:datetime-bce-timezone-rollover :datetime-leap-second-context)
(token-position token) "DateTime normalization is outside the supported BCE/leap-second calendar context."))
(values value (topic-maps.tmdm:datetime-datatype))))
- (:excluded-literal
- (problem 'ctm-unsupported (if (eq context :variant) :variant-literal-datatype :occurrence-literal-datatype) (token-position token)
- "The separate CTM '*' literal is outside this slice."))
+ (:ctm-integer
+ (consume parser)
+ ;; CTM 3.4.1 [33]: opaque lexical value, unrelated to Wildcards/counters.
+ (values "*" "http://psi.topicmaps.org/iso13250/ctm-integer"))
(:string
(consume parser)
(if (punctuation-p parser #\^)
(attach-reifier role reifier marker))))
(attach-reifier (topic-maps.tmdm:canonical-association association) r position))))))
+(defun read-prolog (parser)
+ "CTM 3.5: one initial encoding header, then one optional version 1.0."
+ (when (and (eq :directive (token-kind (look parser)))
+ (string= "encoding" (token-value (look parser))))
+ (let ((token (consume parser)))
+ (unless (zerop (token-position token))
+ (problem 'ctm-rejection :misplaced-encoding (token-position token)
+ "Encoding declaration must be the first characters of the CTM stream."))
+ (multiple-value-bind (name end position)
+ (ctm-encoding-name *source-text* (token-position token) :character)
+ (unless (string-equal name "UTF-8")
+ (problem 'ctm-unsupported :unsupported-encoding position "Only UTF-8 CTM input is implemented."))
+ (setf (lexer-cursor (parser-lexer parser)) end))))
+ (when (and (eq :directive (token-kind (look parser)))
+ (string= "version" (token-value (look parser))))
+ (consume parser)
+ (let ((lexer (parser-lexer parser)))
+ (skip-layout lexer)
+ (let ((start (lexer-cursor lexer)))
+ ;; Match the fixed terminal before the numeric-literal lexer. A Topic
+ ;; can follow immediately; extra numeric spelling is not version 1.0.
+ (unless (and (ctm-ascii-prefix-p *source-text* start "1.0")
+ (let ((end (+ start 3)))
+ (not (and (< end (length *source-text*))
+ (find (char *source-text* end) "0123456789.+-")))))
+ (problem 'ctm-rejection :invalid-version start "The only CTM version is the exact spelling 1.0."))
+ (setf (lexer-cursor lexer) (+ start 3))))))
+
(defun deserialize-context (text &key document-iri source-name (wildcard-iri document-iri) (wildcard-counter 0)
resource-resolver (max-resource-depth 32))
"Read the documented CTM subset, with per-document prefixes and qualified strings.
(topic-maps.tmdm:core-subject-error
(lambda (c) (problem 'ctm-rejection (topic-maps.tmdm:core-subject-error-code c)
(lexer-cursor (parser-lexer parser)) (princ-to-string c)))))
+ (read-prolog parser)
(loop while (eq :directive (token-kind (look parser))) do (read-directive parser map))
(multiple-value-bind (r position) (read-reifier parser map) (attach-reifier map r position))
(loop until (eq :eof (token-kind (look parser))) do
:resource-resolver resource-resolver :max-resource-depth max-resource-depth)
(values map counter)))
-(defun deserialize-file (pathname &key document-iri source-name (wildcard-iri document-iri) (wildcard-counter 0)
+(defun deserialize-file (pathname &key document-iri source-name encoding (wildcard-iri document-iri) (wildcard-counter 0)
resource-resolver (max-resource-depth 32))
- "Read a UTF-8 CTM file, with optional UTF-8 BOM, using an explicit Document IRI.
-Other encoding directives are UNSUPPORTED. File I/O errors propagate unchanged."
+ "Read CTM octets using an explicit Document IRI and optional supplied ENCODING.
+Check supplied encoding/BOM/declaration conflicts before strict UTF-8 decoding.
+Other charsets remain UNSUPPORTED. File I/O errors propagate unchanged."
(let ((*source-name* source-name) (*source-text* ""))
(require-document-iri document-iri)
(require-wildcard-environment wildcard-iri wildcard-counter)
(let ((bytes (make-array (file-length input) :element-type '(unsigned-byte 8))))
(unless (= (length bytes) (read-sequence bytes input))
(error "Input file changed or was truncated during reading."))
- (deserialize-string (decode-utf8 bytes :strip-bom t)
+ (deserialize-string (decode-ctm-octets bytes :encoding encoding)
:document-iri document-iri :source-name source-name
:wildcard-iri wildcard-iri :wildcard-counter wildcard-counter
:resource-resolver resource-resolver :max-resource-depth max-resource-depth)))))