Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror
×
Programming IT Technology

QT 3.2 Released 95

GlennZ writes "Today, Trolltech has released version 3.2 of QT. This release includes a completely rewritten, faster font-rendering engine and a lot more. Go download it today!"
This discussion has been archived. No new comments can be posted.

QT 3.2 Released

Comments Filter:
  • Not out yet :(.
  • font rednering (Score:1, Interesting)

    by Anonymous Coward
    why not just use freetype and be done with it?
  • Splash Screen (Score:4, Insightful)

    by Per Wigren ( 5315 ) on Wednesday July 23, 2003 @10:32AM (#6511717) Homepage
    * QSplashScreen (add splash screens to applications)

    I really hope that one will be able to disable this by a standard Qt-argument (--no-splash) or by a env-variable because I'm afraid this will be abused.. :P
    • Re:Splash Screen (Score:4, Informative)

      by Arandir ( 19206 ) on Wednesday July 23, 2003 @01:08PM (#6513235) Homepage Journal
      The QSplash class is trivial. So trivial I wrote my own in an hour that turned out to 95% identical to Qt's before I knew that they were going to have one. It simply puts a pixmap widget on the screen, starts a timer, and waits for either a timeout or a mouseclick.

      QSplash is not going to check the command line arguments. This is an instance where it is the programmer's job to determine if the user wants a splash screen or not. But don't worry too much, every instance of a splash screen in a Qt application I have ever seen (Kdevelop, Quanta, etc) allows you to turn it off. Using QSplash instead of a homegrown splash isn't going to change this.
      • Re:Splash Screen (Score:4, Informative)

        by cyb97 ( 520582 ) <cyb97@noxtension.com> on Wednesday July 23, 2003 @01:50PM (#6513625) Homepage Journal
        It's probably gonna make more developers create splash screens as it's timesaving... While it's a waste spending about an hour programming a splash-class if your program lacks in other divisions; spending 5-6 minutes implementing this new class is more appropriate...
        If lack of time is the reason it's implemented in the first place, I guess implementing a --no-splash is probably not in the .plan for some future ;-)

        But I can't really see how more choice can be worse...

        • Re:Splash Screen (Score:4, Informative)

          by Arandir ( 19206 ) on Wednesday July 23, 2003 @05:00PM (#6515992) Homepage Journal
          I guess implementing a --no-splash is probably not in the .plan for some future

          Well, it's also goes against the Way of Doing Things. Standard Qt command line options should not modify the behavior of the software. So the command line options you see are stuff like fonts, colormaps and the like. You don't see stuff like "--cancel-button-left" because that's not Qt's job, it's the developer's.

          KDE is another story, because part of the purpose of KDE is to provide a unifrom look and feel for the desktop. I can easily imagine a KSplash class that checks for a --no-splash option.
  • by TheSHAD0W ( 258774 ) on Wednesday July 23, 2003 @11:31AM (#6512337) Homepage
    Isn't Quicktime 4 already out? And who's Trolltech? I thought Apple made Quicktime...
  • Why moc? (Score:3, Interesting)

    by dalleboy ( 539331 ) on Wednesday July 23, 2003 @02:27PM (#6513928) Homepage
    When are they dumping the Meta Object Compiler (moc) [trolltech.com] and switch over to Boost.Signals [boost.org]?
    • Try actually using both. Qt signal/slots are much more flexible than Boost.Signals. But beyond that, I did a test once (for a class) where I used Qt signal/slots versus Boost.Signals for the exact same test application. The Qt version was 17k statically linked. The Boost version was 60k statically linked. In addition, the syntax was significantly gnarlier with Boost than with Qt.
      • The moc is not just about signals. It is used for type identification, handles properties and some other stuff as well, these capabilities are used for example by Designer.

        Also, boost signals aren't exactly the same as Qt signals (e.g. in Qt you can connect an X(int) to a X(void) slot, and there are some automatic conversions as well). Changing to boost would impact existing Qt applications, and the Trolls are quite proud of their binary and source compatibility.

        As pointed out by other posters, some of th
      • Sure the boost syntax is a bit gnarlier, but not much. As on the flexibility of Qt it has a lot of limitations (http://doc.trolltech.com/3.2/moc.html).

        // Boost.Signals usage:

        class boost_signal_holder
        {
        public:
        void send_signal(int i)
        {
        boost_signal(i);
        }

        void connect_slot(const boost_signal_type::slot_type& slot)
        {
        boost_signal.connect(slot);
        }

        typedef boost::signal<void (int)> boost_signal_type;

        boost_signal_type boost_signal;
        } boost_signal_holder_object;

        class boost_slot_holder : public b

Your own mileage may vary.

Working...