Release Notes for RIPE Whois Server version 3.2.0
*************************************************

The most important difference between 3.1.1 and 3.2.0 is the dbupdate
program: It was rewritten almost from scratch, to provide better
configurability, cleaner acknowledgements and notifications and
modularity.

People upgrading from 3.1.1 need to read the file UPDATE-DB-3.2.0

whois_rip (query and update core) changes:

o Transaction support is added on MySQL side to remedy occasional 
  queries that take very long time. InnoDB is used for this purpose
  and it is optional.
o ER module is decomissioned. LG module was created instead which has
  a much simpler API.
o Now the server limits the number of concurrent connections per IP.
o Revised query parsing:
  http://www.ripe.net/ripencc/pub-services/db/whois-query-2002.html
o Various bug fixes:
   - Referral queries did not pass the effective IP.
   - Syntax check of inet6num was incorrect.

dbupdate (mail processor) changes:

o dbupdate has been rewritten almost from scratch.
   ( see http://www.ripe.net/ripencc/pub-services/db/dbupdate/
    for more information )
o LU module was created to encapsulate whois look-ups.
o AU and UP modules rewritten.
o AK module is decomissioned. Its functionality is incorporated into
    NT module. Static texts for acknowledgements, notifications
    and forward messages are taken out of C source files and put into
    xslt templates whose names are specified in configuration file.
o PA module is decomissioned.
o EP module is decomissioned.
o MM module is decomissioned.
o SY module is decomissioned.
o CR module is introduced.
o KM module is introduced.
 (see modules/MODULES for short descriptions of modules)


Release Notes for RIPE Whois Server version 3.1.1
*************************************************

This release contains new features.  Please see the changes documented
on the specific software (below) for details.


People upgrading from 3.1.0 need to do the following:

- Run the normal "build.sh" and "make". 
- Use "make upgrade" to install *only* the new binaries and SQL 
  statements.
- Run the utils/ip6arpa-fix.pl script.  This fixes the data in the
  database which was incorrect for certain ip6.int and ip6.arpa domain
  entries.
- Apply an SQL patch to the schema:
    $ mysql RIPADMIN < bin/load/SQL/create.RIPADMIN.sql.patch
- Add the AUTHMETHOD, AC_AUTO_SAVE, AC_LOAD, and AC_SAVE_INTERVAL
  variables to your configuration files.

We use:

    # authentication methods supported - there is also MAIL-FROM
    AUTHMETHOD NONE
    AUTHMETHOD CRYPT-PW
    AUTHMETHOD MD5-PW
    AUTHMETHOD PGPKEY
    
    # whether to save client query access, and how often
    AC_AUTO_SAVE 1
    AC_SAVE_INTERVAL 1200

    # whether to load saved client query access on start
    AC_LOAD 1


whois_rip (query and update core) changes:

o User query tracking (access tree) is now persistent.  It is stored
  in an SQL table so this information is not lost when the server
  restarts.

o Socket input buffered.  Running as a mirror client and domain
  referrals are now much more efficient.

o Various bug fixes.
  - IPv6 DNS name mapping (ip6.int and ip6.arpa) was incorrect for
    domains that did not end on a 16-bit boundary, e.g. 2001:11::/24.
    Queries would not work properly.  This has been fixed.
  - MySQL would crash when given a name query with more than 31 names.
    Using a newer compiler fixed this bug, but it would still not
    query correctly.  Fixed.
  - If a VNRTM connection to a mirror server became unavailable (e.g.
    hardware failure, network bounce) the client would never detect
    this.  KEEPALIVE is now set on the connection, which fixes this
    problem.

dbupdate (mail processor) changes:

o Support for MD5-PW authentication added.

o The type of authentication supported by the server is now
  configurable.  For example, if you do not wish to support MAIL-FROM,
  you can prevent this.  The AUTHMETHOD configuration variable
  controls which methods are supported.

o Multiple "source:" definitions are supported for updates.  To do
  this, put an UPDSOURCE line in your rip.config for each source.

o IPv6 "status:" attribute has been changed from "generated" to the
  new community-agreed values.  Additionally, the server now enforces
  correct hierarchy for IPv6  "status:" attributes.

o Various bug fixes.
  - Support for nesting PGP and  "password:" lines now works.
  - Hierarchical filter-set, prng-set, and rtr-set names now
    properly supported.
  - dbupdate passed the authorisation checks if it could detect the
    key ID of a PGP key that signed the message, even if the signature
    itself wasn't good.  Fixed.
  - key-cert now works with networkupdate.
  - Updates from STDIN are now properly logged.
  - Various authorisation attributes, e.g. "mnt-by:", now work as
    lists.
  - routes with exact matches now properly check "mnt-by:".


utility changes:

o whois3 (client) now reads input line-buffered for efficiency.

o text_export now only produces RPSL output by default.  If you still
  need the "short" output, you can request it via the "-s" flag.


