[RFC] Full float128, third iteration

[FX]

Dear all,

Please find the latest (third?) iteration of the float128 support patch. To build:

  * Get libquad from here: http://quatramaran.ens.fr/~coudert/tmp/libquad.tar.bz2
  * Compile it (with a recent GCC, maybe a previous build of trunk!): ./configure --prefix=/foo && make && make install
  * Patch GCC, configure it with --with-quad=/foo and build

Problems solved:
  * libquad should build on cygwin (and targets with sqrtl/cbrtl); the sqrtq/cbrtq implementations are more robust now
  * the output precision issue (lots of zeros showing up) is fixed

Existing issue:
  * array intrinsics; that's next on my to-do list

Things that need to be tested:
  * I/O, again and again and again :)
  * what happens with flags like -fdefault-real-8, i.e. do we follow the documentation correctly


Thanks to everyone who tested the two previous patches on their systems, I hope we're now close to the final integration (at least from the technical point of view!).

FX



-------------- next part --------------
A non-text attachment was scrubbed...
Name: full_testing_3.diff
Type: application/octet-stream
Size: 30282 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/fortran/attachments/20100909/f461e694/attachment.obj>