root / pykota / trunk / NEWS @ 3086

# $Id$

PyKota - Print Quotas for CUPS and LPRng

(c) 2003, 2004, 2005, 2006 Jerome Alet <alet@librelogiciel.com>
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.

============================================================

PyKota NEWS :
       
    - 1.25alpha18 (2006-11-30) :
    
        - Fixed a severe bug in the LDAP backend wrt the overcharging factor.
        
    - 1.25alpha17 (2006-11-27) :
    
        - pkrefund can now create receipts in the PDF format.
        
        - The code was changed to accomodate ReportLab 2.x in pkbanner
          and pkinvoice. ReportLab v1.x is now only supported with 
          an english or C locale.
          
        - pknotify now accepts the --noremote command line option to
          decide what has to be done when there's no PyKotIcon running
          or accepting connections from the print server on the client host.
          
        - pkprinters now only modifies CUPS' printers.conf when the
          -C | --cups command line switch is used.
          
    - 1.25alpha16 (2006-11-12) :
      
        - Several fixes to CGI scripts wrt i18n.
        
        - Introduced the pkrefund command line tool to refund jobs.
        
        - Several i18n fixes.
        
        - Several minor fixes to documentation and translations.
        
    - 1.25alpha15 (2006-11-06) :   
    
        - pkprinters --add and --delete now automatically reroutes CUPS
          print queues through PyKota or through CUPS only.
          
        - Fixed date and time parsing problems.
        
        - Finally simplified the locking code again, and checked that no
          stalled lock can persist. The lock files remain on disk though,
          but it is a small price to pay.
          
    - 1.25alpha14 (2006-10-13) :
      
        - Removed some locking code which would have prevented multi-server
          configuration to work correctly with a shared lock directory.
          
        - First useable release with the computation of ink
          usage, through the introduction of the 'ink(cspace, resolution)'
          value for the 'preaccounter' and 'accounter' directives in
          pykota.conf
          
    - 1.25alpha13 (2006-10-04) :
     
        - The dumpykota.cgi script now does basic form checking
          to immediately avoid incompatible dump options and
          alerts the end user if Javascript is enabled.
          
        - The filter keys in dumpykota can now be specified in
          lower or upper case.
          
        - Introduced the 'noprintingmaxdelay' directive to workaround
          some buggy printers when hardware accounting is used.
          
        - Improved the reliability of hardware accounting.  
          
    - 1.25alpha12 (2006-09-15) :
    
        - Fixes encoding of username and billingcode when they
          are overwritten by a subprocess.
          
    - 1.25alpha11 (2006-09-06) :
    
        - Now fails in an helpful manner if python-mysqldb is too old.
        
        - Now exports the original username and billing code (i.e.
          their values before they are eventually overwritten) in
          two additionnal environment variables.
          
    - 1.25alpha10 (2006-09-03) :
    
        - Rewrote the locking mechanism to try to fix the problem
          some people reported.
          
        - Permissions on cupspykota are now correctly set after
          installation.
          
        - Added a workaround to incorrect charset through the add
          of the python-chardet module as a dependency.
          
    - 1.25alpha9 (2006-08-10) :
    
        - Fixed a problem when an SMTP error occured and the recovery
          routine failed to handle it correctly.
          
    - 1.25alpha8 (2006-07-11) :
    
        - Fixed a problem in the MySQL schema.
        
    - 1.25alpha7 (2006-06-20) :
    
        - Fixed some date formatting problems when retrieving the 
          printing history with MySQL.
          
        - Began code cleaning.  
          
    - 1.25alpha6 (2006-06-13) :
    
        - pkusers now supports the --email command line switch.
          The old method to set the email address is still supported,
          but the new one allows the email address to be modified
          as well.
          
    - 1.25alpha6 (2006-06-09) :
    
        - Fixed the incorrect overwriting of the original billing
          code if the overwrite_jobticket directive was used.
          
    - 1.25alpha5 (2006-06-03) :
    
        - Fixed minor issues with pkipplib's usage.
        
        - Fixed MySQL schema to workaround some "features"...
        
    - 1.25alpha4 (2006-05-23) :
    
        - Introduced the 'askconfirmation' directive to interact with
          the end user through any script (pknotify is recommended)
          and see if he really wants to print when knowing the cost
          of the job.
          
