Build is verified!

Ryan.Oliver at pha.com.au Ryan.Oliver at pha.com.au
Thu Mar 20 20:08:42 PST 2003


> >
> > No, not cross compiling... cross linking.
> > Host arch and OS are the same
>
> well, functionally equivalent to a cross compile where host == target,
> then :)

Hehehe, I'll give you that ;-)

> > the inability to run the freshly compiled app during the course of the
> > build therefore we would be using the host system binaries up until
> > the point of chroot...
>
> Irrelevant, since that would be stuff like sed/awk. the toolchain used
> and libs etc. would all be the cross compiled ones...

It depends on what the base system you are coming from is, and whether the
tools supplied on the host OS have all the available functionality
required.
Not always the case...

Less margin for error the current (PLFS) way...

> > Probably more but thats all I can come up with
> > off the top of my head...
>
> You half succeeded in convincing me of your point. ;-) (one day, I
> intend to cross compile LFS from ARM to X86, just to see what breaks ;-)

Lots... I started down this road a while ago, still trying to get something
relatively workable building from Solaris 2.8 to test the build on sparc (
I could just stick a RH on there but where's the fun in that ;-) )

A hint, save yourself a lot of pain, use ash not bash :-)

> well the ch5 gcc had better build OK in either scenario - if it doesnt,
> theres not a lot you can do

Thats true, hence the usage of "make check" throughout the build... if the
first gcc binutils and glibc compile OK you can be pretty comfortable in
the fact that the rest will... as long as there aren't any weird bugs in
those packages ( out of our control )

> , as you may well only find out when you
> compare the stage 2/3 ones ;)

That comes down to performing the iterative checks and squishing the
differences before the next run...

> > That'd be far too convoluted for the book...
>
> maybe...

You could never write a generic cross-compile build, only a guideline on
what steps to take and what to look out for to achieve it... far too many
variables...

> > Good questions mate, hope this clarifies things a little...
>
> Thanks. I appreciate the time you took to answer my questions. Perhaps I
> generated some food for thought too, eh? :)

Hehehehe, good to get the brain working again on something other than shell
scripts :-)

Regards
Ryan

-- 
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-dev' in the subject header of the message



More information about the lfs-dev mailing list