| 1 | * Validate that divrem in jit_x86-cpu.c is not modifying |
| 2 | the non result arguments. This is not verified by clobber.tst, |
| 3 | as it only checks registers not involved in the operation |
| 4 | (because it does not know about values being set as input |
| 5 | for the the operation). |
| 6 | |
| 7 | * Write a simple higher level language implementation generating |
| 8 | jit with lightning, that could be some lisp or C like language. |
| 9 | |
| 10 | * rerun ./configure --enable-devel-get-jit-size and regenerate |
| 11 | the related jit_$arch-sz.c for the ports where nodata is |
| 12 | meaningful: |
| 13 | hppa (done) |
| 14 | i586 (done) |
| 15 | ia64 |
| 16 | mips o32 (done) |
| 17 | mips n32 |
| 18 | mips n64 |
| 19 | powerpc 32 (done) |
| 20 | powerpc 64 (done) |
| 21 | ppc |
| 22 | s390x (done) |
| 23 | sparc (done) |
| 24 | x86_64 (done) |
| 25 | Missing ones are due to no longer (remote) access to such hosts |
| 26 | and may be broken with jit_set_data(..., JIT_DISABLE_DATA). |
| 27 | (ia64 hp-ux or linx), (irix mips for 32 or 64 abi), and |
| 28 | (darwin ppc). |