linking headers and sources
Hi,
I'm curious about the process that a standard compiler would use to "find" the source file for a given header. Prompted by this thread (http://www.cplusplus.com/forum/beginner/30207/), I was reading about function and class templates, and trying to compare the pros and cons between a plain old "function library" header/source, and using a function template instead.
With a function library setup, we would have a header and a source file, one for the prototypes and the other for the implementations. From any project we want to use those functions in, we need only #include the header file (correct?) and the linker will somehow link the appropriate source file for the function implementation (do I understand this correctly? I think so...).
My question is how the compiler/linker knows which source file contains the implementations of the header prototypes. It can't be based off the #include statement, because many of the sources in a project may include that. Is it the like filename? x.h always links with x.cpp, y.h always links with y.cpp? It seems a fragile thing to be linking on, so I wonder if there's more behind the scenes.
Reading on templates here in the tutorials, I noted this particular statement:
Why not, for a library of functions, just include both the function prototype and implementation in the header file, and do away with the source?
Thanks for any insight,
Jack
I'm curious about the process that a standard compiler would use to "find" the source file for a given header. Prompted by this thread (http://www.cplusplus.com/forum/beginner/30207/), I was reading about function and class templates, and trying to compare the pros and cons between a plain old "function library" header/source, and using a function template instead.
With a function library setup, we would have a header and a source file, one for the prototypes and the other for the implementations. From any project we want to use those functions in, we need only #include the header file (correct?) and the linker will somehow link the appropriate source file for the function implementation (do I understand this correctly? I think so...).
My question is how the compiler/linker knows which source file contains the implementations of the header prototypes. It can't be based off the #include statement, because many of the sources in a project may include that. Is it the like filename? x.h always links with x.cpp, y.h always links with y.cpp? It seems a fragile thing to be linking on, so I wonder if there's more behind the scenes.
Reading on templates here in the tutorials, I noted this particular statement:
| Because templates are compiled when required, this forces a restriction for multi-file projects: the implementation (definition) of a template class or function must be in the same file as its declaration. That means that we cannot separate the interface in a separate header file, and that we must include both interface and implementation in any file that uses the templates. |
Why not, for a library of functions, just include both the function prototype and implementation in the header file, and do away with the source?
Thanks for any insight,
Jack
http://www.lurklurk.org/linkers/linkers.html
Nevermind, I found it (maybe not fully understand yet), but there's a link on basic linker operation which answers the question about how a linker finds the correct source file for a given header.
I'm still slightly curious why we seperate the prototypes and definitions into sperate files though...
Nevermind, I found it (maybe not fully understand yet), but there's a link on basic linker operation which answers the question about how a linker finds the correct source file for a given header.
I'm still slightly curious why we seperate the prototypes and definitions into sperate files though...
Topic archived. No new replies allowed.