CVS log for java/classes/org/w3c/rdf/examples/ARPServlet.java

Up to [Public] / java / classes / org / w3c / rdf / examples

Request diff between arbitrary revisions

Keyword substitution: kv
Default branch: MAIN

OBSOLETE - see http://dev.w3.org/cvsweb/2006/RDFValidator/ for the latest RDF validator code.

syntax fix

a bit of escapism, and further changes to layout

layout changes

rearranged commented fonts for some sanity.
normalized a few comments.
set UTF-8 in doPost.

i18n changes accumulated over the last 17 months

patch from ammanuel <3F8668F0.3050703@w3.org> to fix n3 datatypes

ryan's work re-ordering output, added base URI for relative URI resolution

enhanced error message provided by Martin related to undecodable data when trying to read URIs

fixed a well-formedness problem in the generated XHTML result page (/body/html was appearing before the bottom nav bar)
made applet param tag attributes lower case and closed them so as to have a fully conformant XHTML 1.0 transitional page in all cases (tested with NS7.02, ie6.0, mozilla1.4alpha)

now printing the datatype and lang information for literals in the triple table (not in the graphical representation)

fixed a bug in the previous version:
SVG files for SVG - embed and IsaViz/ZVTM were badly referenced (pointing to jiggles.w3.org:8005 instead of www.w3.org/RDF/Validator/)

fixed problem reported by Dave Beckett (non-RDF documents reported as valid whereas they do not contain any rdf statement)

implemented support for the IsaViz plug-in (applet)
added SVG-embedded (in an object tag)
migrated to ARP2-alpha-1

fixed a bug in format() for warning/error messages (special XML chars were not escaped)

update to work with Jena1.6.1
i18n code from Martin is working
some modifications to the output
bug fix (does not crash any longer when there is no XML header and tries to find out the charset)

allowed GRAPH_VIZ_FONT_DIR and GRAPH_VIZ_PATH to be absolute

tweaking border styles (for IE, Netscape, Opera), correcting html validation problems, removing some unneeded code

added header and links to jump to graph,...

set output to GraphViz to utf-8
aligned defaults (description and actual behavior to form)

changed stream to Writer in comments

use default for FORMAT == null

check for ORIENTATON == null

- some fixes in printDocumentFooter (& -> &, \> -> />)
- removing uses of sURI after data read into sRDF

fixed doGet for testing, reorganized code (moved from doGet/doPost to process)

one more bug fix

bug fix

replacing & with & on RDF output

added 'meta charset' to output

Added new parameter PARSE to avoid the clumsy "NOTE:   if a URI is entered,
it will take precedence so it must be cleared to parse the RDF.".
This code is not yet tested (because jiggles.w3.org is down).
Requires changes to the validator page:
   <input type="submit" value="Parse RDF" name="PARSE" />
and
   <input type="submit" value="Parse URI: " name="PARSE" />
Is written in a way that should work even without these changes.

After a long, slow detour input from URI finally works.
Lessons:
- Beware of String == String
- Beware of null.toUpperCase
- Beware of syntax details of org.apache.regexp.RE, in particular [-\\.]

added getRDFException for more detailled error reporting

eliminated multiple reading of URIs; URIs are now only read in
getRDFfromURI, and stored to sRDF.

the whole mess! uri submission doesn't work anymore at all

some updates

exception experiment

bug fixes

tweaking exceptions

fixing exceptions

bug fix

limiting exceptions

bug fix

debugging/error messages

experimenting with error messages

more detailled error messages

changed error messages

bug fix

avoid getting RDF more than one time from URI

bug fixes

rewrote getRDFfromURI, with all the trimmings

more bug fixes

small fixes

getting Content-Type/charset of URI

renamed getByteStream to getRDFfromURI to better reflect what it's supposed to do.

small fixes

removed test string

more exception cleanup

more exception cleanup

fixing more exceptions

fixing exceptions

small bug fixes

fixed a misplaced }

changed ServletOutputStream to PrintWriter everywhere

setContentType to text/html;CHARSET=UTF-8

small adjustments

doing conversion from iso-8859-1 to utf-8 by hand (for test purposes)

tried to make sure input is interpreted as UTF-8
Made sure non-ascii characters are output correctly in Ntriples

check request parameter encoding and set to UTF-8 if not set.
Doesn't compile for me because of some missing packages.

Improve documentation; remove some unused code

Bug fix: if the object of a triple is an anonymous node, the
corresponding arc (predicate) was un-labeled.

Fix support for making anon node empty/blank labels.

Bug fix: if two objects had identical Literal values, the
graph only had a single node for the two Literals.  This fix
gives every Literal its own node in the graph.

If the graph format is SVG, use the Dot's [undocumented] URL field
to make URI when the subject/predicate/object is a [non-anonymous]
URI.

New options: allow graphs with anon nodes to empty - just like they
are in the M&S spec; allow the user to just get a graph - and no
triples.

Fix typo in the HTML that is used to submit an error report.

Added support for turning on ARP's embedded RDF mode.  This must
be set to false [default is true] if the RDF is NOT enclosed in
<RDF>...</RDF>.

Fix typos in problem reporting form.

Improve some error logging; minor format changes.

Make the Literals be enclosed in rectangles and Resources
enclosed in ellipses.

Format changes.  Also check for file: URI schemes and
do not allow them.

Improve error reporting.

Only use a single StatementHandler.

Bug fixes; replace StringTokenizer with Apache's RE (Regular
Expression) class.

If an object is a Literal, enclose it in a box.

Servlet for an ARP-based online RDF validation service.
ARP: http://www-uk.hpl.hp.com/people/jjc/arp/