OSCL/CXML
2
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

kommentaraenderungen

Browse Source
This commit is contained in:
dlichteblau
2005-11-28 22:22:49 +00:00
parent e45c6465da
commit e688f34235
3 changed files with 39 additions and 57 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
README.html
View File

@ -65,8 +65,10 @@
</ul>
<p>
CXML currently implements a namespace-aware, validating SAX-like
<a href="http://www.w3.org/TR/2000/REC-xml-20001006">XML&nbsp;1.0</a>
CXML currently implements a <a
href="http://www.w3.org/TR/REC-xml-names/">namespace-aware</a>,
validating SAX-like <a
href="http://www.w3.org/TR/2000/REC-xml-20001006">XML&nbsp;1.0</a>
parser as well as the <a
href="http://www.w3.org/TR/REC-DOM-Level-1/level-one-core.html">DOM&nbsp;Level&nbsp;1&nbsp;Core</a>
interfaces.

33
runes/encodings.lisp
View File

@ -107,6 +107,39 @@
(defmacro %< (&rest xs) `(fx-pred < ,@xs))
(defmacro %> (&rest xs) `(fx-pred > ,@xs))
;;; Decoders
;; The decoders share a common signature:
;;
;; DECODE input input-start input-end
;; output output-start output-end
;; eof-p
;; -> first-not-written ; first-not-read
;;
;; These decode functions should decode as much characters off `input'
;; into the `output' as possible and return the indexes to the first
;; not read and first not written element of `input' and `output'
;; respectively. If there are not enough bytes in `input' to decode a
;; full character, decoding shold be abandomed; the caller has to
;; ensure that the remaining bytes of `input' are passed to the
;; decoder again with more bytes appended.
;;
;; `eof-p' now in turn indicates, if the given input sequence, is all
;; the producer does have and might be used to produce error messages
;; in case of incomplete codes or decided what to do.
;;
;; Decoders are expected to handle the various CR/NL conventions and
;; canonicalize each end of line into a single NL rune (#xA) in good
;; old Lisp tradition.
;;
;; TODO: change this to an encoding class, which then might carry
;; additional state. Stateless encodings could been represented by
;; keywords. e.g.
;;
;; defmethod DECODE-SEQUENCE ((encoding (eql :utf-8)) ...)
;;
(defmethod decode-sequence ((encoding (eql :utf-16-big-endian))
in in-start in-end out out-start out-end eof?)
;; -> new wptr, new rptr

57
xml/xml-parse.lisp
View File

@ -11,6 +11,7 @@
;;; © copyright 2003 by Henrik Motakef
;;; © copyright 2004 knowledgeTools Int. GmbH
;;; © copyright 2004 David Lichteblau
;;; © copyright 2005 David Lichteblau
;;; This library is free software; you can redistribute it and/or
;;; modify it under the terms of the GNU Library General Public
@ -91,29 +92,8 @@
;; :cdata <rod>
;;; NOTES
;;
;; Stream buffers as well as RODs are supposed to be encoded in
;; UTF-16.
;; where does the time go?
;; DATA-RUNE-P
;; CANON-NOT-CDATA-ATTVAL
;; READ-ATTVAL (MUFFLE)
;; CLOSy DOM
;; UTF-8 decoder (13%)
;; READ-ATTVAL (10%)
;;
;;; TODO
;;
;; o Improve error messages:
;; - line and column number counters
;; - better texts
;; - better handling of errors (no crash'n burn behaviour)
;;
;; o provide for a faster DOM
;;
;; o morph zstream into a context object and thus also get rid of
@ -132,7 +112,7 @@
;;
;; o max depth together with circle detection
;; (or proof, that our circle detection is enough).
;; [gemeint ist wohl zstream-push--david]
;; [gemeint ist zstream-push--david]
;;
;; o better extensibility wrt character representation, one may want to
;; have
@ -260,39 +240,6 @@
(defmacro with-open-xfile ((stream &rest open-args) &body body)
`(call-with-open-xfile (lambda (,stream) .,body) .,open-args))
;;; Decoders
;; The decoders share a common signature:
;;
;; DECODE input input-start input-end
;; output output-start output-end
;; eof-p
;; -> first-not-written ; first-not-read
;;
;; These decode functions should decode as much characters off `input'
;; into the `output' as possible and return the indexes to the first
;; not read and first not written element of `input' and `output'
;; respectively. If there are not enough bytes in `input' to decode a
;; full character, decoding shold be abandomed; the caller has to
;; ensure that the remaining bytes of `input' are passed to the
;; decoder again with more bytes appended.
;;
;; `eof-p' now in turn indicates, if the given input sequence, is all
;; the producer does have and might be used to produce error messages
;; in case of incomplete codes or decided what to do.
;;
;; Decoders are expected to handle the various CR/NL conventions and
;; canonicalize each end of line into a single NL rune (#xA) in good
;; old Lisp tradition.
;;
;; TODO: change this to an encoding class, which then might carry
;; additional state. Stateless encodings could been represented by
;; keywords. e.g.
;;
;; defmethod DECODE-SEQUENCE ((encoding (eql :utf-8)) ...)
;;
;;;; -------------------------------------------------------------------
;;;; Rechnen mit Runen
;;;;