From: Charles Hixson Date: 2002-07-17T03:40:37+09:00 Subject: Re: TkText no -textvariable workaround? Hal E. Fulton wrote: >----- Original Message ----- >From: "Dave Thomas" >To: "ruby-talk ML" >Sent: Tuesday, July 16, 2002 1:01 PM >Subject: Re: TkText no -textvariable workaround? > > > > >>Dossy writes: >> >> >> >>>I did say "academically" -- in a pure OO language where everything >>>is an object, the receiver would be the object a, and the method -- or, >>>the message being sent, would have been "=(x)" ... >>> >>> > >[much snippage] > >Very interesting and enlightening. This is probably >the real reason, I guess, that Ruby works this way. > > > >>Now we have something interesting. The Fixnum '1' that is currently >>'a' now becomes the Fixnum '2'. But Fixnums are immediate values, so >>how are we to implement this? Do all '1's become '2'? >> >> > > >OT: This stirs a dim memory in my mind. Sorry I keep >reminiscing and regressing to the past -- it's a >quirk of advanced age (I recently turned 41). > >In older versions of FORTRAN, there was a simple way of >doing just that -- I forget how it worked -- a COMMON >block or something? You could have a stored "constant" 1 >and change it (say) to 2... upon which every 1 referenced >as a constant in the program would be changed to a 2 >internally (since the compiler pooled its constants). So >a statement MYVAR = 1 would then assign the value 2 when >it was executed... amazing language, FORTRAN. Usually when >you're given enough rope to hang yourself, you gain >flexibility and power as a benefit. Not in this case! > > >Hal Fulton > > Was that run76, or was it FTN? Or did more than one Fortran compiler have that bug? -- -- Charles Hixson Gnu software that is free, The best is yet to be.