On 12/02/13 13:50, Jeroen Kransen wrote:
Is it ok to report bugs here? I just downloaded the release candidate to
try and see if it solves an issue we're having. It does not.
The project uses JIRA:
https://issues.apache.org/jira/browse/JENA
If you could provide a complete, minimal example (ideally with Derby or
H2 for practicality reasons), someone might have a look. Patches welcome.
In SDB, when we try to delete a named graph, it does not delete it. It
seems to make a switch from named graph to default graph and try to delete
the triples there. I tried to debug what happened. In the class
DatasetGraph line 165 it explicitly converts triples to quads using the
default graph uri. I suspect this is wrong, as it tracked the graph uri up
till this point.
Even in layout2?
When debugging I also had a suspicion that in SDB the triples are deleted
one by one when deleting a whole graph. Is this true? Because I would think
it is much more efficient to do a "delete where graph = :graph" instead. I
hope I'm wrong on this.
In a previous version of TDB I recently mentioned a bug where disk space is
not freed when named graphs are deleted. Is this solved by any chance in
this release candidate? It is too much work for me to switch back just to
try it at this moment.
No changes in this area.
The issue is two fold:
1/ across restarts (index blocks should be recycled within a JVM
2/ the node table isn't GC'ed
(but it isn't in SDB either nor various other RDF databases)
I'd like to add a "compact" operation but I have there are higher
priorities.
Could SDB also be added to the aggregated build?
The testing cycle for SDB is much longer so it would have a major impact
on the rest of Jena.
We have been keeping the SNAPSHOT up to date and it runs the test suite
nightly on Derby (which is quite picky so a good test SQL DB and it also
can execute tree-joins queries, which H2 and friends can't).
I am not comfortable with untested software being released. I used to
have access to a machine with every database installed for SDB testing
but I don't now. I don't use SDB at work currently so my work time
isn't going to go on it.
If people want to join and help systematic regular testing of SDB that
would be ideal. It does not have to be down to the project committers.
The community is everyone interested.
The brutal bottom line is that if there is no help, it does not get
tested. It's as simple as that.
Andy
Jeroen
2013/2/12 Adrian Gschwend <[email protected]>
On 12.02.13 14:11, Claude Warren wrote:
Is that information on a webpage somewhere? I'd like to be able to point
to it in any posts.
https://mail-archives.apache.org/mod_mbox/jena-users/201302.mbox/%3C511A076A.2030907%40apache.org%3E
:-D
cu
Adrian
