Re: Building a cross compiler, and unsure about the use of --without-headers

[David Paterson <dnpaterson at gmail dot com>]

On Wed, Mar 16, 2011 at 10:40 AM, Kai Ruottu <kai.ruottu@wippies.com> wrote:
> 16.3.2011 11:15, David Paterson kirjoitti:
>>
>> On Wed, Mar 16, 2011 at 12:47 AM, Ian Lance Taylor<iant@google.com>
>> ?wrote:
>>>
>>> David Paterson<dnpaterson@gmail.com> ?writes:
>>>
>>>> Based on various examples of building cross systems around the net, I
>>>> have a procedure where the compiler gets built twice, once as a basic
>>>> version to allow building of Newlib, and then again as the full cross
>>>> compiler version. ?Looking at the esamples, the first, basic, build
>>>> always seems to use --without-headers (and --with-newlib). ?For the
>>>> second build however, some sites have both of these again, while
>>>> others have only --with-newlib, so I'm not sure whether I should
>>>> include this option in my second stage build or not...
>>>>
>>>> A search turns up a few references to this option, but none of them
>>>> say much beyond something along the lines of "leaving it out will
>>>> disable some features". ?Does anybody have any thoughts on the wisdom
>>>> or otherwise of using this option?
>>>
>>> --without-headers is the default, so it doesn't really matter whether
>>> you specify it or not.
>>>
>>> Ian
>>
>> Ah, I didn't think to check the defaults... ?Thanks for the quick
>> response Ian :-)
>
> Needing to build twice is of course also very weird suggestion :(
>
> Producing newlib-based GCCs for embedded systems from scratch has
> succeeded in only one stage during the last 20 or so years, I
> remember building my 1st one in early 1990's and then without any
> know-how I maybe had to do more than one "iteration" :(

Ah well, this is my first attempt, so I'm learning as I go :-)

I found quite a few descriptions and examples on the net, but they all
used a similar approach of building a basic "bootstrap" compiler, then
Newlib, then the full compiler.  My "configure" options choose only
the C compiler in the first stage, and then both C and C++ in the
second.  Could I build both in the first stage, or even add others
like Ada and Fortran?  I thought until Newlib was built I couldn't
create more than a basic C compiler...

> (snip) .... Or the whole 'newlib' and 'libgloss' sources in the
> main GCC source dir if wanting to build newlib at the same time
> with the GCC... Then the whole GCC, C & C++ compilers, libgcc,
> libstdc++ etc. will be built in only one build stage! Without or
> with newlib...

Ah - would I copy the Newlib sources into a specific location?  Or
just move the whole directory?

> So I guess the instructions being written by newbies who have just
> succeeded with their first crosstoolchain and want to share their
> "know-how" with others. Not by people who have built 100's of
> newlib-based crosstoolchains for some reasons, just-for-a-fun or
> for more serious ones...

There aren't actually a lot of examples or instructions around that I
could find, and all of the ones I found ran a full "configure, make,
install" sequence for both the bootstrap and final compiler builds.
I'm sure that, as you say, there a lot of people building cross
toolchains, for a lot of different systems, but the information is
quite hard to find.  I don't think I've seen one that does a
single-stage approach as you describe.

> The only thing hinting to a need to produce 'libstdc++' etc. after
> having newlib built and installed is that the generic newlib headers
> aren't the final ones. Here a diff made from the 'sparc-elf' case
> during the newlib-1.13.0 time :

> (snip)
Interesting - I didn't realise this was happening.  I'll spend a bit
more time investigating and experimenting to see exactly what goes on
during configuration and building of each of the components.

> During my last 'sparc-elf' target build (just checking), I only
> copied the generic newlib headers (with a single 'cp -r') and
> built GCC, libgcc, libstdc++ etc. with only them... Building and
> installing newlib would give the diff for the current situation
> and if then rebuilding 'libgcc', 'libstdc++' again would tell
> there being any difference to the "with bare generic headers"
> results, maybe then there would be some idea in that 2-stage
> process...
>
> So although the build with the generic newlib headers will work,
> there is that suspect: "It cannot be this easy, something must
> be wrong!". For the paranoids there are the steps :
>
> 1. build and install only GCC and libgcc
>
> 2. build and install newlib
>
> 3. continue the GCC build process via 'make ; make install'

Hmmm, now one of the reasons I've followed the two stage approach is
that I ran into a build problem doing something like this.  After a
bit of googling and experimenting, the solution appeared to be to
delete everything in the build directory before rebuilding the
compiler.  (I can't remember exactly what the error was, but I can try
to reproduce it if you're interested.)

> But also these steps will produce the GCC only once, continuing
> after a short coffee break for the paranoid's nerves...
>
> Of course if one puts the 'newlib' and 'libgloss' among the GCC
> sources, then everything should appear during the first build
> and be fully right!

OK - I may try this, but as I mentioned above, which directory should
I put the sources in?

Thanks for the help and info Kai - very interesting stuff.  My brain
may melt by the end of this :-)

And I'm still not sure exactly what "--without-headers" does, as I get
different sized gcc.exe's with and without it, but both seem to
work...

David