[lldb] Tweak check for CommandLineTools in ParseXcodeSDK

[kastiglione]

Follow up to #128712

[llvmbot]

@llvm/pr-subscribers-lldb

Author: Dave Lee (kastiglione)

Changes

Follow up to #128712

---

Full diff: https://github.com/llvm/llvm-project/pull/154574.diff

1 Files Affected:

• (modified) lldb/source/Plugins/SymbolFile/DWARF/SymbolFileDWARF.cpp (+3-2)

diff --git a/lldb/source/Plugins/SymbolFile/DWARF/SymbolFileDWARF.cpp b/lldb/source/Plugins/SymbolFile/DWARF/SymbolFileDWARF.cpp
index 9958af26379b9..163cc19943683 100644
--- a/lldb/source/Plugins/SymbolFile/DWARF/SymbolFileDWARF.cpp
+++ b/lldb/source/Plugins/SymbolFile/DWARF/SymbolFileDWARF.cpp
@@ -8,6 +8,7 @@
 
 #include "SymbolFileDWARF.h"
 #include "llvm/ADT/STLExtras.h"
+#include "llvm/ADT/StringRef.h"
 #include "llvm/DebugInfo/DWARF/DWARFAddressRange.h"
 #include "llvm/DebugInfo/DWARF/DWARFDebugLoc.h"
 #include "llvm/Support/Casting.h"
@@ -998,12 +999,12 @@ XcodeSDK SymbolFileDWARF::ParseXcodeSDK(CompileUnit &comp_unit) {
   const char *sdk = cu_die.GetAttributeValueAsString(DW_AT_APPLE_sdk, nullptr);
   if (!sdk)
     return {};
-  std::string sysroot =
+  llvm::StringRef sysroot =
       cu_die.GetAttributeValueAsString(DW_AT_LLVM_sysroot, "");
 
   // RegisterXcodeSDK calls into xcrun which is not aware of CLT, which is
   // expensive.
-  if (sysroot.find("/Library/Developer/CommandLineTools/SDKs") != 0) {
+  if (!sysroot.starts_with("/Library/Developer/CommandLineTools/SDKs")) {
     // Register the sysroot path remapping with the module belonging to
     // the CU as well as the one belonging to the symbol file. The two
     // would be different if this is an OSO object and module is the

[bulbazord]

Nice :)

[Michael137]

Thanks!

[Michael137]

Now that this is a StringRef we don't need the move down on line 1021. Though I'd double-check that the lifetimes make sense. I'm speculating because i dont have the source in front of me, but suspect I used a std::string+move here just because FileSpec takes a std::string in its constructor? Either way, it doesn't buy us much so the readability win here is worth it

[kastiglione]

thanks for pointing out the move.

double-check that the lifetimes make sense

It appears fine, and the tests pass. Is that enough or should we test this more in some way?

[Michael137]

Just checked and FileSpec takes a StringRef anyway, so we're good