Re: a problem with GNU make

[a R T u n]

Paul D. Smith wrote:

> I should point out that the version of GNU make that comes with the
> Cygwin toolkit has been modified by the Cygwin folks: it's not the same
> as the version that comes from the FSF.

That is interesting because in Cygwin it shows that GNU Make is
copyrighted to FSF.

>   artun> The makefile that I am using for this is:
>
>   artun> %.obj : %.cpp
>   artun>      @$(CC) $(CFLAGS) $<
>
> The above pattern is not correct for what you want to do.  If you're
> trying to compile $(OBJ_DIR)/foo.obj from $(SRC_DIR)/foo.cpp, your
> target pattern needs to be:
>
>     $(OBJ_DIR)/%.obj : $(SRC_DIR)/%.cpp
>
> You will probably find the doc on "How Not to use VPATH" on my web site
> (in my sig) interesting reading.

I've read that doc and learned not to use VPATH for the output. But I
changed the erroneous pattern rule to be:

$(OBJ_DIR)/%.obj : %.cpp
        @$(CC) $(CFLAGS) $<

giving vpath right after directory definitions as:

vpath %.cpp $(SRC_DIR)

As I pasted here I preferred "vpath directive" instead of "VPATH
variable". The reason was that I wanted Make to grab only cpp files. It
was as written in the O'reilly book named GNU Make. I thought it is
true since it compiled with no error. I mentioned it to discuss the
difference.

>   artun> $(OUTFILE) : $(OBJ_DIR)$(OBJECTS)
>
> That won't work; if OBJECTS contains more than one value OBJ_DIR will
> apply only to the first one.  You mean something like:
>
>     $(OUTFILE) : $(addprefix $(OBJ_DIR)/,$(OBJECTS))
>
>   artun> But this time I got an error from GNU make that "No rule to
>   artun> make target firstfile.obj".

Your offer worked exactly with no error expect a problem. Let me first
paste my modification:

$(OUTFILE) : $(addprefix $(OBJ_DIR)/,$(OBJECTS))
        @$(LD) $(LFLAGS) $^

I needed to change $< to $^, to let Make to link all the object files,
because when I wrote $< only the first obj file was linked although
there came no error. I guess of course only the first obj file is
linked, because the output file was exactly 2KB although it should have
been hundreds of KBs.

Thanks in helpful advices...