        - Now uses pkipplib if present to directly retrieve the job
          information from the CUPS server over IPP.
          
    - 1.25alpha3 (2006-05-11) :
    
        - Added a lock to ensure that a single cupspykota backend
          is processing datas for a particular print queue at any
          given time.
          
        - Now exports more environment variables for the subprocesses.
          
    - 1.25alpha2 (2006-05-03) :
    
        - Fixed some problems with DateTime objects.
        
        - Added support for PySNMP v4.x in addition to 3.4.x.
        
    - 1.25alpha1 (2006-04-14) :
    
        - Make the MySQL schema work with old releases of MySQL.
        
        - Fixed several character encoding problems.
          
        - Ensures that the denied banners counter is properly initialized
          with relationnal databases.
          
    - 1.24 (2006-04-08) :
    
        - Improved documentation and packaging stuff.
        
    - 1.24beta (2006-03-26) :
    
        - Finalized the pknotify command.
        
        - Updated the messages catalogs for translations.
        
    - 1.24alpha18 (2006-03-15) :
    
        - Now uses MIME encapsulation for email messages to avoid
          some problems with 8 bits messages being rejected.
          
    - 1.24alpha17 (2006-03-06) :
    
        - Fixed a major problem in the PostgreSQL backend.
        
        - pknotify can now check a remote user's name and password.
        
    - 1.24alpha16 (2006-03-02) :
    
        - Introduced the pknotify command line tool to dialog
          with remote PyKotIcon applications.
          
    - 1.24alpha15 (2006-03-02) :
        
        - optimized pkbcodes, pkprinters and pkusers using edpykota as 
          the model.
        
        - 'edpykota --list' and 'pkusers --list' are now authorized 
          to mere mortal users, but the list is restricted to informations
          about the current user.
        
        - Extended the functionnality of the 'onbackenderror' directive to
          allow for configurable retries.
          
    - 1.24alpha14 :
    
        - Finished optimizations in edpykota and fixed some minor
          problems with LDAP.
          
    - 1.24alpha13 :
    
        - Introduced the pkusers command line tool to manage users and  
          users groups. This command doesn't deal with print quota entries,
          which continue to be managed through the edpykota command line
          tool.
          
        - Major rewrite and incompatible changes to command line options
          in edpykota. This command now only deals with print quota
          entries. Printers, users (or users groups) now have to be created
          first using pkprinters or pkusers.
          
    - 1.24alpha12 :
    
        - Introduced the 'duplicatesdelay' directive to specify after
          which duration two duplicate jobs are not considered duplicate
          jobs anymore.
          
        - Introduced the 'balancezero' directive to specify any value  
          as the balance limit instead of 0.0. This can allow admins to
          give a few free credits to users, for example.
          
    - 1.24alpha11 :
    
        - Huge speed improvements when modifying billing codes and
          printers.
        
        - Both pkbcodes and pkprinters now display the percent done
          for all actions.
          
        - The pkinvoice invoice generator command line utility is now
          included.
          
    - 1.24alpha10 :   
    
        - Huge speed improvements when deleting users, printers, and
          billing codes.
          
    - 1.24alpha9 :
    
        - Preliminary support for MySQL, thanks to Matt Hyclak.
        
    - 1.24alpha8 :
    
        - The 'preaccounter' configuration directive was introduced to control
          which parser to use for precomputation of the job's size. When unset,
          the internal parser is used just like it was previously. See
          conf/pykota.conf.sample for more details.
          
        - Now subprocesses launched through the overwrite_jobticket directive
          can also output 'CANCEL', to represent the user's own choice to
          cancel the print job.
          
    - 1.24alpha7 :
     
        - pkbanner now handles new print limiting modes correctly.
        
        - The cupspykota backend wrapper now accepts to be interrupted
          with a kill -INT. This causes an immediate and clean abort
          of the current print job.
          
        - Matt Hyclak contributed the possibility to store the printing
          history in a MySQL database. See the contributed/ directory
          for details.
          
        - Added 2006 to the copyright's years.
        
    - 1.24alpha6 :
    
        - Fixed internal SNMP and PJL handling code for very fast
          printers : PyKota didn't see them going from 'idle'
          to 'printing' to 'idle' again.
          
