[Top] [All Lists]

Re: Editheader issues

2007-04-25 12:10:48

On Tue, Apr 24, 2007 at 05:41:31PM -0600, Philip Guenther wrote:

On Wed, 25 Apr 2007, Kjetil Torgrim Homme wrote:
On Tue, 2007-04-24 at 14:21 -0700, Ned Freed wrote:
We also need to specify what happens if :last is specified in
deleteheader without :index. I suggest saying that it will be ignored.
The alternative is for it to assume an index of 1, but clarity should
trump having a more concise way of specifying "delete the last field"

I think that's a syntax error according to the draft.  if :index <n>
and :last were independently optional, I would agree with your analysis.

Right.  SMI's implementation has always considered
    deleteheader :last "foo";

to be a syntax error, as reflected in the grammar.

Hmm, so it is.  I have explicitly detected the missing :index and
defaulted to 1, but I'm happy to correct that... guess I will do it
now while I'm thinking about it.

I have no problem with this being an error, but I'm a little leery about
depending on the grammar to call it out as such, especially since the grammar
literally implies that :last has to follow :index, which of course it doesn't
since tagged arguments can be given in any order. And yes, I'm well aware that
the practice of using ABNF-like rules but saying "the order imposed by the
grammar is not actually required", but I worry about people inferring that
other grammar-imposed restrictions are simply artifacts.

-mm-  (then again, I still have "replaceheader")

Yeah, me too... And what a pain it was to implement!


<Prev in Thread] Current Thread [Next in Thread>