diff options
author | Joshua Haberman <joshua@reverberate.org> | 2011-02-14 10:24:33 -0800 |
---|---|---|
committer | Joshua Haberman <joshua@reverberate.org> | 2011-02-14 10:24:33 -0800 |
commit | 4f9aeee6c7414bec2e1d8a9d3508dbd21f4d3394 (patch) | |
tree | 9c612f70ba57a593b7885f8e7478cb0399d2c37a /src/upb_def.c | |
parent | 6117730c85e5d64239337f0e8514109054202f5a (diff) |
More completely fixed the 0-key thing.
Unfortunately this degrades hash table lookup performance by
about 8%, which affects the streaming benchmark for googlemessage1
by about 5%. We could get this back at the cost of some memory,
but it would be nice to avoid that.
Diffstat (limited to 'src/upb_def.c')
-rw-r--r-- | src/upb_def.c | 6 |
1 files changed, 0 insertions, 6 deletions
diff --git a/src/upb_def.c b/src/upb_def.c index 879f1e8..7e962c8 100644 --- a/src/upb_def.c +++ b/src/upb_def.c @@ -476,12 +476,6 @@ static upb_flow_t upb_enumdef_EnumValueDescriptorProto_value(void *_b, break; case GOOGLE_PROTOBUF_ENUMVALUEDESCRIPTORPROTO_NUMBER_FIELDNUM: b->number = upb_value_getint32(val); - if (b->number == 0) { - upb_seterr(&b->status, UPB_ERROR, - "Enums with a value of 0 are currently broken; see " - "http://code.google.com/p/upb/issues/detail?id=1."); - return UPB_BREAK; - } b->saw_number = true; break; default: |