[email protected] has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/18098 )

Change subject: KUDU-3197 [tserver] optimal Schema's memory used
......................................................................


Patch Set 5:

(2 comments)

http://gerrit.cloudera.org:8080/#/c/18098/4//COMMIT_MSG
Commit Message:

PS4:
> Started looking briefly into this, but it seems like a lot of the test-only
Yes, I agree with you. And that's my thought when start the work. But changes 
out of control.

If we use std::shared_ptr<Schema>, the changes will be less than 
scoped_refptr<Schema>.
And Kudu recommends scoped_refptr<Schema>.
Once we choose Schema extends RefCountedThreadSafe, the commits are the result,
the changes are accumulated little by little.

Another way is wangningito's way, that's a simple method but not strict, maybe 
small proability core.
wangningito's commit at: https://gerrit.cloudera.org/c/16508/


http://gerrit.cloudera.org:8080/#/c/18098/4/src/kudu/client/schema.h
File src/kudu/client/schema.h:

http://gerrit.cloudera.org:8080/#/c/18098/4/src/kudu/client/schema.h@732
PS4, Line 732:  /// Convert a KuduSchema to a Schema.
             :   ///
             :   /// Private API.
             :   ///
             :   /// @param[in] kudu_schema
             :   ///   The KuduSchema to convert
             :   /// @return The converted Schema
             :   static scoped_refptr<Schema> ToSchema(const KuduSchema& 
kudu_schema) KUDU_NO_EXPORT;
> This change doesn't seem necessary to address KUDU-3197
Yes.

But Schema extends RefCountedThreadSafe, then the Schema's usage is changed. I 
have to change it.

Schema s;    is not valid because deconstruction will check fail.
So Schema should use it like this  scoped_refptr<Schema> s(new Schema),  or use 
Schema&.

The functions which return Schema would cause the problem when deconstruction.

A few functions below has changed like this.



--
To view, visit http://gerrit.cloudera.org:8080/18098
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I4c9cf24475de0b787542761522915a86db6564d7
Gerrit-Change-Number: 18098
Gerrit-PatchSet: 5
Gerrit-Owner: Anonymous Coward <[email protected]>
Gerrit-Reviewer: Andrew Wong <[email protected]>
Gerrit-Reviewer: Anonymous Coward <[email protected]>
Gerrit-Reviewer: Kudu Jenkins (120)
Gerrit-Reviewer: Yingchun Lai <[email protected]>
Gerrit-Comment-Date: Wed, 22 Dec 2021 11:58:32 +0000
Gerrit-HasComments: Yes

Reply via email to