Also, this release makes it easier to customise the code for different
implementations.  There is now a defs/variants/ directory that
contains definitions of the RIPE, APNIC, and IETF variations of RPSL.
There is more work to be done making the code fully customisable,
which will be completed in future releases.

To use an alternate variant change the RPSL_VARIANT variable in the
build.sh script.



Release Notes for RIPE Whois Server version 3.1.0
*************************************************

This release adds several user-visible improvements:

- New status type for inetnum objects: LIR-PARTITIONED.
- New object type, irt, and associated new attributes: "irt:",
  "mnt-irt:", "irt-nfy:", "signature:", and "encryption:".

The SQL database schema has been changed to support these new attributes
and class.  A patch for existing databases is provided in the file
bin/load/SQL/create.irt-tables.sql.patch.  Optimised indexing has been
added, and is provided in the file bin/load/SQL/create.tables.sql.patch.

Additionally, the RPSL parsing has been totally rewritten.  The new
mechanism makes adding new classes and attributes much easier.
Modifying the syntax definitions of classes and attributes is also
much easier, and may be done in many cases by edits to XML files.
Please see the file defs/HOWTO for details on adding new classes or
changing definitions.

Some new options were added to the configuration.  If you are
upgrading, please see the HELPHEADER, MTFWHEADER, and MTFWTXT message
texts in the same configurations and add equivalent messages to your
own configuration.

People upgrading from 3.0.x need to do the following:

- Run the normal "build.sh" and "make". 
- Use "make upgrade" to install *only* the new binaries and SQL 
  statements.
- Apply the SQL patch to the schema:
    $ mysql SAMPLEDB < bin/load/SQL/create.irt-tables.sql.patch
    $ mysql SAMPLEDB < bin/load/SQL/create.tables.sql.patch
  The second patch may take a while (30 minutes or so), as it reindexes
  many tables.
- Add the HELPHEADER, MTFWHEADER, and MTFWTXT message texts to their
  configuration files.


whois_rip (query and update core) changes:

o RPSL object library takes care of object parsing and syntax checks.
  Uses CORE syntax set.  As a consequence, the server has improved
  handling of malformed objects and garbage.  Use of mirror reflector is
  now optional in cases where the NRTM server supports RPSL, even if the
  implementation uses extensions to RPSL, or is not 100% RFC-compliant.

o Error reporting in IP module has been improved, along with improved
  checks for meaningfulness, proper alignment for IP address related
  attributes (inetnum, route, etc.).

o IP6.ARPA domains are stored in the radix tree and available for search
  with -d flag.

o A new query flag, "-c" has been added to the query path to support
  searches for inetnum or inet6num objects with "mnt-irt:" attributes.

o Fix occasional server abnormal shutdown when client disconnects
  unexpectedly.

o "-i person" flag now works as "-i pn", both now also check for
  "cross-nfy:" references.

o Now reads defaults from /etc/my.cnf, needed for connecting to
  non-standard MySQL configurations.

dbupdate (mail processor) changes:

o dbupdate has been substantially changed to use the new API parsing
  library.

o It has also been converted to plain C instead of a mix of C and C++.
  All the old C++ files have now been obsoleted and no longer form
  part of the release.  This includes the whole of the src/ sub
  directory in the "up" module which contained all the old C++ parsing
  functions.  References to src/ have been removed from the Makefiles.

o Where the code has been changed to use the new API library, the
  logic is basically the same, but the implementation has changed.  A
  line by line review ofthe code from the 'process_file' function in
  dbupdate.c downwards has been carried out during this change.
  Numerous, previously un-noticed, little bugs were caught and fixed
  in this process.  The basic structure of dbupdate is un-altered.

