On 25/08/18 23:44, Luc Van Oostenryck wrote: > On Sat, Aug 25, 2018 at 10:11:57PM +0100, Ramsay Jones wrote: >> >> >> On 25/08/18 16:43, Luc Van Oostenryck wrote: >>> Opcodes are defined in linearize.c:enum opcode. >>> The file opcode.c also contains a table with opcodes properties. >>> >>> Centralize these definitions into a single file: opcode.def that >> >> opcode.def? Hmm, I think opcode-def.h may be better. > > Well, it's not a header file, it can't be used as one. I then prefer Heh, it is no more or less a header file as any other in the project! The fact that the macros called in it are defined in another file is also not that uncommon! > to not use the .h extension for it (and using the extension .def Windows folk may get confused. ;-) > for this sort of file (containing definition that must be expanded > where used) is relatively common). > >>> diff --git a/opcode.def b/opcode.def >>> new file mode 100644 >>> index 000000000..7add0b879 >>> --- /dev/null >>> +++ b/opcode.def >>> @@ -0,0 +1,114 @@ >>> +// OPCODE negated swaped float >>> + >>> +OPCODE(BADOP, BADOP, BADOP, BADOP) >> >> Heh, something is wrong with \t chars in this table. This line looked >> fine in the original email, but when 'quoted' with '> ', it looks >> wrong. However, some other lines ... > > It's because I've used tabs to align the second column. The real file is > fine. I suppose I should convert those into spaces. Hmm, the original unquoted email had problems also. ATB, Ramsay Jones