
Introduction into OcamlNet
|
|
|
The library component "netstring" A short overview: The Netchannels module wraps I/O channels into
classes. This increases the flexibility of channels a lot. Netchannels
are used throughout the rest of the library. - There is also a variant
called Netstream supporting an intelligent buffering that
allows look-ahead.
There are a number of quite low-level modules that are
very useful, too. Netbuffer is an enhanced version of the
Buffer module of the standard library that allows deletions within buffers.
- There are two wrapper modules for regular expression engines:
Netstring_str and Netstring_pcre. These have almost
identical interfaces, but differ in one point. Netstring_str accepts
Str-like regular expressions, while Netstring_pcre prefers the Perl/PCRE
syntax. The current implementation of these modules bases on PCRE in
both cases. The idea of these wrapper modules is that we can change the
underlying regexp engine while retaining the old interfaces. The Netencoding modules deals with some frequently
found encodings of characters: Base64, QuotedPrintable, URL-encoding,
HTML-encoding. Strings can be encoded and decoded. For Base64 and
QuotedPrintable it is even possible to encode or decode while reading
from or writing to a netchannel. The algorithm works chunk by chunk,
and thus supports large data streams. Another kind of encoding is addressed by Netconversion:
This module can recode between character sets. Again, conversion can
be applied to strings or netchannels. The focus of this module is
conversion between character encodings, but not Unicode support.
So you do not find a special data structure for Unicode strings,
and you do not find functions for locales. It is just what network
protocols need. One of the most important data structures are mail-style
messages, also called RFC822 messages. The module Netmime
supports these messages, including (as the name says) the MIME extensions.
Messages are represented as classes, and can be processed in a rather
convenient way. If needed, there is also the low-level Mimestring
module. There are two special-purpose parsers: Netdate
for date strings, and Netaddress for mail addresses. The module Netsendmail allows the composition
of mails with attachments, and calling the MTA. This is very useful
for daemons that want to send mail. (An SMTP implementation is still
missing, though.) There is a parser and printer for HTML text: Nethtml.
The HTML document is represented as recursive data type. The HTML parser
can cope with errorneous HTML to a certain extent and tries to recover
from errors. The parser is parameterized with a "simplified DTD" that
is suffient for HTML. It is possible to use this parser for other
SGML document types, as long as the simplified DTDs contain enough
information to find out where elements begin and end. The module Neturl provides an abstract data type
for URLs, allowing users to parse URLs, modify them, and print them.
The module is known to be very strict regarding the URL syntax, but
nevertheless very useful.
|