Janus wrote:
> i did that yesterday and it didn't work. now things have changed a bit.
> it works but i don't know why. i have moved some declarations and
> includes in the meantime (basically i cleaned up a bit) but i can't see
> how this effected this issue.
Doing a 'make clean;make' causing things to work is likely
due to changes being made to your apps own #include files,
like the redefinition of classes or structures.
When you change the definition of classes/structs,
all code using them has to be rebuilt. If you don't,
prebuild modules (.o's) still assuming classes/structs
are of the old size and arrangement will crash or go wild
when the rebuilt main() version passes pointers to classes/structs
with the new size down to the old un-rebuilt modules.
Rebuilding all the code makes sure all the modules agree
on class/struct sizes, and is probably why your code suddenly
works when you make clean;make.
A common similar thing is having a hard bug that you can't figure
out, then adding a printf() during debugging makes it work, then
you take the printf away and it still works.
All basically the same issue; a problem in the makefile's dependencies.
If the makefile checks .H file date stamps against /all/ code that uses
them, then this won't happen. But that would involve using makedepend
(or whatever it is) which most folks don't use (myself included), so
I just know to do a 'make clean; make' whenever I diddle around with .H
files..
_______________________________________________
fltk mailing list
[email protected]
http://lists.easysw.com/mailman/listinfo/fltk
