Re: gnustep-make experiment

[Nicola Pero]

>> You can have non-flattened multi-platform installations that are 
>> mounted from 
>> the network; the same gnustep-make will then use different 
>> compilation 
>> flags/tools for the different hosts (keep in mind that each host 
>> might also 
>> have a different
>> filesystem configuration, eg, they could be sharing a network-mounted 
>> System 
>> domain,
>> while having different domains in different locations).
>
> this is also possible if you split up the .pc files one for each 
> domain configuration
> then the different hosts can configure their PKG_CONFIG_PATH's to
> different search orders so each machine/user gets different 
> configurations for different.

What I was trying to explain is that pkg-config only prints a set of fixed 
flags from
a text file.  But that's the very easy bit! ;-)

The hard part in the implementation is actually getting the flags.  The only
easy way I can think of doing that is by having a quick gnustep-make run with 
some special targets that just print the values.  This can be embedded
into a gnustep-config script, since we are free to put whatever shell code
we want into gnustep-config. :-)

So, in all cases we need to have a gnustep-config script that outputs the flags,
and implemented as explained (that script will also work on all platforms).

I suppose using this gnustep-config script you could then generate .pc files
and put them somewhere ... so people can use pkg-config instead of 
gnustep-config 
to print the flags on Linux. ;-)

I can't see why they would want to do that though, as the tools could have a 
similar syntax, 
but gnustep-config would be automatically available everywhere gnustep-make is, 
while 
pkg-config wouldn't.  So anyone who understands what they are doing will always 
want
to use gnustep-config for portability. ;-)

This is not because of some evil prejudice against pkg-config, it's just that 
the building
system we have works in a different way.  Getting the compile/link flags so 
that you're able
to compile/link things without using gnustep-make is certainly something we 
want to be able
to support, but the fact that pkg-config can print flags reading from a text 
file doesn't 
really help in getting there.  Because we don't even store fixed sets of flags, 
but we 
compute them dynamically, using pkg-config to print them is more of an 
additional problem
than a solution.

And then adding an external dependency - which is a massive pain for users, 
developers
and maintainers - just to do the equivalent of 'echo $CFLAGS' is somehow hardly 
attractive.

Thanks