  Checks for valid sources, countries and nic-hdl suffixes are now
  handled by additional functions in UP_extrnl_syntax.c.  These were
  previously handled by the old parsing routines.  *The valid nic-hdl
  suffixes are hard coded in dbupdate.c for the moment.*

o Additional code has been added to handle the LIR_PARTITIONED status
  and irt objects.

o PGP fixes: 
  - A bug that allowed creation of key-cert objects that contain
    multiple PGP keys has been fixed.
  - Works with emacs PGP mail (emacs being the *only* RFC-compliant
    MUA).

o Various bug fixes:
  - A bug was fixed where inet6num generated attributes were only
    generated on updates and not on creation.
  - Multiple lines are now handled correctly in any of the email
    header fields. Also the Cc: field is shown in the comment block
    section of an ack message.
  - A help request no longer returns error messages in the ack
    message. This also required a change to the HELPHEADER message
    texts in rip.config files.
  - Some corrections were made to the authorisation checks of route
    creation where it deviated from the flow chart shown at:
      http://www.ripe.net/ripencc/faq/database/route-creation-checks.html
    There are still some minor deviations from this flowchart and this
    will be looked at more closely for a future release.
  - A lot more of the malloc'd memory is now freed than before.  This
    does not really matter as dbupdate always terminates after each
    update, but it is more correct to free the memory whilst the
    program is running.
  - Lots of point fixes, e.g. use of mail text as format specifier in
    printf-style command, memory allocations too small, glist double
    free, etc.

o A whole new test suite has been developed for extensive testing of
  dbupdate.  A series of tests have been carried out using pre-defined
  objects with expected results and also comparison testing between
  the new and previous releases using real data from the update
  logs.  This testing has been done more thoroughly and systematically
  than with any previous release of dbupdate.  This test suite and
  pre-defined data sets need some more work on them before we can
  include them in the release, but we will be able to include them in
  the next release.

o Some improvements may be noticed in the error reporting, but a major
  overhaul of error reporting will be done for a future release.


whois3 (client) fixes:

o Bug where two blank lines caused client to stop output fixed.

mr (mirror reflector) fixes:

o Removed bug attempting to write to non-connected socket on name lookup
  failure.


Release Notes for RIPE Whois Server version 3.0.2
*************************************************

This release is intended both as a bug fix release, and also to offer
much-improved ease of installation and configuration.

whois_rip bug fixes:

o Arbritrary sprintf() calls changed to g_string_sprintf() to prevent
  buffer overflows.

o Fixed double-close of file in NRTM code causing numerous errors
  throughout the application.

o Several queries fixed:
  - The "-i nserver ip-address" query now works properly.
  - The "-K as-set-name" query now works properly.
  - The "-?" option now presents help.

o Query keys are no longer truncated to 63 characters.

o The nic-handles of dummy objects (objects that were created in the
  database to resolve non existing references) were not stored in the
  NIC handle repository (NHR). So sometimes an AUTO NIC handle request
  failed with an error 32 (conflicting operation): whois_rip had picked
  up the next available NIC handle, but it was used in a dummy object
  already, thus the creation mysteriously failed. This was fixed by
  putting the NIC handles of dummy objects into the NHR.

o All compiler warnings reviewed, most fixed.

o Numerous fixes for Linux and FreeBSD portability.


dbupdate bug fixes:

In general, better error reporting in acknowledgement messages:

o When the update message contained no objects, we didn't print a
  warning about this in the acknowledgement message.

o When a name change of a person/role object was attempted, the error
  generated was "Authorisation failed". The error message is now
  clearer.

o The status of update was not mentioned (SUCCEEDED or FAILED) in the
  subject line of the acknowledgement.

o Sometimes empty cross-notification messages were sent.

o We now filter out the paragraphs which are obviously not objects (like
  signatures etc.) so that the acknowledgement messages won't overflow
  with syntax error messages.

Some other fixes in dbupdate:

o Multiple-line "From:" in the update mail header used to confuse
  dbupdate.

o Updating a person/role object whose NIC handle appears in some other
  person/role object's name was sometimes not possible.

o Sometimes it didn't send maintainer requests to HUMAILBOX.

o An initialisation problem in MM module solved which caused core dumps
  in rare conditions.

o Numerous fixes for Linux and FreeBSD portability.


Configuration:

o The initial installation and configuration has been much simplified
  and in other ways improved.  If you have used previous releases,
  please review the new INSTALL guide.


