[
https://issues.apache.org/jira/browse/HBASE-10824?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13946518#comment-13946518
]
Anoop Sam John commented on HBASE-10824:
----------------------------------------
LGTM
Still if the protobuf generate a line of code in which "com.google.protobuf" or
"hbase.protobuf.generated" is not coming and line length >100, it might show us
wrong report. But normally this wont happen I guess.. With this patch we are
in better position. +1 for commit.
> Enhance detection of protobuf generated code in line length check
> -----------------------------------------------------------------
>
> Key: HBASE-10824
> URL: https://issues.apache.org/jira/browse/HBASE-10824
> Project: HBase
> Issue Type: Task
> Reporter: Ted Yu
> Assignee: Ted Yu
> Priority: Minor
> Attachments: 10824-v1.txt
>
>
> In review of HBASE-5175, Anoop found that long line detection wasn't very
> effective for protobuf generated code.
> Here is one example:
> {code}
> -1 lineLengths. The patch introduces the following lines longer than 100:
> + private DoubleMsg(boolean noInit)
> { this.unknownFields =
> com.google.protobuf.UnknownFieldSet.getDefaultInstance(); }
> {code}
> This is current filter:
> {code}
> lines=`cat $PATCH_DIR/patch | grep "^+" | grep -v "^@@" | grep -v "^+++" |
> grep -v "import" | grep -v "hbase.protobuf.generated" | awk -v
> len="$MAX_LINE_LENGTH_PATCH" 'length ($0) > len' | head -n 10`
> {code}
> 'com.google.protobuf.' should be considered as well.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
