Adar Dembo has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/12592 )

Change subject: KUDU-2744: Expose RPC interface for diff scan
......................................................................


Patch Set 6:

(5 comments)

http://gerrit.cloudera.org:8080/#/c/12592/6/src/kudu/tserver/tablet_server-test.cc
File src/kudu/tserver/tablet_server-test.cc:

http://gerrit.cloudera.org:8080/#/c/12592/6/src/kudu/tserver/tablet_server-test.cc@2364
PS6, Line 2364:   auto iter = keys.begin();
keys is just a vector so you could index into it with i.

Below too.


http://gerrit.cloudera.org:8080/#/c/12592/6/src/kudu/tserver/tablet_server-test.cc@2368
PS6, Line 2368:     CHECK_OK(row.SetInt32(0, key));
Doesn't this suggest that key, keys, i, new_val, and all the other int64_t 
types here should be int32_t? I imagined that you used int64_t to avoid 
overflow, but then aren't these SetInt32 calls vulnerable to overflow?

Below too.


http://gerrit.cloudera.org:8080/#/c/12592/6/src/kudu/tserver/tablet_server-test.cc@2455
PS6, Line 2455: // Send requests with timestamps past the ancient history mark. 
We should get
              : // reasonable failure messages.
              : TEST_F(TabletServerTest, TestDiffScanPastAHM) {
May want to rename the test (and update the comment) to account for the other 
negative tests going on here.


http://gerrit.cloudera.org:8080/#/c/12592/6/src/kudu/tserver/tablet_server-test.cc@2506
PS6, Line 2506:   NO_FATALS(req_assert_invalid_argument(
              :       "scan start timestamp is only supported in ORDERED order 
mode"));
TabletServerErrorPB::INVALID_SNAPSHOT seems like a strange case here.


http://gerrit.cloudera.org:8080/#/c/12592/6/src/kudu/tserver/tablet_service.cc
File src/kudu/tserver/tablet_service.cc:

http://gerrit.cloudera.org:8080/#/c/12592/6/src/kudu/tserver/tablet_service.cc@1973
PS6, Line 1973:     if (snap_start_timestamp->CompareTo(snap_end_timestamp) > 
0) {
Note that this comparison allows for the two timestamps to match, which the 
error on L1975 doesn't mention. FWIW I do think we need to let them match as 
that represents a range of a single "point".

So maybe just reword the error message to say that the start timestamp was not 
less than the end timestamp, or something to that effect?



--
To view, visit http://gerrit.cloudera.org:8080/12592
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I9602b549dc499f92b2069b297e66f366e51fb312
Gerrit-Change-Number: 12592
Gerrit-PatchSet: 6
Gerrit-Owner: Mike Percy <[email protected]>
Gerrit-Reviewer: Adar Dembo <[email protected]>
Gerrit-Reviewer: Grant Henke <[email protected]>
Gerrit-Reviewer: Greg Solovyev <[email protected]>
Gerrit-Reviewer: Kudu Jenkins (120)
Gerrit-Reviewer: Mike Percy <[email protected]>
Gerrit-Comment-Date: Wed, 20 Mar 2019 04:27:06 +0000
Gerrit-HasComments: Yes

Reply via email to