Ian Maxon has posted comments on this change.

Change subject: ASTERIXDB-1479: Change storage valid int and add explicit 
version
......................................................................


Patch Set 8:

(5 comments)

https://asterix-gerrit.ics.uci.edu/#/c/919/8/hyracks-fullstack/hyracks/hyracks-storage-am-btree/src/main/java/org/apache/hyracks/storage/am/btree/frames/BTreeFieldPrefixNSMLeafFrame.java
File 
hyracks-fullstack/hyracks/hyracks-storage-am-btree/src/main/java/org/apache/hyracks/storage/am/btree/frames/BTreeFieldPrefixNSMLeafFrame.java:

Line 464:         strBuilder.append("TUPLE_COUNT_OFF:             " + 
tupleCountOff + "\n");
> Should we revert this?  Not sure why we want to change the name of the disp
Yeah I can revert this, it was just the IntelliJ refactor plugin being a bit 
overzealous.


https://asterix-gerrit.ics.uci.edu/#/c/919/8/hyracks-fullstack/hyracks/hyracks-storage-am-common/src/main/java/org/apache/hyracks/storage/am/common/freepage/LinkedMetaDataPageManager.java
File 
hyracks-fullstack/hyracks/hyracks-storage-am-common/src/main/java/org/apache/hyracks/storage/am/common/freepage/LinkedMetaDataPageManager.java:

Line 475:             return (long)(metadataPageNum * 
bufferCache.getPageSize()) + LIFOMetaDataFrame.LSN_OFF;
> Should one of the operands to the multiply be casted to a long rather than 
Well it's fixed as far as sonarQube is concerned :P it did complain at first. 
It only complained because I changed the name here, so it's been a bug for a 
while apparently. Doesn't the output of metadataPageNum*getPageSize() being 
casted to a long guarantee the result of the subsequent addition is also a long?


https://asterix-gerrit.ics.uci.edu/#/c/919/8/hyracks-fullstack/hyracks/hyracks-storage-am-lsm-common/src/main/java/org/apache/hyracks/storage/am/lsm/common/impls/AbstractLSMIndexFileManager.java
File 
hyracks-fullstack/hyracks/hyracks-storage-am-lsm-common/src/main/java/org/apache/hyracks/storage/am/lsm/common/impls/AbstractLSMIndexFileManager.java:

Line 53:     public enum TREE_INDEX_STATE {
> enums should be named like classes (e.g. TreeIndexState)- not sure why this
Done


Line 55:         WRONG_VERS,
> Not loving this abbreviation- can we have WRONG_VERSION or perhaps VERSION_
Done


Line 105:                     return TREE_INDEX_STATE.WRONG_VERS;
> We may have to revisit this in the future- we're going to need to know whic
The index metadata can still be opened though to read what version it is. Right 
now the error/version mismatch behavior is actually that the LSM harness will 
not open the index because there are a mismatched # of bloom filter and  index 
files (because the _b part of the component won't open due to this code).


-- 
To view, visit https://asterix-gerrit.ics.uci.edu/919
To unsubscribe, visit https://asterix-gerrit.ics.uci.edu/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I888ff0eacf5b3cb6ad7ec002c74f113c6ffcd496
Gerrit-PatchSet: 8
Gerrit-Project: asterixdb
Gerrit-Branch: master
Gerrit-Owner: Ian Maxon <[email protected]>
Gerrit-Reviewer: Ian Maxon <[email protected]>
Gerrit-Reviewer: Jenkins <[email protected]>
Gerrit-Reviewer: Michael Blow <[email protected]>
Gerrit-Reviewer: Till Westmann <[email protected]>
Gerrit-Reviewer: abdullah alamoudi <[email protected]>
Gerrit-HasComments: Yes

Reply via email to