Ensure rdf:first and :rest properties replace previous ones (removing old ones before adding) rather than duplicating them; write follow-property whenever previous child is there regardless of follow-siblings policy
9df54f49e7f5 — Chris Cannam 9 years ago
Avoid using blank nodes for "top-level" objects picked up through object-tree following (use them only for property objects)
576c3ac0c1b9 — Chris Cannam 9 years ago
Some documentation notes
f619be7df6e6 — Chris Cannam 9 years ago
Rationalise generation of blank nodes in object storer.  This now passes all our tests, but it generates a "spare" blank node in test-object-mapper-2.ttl, so test coverage clearly isn't adequate yet
ad45098bb046 — Chris Cannam 9 years ago
Some obvious fixes
6ddbeaa1aa5a — Chris Cannam 9 years ago
Get ObjectStorer compiling again; but this will take a lot more review
38c70f2b9d61 — Chris Cannam 9 years ago
Rename loadSingle to allocateSingle -- it's an allocate phase function not a populate phase one
fb9f3d65be59 — Chris Cannam 9 years ago
Start implementing the collect pass.  Does not currently build
c9fb35e9853a — Chris Cannam 9 years ago
Start thinking about restructuring ObjectStorer in the same way as ObjectLoader previously was -- have run into a requirement for this when writing follows-properties using object mapper
f0a79d02f132 — Chris Cannam 9 years ago
More debug
3efbde7e1335 — Chris Cannam 9 years ago
Add file required by previous commit (oops)
49f81a41f90b — Chris Cannam 9 years ago
Add RDFException constructors using node and triple; use these to make the changeset change/revert exceptions more comprehensive
4d2364b3fda3 — Chris Cannam 9 years ago
Show interpretation of Serd status code as error message
67edaa336785 — Chris Cannam 9 years ago
Split blank-node policy out into three options -- use blank nodes, don't use them for object property nodes, and don't permit them anywhere (not even as list nodes)
ce051c9275d7 — Chris Cannam 9 years ago
Clarify blank node policy
c22426e495a7 — Chris Cannam 9 years ago
Split the ObjectLoader load callbacks into two, immediate and final callbacks (final is what we had before, immediate is what we need in the case of e.g. providing context for a document object before its constituents are added to it)
3fc91d43b04a — Chris Cannam 9 years ago
Fix up the Qt widgets test a bit -- reworking this for new object mapper is next on the list
a4f12f103e6e — Chris Cannam 9 years ago
Fix failure to call load callbacks
d3219f6324fe — Chris Cannam 9 years ago
Added tag v0.8 for changeset 39682ebe2dc8
39682ebe2dc8 — Chris Cannam v0.8 9 years ago
Update change log
Next