On Sun, Apr 06, 2003 at 12:39:46PM +0200, Andreas Aardal Hanssen wrote:
> It strips CR from the data that is stored to disk, but that's just because
> qmail-local does the same when it stores messages, so instead of having
> some messages CRLF and some LF in the Maildir, they are all LF with
> APPEND's behavior. But I do not know of any requirement in Maildir to
> strip those CR's.
Okay, that makes perfect sense and I agree it is "good" behavior in that it
maintains sanity and consistency in the mail store. The only thing that I
can see as a problem, is what you mentioned regarding CR CR LF being
converted to CR LF CR LF upon retrieval.
I would say, though, that the GIGO principle doesn't cleanly apply here
because, even though it would be garbage, the RFC is quite clear that any
8->7 bit conversion (which this isn't per se) MUST be 'reversibly encoded',
yet clients only SHOULD append RFC-822 messages. Therefore, it is legitimate
(though ill advised) for a compliant client to upload unencoded binary
messages where the conversion of CR would break the contents. Via APPEND
all sorts of garbage could be legitimately uploaded to a mail store.
> Binc strips all CR from documents stored in the depository. If it sees an
> LF, it outputs CRLF. So the above sequence would be converted to CR LF
> only. From what I understand, this is an ok way to do things.
I agree that this behavior is desirable. I'll be sure to not change it in
the patch for int->char conversion in readChar.
As a side note, at the moment, InputException is used in at least one place
to also mean an OutputException during a write. I went ahead and
slipped in two more Exception classes, IOException and OutputException,
moving InputE and OutputE as subclasses of IOE. This made the handling
of certain copy operations easier. I'll be sending a patch along as soon
as I reach the next milestone on another project I'm working on.
C=)
--
--------------------------------------------------------------------------
Better the hard truth than the comforting fantasy. -- Carl Sagan
--------------------------------------------------------------------------
Caskey <caskey*technocage.com> /// TechnoCage Inc.
--------------------------------------------------------------------------
A presumption on your part does not constitute an obligation on my part.
