Mail Archives: djgpp/2002/11/07/13:45:10
Charles Wilkins <chas AT pcscs DOT com> wrote:
>>You just
>>have to make sure that some implementation exists, and is found by the
>>compiler / linker.
> It is not that simple when we are talking about a cross compiler that
> creates a binary for a target different than the host. Is it?
It is, as long as you can find some source for the necessary runtime
support files. Only if you can't (e.g. if your cross-compiler is the
only compiler you have for that target system), you really have to
bite that bullet. And it can be quite a bullet to bite the first few
times round. I once did a cross-build with a completely alien
platform as the host (DEC 64bit Alpha running Digital Unix), and DJGPP
as the target. It's not an exercise I'd redo from scratch just for
the fun of it...
> The archive libstdc++.a which is created by a typical native install
> of gcc-3.2 is _not_ what the cross compiler / linker seeks.
I was speaking of the *DJGPP* native build. I.e. the *.a files you
find in gpp32b.zip on your friendly DJGPP ftp site.
> It does matter where the library comes from and how it is created.
Not really. It only matters that wherever it came from, it was built
correctly, for the right target, and not damaged on its way from
"there" to "here.
> One cannot simply take libstdc++.a and plop it into place for a cross
> compiler that creates executables for a different target. Can they?
You can, if you take it right *from* that target.
--
Hans-Bernhard Broeker (broeker AT physik DOT rwth-aachen DOT de)
Even if all the snow were burnt, ashes would remain.
- Raw text -