[Lldb-commits] [PATCH] lldb handle breakpoints on ARM64
Todd Fiala
tfiala at google.com
Tue Aug 19 10:16:15 PDT 2014
Minor style nit - parens around sizeof arg to match rest of code.
When you address that, if you wanted to fix the Error idiomatic usage I pointed out (which is due to me - you're just following my non-idiomatic usage of Error), feel free to clean that up in those two routines. If not, I'll catch up with it later in another pass.
Other than that, LGTM.
Tested:
Ubuntu 14.04 x86_64, llvm-3.5-built lldb, all tests passed except for unrelated TestCallStopAndContinue.py, which fails intermittently on my system.
================
Comment at: source/Plugins/Platform/Linux/PlatformLinux.cpp:429
@@ +428,3 @@
+ trap_opcode = g_aarch64_opcode;
+ trap_opcode_size = sizeof g_aarch64_opcode;
+ }
----------------
style: use parens around sizeof's argument to match other code.
================
Comment at: source/Plugins/Process/Linux/NativeProcessLinux.cpp:2893
@@ +2892,3 @@
+ actual_opcode_size = static_cast<uint32_t> (sizeof g_aarch64_opcode);
+ return Error ();
+
----------------
Add parens around sizeof arg (i.e. sizeof (g_aarch64_opcode)).
Idiomatic usage: (this is my fault): we generally define an Error at the top of the method, then set it and return the local error value rather than create a new one.
If you want to put an Error object at the top:
```
Error error;
...
return error;
```
and replace the other Error usages in the function, that would be fine. Otherwise, I'll do a clean sweep later and fix that.
================
Comment at: source/Plugins/Process/Linux/NativeProcessLinux.cpp:2927
@@ +2926,3 @@
+ trap_opcode_bytes = g_aarch64_opcode;
+ actual_opcode_size = sizeof g_aarch64_opcode;
+ return Error ();
----------------
parens around sizeof arg.
================
Comment at: source/Plugins/Process/Linux/NativeProcessLinux.cpp:2928
@@ +2927,3 @@
+ actual_opcode_size = sizeof g_aarch64_opcode;
+ return Error ();
+
----------------
Idiomatic usage. (again my original sin). You don't have to clean this up, I'll make a pass on it at another time if not.
http://reviews.llvm.org/D4969
More information about the lldb-commits
mailing list