Of course offline instrumentation does change classes and it does this almost in exact the same way as during on-the-fly instrumentation. In later case you just don't see this change as it happens on-the-fly. Nothin weird here. I have no idea what you mean by "/* compiled code */". The only wild guess is that your decompiler is not capable to decompile such classes and so shows this, but that's completely different story.
Now the most important thing: in case of offline instrumentation you must run generation of report against original non-instrumented class-files, otherwise you'll be getting exact same error about missmatch of classes. On Monday, June 19, 2017 at 7:04:36 PM UTC+2, [email protected] wrote: > > Hi Evgeny, > > Thanks for the quick reply, I will try it. > > Another thing I notice that the instrument seems sometime will change the > classes file in a wired way, after the off line instrument, a lot of code > become /* compiled code */. And the data from this class can't be correctly > generated, do you have any idea on it? > > Yutian Song > > On Monday, June 19, 2017 at 9:49:28 AM UTC-7, Evgeny Mandrikov wrote: > > There are known cases when class-files are different between executions > of exact same compiler in exact same configuration. Hence the only thing > that you must guarantee and hence check is not a configuration, but exactly > the fact that class-files are byte-to-byte identical on both machines. > > > > On Monday, June 19, 2017 at 6:27:53 PM UTC+2, [email protected] > wrote:I have tested with the Jacoco a lot in the past month. Yesterday I > notice there is a class do no match with execution data, I search for > several discuss and also find the explanation > http://www.eclemma.org/jacoco/trunk/doc/classids.html. > > In my case I run try to using different machine to run the test and > generate report by Jenkins, I think they share the same config but I will > check. So even though the classes files are from different machines I > thought they are same. Anyone has any idea of the config I could check on > the machines other than: > > Different compiler vendor (e.g. Eclipse vs. Oracle JDK) > > Different compiler versions > > Different compiler settings (e.g. debug vs. non-debug) > > -- You received this message because you are subscribed to the Google Groups "JaCoCo and EclEmma Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/jacoco/ad61024f-b95c-4986-9066-38a376f9c138%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
