Commit 35bc7ce1 35bc7ce1d005241ff5415476a51ee94012ccf164 by Alain Magloire

Changes propose by Sam Roberts, They look valid and interesting.

1 parent 2bb68e42
@code{#include <mailutils/mailbox.h>}
@deftp {Data Type} address_t
The @code{address_t} object is used to hold information about a parsed
RFC822 address list, and is an opaque
data structure to the user. Functions are provided to retrieve information
about the address list. An address list is a sequence of addresses, each
of which has four components: an optional phrase describing the address, the
local-part of the address, the domain of the address, and an optional
comment field.
If the phrase is present it is probably the personal name
associated with the the address, if not, the comment may contain some
identifying information.
Since the address-list may contain multiple addresses, they are accessed
by a @strong{one-based index number}, @var{no}.
Comment: Why is this one-based? nobody
familiar with C would ever expect that!
The size of a particular component
may be queried by providing @code{0} for the @var{len} of the buffer,
in which case the buffer is optional and if @var{n} is provided
*@var{n} is assigned the length of the component string.
If @var{len}
is greater than @code{0} it is the length of the buffer @var{buf}, and
as much of the component as possible will be copied into the buffer.
Comments:
The API isn't complete, there are useful things you can't do.
There needs to be a way of accessing the local-part and the
domain of an email address seperately, the syntax of local-part is too complex
to expect somebody to parse it, in particular strchring for '@@'
will fail if there is an '@@' in the local-part, which is valid if
its quoted or escaped.
You can't create an address or an address list.
Creating one should apropriately quote unsafe characters in the local
part and phrase/personal name, not allow a comment (its deprecated), etc.
What about groups? They are easy to parse, but a pain from an api point
of view. _address, instead of being a pure linked list, could
be more tree-like, so each _address also contained and _address* group
member. If there was a group, then email would be the name of the group,
something like "my special group: ;", and group would be the (possibly
empty) list of addresses in the group, and each of those addresses
could in turn be a group, etc.
This data structure reflects the structure of an address ok, but doesn't
address (;-)) what happens if you had a program that just wanted to print
out all the recipients email addresses. The @var{no} could still be
a linear index into the tree, so those not wanting to know could not
know, and the group nodes could silently dissappear.
But what if you wanted to iterate the whole tree?
address_is_group(addr, no)? address_get_group_name(addr, no)?
address_get_group(addr, &addr)?
For creating, maybe the easy way is to splice an address list.
address_create(&addr, NULL)
address_insert(addr, -1, "Big John", "bluesman", "yahoo.com");
// -1 is the end, anything else is position to be inserted at, for ex.
address_create(&group, NULL)
address_insert(group, -1, "Sam", "sroberts", "uniserve.com");
address_insert(group, -1, "Joe @ \"home\"", "joe", "uniserve.com");
address_insert_group(addr, group, "the uniserve boys");
This would be:
@example
Big John <bluesman@@yahoo.com>, the uniserve boys: Sam
<sroberts@@uniserve.com>, "Joe @@ \"home\"" <joe@@uniserve.com> ;
@end example
This is just a sketch of what I think should be conceptually possible
in a complete address parser class. Basically the goals are to be
able to see the structure of an RFC822 address exactly (except for
comments... they get munged together if present, but what the hell,
they're comments!) and to build as complex an address as is
possible. Simple things should still be simple.
Two problems are domain literals, and non-ascii characters. I
think domain literals should be parsed from [127.0.0.1] into
the more commonly groked 127.0.0.1 when somebody does a get_domain()
on one, and that if somebody wants to provide one as a domain, they
can.
Non ascii chars are uglier, perhaps a warning? Perhaps a non-strict
switch that makes it an error, and otherwise if they want umlauts
and utf-8 in the strings, just allow it? The machinery to encode
and decode header fields according to the MIME spec doesn't really
belong here, it's a layer on top of rfc822.
@end deftp
@deftypefun int address_create (address_t *@var{addr}, const char *@var{string})
This function allocates and initializes @var{addr} by parsing the
RFC822 address-list @var{string}.
The return value is @code{0} on success and a code number on error conditions:
@table @code
@item ENOMEM
Not enough memory to allocate resources.
@end table
@end deftypefun
@deftypefun void address_destroy (address_t *@var{addr})
The @var{addr} is destroyed.
@end deftypefun
@deftypefun int address_get_email (address_t *@var{addr},
size_t no, char* buf, size_t len, size_t* n)
Acesses the @var{no}th email address component.
The return value is @code{0} on success and a code number on error conditions:
@table @code
@item EINVAL
@var{addr} is NULL.
@end table
@end deftypefun
......@@ -110,6 +110,7 @@ This document was produced for version @value{VERSION} of @sc{gnu}
* Attribute:: Attribute.
* Stream:: Stream API.
* Authenticator:: Authenticator.
* Address:: Address.
* Locker:: Locker.
* Encoding:: Encoding API.
* URL:: Unified Ressource Locator.
......@@ -210,14 +211,21 @@ For example writing a simple @command{from} command that will list the
@include stream.texi
@node Authenticator, Locker, Stream, Top
@node Authenticator, Address, Stream, Top
@comment node-name, next, previous, up
@chapter Authenticator
@cindex Authenticator
@include auth.texi
@node Locker , Encoding, Authenticator, Top
@node Address, Locker, Authenticator, Top
@comment node-name, next, previous, up
@chapter Address
@cindex Address
@include address.texi
@node Locker, Encoding, Address, Top
@comment node-name, next, previous, up
@chapter Locker
@cindex Locker
......