https://bz.apache.org/bugzilla/show_bug.cgi?id=62887
Alan Silva <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW --- Comment #2 from Alan Silva <[email protected]> --- Sure thing. I'm using a Parameterized Controller (https://jmeter-plugins.org/wiki/ParameterizedController/) to call a custom HLS sampler (an Apple media streaming protocol). I've found the change to be in src/core/org/apache/jmeter/samplers/SampleResult.java in the: public void addSubResult(SampleResult subResult) { addSubResult(subResult, true); } which in turns calls public void addSubResult(SampleResult subResult, boolean renameSubResults) to clobber my custom result label names. I honestly find this design choice pretty strange because: 1 - It breaks labels for users upgrading from Jmeter 4.0 to Jmeter 5.0 and the change is not explicitly documented in the Jmeter 5.0 release notes (https://jmeter.apache.org/changes.html) 2 - Clobbering a user's SampleResult labels by default is ill-advised. I would propose simply doing the following: public void addSubResult(SampleResult subResult) { addSubResult(subResult, subResult.getSampleLabel().isEmpty()); } 1 - If the user's label is empty, cool, use the new renameSubResults feature. 2 - If it's not, respect the given label to the SampleResult object. I wouldn't even make addSubResult(SampleResult subResult, boolean renameSubResults) a public method. It should only be called privately due to the lack of a SampleLabel by a user. The scenario in which a user goes through the effort of naming their SampleResult label and then calling addSubResult(subResult, true) makes no sense to me. The logical thing to do if they if don't care about their label anymore would be simply to remove their call to subResult.setSampleLabel(). What are your thoughts on this? Cheers. -- You are receiving this mail because: You are the assignee for the bug.
