ShapeShifter: Nobody minds you asking basic C++ questions. I (and others) even took the time to answer them, even though those questions are covered by any introduction text to C++, some of which were even recommended here. We provided way more interaction here as you had in the other forum thread you mentioned (which did provide an answer which I found genuinely helpful by the way).
I am new to this whole forum thing (which I am doing almost exclusively in my spare time, I am not paid to answer questions here!) and this forum is new, too. So I am seriously interested in understanding your expectations.
this stylesheet, only modifies the height of the combobox itself, the individual item height in the popup is still defined by the font size. Not sure if we can set some margin for the font itself have to chk that :)
Thanks, Denis. What I'm trying to build now (this is actually a static lib) is already a part of SUBDIRS type .pro file. I thought about making the lib's top .pro file with SUBDIRS template, and putting a separate .pro file in all nested directories with OBJECTS_DIR overridden for each directory, so that object files do not conflict. I will keep this solution in mind.
To my point view, the best solution would be making qmake create the same directory structure under specified OBJECTS_DIR. I mean, let say we have
No one can do the same thing with already built C++ code with an acceptable cost.
iunknwn: Of course this relies on macros under the hood:-)
Runtime detection is usually not what you want in the first place: Usually you need some code to do something platform specific and that code will not compile anywhere but on that one platform. You need macros for that or you will get into trouble everywhere that platform specific code does not work.
Runtime detection is basically only useful if you need to know the exact version of the OS (e.g. Windows 7) you are running on and not just the generic type of OS (e.g Windows).
canFetchMore() will make a round trip to server as well. And it might get uglier than "select count(*)" since the hint indicate you should fetch more records to get accurate record count. More round trip for huge database of 100,000,000 records.
What database are you using? Mysql ISAAM will provide instantaneous response on "select count(*)" as it is cached. Mysql InnoDB and postgresql count the record everytime, thus "select count" is a little slow (Not really that noticeable).
Bottom line is, round trip to server is unavoidable. Franzk's addition by caching, looks marvelous. I will go with that.
[quote author="Franzk" date="1283406222"]Just because the memory footprint of a program is growing, it doesn't mean there is a memory leak. As Denis stated, the WTM is a rather crude tool and can therefore only really be used for a crude overview. [/quote]
I know WTM is not a good tool for that, but I thinks it can give me an idea of memory usage. Maybe this is not a memory leak, but it sound strange to me that a simple do-nothing application use so much memory and that it increases even if I do nothing. The same do-nothing application implemented as dialog (without the subwindow) has a stable memory usage. Increasing memory footprint dones't mean memory leak, of course, but if it isn't a memory leak, it is memory used somewhere by the application. In this case the application does nothing at all, so I suppose is Qt that is doing something with that memory.
Is that a normal behaviour for a Qt application?
My real-application, with about 64 QSpinBox into a subwindow, refreshed every 250ms, easly eats megabytes. At first I supposed there is a memory leak in my code, even it don't use the freestore, but then I extract the piece of the GUI I had posted. Of couse, maybe I have a problem with my code also.
I've managed to incorporate an Event Dispatcher defining an Event Sink Map using ATL to notify events in Qt now. After evaluating the possibilities not using ATL I did not find a feasible alternative. Too bad ATL is not part of the free Windows SDK or VC++ Express distros.
I'd still be pleased to here if anyone managed with ActiveQt or without ATL.
Apologies for replying to myself. After deinstalling the 3rd party codecs and testing for a while it crashed again, but only once, where before it would crash much more often. This seems to happen when I try to play an MPEG-4 H.264 file.
BTW, I don't understand why the file doesn't play without any 3rd party codec on Windows 7. Windows Media Player plays it just fine. Shouldn't everything that works in WMP also work in Phonon::VideoPlayer?
This is a huge problem for the app I'm writing. The customers expect to be able to play common video files, especially if they work in the default media player. WMV can not be used either, since that doesn't work with the Mac version (see "WMV Playback using Phonon::VideoWidget and Flip4Mac fails in 32-Bit builds":http://developer.qt.nokia.com/forums/viewthread/628).
Mixed backward/forward slashes is usually not a problem. Windows handles them pretty OK as it understands urls for local files as well. Also, while the paths are displayed with forward slashes in them, they are converted to backward slashes where necessary.
If you run into this in your code and you are not sure a library can properly work with the forward slashes, use QDir::toNativeSeparators(). It'll sort it all out for you.
Disabled Categories are greyed out
This topic will be moved to the category
Looks like your connection to Qt Forum was lost, please wait while we try to reconnect.