[WIP][DBNode] - Optimize Tag Encoder #1866
Conversation
| @@ -156,42 +195,38 @@ func (e *encoder) Finalize() { | |||
| p.Put(e) | |||
| } | |||
|
|
|||
| func (e *encoder) encodeTag(t ident.Tag) error { | |||
| if len(t.Name.Bytes()) == 0 { | |||
| func (e *encoder) encodeTag(name, value []byte) error { | |||
There was a problem hiding this comment.
Could also inline encodeTag I guess?
There was a problem hiding this comment.
Yeah I tried, didn’t get much benefit as far as I could tell
| tags := srcTags.Duplicate() | ||
| defer tags.Close() | ||
| if sliceIter, ok := tagsIter.(interface { | ||
| SliceIter() []ident.Tag |
There was a problem hiding this comment.
Does the commit log write path have a SliceIter that backs it? Or Is it only the tests that benefit from this speed up?
There was a problem hiding this comment.
Yeah it does I think unless I’m misunderstanding something: https://github.com/m3db/m3/blob/master/src/dbnode/persist/fs/commitlog/writer.go#L137
There was a problem hiding this comment.
I see, should we perhaps just make this another method, i.e. EncodeSlice(...) or similar?
There was a problem hiding this comment.
I think the interface assertion is the cleanest personally (the assertion is very cheap since it only needs to check if one method exists) but we can add a new method if you feel strongly
|
Closing due to the change we merged instead #1898 |
No description provided.