Opened 8 years ago
Closed 8 years ago
#11699 closed enhancement (fixed)
Tracker-2.2.0
| Reported by: | Douglas R. Reno | Owned by: | Douglas R. Reno |
|---|---|---|---|
| Priority: | normal | Milestone: | 8.4 |
| Component: | BOOK | Version: | SVN |
| Severity: | medium | Keywords: | |
| Cc: |
Description
New minor version
NEW in 2.2.0 - 2019-02-20
=========================
* Multiple memory leak and corruption fixes
* Bumped glib minimum version to 2.46, it already was in practical terms
* Test suite improvements
* Restore log domain
Translations: ca, cs, da, kk, lt, pt_BR, sv, zh_TW
Highlighted changes since 2.1.x:
--------------------------------
* New SPARQL parser, able to generate SQL that is generally more readable
and at places performs better. Multiple buglets fixed in the process
* Much improved support of SPARQL1.1 features and syntax that was missing:
- Property paths: Allowing to match connectivity between two resources
by an arbitrary length path. There is a number of supported operators
(alternative, sequence, oneOrMany, ...) that can be combined, e.g:
SELECT ?s ?p { ?s ^(nfo:belongsToContainer*)/(nie:url|nie:title) ?p }
Only the negated path operator (!) is not supported at the moment.
- Support for fully unrestricted queries, eg:
SELECT ?s ?p ?o { ?s ?p ?o } ORDER BY ?o ?p ?s
Queries with unrestricted predicate (?p in the example above) were
just supported in a very restricted set of situations. All those
limitations are gone.
- MINUS allows subtracting the solutions that match the given triples
template, eg:
SELECT ?s { ?s a nfo:Media } MINUS { ?s a nfo:MusicPiece }
* Support for prepared statements. TrackerSparqlStatement can be built
with SELECT queries containing (custom) ~var syntax, and updating
their values before obtaining a cursor.
* Many tests were added, and Tracker is generally much better tested thanks
to CI.
* tracker-store now automatically shuts down on inactivity.
NOTE TO PACKAGERS: This release might trigger ABI checks, as
TrackerSparqlConnectionClass struct size grew. This object can only
be subclassed within Tracker and instantiated through Tracker API, so
the change is inconsequential.
Better tested? Sounds good to me. :-)
Change History (2)
comment:1 by , 8 years ago
| Owner: | changed from to |
|---|---|
| Status: | new → assigned |
comment:2 by , 8 years ago
| Resolution: | → fixed |
|---|---|
| Status: | assigned → closed |
Note:
See TracTickets
for help on using tickets.

Fixed at r21221