Hi, Pierre-Yves
When you create an M datastream, Fedora stores the content in a separate
location with byte-for-byte accuracy. It stores exactly what you send
without modification. With inline XML (and FOXML in general), the whitespace
(indenting, spaces between attribute values, etc) may be
There is a bug report..
https://fedora-commons.org/jira/browse/FCREPO-190
Currently, Fedora transforms the XML into a non-standard canonical form
before calculating the checksum. If you wanted to verify the checksum today,
you would have to use the same non-standard transformation rules on the
I have checked the Fedora code. There are 32 usages of the reference
Constants.MODEL.HAS_MODEL, to get the content models of a object. To me,
this feels like something that should be collected in a general method,
like getContentModels(pid) or the like.
If I remember correctly, that was the
On Thu, 2009-07-30 at 12:14 -0400, Chris Wilper wrote:
For notification of relationship-level (err.. statement-level)
changes, currently apps can subscribe to the appropriate RELS-*
changes, but as you point out, determining the diffs is difficult
because you have to look at the previous
A few quick comments (I am in meetings for the rest of the week, and
will have more time to think about this next week).
I think I imagine these stores to live alongside the HighlevelStorage
system, in that the HighlevelStorage is just one implementation. But
they might also live below the
Is the policy manager service used for anything
other than
bootstrap (and unit tests :)? (or intended to be?)
used for - no (apart from the tests!). intended to be - it was in Muradora,
it's an open question as to the form of a policy management API
I see. I think this is not
I think the biggest hang-up in the whole process was figuring out
which exclusions had to be done in the root pom for which libraries,
so that the real log4j and commons-logging libraries wouldn't be
accidentially transitively included. I found mvn dependency:tree
helpful along the way.
Is
Hi, All
Hi, All
There may or may not be any Metro service in Madrid in time for our
meeting. As far as I understand, a subway strike was postponed through
this weekend (I had no problem taking the metro from the airport to the
my hotel). On Monday, it looks like there is supposed to be
Hi all,
Chris and I just merged the fcrepo-837 branch to master. This change,
which was also applied to the maintenance-3.4 branch, standardizes on
Unix-style (LF) line endings for all text in the repository.
Formerly, there was a mix of DOS-style (CRLF) and Unix-style line
endings, and the move
I get the same message. It fails when trying to assemble the installer.
The utilities-main jar seems that it is neither created as an artifact
in the fcrepo-server/target directory, nor installed to the local maven
repo. Other artifacts in fcrepo-server are created just fine (e.g. the
rmi
would it make sense to do away
entirely with this caching code, and leave management of credentials to
the upstream authnz providers?
I remember this particular caching layer causing a one or two insidious
bugs in the past, all of which were nearly impossible to figure out. I
would be
Hi all,
I just committed a working proof of concept branch
of fcrepo-504. As mentioned on the committers call, it tries to remove
security configuration from web.xml and into Spring. The ultimate goal
is to have fedora produce a generic war file without any custom,
baked-in configuration (i.e
Hi Frank,
I have some unpublished code that served as a proof of concept while
initially proposing HighLevelStorage. Let me see if I can put it up on
github so you can take a look. It an experiment in using HBase, and in
investigating what would be required for Fedora to be used in a
One possible direction forward here would be to start by offering a
configurable option to let Field Search sit over a GSearch instance for a
given repository (or pool of repositories)
If FieldSearch is to remain part of Fedora's external APIs, I think this
would be the most logical way
Hi Frank,
I put my initial HighlevelStorage proof of concept up on github, in my
forked fcrepo repo:
https://github.com/birkland/fcrepo.git
branch: hlstore_hbase_poc
(you can see it as part of the fcrepo network view:
https://github.com/fcrepo/fcrepo/network)
Everything of interest is in the
Does this apply to the DORegistry.register() method as well? i.e. do i
have to check that the DigitalObject's PID passed to the register()
method is not part of the neverGeneratePid collection,?
I might be missing something, but I cannot seem to find a
DORegistry.register() method. There
- The role of the DeploymentManager is to find and return the
representations created by a DisseminationService.
If you're referring to DeploymentManager in the hlstorage POC, its
purpose is to look up which ServiceDeployment object is associated with
a given ServiceDefinition in the context
Yuck. Would excluding them from the war file really be enough? I know
we've gone through a few efforts in the past to try to get Fedora
working properly in app servers, but it seems like a very difficult
proposition to get *any* sufficiently sophisticated webapp to deploy
in multiple kinds
I'd like to suggest a different route, one already formally endorsed by
Fedora. Moving the application to the OSGi framework will enable it to be
deployed in almost any container (including many totally off the JEE specs),
helped by the clean, stringent OSGi classloading architecture. It's
Hi Frank,
I just confirmed that Steve was right. This exact exception is thrown
if the schema is inaccessible somehow. Unfortunately, time constraints
prevented us from implementing a better solution such as validating
against static schema files in the classpath. The workaround,
I thought I added all the schemas, and an entity resolver for them,
into the integration tests. Hmm.
https://github.com/fcrepo/fcrepo/commit/b05bf836e471ed5b46adbdfc8c46ce7788848343
I guess I missed the BatchModifyParser- though it doesn't have access
to that resolver, anyway.
Yeah, I
If you partition across full repository instances as described, you have set
yourself one set of scaling problems. If you instead partition your storage
layers (Low Level Store, RI, SQL DB, etc.) and provide one Fedora web
application (one suite of web services) over them, you can set
when one reads the replication mirroring documentation, the mechanism using
RMI is described as if one is using RMI as a transport mechanism, implicating
that a shared directory is not needed. Is this a correct interpretation?
Yes, this is correct. The RMI transport mechanism was designed
So here's a provocative question to start: Assuming for a moment that
the core Fedora object model (versioning warts and all) stays the same
for 4.0, would something like this interface actually be compatible
with the major objectives we've talked about with respect to High
Level Storage?
24 matches
Mail list logo