Portability:

o We have tested the release on Solaris 8, Red Hat Linux 7.2, and
  FreeBSD 4.3 RELEASE.  Please refer to the file README.solaris,
  README.linux, and README.freebsd for platform-specific notes.



Release Notes for RIP version 3.0.1
***********************************

This is primarily a bug fix release.

whois_rip bug fixes:

o In update mode when stopping updates dbupdate connections in the
  backlog queue were not processed and dropped later when server was
  stopped.

o '\t' was not recognized as a valid whitespace. So some attributes
  containing several words were not stored correctly in the back-end
  database.

o When in VNRTM mode the mirror client enters a loop requesting serials
  without any delay between the attempts, when a requested range doesn't
  exist or server returns an error. Thus flooding the mirror server with
  requests.

o Object templates were corrected in accordance with RIP object
  definitions.
  
o Server will now timeout on socket writes, if client is unresponsive.

o Lookups for netname, mntner, and limerick now corrected to allow
  underscore.


dbupdate bug fixes:

o The forward message of a networkupdate didn't contain the IP of the
  updating client.

o When the update failed with authentication failure, the
  acknowledgement message didn't show the object.

o Classless inetnums were accepted. Not any more.

o When updating key-certificate objects, generated attributes used to
  disappear.

o E-mail address from the "Reply-To" field of an update message
  was not extracted correctly.

o The parantheses were not taken as comments in e-mail addresses.

o In cross-notifications, the "exact matches" and "less specifics"
  parts listed the route itself.

o A '\'', '.' or '`' character wasn't possible in person/role
  objects.

o Dates in changed attributes were not checked properly.

o dbupdate core dumped when deleting some syntactically incorrect legacy
  objects.


Features:

o A mirror server never returns the latest serial when serving mirroring
  request. So if the latest serial causes server crash and database
  corruption, it never propagates to the secondary servers. The latest
  serial is not displayed with -q sources query as well.


Tools:

o A new Whois client, whois3, is now included with the release.  This
  client offers ability to query in persistent mode, reading multiple
  queries from standard input and sending them to the server.

o The ripe2rpsl script has had support for optional country-code
  attribute added to most classes.



Release Notes for RIP version 3.0
*********************************
The major changes with this version are:


- CA module
    All configuration variables are currently mandatory.  

- UP module
    Implemented networkupdate.

    Subject line parsing and processing has been implemented.

    Several bug fixes.

- UD 
    Improved logging when performing crash recovery

- SV 
    SIGUSR1 is used to pause/resume updates. Normally, command interface
    is used.

    SVSIGTERM and SIGINT are both used to shutdown the server

- RP/RX
    Reloading of the radix trees for a given source is implemented


- QI
    Query path no longer core dumps when MySQL returns an errors.

    We uncovered a couple of MySQL bugs, one of which caused MySQL to
    incorrectly return an error.  Should this (or a different) bug occur
    in the future, the server will continue to answer queries as best it
    can.

    Blank line removed between banner and copyright message.  For a
    valid query, the banner and copyright appear as one comment block.

    Temporary tables configured automatically.

    Version of MySQL detected at compile and TEMPORARY tables now used
    if possible.

- PW
    Ability to pause queries.  Useful for doing manipulations of the
    MySQL database "under the hood" without taking RIP off-line.


Release Notes for RIP version 3.0beta3
**************************************
There are two major changes with this version:

- With MySQL 3.23 you now have the option to use TEMPORARY tables.
  There are two advantages of this:

    1. The tables are automatically deleted when a connection is closed.

    2. If the TMPDIR environment variable used by the MySQL daemon is
       set to a directory mounted in memory, then there is no disk I/O
       for temporary tables.  This is much faster - we actually can't
       support our query load without it!

  To use this, you must do a quick source modification today.  This will
  be changed to a configure option in the future.  To configure this,
  edit modules/qi/query_instructions.c, and add the line:
 
    #define MYSQL_3_23

  To the top of the file before you build RIP.

- A major memory leak on Solaris 8 has been fixed.  The
  regcomp()/regfree() pair did not free memory.  As a workaround,
  regcomp() is only called at startup.  This should also be faster.

Other bug fixes:

- CLIENTADDRESS referral now properly passes addresses

- AS-SET queries with ':' now work (hopefully)

- The "-m" queries now work.

