Re: [Gcl-devel] Amusing bug

[Camm Maguire]

Greetings!  And thanks for bringing up the -I. issue so clearly.

I have confirmed that taking -I. out of cmpmain.lsp does not adversely
affect the maxima build, and does prevent me from having to rename
float.lisp to float1.lisp.  

I'm going to be bold and commit this, but I'm worried about non-gcc
compilers.  Anyone know the story here?

Thanks again!

Gregory Wright <address@hidden> writes:

> Hi,
> 
> While building maxima from cvs I came across an amusing bug. Here it is:
> 
> The maxima build failed while building float.lisp. It failed with a
> syntax error in a file included from /usr/include/math.h. It turns
> out...
> 
>       ...that math.h includes a file float.h (via #include <float.h>)
> 
>       ...that gcl also makes file float.h and includes it via
>       #include "float.h".
> 
>       ...but gcl calls gcc with an explicit -I. which causes the
>       current directory to be searched before any of the standard
>       locations.
> 
> The end result being that the float.h file included by math.h is the
> local one (the result of compiling a lisp file float.lisp) and not the
> one expected (typically in /usr/lib/gcc-lib...).
> 
> According the gcc doc, files included by #include ".." search from the
> current directory, so an explicit -I. shouldn't be required.
> 
> Is there a subtlety in this that I'm not seeing?
> 
> Best Wishes,
> 
> Greg
> 
> 
> 
> -- 
> 
> Gregory Wright
> Chief Technical Officer
> PacketStorm Communications, Inc.
> 20 Meridian Road
> Eatontown, New Jersey 07724
> 
> 1 732 544-2434 ext. 206
> 1 732 544-2437 [fax]
> address@hidden
> 
> 
> 
> _______________________________________________
> Gcl-devel mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/gcl-devel
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah