labath added inline comments.
================ Comment at: include/lldb/Core/RichManglingInfo.h:36 + /// "a::b". + llvm::StringRef GetFunctionDeclContextName() const; + ---------------- sgraenitz wrote: > I realized that returning `llvm::StringRef` from here may not be the best > idea. To the outside it appears to be an immutable string, but internally it > will be reused and even realloc'ed. I guess returning a `ConstString` is not > acceptable as an alternative performance-wise, as it adds one `memcpy` per > query. I will sketch something here, but it's not easy as it needs to fit > both, the IPD and C++ MethodName. That is a bit unfortunate, but I don't think things would be any better if you left this as `const char *`. AFAICT, the caller converts this to a ConstString anyway, so returning that type might not be such a bad choice, but I would be fine with just documenting that the returned string is valid until the next `GetXXX` call. ================ Comment at: source/Core/RichManglingInfo.cpp:93 + case PluginCxxLanguage: + return get<CPlusPlusLanguage::MethodName>()->GetContext().data(); + } ---------------- These `.data()` calls can now be removed. ================ Comment at: source/Symbol/Symtab.cpp:397 + // The base name will be our entry's name. + entry.cstring = ConstString(base_name.data()); + ---------------- remove `.data()` https://reviews.llvm.org/D50071 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
