* several bugfixes
* Fix xrefs/refs confusion with asm.cmt.refs
* Also check if xref source is valid offset
* Fix ax* to show type of ref
* anal/xrefs: pass the right hashtable to mylistrefs
* r_anal_refs_get should use anal->refs, not anal->xrefs
* anal/fcn: add xref also when there's a CALL to a noreturn function
* projects: use radare2 output of ax command to save/restore xrefs
* core/cmd_anal: when analyzing calls, we should use REF_TYPE_CALL
* core/canal: use xrefs API, because the list is a just a copy
* core/canal: rename "loc." entries to "fcn." when a CALL is found
* sort xrefs to make results consistent
* core/canal: avoid recomputing function every time during `aan`
* anal: move fcn_refs/xrefs functions to the xrefs.c file
* core/canal.c: avoid iterating fcn xrefs list
* anal/xrefs: compute fcn xrefs by analyzing fcn addresses
* remove other unused functions
* anal/xrefs: remove fcn_xrefs_add and fcn_xrefs_deln
* anal/xrefs: remove old fcn->refs/xrefs
* anal/xrefs: directly store RAnalRef objects in the hash table
* libr: prevent memory leaks when using refs/xrefs
* anal/xrefs: merge anal/ref and anal/xref and clean API
Big xrefs/refs refactoring that provides a more uniform and simple API.
It avoids changes to refs/xrefs except through the API and it keeps all
xrefs/refs info in one single place, to improve consistency.
Thanks to:
Riccardo Schirone <sirmy15@gmail.com>
pancake <pancake@nopcode.org>
rene <rlaemmert@gmail.com>
* core: Make segment granularity adjustable
Segments in x86 real mode have 16 byte granularity(4 address bits).
There are however systems (namely the Vautomation/ARC Turbo 186 cores)
that use 256 byte(8 address bits) instead to allow access to 16MiB memory.
Unfortunately fixed granularity is hard coded in lots of places in r2.
This patch adds asm.seggrn configuration and fixes the hardcoded
assumptions in core.
Signed-off-by: Dirk Eibach <eibach@gdsys.cc>
* anal_x86_cs: Make segment granularity adjustable
Fix hard coded segment granularity assumptions in capstone
anal module.
Signed-off-by: Dirk Eibach <dirk.eibach@gdsys.cc>
* Added initial code for file completion - also completes from CWD :(
* Changed to use near-same code as in the 'Pl' command, still shows current directory files??
* Added comparison to select only the similar results. However, they are weirdly displayed.. UTF?
* Previously I pointed tmp_argv to foo, which was beeing freed.. I pointed it to a new string (r_str_newf). Now it works!
* Added check to make sure projectspath is not null.