        - Now waitprinter.sh accepts SNMP version number and 
          community name as arguments, thanks to Brandon
          Holbrook.
          
        - Now command line errors and Ctrl+C repectively cause
          PyKota commands to exit with status -2 and -3.
          
    - 1.24alpha5 :
    
        - All command line tools now accept -A | --arguments filename.args
          as a way to bypass the shell's maximal command line size :
          just put each argument on a line by itself in a file, and
          pass the file's name as the -A | --arguments command line
          option's argument. For now this won't be documented in
          commands' helps.
          
        - dumpykota can now dump the complete database in XML 
          format.
          
    - 1.24alpha4 :
    
        - Added support for the SQLite3 backend.
          See pykota/initscripts/sqlite/README.sqlite
        
    - 1.24alpha3 :
    
        - Fix for mailandpopup.sh : smbstatus can't be run setuid, 
          so we regain priviledges before running an external 
          mailto directive.
        
    - 1.24alpha2 :
    
        - Introduce the "onbackenderror" directive in pykota.conf, to
          control if the user must be charged or not whenever a real
          CUPS backend fails. The default value is 'nocharge'.
          Set it to 'charge' if you think your users can cause a CUPS
          backend to fail.
          
        - Now mark original backends related problems as such in the 
          printing history. Modified printquota.cgi and pykota.css
          to make them be more visible.
          
    - 1.24alpha1 :
    
        - Fixed a problem which caused backend errors to be ignored.
        
        - The sample LDIF file now contains the two necessary user
          entries to have an LDAP PyKota Admin and an LDAP PyKota
          ReadOnly user. The sample configuration files were updated
          to contain the default passwords used in the LDIF file.
          
        - dumpykota is now more strict when checking arguments.   
        
    - 1.23 :
    
        - Many minor bug fixes and code cleaning.
        
        - Improved documentation.
        
        - Added Polish translation.
        
        - Improved French and Swedish translations.
        
    - 1.23beta :
    
        - Now the mailto directive is honored in all cases where the job is
          denied or a warning has to be issued, and the correct reason
          can be sent back to the user.
          
        - The PYKOTAREASON environment variable, if defined, contains the  
          textual reason why the job was denied or a warning has to be
          issued.
        
    - 1.23alpha31 :
    
        - Added support for --maxjobsize, --passthrough and --nopassthrough
          in pkprinters. These modes are now also honored when printing.
          
        - Allows the admin to restrict the number of print queues which will
          be treated by pkturnkey by passing queues' names as additionnal 
          arguments on pkturnkey's command line.
          
        - pkturnkey now supersedes pkhint with the help of the -c | --doconf
          command line option.   
          
    - 1.23alpha30 :
    
        - Database changes : you MUST use upgrade-to-1.23.sql again if you
          use PostgreSQL, or copy the new schema and restart your LDAP
          server if you use LDAP.
          
        - The -C | --comment command line switch was added to edpykota to  
          attach textual information to modifications done to a user's
          account balance.
          
        - The -l | --limitby command line switch to edpykota was modified  
          to allow 'noquota', 'nochange', and 'noprint'. See the edpykota
          command's help or manual page for details.
          
        - The -i | --ingroups command line switch to repykota allows  
          PyKota Administrators to only lists members of the specified
          group or groups (separated with commas).
          
        - The -I | --increase command line switch to edpykota allows
          to increase or decrease page limits.
          
        - Added the precomputed job's size and price to the history
          for controlling purpose.
        
    - 1.23alpha29 :
    
        - Fixed a small problem with billing codes.
        
        - Exports some environment variables earlier, to render the script
          launched by the overwrite_jobticket directive more powerful.
          
        - Now autopykota also creates print quota entries for the user on the
          printers groups the current printer is a member of.
          
    - 1.23alpha28 :
    
        - Added users groups management to pkturnkey.
        
        - Fixed umask in setup script.
        
    - 1.23alpha27 :
    
        - Replaced the --dryrun command line option by the --force
          command line option in pkturnkey, with exactly the opposite
          effect (for security reasons).
          
    - 1.23alpha26 :
    
        - Allows parametrized pjl and snmp hardware accounting methods,
          to be able to respectively set the TCP port and the SNMP
          community.
          
