14 Jun, 2007

4 commits


13 Jun, 2007

3 commits


12 Jun, 2007

7 commits


11 Jun, 2007

5 commits


08 Jun, 2007

1 commit


07 Jun, 2007

4 commits

  • use a privilege of "unindexed-search" instead.  Also, eliminate the
    index-rebuild privilege and fold all of its functionality into ldif-import,
    since having a separate privilege for it doesn't provide much benefit and
    creates additional administrative overhead.
    
    OpenDS Issue Numbers:  1765, 1776
    
    
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2051 41b1ffd8-f28e-4786-ab96-9950f0a78031
    neil_a_wilson
     
  • for entries in private backends (e.g., the root DSE, schema, monitor entries,
    config entries, etc.).  The entryUUID value that will be generated is based on
    an MD5 hash of the entry DN, but this shouldn't be a problem for entries in
    private backends because none of them allow modify DN operations.  User entries
    should have a real entryUUID value generated when the entry is created (either
    via an LDAP add or an LDIF import).
    
    OpenDS Isssue Number:  1775
    
    
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2050 41b1ffd8-f28e-4786-ab96-9950f0a78031
    neil_a_wilson
     
  • element that includes flags that indicate how the entry was encoded.  The
    flags currently defined include:
    
    - Whether to exclude the DN from the encoded entry.  This may be useful for the
      filesystem entry cache, since it will already have a reference to the DN.
    
    - Whether to compress the set of object classes contained in the entry with a
      binary token.
    
    - Whether to compress the attribute descriptions in the entry with binary
      tokens.
    
    These changes can help improve the encode/decode performance, and can reduce
    the entry footprint (by about 30% for entries based on the example.template).
    None of these options are enabled at the present time, but components which
    call the Entry.encode() method will be able to indicate which of them should be
    used for that entry.
    
    OpenDS Issue Numbers:  660, 1675, 1770
    
    
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2049 41b1ffd8-f28e-4786-ab96-9950f0a78031
    neil_a_wilson
     
  • Considering the limited added value that the JMX port will provide on 1.0 and the fact that we do not want to expose it on the QuickSetup, in the admin group we consider to disable this port by default.  Before committing a fix for this maybe we should send a Heads Up message to the users mailing list.
    
    Basically after the changes the JMX connection handler will not be enabled if the server is installed using the graphical setup.  For the command line setup, JMX will not be enabled unless the user explicitly includes de -x option (for the JMX port) when invoking the command-line.
    
    
    
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2046 41b1ffd8-f28e-4786-ab96-9950f0a78031
    jvergara
     

06 Jun, 2007

3 commits


05 Jun, 2007

2 commits


04 Jun, 2007

4 commits


02 Jun, 2007

1 commit

  • server will only allow client connections over loopback interfaces and will
    reject requests from non-root users.  This can be used in cases where it would
    be helpful for the server to be online to address a problem, but there might be
    security risks in having it fully available (e.g., the server detects a
    malformed access control rule on startup, and we don't want to allow normal
    access to the server since that rule might be intended to prevent users from
    seeing sensitive information and not having it interpreted properly could be
    dangerous).
    
    This mode is designed so that server components like the access control
    subsystem can place the server in this mode if a problem is detected, but it
    also includes tasks that can be used to manually place the server into and out
    of the lockdown mode.  These tasks will only be allowed to be invoked by root
    users over a loopback connection.
    
    OpenDS Issue Number:  1758
    
    
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2002 41b1ffd8-f28e-4786-ab96-9950f0a78031
    neil_a_wilson
     

01 Jun, 2007

6 commits

  • a function that creates a backup that's used in the restore1 testcase.
    
    
    
    git-svn-id: https://svn.forgerock.org/opendj/trunk@1999 41b1ffd8-f28e-4786-ab96-9950f0a78031
    tdj_tx
     
  • 
    git-svn-id: https://svn.forgerock.org/opendj/trunk@1998 41b1ffd8-f28e-4786-ab96-9950f0a78031
    mkeyes
     
  • always based on the authentication identity rather than the authorization identity.  This
    means that when the two are different, the result could be incorrect.  One key example of
    this is the use of the proxied authorization control by a root user.  In this case, the
    proxied authorization would not be subject to access control because the authenticated
    user (but not the authorized user) had the bypass-acl privilege.
    
    This change ensures that the proxied-auth privilege is always evaluated as the
    authentication identity, but all other priviliges are always evaluated as the
    authorization identity.
    
    I have also updated a number of test cases that were incorrectly depending on the
    former behavior.
    
    OpenDS Issue Number:  1749
    
    
    git-svn-id: https://svn.forgerock.org/opendj/trunk@1997 41b1ffd8-f28e-4786-ab96-9950f0a78031
    neil_a_wilson
     
  • 
    git-svn-id: https://svn.forgerock.org/opendj/trunk@1996 41b1ffd8-f28e-4786-ab96-9950f0a78031
    mkeyes
     
  • … more user-friendly encode/decode methods:
    
    * the DurationPropertyDefinition encoder is unchanged: it uses the base property's base unit
    * the DurationPropertyDefinition decoder now supports a mixed format in addition to the old format: now users can specify durations using a mixture of units such as 1h30m instead of 90m (both are permitted)
    * the DurationUnit class has a toString(long) method which applications can use to get the above mixed unit representation (e.g. CLI)
    
    The unit tests are updated. During the precommit the password policy test cases failed because they were testing some invalid durations which were not being checked in the DurationPropertyDefinition test suite. I've put these missing test cases in the DurationPropertyDefinition test suite. In addition, I've numbered the PWP test data entries so that they are easier to debug in future.
    
    
    git-svn-id: https://svn.forgerock.org/opendj/trunk@1995 41b1ffd8-f28e-4786-ab96-9950f0a78031
    matthew_swift
     
  • 
    git-svn-id: https://svn.forgerock.org/opendj/trunk@1994 41b1ffd8-f28e-4786-ab96-9950f0a78031
    mkeyes