Implement io.buffer for fast IO (work in progress)
Rewrite util/cache.c and use it form io/buffer.c
Refactor util/ht.c and util/ht64.c to reuse code
Various minor fixes in RIO api
cache.c must be rewritten into r_buf for
- r2 -d gdb://<host>:<port>
- register maps not yet implemented
- some basic step/continue should work in cfg.debug=true
- gdbwrap instance is shared between RIO and RDebug
* RDebug is now arch-sensitive
- Plugins describe which architectures and register sizes are supported
- Native debugger is restricted to local CPU
- Remote debugger (GDB) arch can be specified with -e asm.arch=arm
* Fix some random warnings
* Move R_ASM_ARCH into R_SYS_ARCH
- Helper functions to translate id to string and string to id
are now in util/sys.c (r_util)
- Move all R_SYS_* from r_util to r_types
- Endianness, OS, CPU and regsize is now 'global'
- Kinda smart (150LOC)
- Fork it for 64bit hash keys (make ht64 in libr/util)
* Make RConfig use the RHashTable to resolve by name
- Code cleanup resulting in -30LOC
- O(1) access to config variables (speedup!)
- Make r_list_free and r_list_destroy take sense
