Moving hdfs-dev@ and mapreduce-dev@ to Bcc: to avoid cross-postings. >From what I understand you want to have a test case which demonstrates that whatever socket/stream connections were opened are properly closed in case of IOException? I guess one of many possibilities would be to check if a connection is still open after IOException were thrown.
For the second part of your (quite open-ended) question: if you make a modification that might potentially leave unclaimed resources such as socket connection in case of abnormal control flow then yes - you'd better have a test case to cover that error recovery is done properly. -- Take care, Konstantin (Cos) Boudnik On Mon, Jan 17, 2011 at 06:04, bhallamudi kamesh <bhallamudi.kam...@huawei.com> wrote: > Hi All, > How to write test cases for the scenarios such as throwing an IOException > from the try block and ensuing that corresponding stream's/socket's closure > either in catch block or in finally block. > Do we need to write test cases for these scenarios in order to submit > patches? > > Thanks in advance > > With Regards, > Bh.V.S.Kamesh. > > > >
