Re: TODO #1

[Han-Wen Nienhuys]

address@hidden writes:
> On Apr 27, 2004, at 2:55 AM, Han-Wen Nienhuys wrote:
> 
> > Cool, nice to see! While you're at it, can you double check that
> > post_incr and incr are used? If not, they can be scrapped wholesale.
> >
> 
> Right.  Here's the revised diff.  It turns out that incr() is being 
> used in one place (lily/box.cc).  With this patch I included incr() and 
> other_axis() as inline in the header, so that you can delete axis.cc.  
> Those are the only two functions from axis.cc that were being used, and 
> they're both small enough to be inlined.

Hi ,

I tried to put in your patch, but it rejected. Your mailer seems to
have munged it, adding spurious spaces at the start of the line. 
Can you resend, or send it to me privately?

> +++ flower/include/axes.hh    27 Apr 2004 21:50:11 -0000
> @@ -10,26 +10,28 @@
>   #ifndef AXES_HH
>   #define AXES_HH
> 
> +#include <assert.h>
> +

I have a second TODO that you might want to tackle.

The intention of

>   $(outdir)/%.cc: %.ll
>       $(FLEX) -Cfe -p -p -o$@ $<
> +ifeq ($(shell uname),Darwin)
> +     perl -i -pe 's/^class istream;/#include <iostream>;\nusing namespace 
> std;/' $@

is good, but we want to concentrate platform specific stuff in
configure.in.  Can you try to see if you can work this into a
configure.in test, which sets appropriate variables from
config.make.in.


-- 

 Han-Wen Nienhuys   |   address@hidden   |   http://www.xs4all.nl/~hanwen