Dear all, The attached file is the rebased patch to top of tree. It seems that the dynamic linker in Android only supports .init_array and .fini_array, thus we should enable -fuse-init-array regardless the GCC version. It is essential for C++ global variables initialization on Android X86 and MIPS. Please have a look. Thanks.
Sincerely, Logan On Fri, Aug 31, 2012 at 12:44 PM, Logan Chien <[email protected]>wrote: > Dear all, > > This is the revised patch which renamed "androideabi" to "android" > to avoid confusion. Please have a look. Thanks. > > Sincerely, > Logan > > p.s. Here's the link to the patch: > > https://github.com/loganchien/clang-logan/commit/b16a2b09fc54f21e326c81c7a7e4f61360722dbf > > On Thu, Aug 30, 2012 at 3:15 PM, Logan Chien <[email protected]>wrote: > >> Hi Tim, >> >> Thanks for your reply. androideabi is not for MIPS or X86. >> I have an upcoming patch to rename all of the usage. In fact >> some friend from MIPS suggested me not to use the "EABI" >> to avoid confusion. Thanks. >> >> Logan >> >> >> On Thu, Aug 30, 2012 at 3:05 PM, Tim Northover >> <[email protected]>wrote: >> >>> On Thu, Aug 30, 2012 at 7:38 AM, Logan Chien <[email protected]> >>> wrote: >>> > Sorry, I have attached the wrong patch. This is the correct one. >>> Please >>> > have a look. Thanks. >>> >>> I'm not involved in android development, but is "androideabi" really >>> used on x86 and Mips? It looks like the name involves "EABI" which is >>> a specific ARM ABI. Your test seems to be the first to apply it to >>> non-ARM targets, which may be a sensible thing to do (everyone and his >>> brother uses the ia-64 ABI these days) but seems a little odd. >>> >>> Anyone else know more about Android convention? >>> >>> Tim. >>> >> >> >
android-init-fini-array.patch
Description: Binary data
_______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
