Marius Mauch wrote: >> Sure. Presumably you test packages with standard C-flags as users are >> advised to before bug-reporting? Other than USE flags what else would >> make your packages unsuitable for others? If it's only USE flags, >> then at least the pkg is a start- if others want different settings >> they can compile their own. > > The (well, at least one) problem is that you're only thinking about > individual packages. However to be of any real use you'd need all > packages to use the same system configuration. Otherwise you'll get ABI > breakages and other runtime errors.
Stuart mentioned the ABI problem as well: > The binary packages need to be built as a set, to be sure that there is no > ABI breakage going on. I understand the ABI changes at major compiler upgrades, especially for C++. Is this such a problem for C? I thought that was the whole point of the Linux ABI (so developers can in fact use the same binary for different distros.) I'm guessing you're going to point out all the posts about recompiling your whole system after a toolchain upgrade. So if I understand this right, we can't all compile for the same ABI since it changes according to which version of the C compiler/ glibc you're using. > Oh, and people using those binaries > would ahve to use the same system configuration as well (or at least a > very similar one). > This pretty much rules out devs submitting home-build binary packages > of ebuilds they maintain to a central repository. > Fair do. -- [email protected] mailing list
