Subj : Re: "On the Fly" Hex-numb conversion ... ??? To : Steve From : Klaus Meinhard Date : Mon Mar 04 2002 12:40 am From: "Klaus Meinhard" Steve, Gerald, > When you try to concatenate the values of two (or more) variables, your > schem would have > > %a%%b%%c% > > which is ambiguous - %% is also the representation of a literal %; the > JP scheme > > %[a]%[b]%c > > is unambiguous. Yes. On a purely theoretical basis: since 4DOS doesn't need the second % char, it's occurrence gives rise to some interesting parsing problems: Is %a%b equal to VarA and "b" or VarA and VarB? 4Dos has to check the environment somehow, and that may need some (often negligible) time. In %a%%b% or %a%%b (which is valid too) Is what comes after VarA a percent char( %%), or VarB? 4DOS has to decide, and may not do so consistently across all versions or even across to 4NT etc. In the case of the prior existence of variables %s, %so, %som a construct like set x=%y%sometextfollows has obvious problems. I have already seen 4DOS version changes, in which a slight change to the parser made such (working beforehand) constructs suddenly fail. Facit: while 4DOS supports the use of the %x% variable syntax, it's use is bound to lead to problems. The use of %x, %[x]someothertext, is unequivocal, though it may mean an additional byte now and then. -- Viele Gre, best regards, *Klaus Meinhard* 0736'57" East 5307'52" North Author of the 4XBTM batch collection at http://www.4xbtm.de --- BBBS/NT v4.01 Flag-2 * Origin: Mach2 Systems (1:342/3) .