On Wed, 2006-11-08 at 14:38 +0000, Alexey Melnikov wrote:
Kjetil,
We' ve discussed your suggested text for encoded-character Sieve
extension during the San Diego meeting.
I think people are generally happy with your text (+ Philip's proposal
for generic syntax for "magic things inside {}"), but both Ned and I
don't think that invalid sequences should cause runtime errors. Can you
elaborate on why you think causing runtime errors would be better than
just leaving invalid sequences as is (like in variables)?
hmm, runtime isn't too appealing, but compile time would be nice. this
actually depends on the order we process encode-character and variables.
if we do variables first (so "${hex:${var}}" works), it must be runtime.
if we do encoded-characters first (so "${v${hex:61}r}" works), it can be
compile time.
I don't have any strong feelings about it. it's good to catch as many
errors as possible at upload time, but it may be better to leave the
task of finding probable typos to a Sieve lint program.
--
Kjetil T.