CVS log for Amaya/annotlib/AHTrdf2annot.c

Request diff between arbitrary revisions

Integration of Regis patches
Irene

Integration of Regis Boudin patches.
Irene

When the user has a non-ISO login, annotations and transformations didn't
work.
Irene

Update the copyrigh.
Irene

JK: cleaned the code. Removed all the refs. to the libwww RDF parser
and to the raptor stand-alone parser, as we're compiling with redland
now.

JK: Code cleaning

JK: Fixing a windows bug

-Update to redland 0.9.16
-New bookmark operations (separators, collections, move) and bug fixes.
-Opening multiple bookmark files.

You won't be able to compile unless you install the latest redland:
   http://www.redland.opensource.ac.uk/dist/source/redland-0.9.16.tar.gz

and follow the usual instructions for compiling Amaya with redland.
Windows won't compile after this commit until I update the makefiles.

-jk

Removing _I18N_ within amaya and annotlib directories
Extending the copyright to 2004
Adding needed convertions utf-8 to the default charset.
Irene

compilation error fix when bookmarks enable

This is the Amaya C++ version : ie amaya uses now g++ (c++ version of gcc)

WARNING:
Before compilation do :
    - On debian just install g++ : "apt-get install g++"
    - A cvs update into libwww directory because there is a expat header to update.
    - "rm -rf *"
    - "../configure [options...]"

S. GULLY

JK: API updates for compiling against redland 0.9.14

Revert DC1.1 namespace back to DC1.0 on writes; we have a server bug
that we can't fix in time when querying for replies that use the DC1.1
namespace.  This commit simply puts back revs 1.49, 1.14, and 1.50
respectively.

Change namespace for new annotations to DC1.1.  Still understands
old DC1.0 namespace for existing annotations.

JK: a sigsev when compiling only using raptor

JK: The new raptor included in redland requires a specific init/finish
calls. When using redland, those calls are made directly thru the
redland's init/finish higher level calls.

JK: trying to find a sigsev bug

JK: removed the ugly ugly as it doesn't apply anymore... no more libwww.

JK: Fixed two memory leaks reported by insure

JK: fixed a memory leak reported by Insure

When a Reply has a more specific type than simply
http://www.w3.org/2001/03/thread#Reply we should be saving that type.
Instead, we were saving only the last type we parsed.  Fix this.
RRS

JK: Fixed a sigsev when the value of an object in a triple
is an anonymous node reference.

JK: Added missing convertion for literal values in objects with redland

JK: Intermediate changes that allow to compile with redland/raptor
or with the standalone libraptor (as included in Amaya). Will become
redland only once DB releases it!

avoid acc vio

JK: Activated the handling of anonymous resources when using Raptor.
This is to avoid seeing some author names expressed as "genid#xx".

JK: First try at converting ANNOTschemas.c to use raptor instead
of libwww's XML parser. Some optimizations can still be done.
Removed an include to libwww's stuff.

Permit an annotation to be described using either the Dublin Core 1.1 or
the Dublin Core 1.0 namespace name.  Annotations will coontinue to be
written using the Dublin Core 1.0 namespace
RRS

Bug fixes:
- memory leaks in the fonts handler
- internationalization of annotlib
Irene

[RRS&JK] Fixes for using Raptor parser under Win32

JK: Use the Raptor RDF parser if compiled with option
-DRAPTOR_RDF_PARSER. The raptor library must exist somewhere in the
user's system and the user should point to it thru the Options RAPTOR
section.

JK: Fixing a memory leak and a bad initialization detected by insure.

JK: Be scared! A commit dans l'etat so that I don't lose my job if something
happens to my laptop.
.
Adaption of JamieZ's thread message sorting algorithm.
Date sort of threads (code written for the old sort algorithm, but not yet
adapted to the new algorithm).
Some changes for annotations/thread deletes, updates, posts.
Some bugs remain, but I'm working on them.
Compiles without problems with the ANNOT_ON_ANNOT option. No time to verify
if there's a compiler problem without this flag. Feel free to fix and
commit those changes if that situation happens (or even rollback).

JK: Annotation replies bug fixes, enhancements.
Things missing: sort the threads by date (sorted by reply to done)
Update the local annotations annotates and replies-to info after a post.
UI: show annotations in same window when browsing.