- You can use '-' in the source variable now.

- Improvements in acknowledgement and notification message formats.
  
- Fixed multiple notification problems.
  
- Fixed memory allocation bug in dbupdate.
  
- Fixed bug which prevented the last attributes of the object
    with references to auto NIC handles from appearing in the 
    database.
  
- Refering to invalid auto NIC handles is reported in the
    acknowledgement message as an error.

UD:
- Object parsing is reimplemented improvind performance and fixing
several bugs related to line continuation and list valued attributes.
- New mode of operation is supported when NHR is not updated. This
allows to mirror sources with less strict syntax rules regarding
nic-handles. This also improves performance in such cases.

SK:
- buffer management bug in SK_gets




Release Notes for ripe-dbase version 3.0beta2
*********************************************


This is mainly a bugfix release. There has been however an important
change in the installation procedure: the package has been "autoconfiscated",
i.e. it uses the autoconf facilities. This is a step that will ease
the work on portability, still at its early stage.

Please refer to the INSTALL file for the new instructions on how
to compile the package.


Changes From Last Release
-------------------------


Installation/Maintenance/Administration:

o The package has been autoconfiscated. A configure script is used
  to set several site-dependent parameters.

o The dictionary.txt file is no longer necessary; it has been hardcoded
  inside the source of the configuration module.


Dbupdate:

o Syntax checks of phone & fax attributes have been fixed.

o Checks of changed attribute: the current date is added if
  a date is missing in a 'changed' attribute. The order
  of dates is also checked: if the dates are not in order,
  then the update is rejected.

o Object formatting: The objects are "beautified" before
  being inserted into the database. Extra whitespace is
  added after the type name of the attribute.

o in dbupdate, another switch, '-S', has been implemented. 
  When this switch is used, the notifications and acknowledgements
  are sent to a predefined address (DEFMAIL configuration
  variable).  This is for testing and debugging purposes.

o The MAIL-FROM authentication scheme now accepts
  regular expressions.


Queries:

o Several little bugfixes here and there (referral, radix tree,...)



Known Bugs and Limitations
**************************


Installation/Maintenance/Administration:

o You cannot use "-" characters in paths for the definitions
  of logfiles.

o Several side-utilities still need to be implemented.

o Portability: there hasn't been much work yet on the
  portability issue. This package is known to
  work on Solaris 2.6,7 and 8, on SPARC and Intel
  architectures.


Dbupdate:

o The subject line is not parsed.

o The "networkupdate" utility is to be implemented.

o The modified objects appear doubled in notifications.


Queries:

o The "CLIENTADDRESS" referral does not pass addresses as it should.





Old Release Notes
=================

Release Notes for ripe-dbase version 3.0beta1
*********************************************


Introduction
------------

We are going v3.

The package has been renamed, from RIP to ripe-dbase v3.0 .
The reason is that all the core functionalities of the server
are now in place; what is needed now is bugfixes and work
on the infrastructure. But most of the work is done.

This is the beta of the server which will eventually be put
into production for the RIPE Database and replace the old 
software ripe-dbase v2.3.2 .


Changes From Last Release
-------------------------


Installation:

o "make install" now available.


Queries:

o -F mode now uses a "+ " string in line continuation.
o -d option extends radix-style searches to reverse domains,
  including IPv6 and not-fully-parsable domains.
  Takes IP/prefix/range as argument.


Updates:

o Cross-notifications have been implemented.
o Improved acknowledgements:
  a header summary of the update message is included.
o Implemented crash recovery in both dbupdate and the update server.
  In case of server crash, a full recovery is done, including
  committing/rollbacking the crashed transaction, sending outstanding
  acknowledgements, and bringing the database into consistent state.
o New mode of operation (opMode) for managing a static source with no
  possible updates.


Maintenance/Administration:

o Added admin commands to stop query in progress, unblock an automatically
  denied user, edit the logging filters, pause/resume updates,
  switch to whois mode.
o Almost all parameters and texts are now configurable through the
  configuration file.
o There is now a separate configuration file for sources, with a
  clearer syntax.
o Introduced flexible logging system, configured via the config file.
  The various parts of the software send logs via this system.


