Tim Armstrong has posted comments on this change. ( http://gerrit.cloudera.org:8080/8233 )
Change subject: IMPALA-6002: Add a LLVM diagnostic handler for LLVM linker errors ...................................................................... Patch Set 3: (4 comments) http://gerrit.cloudera.org:8080/#/c/8233/3/be/src/codegen/llvm-codegen.cc File be/src/codegen/llvm-codegen.cc: http://gerrit.cloudera.org:8080/#/c/8233/3/be/src/codegen/llvm-codegen.cc@1628 PS3, Line 1628: LOG(INFO) << diagnostic_handler->error_str_; Include the query id for context, which should be accessible via state_? Will be hard to track down where the log message is coming from otherwise. http://gerrit.cloudera.org:8080/#/c/8233/3/be/src/codegen/llvm-codegen.cc@1634 PS3, Line 1634: (std:: Should be able to drop the std:: - we import it into the namespace in common/names.h (there is other code in the codebase that is inconsistent about this). http://gerrit.cloudera.org:8080/#/c/8233/3/be/src/codegen/llvm-codegen.cc@1635 PS3, Line 1635: error_str_.clear(); > I might be confused about the C++11 stuff, but given that you just std::mov Yup http://gerrit.cloudera.org:8080/#/c/8233/1/tests/query_test/test_codegen.py File tests/query_test/test_codegen.py: http://gerrit.cloudera.org:8080/#/c/8233/1/tests/query_test/test_codegen.py@45 PS1, Line 45: def test_codegen_diagnostic_handler(self, vector): > I don't know LLVM well enough to suggest how to get at these errors. Yeah I think if you directly create a function or global variable with a conflicting name in the module you could induce an error that way. -- To view, visit http://gerrit.cloudera.org:8080/8233 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I907ff1f8975c02d422b4a669afbfc2e536ddf1ff Gerrit-Change-Number: 8233 Gerrit-PatchSet: 3 Gerrit-Owner: Bikramjeet Vig <[email protected]> Gerrit-Reviewer: Bikramjeet Vig <[email protected]> Gerrit-Reviewer: Philip Zeyliger <[email protected]> Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-Reviewer: anujphadke <[email protected]> Gerrit-Comment-Date: Wed, 11 Oct 2017 21:51:13 +0000 Gerrit-HasComments: Yes
