Dan Hecht has posted comments on this change. Change subject: IMPALA-3168: replace HashTable parameters with constants ......................................................................
Patch Set 3: (4 comments) http://gerrit.cloudera.org:8080/#/c/3088/3/be/src/codegen/llvm-codegen.h File be/src/codegen/llvm-codegen.h: Line 267: const std::string& target_name); should we make this private and just ensure we have the interfaces needed to do the translation for C types to IR? That seems a bit easier for the callsites to get the types write (just make sure C types for prototype and constant match). http://gerrit.cloudera.org:8080/#/c/3088/3/be/src/exec/hash-table.cc File be/src/exec/hash-table.cc: Line 91: num_build_tuples_(num_build_tuples), rather than saving this, which we only need for codegen, I think it's slightly better to just pass it to codegen like you already do with stores_duplicates. less redundant state in our objects is better. http://gerrit.cloudera.org:8080/#/c/3088/3/be/src/exec/hash-table.h File be/src/exec/hash-table.h: Line 707: friend class PartitionedAggregationNode; why is this needed? http://gerrit.cloudera.org:8080/#/c/3088/3/be/src/exec/hash-table.inline.h File be/src/exec/hash-table.inline.h: Line 185: } double check that a caller wasn't relying on this function always setting *node. -- To view, visit http://gerrit.cloudera.org:8080/3088 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I483a19662c90ca54bc21d60fd6ba97dbed93eaef Gerrit-PatchSet: 3 Gerrit-Project: Impala Gerrit-Branch: cdh5-trunk Gerrit-Owner: Tim Armstrong <[email protected]> Gerrit-Reviewer: Dan Hecht <[email protected]> Gerrit-Reviewer: Michael Ho Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-HasComments: Yes
