Yeah I downloaded the example zip again from your Touchkit website and downloaded new OF FAT package (with ofxTouch addon added) and test it, and it was still giving me the error. (it doesn’t say its an error though) However pngBrush example works fine in the same setup… and there is only an error in that allocate line.
Umm I will simplify and work off pngBrush example for now. I will update if there is any luck with HelloWorld example compiling…
to the ofxTouchApp constructor in ofxTouchApp.h it worked. So could someone tell me, why the pointers get automatically (and correct) nullified in one example and not nullified in the other? Or is this just randomness?
So could someone tell me, why the pointers get automatically (and correct) nullified in one example and not nullified in the other? Or is this just randomness?
You’ll get undefined behaviour with uninitialized POD class-members. (bools, ints, floats, pointer,etc) This is c+±standard.
Only static variables get nullified by the compiler, not local variables or class members.
Some compiler do init class-members to 0, but only for debug builds. Visual Studio initializes class-members to a special bit-pattern to recognize uninitialized variables / class members, but also only for debug builds.
So do initialize EVERY class-member in the constructor, or you’ll have a hard time debugging nasty bugs / crashes.
But still: When compiling BOTH applications in DEBUG mode, one crashes and the other does not. If this is random, well then I guess I miss random. (sounds like “To understand recursion you have to understand recursion” :? )
But I think thats exactly what you meant with
… or you’ll have a hard time debugging nasty bugs / crashes.
By the way, it has definitely to do with the pointers: when doing something like
‘uninitialized’ means ‘uninitialized’: whatever was at that physical memory location before is still there. for example if you were just watching a movie, it could be fragments of movie frame… so sometimes it will crash, and sometimes it will seem to work fine, but the subtle thing is that when it’s seeming to work fine it’s actually possibly writing random garbage over some important part of your program, which for example might make it crash in totally weird ways about 30 seconds later.
Is the new fix posted as a replacement of old “ofxTouch-Examples-001-xcode.zip”?
I downloaded the zip again and HelloWorld still didn’t compile it fine…umm. I didn’t change the code at all either. (with Latest OF FAT)