[Lldb-commits] [PATCH] Fix handling of CommandInterpreter's events in lldb-mi

[Ilia K]

================
Comment at: source/API/SBCommandInterpreter.cpp:554
@@ +553,3 @@
+{
+    return strcmp (event.GetBroadcasterClass(), SBCommandInterpreter::GetBroadcasterClass()) == 0;
+}
----------------
clayborg wrote:
> ki.stfu wrote:
> > clayborg wrote:
> > > The strings are returns as the "const char *" from a lldb_private::ConstString object, so just do a pointer compare instead of a strcmp.
> > I think it's wrong.
> > ```
> > file1.cpp:
> > 
> > const char s1 [] = "abc";
> > ```
> > 
> > ```
> > file2.cpp:
> > 
> > extern const char s1 [];
> > const char *s2 = "abc";
> > assert (s1==s2); // error
> > assert (!strcmp (s1,s2)); // ok
> > ```
> > 
> > I'd prefer to keep it as is.
> What I am saying is the "const char *" that comes back from both "event.GetBroadcasterClass()" is a "const char *" that comes from a ConstString object. Same thing with the "const char *" that is returned from "SBCommandInterpreter::GetBroadcasterClass()". So of course "file1.cpp" and "file2.cpp" will fail, that is obvious. So your example will work with ConstString objects:
> 
> 
> ```
> file1.cpp:
> ConstString s1("abc");
> 
> ```
> 
> 
> ```
> file2.cpp:
> extern ConstString s1;
> ConstString s2("abc");
> assert(s1.GetCString() == s2.GetCString());
> assert(strcmp(s1.GetCString(), s2.GetCString()) == 0);
> 
> ```
Ok. Thanks for explanation. I'll do it here and for SBTarget/SBThread/SBProcess in another commit.

http://reviews.llvm.org/D8382

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/