Merge fixes for orphanned annots due to new server with structured
property values back into main branch so that local annot indices
won't get corrupted when Amaya is run against an updated server.

Permit the RDF properties describing an Annotation or Reply to arrive
in any order.  This makes things work again with older Annotea servers.

Recognize Reply types again.

Don't create an orphanned annotation when the server returns RDF data
whose subject is something other than an Annotation (e.g. the authenticated
creator of an annotation).  This is a more complete fix to the problem
that was being addressed by AHTrdf2annot.c rev 1.28.2.1.

JK: Iggy fix for threads

JK: Removing compiler warnings when compiling without threads.
Saving annot->author and annot->title using UTF-8 (plus some untested
code for I18N for the same cases).

JK: Several annotation/annotation and thread bug fixes (mostly for local
annotations).

This version includes the ability to use the Raptor RDF parser:

 http://www.redland.opensource.ac.uk/raptor/

instead of the libwww RDF parser.  By default, the libwww RDF parser
is used but the Raptor parser can be used by defining RAPTOR_RDF_PARSER
when compiling this file.

See version 1.18.2.1 of Makefile.in for more information.

JK: A patch for uniformizing the use of file:// regardless of the platform.
This should fix up bugs we had when we wrongly added the file: prefix without
converting the rest of the URL under Windows platform. The patch should be
backwards compatible with existing annotations.
New URLs under Windows will now read: file:///drive:path.

JK: Unicode cleaning

JK: Adding a title to annotations (work in progress).
Window guys, don't forget to recompile the schemas.

JK: Thread changes... now locally saving and downloading annotations
(in progress).

Unicode stuff + code clean up
Irene

Removing the macro TEXT ()
Irene

JK: Fixing compiler warnings
Annotlib: code cleaning (rewrite of all the include statements + use of
annotlib.h)
Changed the Amaya version and date.

Merge exp_AnnotAuthUser branch into main branch.  This will display three
additional fields in the annotation metadata if the server supplies them,
else nothing will display.
RRS

Experimental support for displaying authenticated user information
from an Annotea service.  This adds three fields to the annotation
metadata and changes some interfaces in ANNOTschema.  This is
checked-in to a branch tagged exp_AnnotAuthUser so that Jose can
look it over and decide whether to merge it.
RRS

Capture the namespace name of the annotation schema from annot.schemas.
Relocate all other hard-coded namespace names to annotlib.h to make
future maintenance easier.

JK: Don't print the list of annotations unless we're debugging RDF

Mark orphanned annotations (those whose context XPointer no longer
selects an existing element) and bodyless annotations and write
a log of these in APP_TMPDIR.  This code is on a branch tagged
exp_AnnotErrLog for purposes of concurrent development.
RRS

Free all permanently allocated RDF resources before exiting.

When we have only one annotation, bad things happen (memory trashing,
annotation doesn't re-load).  Make sure to re-initialize lastAnnotURL

barstow: BUG FIX: added parameter checks to the contains() method to
  fix a memory read error that was reported by "Insure".

Use 'Comment' rather than (undefined) 'comment' for default annotation type.
Fix bug when updating local annotation index that caused "#" to be replicated
in annotation type.
Fix display of annotation type that is not from the default schema.
Preliminary support for better selection of annotation type.

Extract a list of annotation types from the parsed schemas.
Still needs work to select just a "preferred" schema.

First (basic) level of support for processing RDF schemas.
This parses the entire schema but only uses it to extract
the human-readable text for the annotation type.

JK: Delete of local annotations

JK: removed unused variables

AB: Bug fix: changed sizeof(string) to strlen(string).

Don't segv if the only type for the annotation is the default type.
The old compatibility context was "#id".

JK: Using XPointer in the context, instead of the Amaya labels

Support multiple annotations

use _RDFDEBUG instead of _DEBUG so main() isn't multiply defined

AB: change the call to HTRDF_parseFile to that function's latest
AP changes.

Miscellaneous code cleanup (e.g. strdup to TtaStrdup).

AB: Changed the call to the RDF parser because its API name changed
recently as well as its return value.

Changed the function comment header blocks to be more consistent
with other files in the annotation's library.

AB: This file contains the RDF parsing code.  As of this version,
the RDF parser being used is the one in libwww.