Very Computer

>> On Thu, 1 May 2003 22:07:41 -0400,

> Hi, I must have a misunderstanding of how make works.
> Here is a simple program to illustrate my problem.
> I have a file called a.cpp and a makefile that looks
> like this:
> CC = g++

> >> On Thu, 1 May 2003 22:07:41 -0400,

> > Hi, I must have a misunderstanding of how make works.
> > Here is a simple program to illustrate my problem.

> > I have a file called a.cpp and a makefile that looks
> > like this:

> > CC = g++

> Usually, CC denotes the C compiler; CXX is used for the
> C++ compiler.

> > all:
> >         ar -r mylib a.o

> all: a.o
>         ar -r mylib a.o

> You need to day that the archive depends on the object.

> (BTW static archives traditionally are called thing.a)

> I'd probably have done:

>     CXX = g++
>     CXXFLAGS = ...

>     OBJECTS = a.o [ b.o c.o ... ]

>     archive.a: $(OBJECTS)
>             ar qv[other flags] archive.a $(OBJECTS)

> hth
> t

> Hi,
> I must have a misunderstanding of how make works.  Here is a simple program
> to illustrate my problem.

> I have a file called a.cpp and a makefile that looks like this:

> CC = g++

> all :
>     ar -r mylib a.o

> If I type make all, I would expect make to not find any file called a.o, but
> notice that there is a file called a.cpp and would then use the built in
> suffix rules to build a.o from a.cpp.  However, make returns an error for
> not being able to find a.o.

> If I first issue the command g++ -c a.cpp which builds a.o and then do make
> all, mylib is built with no problem.

> I must have some misunderstanding of this.  Why does a.o not get built by
> the implicit rules of make?

> Any help in clearing up my misunderstanding, even flaming would be very much
> appreciated.

> Thanks

> Ed Doyle