Community
Participate
Working Groups
As far as I can see, this is not due to "unused variables" as are other cases related to bug 333678. It might be due to some other "fix" in the compiler? Or ... something? I'll attach "decompiled" versions as well as original source. If you do a "diff" you can see the difference is what was 848 iconst_1 849 istore 8 [zzIsFinal] 851 iconst_1 852 istore 9 [zzNoLookAhead] 854 aload_0 [this] 855 bipush 7 857 putfield org.eclipse.wst.xml.core.internal.contenttype.XMLHeadTokenizer.zzState : int [269] 860 goto 6060 863 iload_1 [zzInput] 864 tableswitch default: 996 becomes 848 iload_1 [zzInput] 849 pop 850 iconst_1 851 istore 8 [zzIsFinal] 853 iconst_1 854 istore 9 [zzNoLookAhead] 856 aload_0 [this] 857 bipush 7 859 putfield org.eclipse.wst.xml.core.internal.contenttype.XMLHeadTokenizer.zzState : int [269] 862 goto 6060 865 iload_1 [zzInput] 866 tableswitch default: 996 I've no idea what the significance is ... but ... the fact is, we have different bytes bits with same version/qualifier. So ... technically we should re-tag. I think re-tag is sufficient, since this code is already different from maintenance, and the error report simply says current I-builds differ from our M4 (we used old compiler in M4).
Created attachment 186351 [details] decompiler listing from 3.6 compiler
Created attachment 186352 [details] decompiled output from current (3.7M4) compiler
Created attachment 186353 [details] original source
This no longer shows up in unexpected comparator logs ... not sure if it was "fixed on purpose", or retagged for other reasons, but end result seems to be 'fixed'.