    - 1.23alpha25 :
    
        - Added a --dryrun command line switch to pkturnkey.
        
        - Improved pkturnkey's general usability.
        
        - Added a manpage for pkturnkey.
        
    - 1.23alpha24 :
    
        - Allows different databases to be used depending on the
          user's rights. This allows PyKota admins to directly
          use a master LDAP server for read+write, while letting
          normal users using the replicas for readonly access.
          NB : using different storage backends is also possible,
          but you're on your own there.
          
    - 1.23alpha23 :
    
        - Fixed a problem with banner accounting.
    
        - Introduction of the new 'pkturnkey' command line tool.
        
    - 1.23alpha22 :
    
        - The cupspykota backend was almost entirely rewritten from scratch.
          It's now much more readable, maintainable, etc...
          
        - LPRng support is OFFICIALLY dropped for now. LPRng users are  
          advised to download earlier releases or purchase the
          PyKota v1.22HotFix1 Official release. LPRng support 
          MIGHT be re-added later.
        
    - 1.23alpha21 :
     
        - Fix for IPP requests which contain more than a single 
          job-originating-host-name attribute.
          
    - 1.23alpha20 :
    
        - Added the overwrite_jobticket directive. See pykota.conf.sample
          for how to use it.
          
    - 1.23alpha19 :
    
        - Fixed an LDAP filtering problem when managing billing codes.
        
        - Added the unknown_billingcode directive. See pykota.conf.sample
          for how to use it.
        
    - 1.23alpha18 :
     
        - Added the dump_bfd.php contribution. 
        
        - Several fixes in the internal PJL accounter.
        
        - Now autopykota doesn't create unneeded print quota entries anymore.
        
        - Increased the timeout in the PJL hardware accounting code to
          accomodate some printers (HP5000 at least).
          
    - 1.23alpha17 :
    
        - pkpgcounter is now distributed separately only. This
          means that people must now download and install pkpgcounter
          v1.56 or higher prior to being able to run PyKota.
          
        - Removed all the autoconf/automake stuff which was outdated  
          and confusing.
          
    - 1.23alpha16 :
      
        - Extracted the GPL blurb from all commands' help, and use
          a static version only once. This saves space big time, and
          eases future translations.
          
    - 1.23alpha15 :
    
        - pkbcodes now works with the PostgreSQL backend.
          NB : PLEASE RUN upgrade-to-1.23.sql AGAIN !!!
        
    - 1.23alpha14 : 
    
        - Logs a warning when the user/group doesn't exist when running
          edpykota.
          
    - 1.23alpha13 :
    
        - Minor fix for user groups because of latest database changes.
        
    - 1.23alpha12 :
    
        - Big database changes, use the migration scripts !
        
    - 1.23alpha11 :
    
        - Small fix to the PCL3/4/5 parser for some drivers.
        
    - 1.23alpha10 :
    
        - Fixed a bad exception name in cupspykota when parsing IPP messages.
        
        - Improved the IPP messages module : it can now generate IPP messages
          in addition to parsing them.
        
    - 1.23alpha9 :   
    
        - The denyduplicates directive now accepts a command line as an
          argument beside yes/no. The command is launched and must
          print either ALLOW or DENY to its standard output. Then the
          job gets allowed or denied by PyKota.
          This will allow user interaction in the near future.
          
        - Introduced the new 'striptitle' directive to remove ugly prefixes
          from print job's titles whenever a new print job is submitted.
          
        - Fixed internal SNMP accounter for printers which only have a life time
          page counter, and not a volatile page counter.
          
    - 1.23alpha8 :
    
        - The data dumper can now summarize payments and history with the
          help of the --sum command line option (or checkbox in the CGI).
          
        - The message used when the job's filename, title and options are 
          hidden because of the privacy directive is now translated at
          retrieval time, instead of at saving time.
          
        - Minor fix in the LDAP backend code.  
          
    - 1.23alpha7 :
    
        - Improvements to the internal PJL and SNMP accounter : now should
          support more printers and waste less time.
          
