Language Bindings

Guess that this is caused by the set flags. I don't know. Can I include that somehow as I know it optimizes the code more? Fortran uses C-linkage so any linker should be able to manage it. Note that fortran mandates decoration of the function names itself, because reasons. So if you have: subroutine dummy() end subroutine dummy This is seen by the linker as if it was defined as void dummy_(). If you intend to call fortran from C, you must provide the proper declarations for the compiler in your C code, they can't be deduced automatically. When I use either gfortran or ifort in my code and I'll try to do the cross compilation. Will the fortran part of the code run flawlessly on Mac or Linux? As long as you configure the compiler properly, you can treat fortran code as any other C code, so it should work fine, yes. Tangent: gfortran, I believe, still transpiles the fortran source to C and then runs the C compiler on top of it. In the end why should anyone invest time into keeping a real compiler for a zombie language up to date ...

I'd recommend reading the Qt5 version: https://doc.qt.io/qt-5/signalsandslots.html @GastonMelo You really should take the time to go through Qt's documentation. It's full of tutorials, getting started guides and it also explains all the core concepts provided within the framework.

Hi, Thanks for the feedback ! Can you post the link to said patch ?

Hi, Good question. What does your function look like ?

@JonB Thanks

Just to close this up. Although I posted this question here, to the the PyQt mailing list and on stackoverflow, since I never got an actual answer as to how to write the derived class's constructor to accept the new positional argument while retaining all the existing overloads, in the end I have actually gone for a named argument: class JListWidgetItem(QtWidgets.QListWidgetItem): def __init__(self, *args, value: typing.Any=None, **kwargs): super().__init__(*args, **kwargs) # optional `value` member, passed in as `value=...`, stored in data(Qt.UserRole) # this is a common thing to want, cf. `QComboBox.itemData()` if value is not None: self.setValue(value) and caller goes e.g. JListWidgetItem("text", value=1). Some people have said this is the more "pythonic" way, who knows, but at least it works! Thanks to all.

The first thing you should do is to make Worker inherit both QObject and QRunnable. QObject must come first. Making it a separated class and calling signals on it is wrong. Signals should be emitted from within the class that defines them. Another thing is that you don't set your worker to auto delete so you will end up eating lots of memories depending on how many times plot is called.

@Julian-Guarin-0 said in Web Assembly and decalarative-camera example.: So I guess you are right. Does this confirm what you did see? as i said, i didnt see anything ;) But the configure clearly shows that you might only be possible to play sound files, i don't know if thats really true though for QtWebAssembly.

@aha_1980 said in Pyside QLineEdit entering data from right-to-left: @Joel-Graff your example does not look like valid float number. Do you need exponential format? I'm not sure, if QDoubleSpinbox provides that, otherwise it would be first choice for that task. For the line edit, you could use input masks or hook up an own 'correction function' on every key press. It is, in fact, a valid float. It's called a "station" where 1 station = 100 feet. Thus 999 feet = 9.99 stations, which is commonly annotated 9+99. It's really just a length. The example I gave is actually the input mask I'm using - sorry for the confusion. The problem is, the input mask doesn't really fill out the required placeholders correctly. That is, if I create an input mask that reads: '00009+99.00;_' I would expect it to show ____0+00.__ as a default. Of course, it doesn't, and I admit, I don't really know much about input masks in Qt. Since I can't rely on the input mask to control the data entry, forcing it to enter data from the right (and push the characters to the left) seems the next best option. It's a fairly common data entry technique - I've seen it even in very old terminal-based systems. I was hoping I didn't have to craft my own callback to manage it. Thanks anyway. :)

@aha_1980 thanks. They can add it to the gallery if they like, but the plugin is still work in progress.

**My first problem has been resolved, so I will close this, and open a topic for my new problem. For future people who come to this topic : The solution for "Device not open" was to use "start function" instead of "execute function". Thank you people for helping me.**

Did you also check this https://cgit.kde.org/perlqt.git ?

@aidanikuz please watch following video to install and use pyqt5 https://www.youtube.com/watch?v=yqHiOesB-nI

yes. I guess it might be encoding problem . I set the UTF8 encoding in the options,but it didn't solve the problem. Platform: QT version :Android for armeabi-v7a (GCC 4.9, Qt 5.9.3 for Android armv7) OS: windows 10 Qtcreator version: 4.7.2 Compilers: Android gcc arm-4.9, JDK1.8 Thanks.

@blue34 I haven't tried grpc before. It uses protocol buffers though which I have used and aren't a bad way of doing things. I've never much cared for RPC mechanisms but whatever is easy for you to get implemented in the C++ side and the Java side is a decent way to do things. :) Like I said I always go the other way... native c++ library with the front end in java (android) and ios (objc). I've been trying out djinni for that lately and it makes the jni and what not nice and easy. Wrong direction though so that won't help you out much.

@vivekyuvan As @sierdzio already said, you cannot directly call C++ from C, and therefore you cannot use Qt.. But it is possible to write a wrapper, like discussed here: https://stackoverflow.com/questions/3583353/calling-c-class-member-function-from-c-code From within the C++ class, you can access all kind of Qt stuff again.

I can't explain difference between case 2) and 3). However, what you see in 1) is that signal-slot connections are always executed in order in which connections were made (hint: don't depend on this, though). So, when you connect in your plugin, you always do it after main program already had connected it's signals and slots - and thus your slots are executed after main program's, too. Then in 3) you get immediate execution because (probably) there was no connection made in main program for that particular triggered() signal - you are first in queue. Why 2) and 3) show different behaviour - I don't know.