Community
Participate
Working Groups
The lines in the *.tokens file appear to be in random, probably accidental hash, order. Please apply an alphabetic or numeric sort so that after regeneration there is not a difference to reconcile with CVS.
The order is neither arbitrary nor random but depends only on the order of rules / keywords in your grammar. We haven't had any issues with these files in any of our 130 testlanguages or in our production languages. Do you get any Antlr warnings or errors when you regenerate your language?
Checking the CVS history, it's mostly stable as you say. The M6 development shows gentle change. The regeneration under M7 shows a massive change. Subsequent regeneration under M7 shows another massive change, back to the M6 evolution. The transition to M7 was so painful, that I guess something difficult to reproduce happened. Close as CANT REPRODUCE and I'll reopen if another disproportionate change happens. ?? Could be because i deleted src-gen in an attempt to get rid of MWE2 stack traces ?? No I don't see any ANTLr errors or warnings (in the console). Under M7 the compiler warnings are _much_ better; just need to sort out NLS and @Override.
The contents of src-gen is erased anyway before anything is regenerated so I guess that has nothing to do with the intermediate instable *.tokens file. If you can track this one down to something which can be reproduces, please feel free to open a ticket. Meanwhile, the warnings have been reduced to zero for the src-gen folder.
