A set of extracters for different mimetypes can be activated across AVM land.
The XML extraction sample has been modified to be AVM-specific, and therefore
includes the necessary config to activate the metadata extraction in AVM.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6238 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
AVMNodeService doesn't point at AVMLockingAwareService.
Web Client now points at AVMLockingAwareService.
Forced a couple of places in the Web Client that were going through NodeService
to use AVMLockingAwareService to get proper locking behavior.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6234 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
and what the baseline & latest versions were, as well as the number
of files and links checked in staging. Made LinkValidationAction
use the new apis, and annotated the source code so that it's obvious
what should be done to use the base & latest snapshot version info
(see "Gav TODO" within LinkValidationAction.java).
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6219 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
Fix for AR-1433: Upgrade failures on PostgreSQL
Partial fix for AR-160: Concurrency fix for version counter
Partial fix for AR-1483: Synchronization issues
This is an overhaul of the current state of the upgrade scripts. All redundant scripts have been ripped out and replaced
with fresh 2.1 scripts. Indexes for FK relationships have been put in place where required. Several edge cases have been
dealt with including columns being added to two indexes, Sybase not allowing column additions for NOT NULL columns of
certain types, etc.
These changes have been tested for 1.4.0 -> 2.0 -> 2.1 on
MySQL
Oracle
Sybase
PostgreSQL
Found a bug in the VersionCounter Hibernate classes where the optimistic version support was not implemented.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6216 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
Some databases support transactional changes of the schema, but most don't.
Some don't update the schema metadata until the transaction ends.
To workaround all of these issues, a lock table is created at the beginning of the schema bootstrap
and removed afterwards. Each statement is executed in auto-commit mode. If there is a failure,
there is no alternative but to revert to the original data and try again.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6215 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
These are configurable in linkvalidation-service-context.xml
Also ensured the list of links per file is sorted & unique.
Here's a list of known protocols:
http://www.iana.org/assignments/uri-schemes.html
aaa aaas acap afs cap cid crid data dav dict dns dtn fax file
ftp go gopher h323 http https iax2 icap im imap info ipp iris
iris.beep iris.lwz iris.xpc iris.xpcs ldap mailserver mailto
mid modem msrp msrps mtqp mupdate news nfs nntp opaquelocktoken
pop pres prospero rtsp service shttp sip sips snmp soap.beep
soap.beeps tag tel telnet tftp thismessage tip tn3270 tv urn
vemmi wais xmlrpc.beep xmlrpc.beeps xmpp z39.50r z39.50s
For now, all these URI schemes get a free pass except http & https.
Any URI not qualified by one of these protocol/scheme designators
is presumed to be broken. It would be nice to validate ftp links
for real, but that won't happen for a while.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6189 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
(bad/untrusted certs look like broken links), pretend ftp links
aren't broken, detect old href schemas, and added low-level
feature to drop href info on startup in case we ever need a
sledgehammer. Maybe one day I'll really validate ftp links,
but not now. Also fixed startup logic that got broken earlier
(the config file wasn't getting copied in).
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6183 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
this allows an internal cache to be maintained that permits most calls to
the heavyweight getLatestSnapshotID(String store) AVM API to be skipped.
This addresses a performance problem noticed when the number of stores
is large.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6182 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
Typically this is known as a hack, or to use Jon's words, "We're hiding behind a wildcard".
The Spring config pulls in additional (not extension) bootstrap contexts where it was specifically referencing Link Validation.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6172 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
5910: Web Services getUsers
5913: Web Services admin user password change
5956: LDAP anon simple bind test
6133: WCM-486
6158: Merged V1.4 to V2.0
5600: Split person bootstrap
5642: AR-439 NetBIOS adaptor status request
6160: VersionHistoryPerformance patch with no versionedNodeId
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6166 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
The in-transaction work has to align with the work that will be done by the actual background archival,
but node archival doesn't fully support all model constructs and associated behaviour.
Instead of continuing to hack away at each issue that comes up, a complete archive rethink is in order.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6154 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
The node is renamed and hidden from FileFolderService clients. A background task then performs the archival.
A bootstrap component ensures that failed archivals are picked up again.
Found a bug with the status of nodes archived as part of a hierarchy not being updated correctly.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6148 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
- The queue for background jobs is unlimited
- A smaller number of core threads is maintained
- It is possible to set the threads' priority (default low)
Added hasAspect() method to XPath functions.
Various other cosmetic changes.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6146 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
Mostly working with the webapp, but the webapp needs
to check for an null pointer to ensure that some
result is ready. Other more minor wrinkles exist,
such as capturing the number of files checked and
displaying the version validated; this may be different
from the latest snapshot if the validation gets behind
the checkins for a while.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6136 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
Removed calls to getXXXPaths().
Changed mappings of properties and aspects on AVM Nodes to
favor get all calls.
Added getAspects() and getNodeProperties() which take
AVMNodeDescriptors.
Used these in AVM Indexing to reduce the number of redundant full lookups.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6121 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
This seems to make indexing a bit quicker, and at least doesn't make other things
slower. Bulk import now just sucks; it used to be an order of magnitude suckier.
98% of that is due to Andy's recent changes.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6103 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
Not tested at all yet, but the code & essential algorithm is all there.
Switched off for now to avoid exposing others to potential havoc
until some basic sanity checking is done. Seemingly pretty close.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6081 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
Everything but that code that will bring virtualized staging archives up & apply diffs to href validity info.
Handles abort(), filters junk, graceful start/stop, etc. Not usable yet, but soon (assuming no gaping bugs).
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6079 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
The following filename is valid now: "x ¬ £ % & + ; x.txt"
This was a restriction imposed by WebDAV, but the encoding of the repsonses is working well and these restrictions be removed as a result.
Fixed AR-1281: WebDAV upload was assigning incorrect encoding
I added a bean 'charset.finder', which can be fetched from the MimetypeService.
Various pluggins now exist to decode a stream and figure out what the encoding is.
WebDAV and CIFS/FTP are now hooked into this so that they guess a little better.
Fixed others:
Added retrying transactions to WebDAV.
Read/write transactions for WebDAV.
git-svn-id: https://svn.alfresco.com/repos/alfresco-enterprise/alfresco/HEAD/root@6073 c4b6b30b-aa2e-2d43-bbcb-ca4b014f7261
