Re: [MinGW-dvlpr] GCC configure flags for 4.6.0

[Charles Wilson <cwilso11@us...>]

On 6/8/2011 8:00 PM, John E. / TDM wrote:
> The rest of this patch does address the "-l" problem. :)

Good to know.  I only gave the patch a quick perusal, so I obviously
missed something.

> I should note that using this patch requires the builder of GCC to be
> particularly careful to stage all GCC prerequisites into the --prefix
> directory, or otherwise make them available during the bootstrap.

Well, sure, but that's the case for --with-sysroot as well: you have to
ensure that all the dependencies for the target runtime libs are
installed into that directory (or into the one specified via
--with-build-sysroot).

Compiling a compiler always requires special care. :-)

--
Chuck

Re: [MinGW-dvlpr] GCC configure flags for 4.6.0

[John E. / TDM <tdragon@td...>]

On 6/8/2011 8:38 AM, Charles Wilson wrote:
> Ah, so that's what you meant: you *have* to use sys-root to avoid the
> problem.  I'm not sure that's the best solution...I'm not sure why, but
> /something/ about that bothers me.  It seems like taking a feature
> developed and intended for cross-compilers, and using it with a native
> compiler to take advantage of a side effect, is kinda kludgy.

Also, a sysroot-ed mingw32 GCC lacks one or two search paths that the 
current build uses -- <root>/include among them, IIRC.

> The core part of TDM's relocation patch is the following:
> *snip*
> That is, in this particular function, it simply omits the sysroot
> requirement before allowing recomputation of paths if the binary has
> been relocated.
>
> *snip*
>
> 'Course, it doesn't appear that this patch addresses the -l problem,
> only the -I problem, so perhaps going the full-sysroot-monty IS better.

The rest of this patch does address the "-l" problem. :)

I should note that using this patch requires the builder of GCC to be 
particularly careful to stage all GCC prerequisites into the --prefix 
directory, or otherwise make them available during the bootstrap.

-John E. / TDM