inconsistent declarations of variables as SPECIAL (or SPECVARS) and GLOBAL

[masinter]

looking at another file there's the an example of "PRETTYFLG". It seems like some code treats it as SPECIAL (and binds it) while other code treats it as a globalvar and uses (RESETLST (RESETSAVE PRETTYFLG))

These two usage patterns are inconsistent. I suspect this is a common problem.
If the variable is never bound special, it CAN be treated as a global variable (which is a PUTBASEPTR rather than a Free variable lookup.

My main concern is that the files we have may have been compiled with one set of declarations and we're possibly recompiling with a different set of declarations (e.g., after a GATHER-INFO(ALL).

[masinter]

notes on gather-info:
gather-info(2) sets up property lists with properties block-called-by called-by usedfree-by boundby
to "check up on masterscope" and also for incorrect declarations at compiled time.

• all functions called by any should be defined?
• no variable declared global or used global should bound
• no variable @ cl:consantp should be used free or bound
• variable declarations should be consistent file to file?
• ...

masterscope anomolies?
should never have two different records with same name