Hi Ralf, On Thu, 2019-07-25 at 14:39 +0000, Schmelter, Ralf wrote: > Hi Severin, > > thanks for the review. > > > Perhaps there should be a 'testJstackUTF8Output()' which will use the > > thai symbol. Only that test method would use the special-needing-UTF-8- > > marker. > > If we remove checking for the marker, the output would be ignored > completely and only the return code is checked.
I wasn't suggesting to remove the marker. What I meant was to keep the marker at pre-JDK-8227868 state and use a marker with UTF-8 chars for a "testJstackUTF8Output()" test. > So I would argue to > leave it as it is now. But it would probably make sense to make the > marker string smaller (say ~120 chars). The current length was a result > of the buffer size I originally used (32k, instead of 256 now) and as > we have seen this leads to truncation of the jtr file. +1 to reducing the length as long as it continues to reproduce the problem of JDK-8227868 and JDK-8222491. Thanks, Severin
