Hello, all,
I have compared different books and online sources all creating a cross-compiler (see title). I think all 3's method to generate a cross-compiler are very much OK. However, trying to understand everything in all of the 3 methods, makes me wonder...
Where Yaghmour as well as Dan Kegel are patching (/hacking ?) a cross-compiled libc.so (Kegel also adds libpthread.so and libgcc_s.so because he's more recent ?), I don't see that happen in Hollabaugh's way of working. Because Hollabaugh compiles glibc twice, once with a bootstrap gcc (and a prefix on the host's directory structure), and later a second time, but then with the 'full-blown' crosscompiled gcc (and a prefix of /), installing with a custom install_root.
This last (i.e. Hollabaugh's) method seems the egg of Colombus to me, as it is simple and straightforward, not needing any libhack or other hack. But I guess there must be a reason why (Yaghmour and) Kegel don't use it ? Why would that be ?
The only reason I could think of, is that it's too time-consuming compiling twice the same thing ? Or are there on the contrary maybe other reasons for opting to hack some .so files ?
Kind regards,
PhB.