[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

1.8.0 clause 5.12 vs 5.4.3



1.8.0 is not really inconsistent, here -- 5.4.3 details some formatof
operations; these are 'should'.  5.12 states that the conversions 'shall'
be provided; it does not refer back to 5.4.3, ...

This makes no sense.  5.4.3 provides no details; it mentions the operations
pretty generically and refers to 5.12 for details.  That in turn refers to
subclauses for further details.

... it [5.12] does not refer back to 5.4.3, so this allows implementations
to provide the conversions in whatever manner they choose (they might only
be there for literals in a language, perhaps), although the 5.4.3 operations
remain as a recommendation.

Clause 5.12 quite definitely requires more than that; it even starts with
the requirement to *generate* hexadecimal strings using exact conversions
from binary formats.

As far as I can tell no actual change occurred.  The new "should" in 5.4.3
now simply looks like a mistake.  Somebody saw the "should" in 5.12.3 and
got the wrong impression.

7.12.3 states that language standards 'should' provide the conversions,
which is arguably a separate issue entirely.

With this I agree completely.

Michel.
Sent: 2008-04-09 18:26:51 UTC


754 | revision | FAQ | references | list archive