From: me7@privacy.net   
      
   On Sun, 29 May 2005 19:50:40 GMT, in   
   , "Jason Burgon"   
    wrote:   
      
   >"Jim Higgins"  wrote in message   
   >news:apnj91tg4c897p98coeuu4mb0phivvq6mr@4ax.com...   
   >> On Sun, 29 May 2005 01:36:51 GMT, in   
   >> , "Jason Burgon"   
   >>  wrote:   
   >   
   >> >BP7 and TP7 zero all global VARs, both inside and outside the IDE.   
   >>   
   >> It may do so based on observation, but the manuals - don't ask me   
   >> where right now - state that uninitialized variables are   
   >> undefined as to content.   
   >   
   >> Regardless of what BP/TP7 do, the definition of the Pascal language says   
   >> uninitialized variables are undefined; i.e., you can't depend on   
   >> their contents.   
   >   
   >Quite, but you can in BP7. I was merely stating a fact, and this behavior is   
   >hardly the first "deviation from the standard" that Borland have employed.   
      
   You're absolutely correct you can depend on the value in BP7...   
   and this deviation from standard is harmless.  I have no quarrel   
   with it.   
      
   But what isn't harmless is the bad programming practice of   
   writing code that  *depends*  on this discovered deviation when   
   that code may not always be compiled by a BP/TP7 compiler.  A lot   
   of BP/TP7 code will compile just fine with the TP6 (or other)   
   compilers and if that code depends on initial zeroing of   
   variables there's no telling what result you'll get when running   
   it.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)