxqt/qps/blob/master/src/commanddialog.cpp#L132
This code also checks parent against null first, then uses the pointer as if no check was performed...
The config file should be (like) ~/.config/qps/qps.conf and other standard locations should be supported too..
It's ~/.qpsrc. That's old-fashioned. No app should do it anymore.
Very easy to implement.
By design:
(1) The PID field should always be the first section when the tree is shown; and
(2) The command filed should always be the last section.
Because of a mistake in the code, (2) can be contradicted.
I'll fix that alongside two related problems: (a) if a field is added from the context menu, it'll be placed after the command field, which contradicts (2); and (b) the PID and command fields can be dragged but they shouldn't because of (1) and (2).
Because there are lots of fields but they aren't sorted alphabetically, whether in the right click menu or in the custom fields dialog.
Then, the sort indicator will be moved to the left, while the sorted column is at its right.
Hello. As I've mentioned in #142 the project used https://github.com/lxqt/qps/blob/master/qps.desktop.in for translations. However, we now have also a Japanese translation in another location. Would not it be sane to have all the translations in a single place?
The good thing is - we have the code now in git, the bad thing is - we have no history, so i will re-build a partly history with old releases.
Title says it basically all, already: internationalization / localization would be something nice to have.
I'll fix it alongside #170 (because it's related to cell painting) but I don't have time to explain it. Just set the font size to 20 in Qps's Preferences and you'll see it (although no one uses 20, it shows a deeper problem).
Sorry for the bad report! I need these reports like a TODO list.
Now one can translate only few lines of Qps, not all. And this should be fixed.
In "modern" Qt GUIs (for 10 years, at least), the sorting header section has a sort indicator and isn't sunken. 3 images are worth 3 thousand words. The correct painting is:
With Kvantum:
With Fusion:
With Breeze:
And so on with any other widget style...
In these images, whose code will be included in a PR, there's no false assumption about header sizes: all sizes are taken from the active widget style.
Sorry, I can't explain it in English but the file "qttableview.cpp" contains suspicious codes and the compiler rightly issues warnings like: ‘cellMaxX’ may be used uninitialized in this function...
The dragged section (column) should be dropped exactly at the cursor position.
That doesn't happen. Sometimes it's positioned as the next/previous column, sometimes it isn't moved at all.
The fix is easy. I'll make a PR after #165 is merged.
As a matter of fact - the latest changes were big and great - with a bit of a but.
Process sorting is much better now, but still a bit crappy sometimes - sorry, has to be blunt and straight. I want the visual output of qps reproducible and deterministic, right now it isn't
Process sorting is not predictable right now fully - we should change it, the current state is not really a 2.0, but very close.
fixing it
I have some problems to write up steps to reproduce reliable, because when changing the views filters/sorting are not fully set/reset/applied - so the only thing i can bring are several screenshots:
the output i would expect first hand:
Reason - it would look a bit irritating otherwise:
(like this)
As i wrote, great work, but a little fine tuning is still needed. To be honest - we can release as 2.0 without a problem, but if it would be easy to solve as a fast shot i would prefer it :D
QPS is running and valgrind is checking it
segfault on qps->show();
valgrind --tool=memcheck --leak-check=yes ./build/src/qpsI was trying to work on memory leaks (the app leaks approximately one megabyte per second on my machine)
I suggest adding the strings from the .desktop to be translated to the .ts files so translators can easily translate them.
About on qps lists:
which is not found.
Guess it should direct to github.
Also, it may be useful to bump the version number.
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
Django
The Web framework for perfectionists with deadlines.
Laravel
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft
Open source projects and samples from Microsoft.
Google
Google ❤️ Open Source for everyone.
Alibaba
Alibaba Open Source for everyone
D3
Data-Driven Documents codes.
Tencent
China tencent open source team.