Qt Contribution

It is all working now - I had to do some finagling using git remote to correct the entries for gerrit to use my userid rather than the email.

Have you followed the steps described here? At the end of the day it doesn't matter how the project is configured, as long as you pass the right arguments and the build path is clean. I don't know which version of Qt you are building on which operating system, but it looks like the build tree has been configured before with some flags that force it into a debug build.

@Paul-Breeuwsma said in Idea for not-null pointer type support in the Qt API: passing null would give undefined behaviour 99.9999999% of the times passing null is segfault. QUndoGroup::addStack is a great example. Its an easy crash to debug I feel like the concept of "not null pointer" already exists in C++, it's called a reference. Qt already uses references as never-null-pointers internally in the D-Pointer architecture (e.g. QObject::QObject(QObjectPrivate &dd, QObject *parent) where the first argument is actually a pointer that is never null). It just turns out sometimes it's more confusing to use the reference than having an explicit pointer

https://github.com/hypre-space/hypre may be interesting as well.

@Engelard there's no direct integration, so yes, you've to file things anew.

@xxuan You could also ask on Qt Developers mailing list where you will find actual Qt developers. This forum is for Qt users.

@IDontKnowHowToCode PyQt is produced by Riverbank Computing. If you would like to see them implement something in PyQt which would help you debug you can ask them. Or this might be a general issue in Python code which calls C libraries.

I think this is a great idea. It would be great to have more control over the serialization of JSON data. It would make it easier to customize the output and make sure that the data is properly formatted. It would also make it easier to debug any issues that may arise. I think this would be a great addition to the Qt library.

My steps: cd /home/user8937666/QtProj/key/ ssh-keygen -t ed25519 -f /home/user8937666/QtProj/key/my_qt_ssh_key -N "" -C myKeyForQt <<< $'\ny' >/dev/null 2>&1 ssh -vvv -p 29418 user8937666@codereview.qt-project.org -i /home/a/QtProj/key/my_qt_ssh_key Now works (thanks). But with default RSA (without -t ed25519 ) it doesn't work (error "no mutual signature algorithm"), why?

Hello. I'm using a different account now. What happened was I subscribed to the mailing list and sent the message, but I got a reply saying that I can't send messages to the mailing list because I'm not a Qt developer. I don't know why but after that I just gave up on the idea and tried to create an external binding, and I completely forgot about this untill now. It's probably a good thing though, because I've gained a lot more experience in C++ since then and I think I wouldn't have been able to contribute anything without it. I've created a gerrit account by following the "Contributing to Qt" wiki, and I'm going to try to send a message again. Just wanted to appologize for forgetting about the thread half-way through the discussion. I'll mark it as solved. EDIT: I can't mark it as solved on this account, but I'll sign in later today and do that.

Hi, and welcome! @Xarzu said in Typo found online: Hello friends. I want to help. I found a typo online (https://wiki.qt.io/Building_Qt_5_from_Git#Getting_the_source_code) under the caption "Cloning specific Qt submodule", the first word "Ypu" should be "You". I only want to help. Thanks for reporting this. I have fixed it. In the future, you can edit Wiki articles yourself. Just log in with your Qt Account. @jsulm said in Typo found online: You should file a bug report here: https://bugreports.qt.io/secure/Dashboard.jspa This is user forum without many developers working on Qt/documentation. This is true for articles at https://doc.qt.io/

For future reference, I am going to answer my own question: pyside6-lupdate is the same thing as the main Qt Linguist lupdate, so its source is located at https://code.qt.io/cgit/qt/qttools.git, and more specifically https://code.qt.io/cgit/qt/qttools.git/tree/src/linguist/lupdate . This appears to be based on the original "pylupdate" code, though it's been refactored a fair bit.

@SGaist said in How is QSqlTableModel ok?: the model shall be deleted before the QCoreApplication and thus its private part as well and thus the db object. Not necessarily... for example #include <QApplication> #include <QTableView> #include <QSqlTableModel> #include <memory> int main(int argc, char *argv[]) { std::unique_ptr<QSqlTableModel> model; QApplication a(argc, argv); QSqlDatabase db = QSqlDatabase::addDatabase(QStringLiteral("QSQLITE")); db.setDatabaseName(QStringLiteral("/path/to/mydb.sqlite")); Q_ASSUME(db.open()); model = std::make_unique<QSqlTableModel>(); model->setTable(QStringLiteral("MyTable")); Q_ASSUME(model->select()); QTableView tableView; tableView.setModel(model.get()); tableView.show(); return a.exec(); } Now we can agree that, in practice, it seldom happens but I still don't understand if the design choice opens up the problem without safeguards or warnings or if I'm just missing something that makes it ok.

Hi, @davecotter said in How to have my .pro depend on qt.pro ?: Third: what's the recommendation about -developer-build ? I understand that regardless of "-prefix" setting, it will build within the Src tree, which is exactly what you do NOT want assuming you're using Git to submit your changes? Out of source builds are the key. @davecotter said in How to have my .pro depend on qt.pro ?: Fourth: the .pro.user file is created within the Src tree, which is not okay with Git. How do i force Qt Creator to put that somewhere else? Usually ignored by the .gitignore file. @davecotter said in How to have my .pro depend on qt.pro ?: Fifth: assuming all the above gets straightened out, when i build in Qt Creator, note i do not do "make install". does that happen automatically? do i need a "post build step" on the "qtmultimedia/src" project to stage that into place for debug builds? You need a custom build step indeed.

AFAIK, it should not. Did you try to do a clean rebuild ?

the answer was i was building the wrong branch. switched to 5.15.2 and all is well.