Re: [RFA] Add support for --without-zlib (take 2)

Re: [RFA] Add support for --without-zlib (take 2)

[Paolo Bonzini]

On 11/19/2009 10:13 PM, Joel Brobecker wrote:
> Hello Paolo,
>
> The following patch got approved for binutils and GDB, but I just
> realized that config/zlib.m4 should probably be approved by a GCC
> maintainer.  I was just wondering if ou wouldn't mind taking a look
> at the new config/zlib.m4?

As long as it is not used by GCC configury, I don't think people mind 
here.  The only suggestion I have is to have [default=auto] somewhere in 
the help string.

BTW, zlib is always distributed with GCC so we always include zlib 
support (and instead of --with-zlib, there is a --with-system-zlib flag 
that does what you expect).  Maybe you could do the same for binutils 
and gdb?  In that case, centralizing the test in AM_ZLIB is certainly 
desirable, and gcc may pick up the macro later or parts of it.

(Though there's no zlib in src, so probably you'd have to contact DJ 
Delorie to have it included in his list of autosynced directories).

Paolo

Re: [RFA] Add support for --without-zlib (take 2)

[Joel Brobecker]

> As long as it is not used by GCC configury, I don't think people mind  
> here.  The only suggestion I have is to have [default=auto] somewhere in  
> the help string.

OK - I will add it.

> BTW, zlib is always distributed with GCC so we always include zlib  
> support (and instead of --with-zlib, there is a --with-system-zlib flag  
> that does what you expect).  Maybe you could do the same for binutils  
> and gdb?  In that case, centralizing the test in AM_ZLIB is certainly  
> desirable, and gcc may pick up the macro later or parts of it.

Yes, Tom Tromey also mentioned that. Right now, I am more interested
in being able to build without zlib (we find its usefulness in GDB to be
marginal for us), but maybe the day will come when GDB/binutils must be
built with zlib, in which case I'll adjust zlib.m4 accordingly.

Thank you!
-- 
Joel