Keep bison-generated files in sync.

Keep bison-generated files in sync.

Use a single rule to build header and source files from
the syntax file, thus avoiding any possibility to have
`make' rebuild one but not the other, which might happen
if the syntax and target files have almost the same but
slightly differing timestamps.

See http://www.mail-archive.com/bug-lilypond@gnu.org/msg37081.html

[Julien Rioux, 2013-09-02 10:59:10]

Add explicit dependencies on parser.hh

[lemzwerg, 2013-09-02 17:18:59]

LGTM.

[dak, 2013-09-02 17:46:15]

https://codereview.appspot.com/13466043/diff/2001/stepmake/stepmake/c++-rules...
File stepmake/stepmake/c++-rules.make (right):

https://codereview.appspot.com/13466043/diff/2001/stepmake/stepmake/c++-rules...
stepmake/stepmake/c++-rules.make:16: $(BISON) -o $(subst .hh,.cc,$@) -d $<
Han-Wen's version was slightly different:

       $(BISON) -d -o $(subst .hh,.cc,$@)  $<

but I don't consider it graceful.  I'd try

       $(BISON) -d -o $*.cc  $<

instead.  I _think_ this should take the right directory.

https://codereview.appspot.com/13466043/diff/2001/stepmake/stepmake/c-rules.make
File stepmake/stepmake/c-rules.make (right):

https://codereview.appspot.com/13466043/diff/2001/stepmake/stepmake/c-rules.m...
stepmake/stepmake/c-rules.make:16: $(BISON) -o $(subst .h,.c,$@) -d $<
Do we need this rule at all?  We don't have .y files in the tree and the only
actual C file would appear to be python/midi.c anyway.

[Julien Rioux, 2013-09-02 17:59:43]

See http://code.google.com/p/lilypond/issues/detail?id=3528

https://codereview.appspot.com/13466043/diff/2001/stepmake/stepmake/c++-rules...
File stepmake/stepmake/c++-rules.make (right):

https://codereview.appspot.com/13466043/diff/2001/stepmake/stepmake/c++-rules...
stepmake/stepmake/c++-rules.make:16: $(BISON) -o $(subst .hh,.cc,$@) -d $<
On 2013/09/02 17:46:15, dak wrote:
> Han-Wen's version was slightly different:
> 
>        $(BISON) -d -o $(subst .hh,.cc,$@)  $<
> 
> but I don't consider it graceful.  I'd try
> 
>        $(BISON) -d -o $*.cc  $<
> 
> instead.  I _think_ this should take the right directory.

The ordering of -d and -o options doesn't matter, so it was exactly the same. We
can use $* if you prefer, but let's not forget the output dir, so
$(BISON) -d -o $(outdir)/$*.cc $<
OK?

https://codereview.appspot.com/13466043/diff/2001/stepmake/stepmake/c-rules.make
File stepmake/stepmake/c-rules.make (right):

https://codereview.appspot.com/13466043/diff/2001/stepmake/stepmake/c-rules.m...
stepmake/stepmake/c-rules.make:16: $(BISON) -o $(subst .h,.c,$@) -d $<
On 2013/09/02 17:46:15, dak wrote:
> Do we need this rule at all?  We don't have .y files in the tree and the only
> actual C file would appear to be python/midi.c anyway.

It is part of stepmake, and it is conceivable (although I wouldn't wish it on
anyone) that somebody has decided to use stepmake for their own project.
(lilypond doesn't use this rule)

[dak, 2013-09-02 18:21:42]

https://codereview.appspot.com/13466043/diff/2001/lily/GNUmakefile
File lily/GNUmakefile (right):

https://codereview.appspot.com/13466043/diff/2001/lily/GNUmakefile#newcode69
lily/GNUmakefile:69: $(outdir)/score.o: $(outdir)/parser.hh
The additional dependencies here appear all wrong.  All of the additional files
include lily-parser.hh rather than parser.hh, and lily-parser.hh does not in
itself include parser.hh.

[Julien Rioux, 2013-09-02 18:42:50]

Respond to review comments.

[dak, 2013-09-03 11:10:17]

https://codereview.appspot.com/13466043/diff/12001/stepmake/stepmake/c++-rule...
File stepmake/stepmake/c++-rules.make (right):

https://codereview.appspot.com/13466043/diff/12001/stepmake/stepmake/c++-rule...
stepmake/stepmake/c++-rules.make:16: $(BISON) -d -o $(outdir)/$*.cc $<
It's somewhat disillusioning, but if I do a diff on my multi-commit series of
patches for this issue, I arrive at exactly the same one-liner.  Except that I
still have an additional space before $< like in the original.  But then the
other uses of $< in this file don't have this space either, so LGTM.