On Fri, 19 Mar 2010, Doug Kwan (Ãö®¶¼w) wrote: > Hi, > > I would like to get some advice on handling Android for x86 in > tool configurations. Android is based on Linux but it is exactly the > same, so some customizations are required. There was a discussion > among people working on Android and someone suggested using > i*86-unknown-android. My question is whether this is a good thing to > do. I would also like to hear from the maintainer of the i*86 > backend. Eventually we would like to see x86 Android modifications to > be pushed up-stream.
My inclination is that Android should use the same arrangements as the existing support I added for different C libraries under the Linux kernel. *-*-linux-gnu* means a system defaulting to using the GNU C library. *-*-linux-uclibc* means a system defaulting to using the uClibc library. So *-*-linux-android* would mean a system defaulting to using Android's C library. -mglibc selects a multilib based on the GNU C library in a toolchain defaulting to uClibc; -muclibc selects a multilib based on uClibc in a toolchain defaulting to the GNU C library. So -mandroid would be added to these options, and *-*-linux* configurations could support three C libraries. This means that configure tests in target libraries that care about which C library is in use cannot be based on the target triplet; they must check features or preprocessor macros (existing code in libstdc++-v3 checks __UCLIBC__, for example). All the triplets above could have existing suffixes, so i686-pc-linux-android, arm-none-linux-uclibceabi, powerpc-none-linux-androidspe (for example) would be possible combinations. The __linux__ preprocessor macro would be defined for all configurations using the Linux kernel. __gnu_linux__ should only be defined for those using the GNU C library (it's defined at present for those using uClibc as well, but I think that's a bug). -- Joseph S. Myers jos...@codesourcery.com
