there are 2 places that reads Opcodes.h, m-c/js/src/vm/jsopcode.py and jsparagus/update_stencil.py.
m-c/js/src/vm/jsopcode.py's pattern ignores the extra macro, and those opcodes are included in the return value.
jsparagus/update_stencil.py's pattern filtered out the opcodes with extra macro.
So that results in inconsistency, and the first patch here fixes it by filtering out Record and Tuple opcodes from the value returned from m-c/js/src/vm/jsopcode.py.
(eventually we could drop the code that parses Opcodes.h from jsparagus/update_stencil.py, but that's for other bug)
https://bugzilla.mozilla.org/show_bug.cgi?id=1730843 modified
Opcodes.h
, with extra macro to conditionally enable those opcodes for Record and Tuple.there are 2 places that reads
Opcodes.h
,m-c/js/src/vm/jsopcode.py
andjsparagus/update_stencil.py
.m-c/js/src/vm/jsopcode.py
's pattern ignores the extra macro, and those opcodes are included in the return value.jsparagus/update_stencil.py
's pattern filtered out the opcodes with extra macro. So that results in inconsistency, and the first patch here fixes it by filtering out Record and Tuple opcodes from the value returned fromm-c/js/src/vm/jsopcode.py
. (eventually we could drop the code that parses Opcodes.h fromjsparagus/update_stencil.py
, but that's for other bug)