vsk added inline comments.
================
Comment at: source/Plugins/Disassembler/llvm/DisassemblerLLVMC.h:81
+ struct Guard {
+ DisassemblerLLVMC *m_instance;
+ Guard(DisassemblerLLVMC *instance, InstructionLLVMC *inst,
----------------
This is nice. Do you think it might be even safer to have the guard own the
disassembler shared_ptr instance? Users could then access the disassembler via
the guard's operator->, and there's less chance of the shared_ptr escaping /
being abused. We could even have GetDisassembler() return a guard instance.
================
Comment at: source/Plugins/Disassembler/llvm/DisassemblerLLVMC.h:85
+ : m_instance(instance) {
+ m_instance->Lock(inst, exe_ctx);
+ }
----------------
It doesn't make sense to me that DisassemblerLLVMC's "Lock" method conflates
locking and initialization. If you decide to have a guard own the disassembler
instance, it'd make sense to split the initialization step out.
Repository:
rLLDB LLDB
https://reviews.llvm.org/D51319
_______________________________________________
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
