Hi Luigi,
thanks for your elaborated answer.
> For example I use: #include "my_class.hpp" and #include <iostream>; so
> that if
> I have to take some program to another system, I'll copy the file
> "my_class.hpp" too, instead I won't copy "iostream", because I know that
> the target system has its own copy of that file.
Normally I do it the same way. Command is a class which is needed in
different projects und itself it is work in progress. So instead of
changing, copying and asking myself if I put it everywhere where it is
needed I decided to put it in a private lib-dir and add this to the
automaticly by dev-c++ searched paths.
> In light of the above, the function "yy_flex_alloc(unsigned int)" is
> somehow missing. Since the file lex.yy.c is already in the project,
> probably you forgot to write the definition of the function
> "yy_flex_alloc", so there is only a declaration of the function,
> probably in the header file "lex.yy.h" (that is: the linker only has a
> prototype).
flex does generate the declerations but not in a seperate file which
should be no problem as far as I understood your example.
By the way thanks for that. That's it what's missing in all the
tutorials, they all end up with '... and then compile it to a running
executable'. But as I am learning right now it is a long way from
source-code over finding/including the right header files over getting
compiled object-code to a well linked program.
> Let me know if what I said did help.
It did :-)
For now it looks like I trapped into a bug of mingw and flex, see:
http://www.cygwin.com/ml/cygwin/1999-12/msg00382.html
I (still) try to switch from mingw to cygwin gcc which unfortunately
isn't done by clicking an dev-c++ option. :-(
So if someone is experienced in this case I would appreciate for some tips.
Gregor (pe...@in...)
|
