----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/61715/#review183939 -----------------------------------------------------------
ambari-common/src/main/python/resource_management/libraries/script/script.py Line 455 (original), 455 (patched) <https://reviews.apache.org/r/61715/#comment259985> it looks that this function is cross-linux version, could we move it to https://github.com/apache/ambari/tree/trunk/ambari-common/src/main/python/resource_management/core/providers/package to respective os provider? - Dmytro Grinenko On Aug. 28, 2017, 3:49 p.m., Dmitro Lisnichenko wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/61715/ > ----------------------------------------------------------- > > (Updated Aug. 28, 2017, 3:49 p.m.) > > > Review request for Ambari, Dmytro Grinenko and Vitalyi Brodetskyi. > > > Bugs: AMBARI-21744 > https://issues.apache.org/jira/browse/AMBARI-21744 > > > Repository: ambari > > > Description > ------- > > Due to the issue with regex (missing ^ and $ boundaries), > resource_management.libraries.script.script.Script#get_package_from_available > may return wrong package. > {code} > >>> list=['hbase_3_0_0_0_229-master', 'hbase_3_0_0_0_229'] > >>> if re.match('hbase_(\d|_)+', 'hbase_3_0_0_0_229-master'): > ... print 'YES' > ... > YES > >>> if re.match('hbase_(\d|_)+', 'hbase_3_0_0_0_229'): > ... print 'YES' > ... > YES > {code} > > In this case, the first package name from a list of available packages will > be returned. > The impact of bug is that we may install a wrong package if it's simillary > named and goes first at list. Patch is a single-line fix. > > > Diffs > ----- > > > ambari-common/src/main/python/resource_management/libraries/script/script.py > 313b666878 > > > Diff: https://reviews.apache.org/r/61715/diff/2/ > > > Testing > ------- > > Will fix unit tests a bit later > > > Thanks, > > Dmitro Lisnichenko > >
