Hello Tomasz,

Regarding calminstruction?.xcall? (xcalm.inc),
I just realized it doesn't work with (symbolic) literal empty value,
which I think could be very easily adjusted by minor change to the code :
replacing line

with


can you pls consider this small adj. in utility package ?

Thx

Done.

Many thx !!

Hihi

Another one :
since call allows argument to be "omitted when the definition of called instruction allows it",
I think this mandates another slight tweak in xcall implementation :
.... i suggest changing the below

to


Tomasz, wud u pls consider ?


As a sidenote, i noticed u apparently did small change to CALM match - namely removing the restriction that bracket shud not appear in pattern - when using 3rd argument... but unless mistaken, I can't seem to find any note on this change anywhere ? Wud u pls elaborate ?

Many thx
Rgds

Yes, seems sensible.

That restriction was a remnant of an earlier iteration of this feature, and it stayed there because when changing the design I was not sure whether it was still needed. But after revisiting it recently, I concluded that it was an artificial restriction, unnecessary code that could be removed to clean things up.

tx !

My bad, that restriction was still necessary and I need to bring it back. When the chosen bracket character occurs in the pattern, it may force the matcher to require multiple fallback points, which is what fasmg's implementation of MATCH does not allow. This causes a pattern like: to fail to find the match on this text: (I'm posting this so that it can serve as a reminder why the restriction is there, as apparently I forgot myself.)

At least the restriction-enforcing code can now be much simpler, thanks to the changes I made to MATCH when implementing wildcard modifiers.