[
https://issues.apache.org/jira/browse/HADOOP-10163?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13846765#comment-13846765
]
Ted Yu commented on HADOOP-10163:
---------------------------------
Here is proposed change from HBASE-10044 :
{code}
+ relativePatchURL=`$GREP -o '"/jira/secure/attachment/[0-9]*/[^"]*'
$PATCH_DIR/jira | $EGREP '(\.txt$|\.patch$|\.diff$)' | sort | tail -1 | $GREP
-o '/jira/secure/attachment/[0-9]*/[^"]*'`
{code}
Suppose patch A was tested on day 1 - JIRA stayed in Patch Available status.
Some more comments were then added.
Tread dump named thread.out was attached.
QA bot would be triggered due to the new attachment. However the above change
would filter out the dump file and go back to patch A which has been run.
The attachment ID of last patch should be known so that test-patch.sh can
decide that no new patch has been attached and bail out.
> Attachment Id for last tested patch should be passed to test-patch.sh
> ---------------------------------------------------------------------
>
> Key: HADOOP-10163
> URL: https://issues.apache.org/jira/browse/HADOOP-10163
> Project: Hadoop Common
> Issue Type: Improvement
> Reporter: Ted Yu
>
> In HBASE-10044, attempt was made to filter attachments according to known
> file extensions.
> However, that change alone wouldn't work because when non-patch is attached,
> QA bot doesn't provide attachment Id for last tested patch.
> This results in the modified test-patch.sh to seek backward and launch
> duplicate test run for last tested patch.
> If attachment Id for last tested patch is provided, test-patch.sh can decide
> whether there is need to run test.
--
This message was sent by Atlassian JIRA
(v6.1.4#6159)
