Date: Fri, 27 Nov 1998 19:40:05 +0000 From: "Paolo Di Francesco" <paipai@tin.it> To: freebsd-sparc@FreeBSD.ORG Subject: Re: [Ultra] Compiler, again Message-ID: <19981127183748.EERI23855.fep04-svc@winworkstation> In-Reply-To: <Pine.BSF.4.05.9811271140390.1605-100000@bright.fx.genx.net> References: <Pine.BSF.4.05.9811270629001.96971-100000@ender.sf.scient.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> >
> > I agree emphatically with staying with BSD make. This -is- BSD that we're
> > porting after all. Starting down the path of gmake will only make it more
> > difficult later.
>
> *nod*
>
What is nod? 8)
[toolchain]
>
> I've been pulling my hair out trying to do this, i downloaded the source
> rpms like you suggested... they both come with several patches which i
> applied, but then it got all confusing.
> Basically, each source rpm comes with a huge patchfile along with several
> other patch files, the huge patchfile i apply in the directory above the
> source tree, and it patches the source tree, however it also makes a
> directory where there are duplicate files of the tree.
> If i merge this directory into the extracted directory then
> configure/compile it seems to work as then the binutils compiled fine,
> however i think i messed up the flags i ghave to configure because while
> this "method" of building worked for binutils, it seems that the same
> flags passed to configure for gcc produce an empty makefile. :(
> I was hoping to make a pkg but I can't seem figure this out.
We need a "black box". Something you put the kernel code in and you obtain the
kernel out.
+-----------+
| |
Kernel code --> | Black Box | --> kernel
| |
+-----------+
So, if someone is using this "black box" please put it on-line somewhere or
tell us how to oabtain it.
Another problem is what to use for this "black box". Someone is using the
toolchain, someone is waiting (like me) for a good crosscompiling enviornment,
so if you have suggestions please tell us what to do!
[I have an Intel-box, no Ultra]
Please note we have no compatibility constrains about experiments, and for the
gcc version, when will ship with the freebsd-sparc who knows what version of
gcc we'll include... maybe 2.9, maybe 3.0, maybe 10.0 ;)
So what about using gcc2.8 or newer version? The important thing is to have the
easiest "black box". (easiest means, the easiest way of building it...)
Another important thing is to have specific "black boxes". I mean we can use
many different enviornment, but everyone of them must be "well fixed".
For example:
Black-box #1: running on i386, used gcc2.7.2 + toolchain, etc
Black-box #2: running on i386, used gcc2.8 + ????
Black-box #3: running on i386, used egcs1.1b + ????
Black-box #4: running on i386, used gccX.Y.Z and emulator under i386 + Solaris bin...etc...
Black-box #5: running on USparc, using gccX.Y.Z
...
...
...
Let me underline it again: we need packaged things so we can do experiments
plugging-in different black-boxes, and playing with code easly! 8)
If you don't help in this there will be not so many kernel developers here...8(
Ciao Ciao
Paolo Di Francesco
_
->B<- All Recycled Bytes Message ...
~
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-sparc" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19981127183748.EERI23855.fep04-svc>