[Info-vax] yet another sys$qiow question

[Stephen Hoffman]

On 2015-08-19 17:42:56 +0000, John Reagan said:

> Anytime you change code generators/optimizers, you run the risk of 
> shaking loose broken programs.  We saw lots of invalid programs from 
> VAX that got wobbly on Alpha.  We even saw it with various improvements 
> we made to GEM over the years.  Invalid programs have undefined 
> behavior.  You might like the current undefined behavior for years, but 
> it is still undefined.  Like uninitialized variables that appeared to 
> be zero or false depending on some sequence of code that executed prior 
> to you (ie, things the image activator may have left on the stack).

Related:
http://blog.llvm.org/2011/05/what-every-c-programmer-should-know.html
http://css.csail.mit.edu/stack/

But then code still using VAX C — and using the str non-l, non-r calls, 
gets, etc — is just asking for the nasal demons; undefined, 
implementation-dependent behavior.  VAX C should have been tied off and 
deprecated an eon ago.   Nasal demons: 
https://groups.google.com/forum/?hl=en#!msg/comp.std.c/ycpVKxTZkgw/S2hHdTbv4d8J 



-- 
Pure Personal Opinion | HoffmanLabs LLC