This week I merged cs2.1 into r2, so we'll be shipping it by default in the next release and now we are going to perform lot of testing to ensure we are not introducing any regression by a bug in capstone or a missuse of it.
The first stage is ready for testing and our build farm raised few issues. - some c99 code (patch send and merged) - android mips builds failed (workarounded), looks like 'mips' is defined to 1 somewhere which causes cs build fail. I had to CFLAGS=-Dmips=mips - coverity scan passed reporting some issues in cs code Actually we are taking capstone from latest github tarball and then applying the c99 patch. It just builds the .a archive. Also, i have noticed that the windows library archive file name is .dll.a, so i was not able to use the same makrfile for w32 and linux builds... Wondering if .dll.a is correct, so i just passed AR_EXT=a to the make arguments. The second stage of the capstone integration consists in checking all the regression testsuite (about 800 tests) to verify that the output of the disassembler is correct in all the corner cases we found during r2 development. The third stage consists in swapping all the capstone plugins as default and remove all the gnu plugins from the main repo. The main reason to switch to capstone for us are: - well maintained disassembler library - no license issues with LGPL - complete instruction introspection We'll keep you informed about how the second and third stages are going, join the irc (#radare at freenode) for more information. --pancake
_______________________________________________ radare mailing list [email protected] http://lists.nopcode.org/listinfo.cgi/radare-nopcode.org
