Re: NetBSD-6 firefox requires gcc 4.6, why not build 4.8?

[David Brownlee <abs%absd.org@localhost>]

On 22 October 2014 20:03, Jonathan Perkin <jperkin%joyent.com@localhost> wrote:
> * On 2014-10-22 at 11:48 PDT, David Brownlee wrote:
>
>> In order to build firefox on netbsd-6 gcc 4.6 or later is required.
>>
>> Currently gcc 4.6 does not build on netbsd-6
>>
>> /var/obj/pkg/lang/gcc46/work/gcc-4.6.4/libiberty/alloca.c: In function
>> 'C_alloca':
>> /var/obj/pkg/lang/gcc46/work/gcc-4.6.4/libiberty/alloca.c:200:5:
>> warning: implicit declaration of function 'XNEWVEC'
>> [-Wimplicit-function-declaration]
>> /var/obj/pkg/lang/gcc46/work/gcc-4.6.4/libiberty/alloca.c:200:43:
>> error: expected expression before 'char'
>> Makefile:504: recipe for target 'alloca.o' failed
>>
>> However, that does raise an interesting point - if an application
>> requires a version of gcc newer than that shipped with a given system,
>> why not default to building the highest supported gcc version, rather
>> than the lowest?
>>
>> If I'm going to have to build a gcc on netbsd-6, it probably makes
>> more sense to build 4.8 rather than 4.6, on the grounds that something
>> else I build later might need 4.8, and one would hope 4.8 would be
>> generating at least a good code as 4.6...
>>
>> What do people think?
>
> There are a few things in gcc.mk we need to cleanup, so it may make
> sense to add support for this at the same time - it seems reasonable
> as something we should support, a GCC_REQD_DEFAULT user var which is
> unset by default or something.

That would definitely also be a useful feature - what do you think of
changing the behaviour to build the most recent gcc version (if the
system determines that a gcc version needs to be built at all)