Re: [Nmh-workers] corruption when attaching all-null files

[Ken Hornstein]

>Better to add a bit that says 'containsnulls'.  MIME requires specific
>behaviour in the face of NULs, as does SMTP.
>
>I don't like this 'isvalid7bit' idea.  The suitability of a body part
>for 7 bit transport can – and should – be inferred from contains8bit
>and containsnulls.  Adding 'isvalid7bit' makes way for conflicting
>assertions as to whether the code is 7 bit or not.  You could always
>'#define ISVALID7BIT(foo) (!(foo->contains8bit || foo->containsnulls))'.
>And similarly for ISVALID8BIT.

You're right ... I was thinking that text/plain US-ASCII technically
prohibits anything < 32, with the exception of CR, LF, TAB, and NP (FF).
But that's not what we're talking about here; for a 7bit CTE the proper
limit is 0 < c < 128.  So yes; containsnuls (don't some greybeards get
grumpy if you use NULL where you meant NUL?) is the right term.

--Ken