Infrastructure:

o Introduced hs_cleanup, a utility to archive and clean up the SQL
  database of old serials and history entries.
o Introduced text_export, a utility to export a database in flat
  files that can then be used to load the database on another server.



Release Notes for RIP version 0.5.0beta1
****************************************


Changes From Last Release
-------------------------


Queries:

o -k quits session only if called with no arguments
  (required by RtConfig)
o -R referral fix: the local object containing "refer:"
  attribute is returned only when -R is specified
o -i upd-to and -i mnt-nfy have been implemented


Updates:

o The PGPKEY authentication scheme is now supported.
  PGP messages are handled by the software.
o There are no more hardcoded values in dbupdate.h .
o The "Reply-To:" fields of incoming messages (if they
  exist) are used to send the acknowledgements.

Plus numerous bugfixes.



Release Notes for RIP version 0.4.2beta1
****************************************


Changes From Last Release
-------------------------


o This is mainly a bugfix release. Several
  bugs have been fixed, including memory leaks.


Release Notes for RIP version 0.4.1beta1
****************************************


Changes From Last Release
-------------------------


o This is mainly a bugfix release.
o A basic whois client, whoisRIP.c, has been added
  to ease the task of querying a RIP database which
  supports new flags in respect to the allowed flags
  of a "classic" RIPE whois client.



Release Notes for RIP version 0.4.0beta1
****************************************


Changes From Last Release
-------------------------


Queries:

o query cancellation: closing the socket aborts the search to save
  resources. Note for -k: all data entered before the server finishes
  answering the previous query will be discarded
o query read timeout: the server closes after no data has been received
  for a specified period of time
o -i member-of displays only members validated by the mbrs-by-ref
  attribute
o -q sources displays the available sources
o thread accounting: the currently active threads can be displayed in the
  admin interface with "show threads"
o input filtering: disallowed characters are dropped from the query
o closing session on error: -k mode connection is closed on error


RIP server:

o Multiple repository support by a single server. 
  That means that one can e.g. NRT-mirror RIPE and RADB 
  and update TEST database running only one [super]server. 
  Operational parameters are defined in configuration file.
o Improved support of RPSL syntax (line continuation, 
  End-Of-Line comments, etc.).
o The server can mirror _only_ RPSL stream now. If someone wants to
  NRT-mirror a RIPE181 database, he needs to run a "mirror reflector"
  and use it as a NRTM server. This tiny program (bin/mr) connects to
  a specified NRTM server and accepts requests from RPSL NRTM client. 
  After receiving a request it converts the data stream on fly using
  a specified convertor.
o While mirroring failed transactions are stored for re-distribution
  anyway. This should improve NRTM server reliability and allow
  re-distribution of data from other repositories not fully compliant with
  current implementation (e.g. different level of ref.integrity, different
  NRTM protocol).
o Different handling of deletion of a set. Now this operation is not
  blocked even if the set is referenced from other objects by member-of
  attribute. In order not to violate DB integrity, the set is actually not
  deleted but converted into a dummy object, not visible outside.


Updates:

o More informative acknowledgement messages.
o MIME support added.
o Notifications (except cross-notifications) are in place.
o The "test" mode is implemented. In this mode, users can create mntner 
  objects without overriding. 
o Most of the configuration variables implemented.


Configuration:

o The configuration module is now in place. It is not fully integrated
  in the modules yet, so in this release the configuration is mixed:
  some variables must still be set in the PROPERTIES file, and a small
  number is still to be modified in the header files.




Release Notes for RIP version 0.3.1beta1
****************************************


Changes From Last Release
-------------------------

o There have been big changes in the Makefile structure.
  The hardcoded variables have been much reduced, and
  currently there is only one site-dependent Makefile
  where to change variables and paths for the user.
  Only the update module still needs some changes 
  in a particular header file in order to work in
  your environment. The up module unly affects dbupdate.
  This makes the package easier to compile and run.
o A number of bugfixes.
o Backend work.
o as-block queries are now supported.




Release Notes for RIP version 0.3.0beta2
****************************************


Changes From Last Release
-------------------------

o All the queries for the "old" object types
  are now supported. IPv6 queries have been implemented.
o in-addr.arpa queries are now supported.
o Full support of -K, -t and -v queries.
o The Access Control System is fully functional.
o The expansion of person objects referenced by name
  is working.
o The update functionality has been integrated.
o Lots of bugfixes and backend improvements.



