Re: Any plans to integrate just binutils‑gdb into GCC?
Arsen Arsenović <[email protected]> Thu, 23 Jul 2026 12:14:41 +0200
| Newsgroups | gmane.comp.gcc.help,gmane.comp.gnu.binutils |
|---|---|
| Organization | BayLibre |
| Message-ID | <[email protected]> |
--=-=-= Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable unlvsur unlvsur via Gcc-help <[email protected]> writes: > The cross compiler approach (aka $target-gcc) ads another dimension which= is mathematically a mess. > > For examples if you have N platforms, to build every possible combination, > clang only has N combos (only needs the host). While gcc needs N^2 > (host*target). It just wastes computing resources to build each of them a= nd it > is a terrible mess. Just look at this. And this has been proven not only = in > theory but also in reality. Having a compiler capable of generating code for a target does not suffice to cross-compile to that target. Indeed, LLVMs build system contains logic for building the runtime libraries for other targets in much the same way as GCC does. It is a waste of time to build all N^2 (or even N^3) combinations. So don't do that. =2D-=20 Arsen Arsenovi=C4=87 --=-=-= Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- iQEKBAEWCgCyFiEE/uKz0RP8AKMWLWBhUsKUMB6ixJMFAmph6ZEbFIAAAAAABAAO bWFudTIsMi41KzEuMTIsMiwyXxSAAAAAAC4AKGlzc3Vlci1mcHJAbm90YXRpb25z Lm9wZW5wZ3AuZmlmdGhob3JzZW1hbi5uZXRGRUUyQjNEMTEzRkMwMEEzMTYyRDYw NjE1MkMyOTQzMDFFQTJDNDkzGBxhYXJzZW5vdmljQGJheWxpYnJlLmNvbQAKCRBS wpQwHqLEk8VVAQCjrD6dQxbjENzTBqUEMda+goVg7cqefP/mC4MDuQFArAD+IJOi F3d7W4Iuj/vxUPg2NfXU6eGMeuStxeA52TYMfwE= =FuGm -----END PGP SIGNATURE----- --=-=-=--