    - 1.23alpha6 :
    
        - The data dumper's start= and end= date filtering options now
          accept a positive or negative offset to be added to the special
          keywords 'yesterday', 'today', 'now', 'tomorrow'. 
          e.g. : dumpykota --data history start=yesterday-5 end=today
          
    - 1.23alpha5 :
    
        - The data dumper's start= and end= date filtering options now accept 
          the special keywords 'yesterday', 'today', 'now' and 'tomorrow'.
          
    - 1.23alpha4 :
    
        - The data dumper now accept start= and end= to specify the starting
          and ending dates to dump the history.
          
    - 1.23alpha3 :
    
        - The destination link for the logo used in CGI scripts is 
          configurable.
          
        - Data dumper improved : it's faster and now datas are correctly 
          escaped.
        
        - The URL to the logo displayed in the CGI scripts is configurable.
        
    - 1.23alpha2 :
    
        - Fixed edpykota --prototype with balance.
        
        - Major rewrite of the IPP message parsing code to definitely
          fix the job-originating-hostname IPP attribute problem.
        
    - 1.23alpha1 :   
    
        - Sets current username to the user CUPS runs as instead of "root"
          when printing test pages from CUPS' web interface (because
          the username is empty in this case).
          
        - Fix for strange values in job-originating-hostname
        
    - 1.22 : 
    
        - Big improvements to the documentation.
        
    - 1.22beta :
    
        - Improved stability in pykotme and pykotme.cgi.
        
        - Fixed a permission problem in pykotme and pykosd.
        
        - Fixed an inconsitency between SQL and LDAP backends'
          behavior in edpykota.
          
        - Improved stability and allowed tracebacks in CGI scripts.
        
        - Improved pykotme.cgi
          
    - 1.22alpha8 :
    
        - Fixed history retrieval for nonexisting printers or users.
        
        - Clicking on the billing code in printquota.cgi's job history
          now shows all jobs matching this billing code.
          
        - Fixed the dump in CUPS' page_log format which didn't
          dump the oldest entry.
          
        - The data dumper now accepts billingcode= and hostname= as
          filters when dumping the job history.
          
        - The job billing code if present is now saved into the database,
          and exported to subprocesses as the PYKOTAJOBBILLING environment
          variable.
          
        - Ctrl+C now makes PyKota commands exit without a traceback.
        
        - Fixed an i18n problem with Python 2.1
        
    - 1.22alpha7 :
    
        - Fixed some problems with the retrieval of the history from LDAP.
        
    - 1.22alpha6 :
    
        - Early errors now produce meaningful error messages (finally !)
        
    - 1.22alpha5 :
       
        - SNMP and PJL improvements with regard to PowerSave / StandBy.
        
    - 1.22alpha4 :
    
        - Fixed charset problem under MacOSX.
        
        - Allows pkbanner to accept more info to display on its 
          command line.
          
    - 1.22alpha3 :
    
        - Fixes some case related problems with some LDAP backends.
        
        - Fixes some problem for printers which directly go from
          "idle" to "warmup" and never to "printing" before
          returning to "idle".
          
        - Some permissions settings may have prevented cupspykota
          from parsing cupsd.conf, this is now fixed.
          
    - 1.22alpha2 :
    
        - Introduction of the web enabled print quote generator.
        
        - Fixed a potential race condition when several backend
          wrappers were present in CUPS' backend directory.
          
        - Fixed dumpykota's output in CUPS' page_log format to better
          match CUPS format.
          
    - 1.22alpha1 :
    
        - Import into Subversion.
        
    - 1.21 :
    
        - Improved documentation.
        
        - Improved dependencies checker.
        
        - Updated some translation files.
        
    - 1.21beta2 :
    
        - Fixed an LPRng specific bug.
        
        - Several improvements on stability issues with manually modified
          LDAP trees.
          
        - Documentation improvements.  
        
    - 1.21beta :
    
        - Prepared translation files for next release.
        
    - 1.21alpha27 :
    
        - Allows software() (no arguments) to the software accounting
          directive. This saves a lot of CPU since the job is not
          parsed again : the precomputed size is used directly in this
          case.
          
        - epykota's --prototype option now works fine with accounts
          limited by balance.
          
    - 1.21alpha26 :
    
        - Fixed the handling of the 'maxdenybanners' directive.
        
        - Introduced the 'denyduplicates' directive.
        
        - Fixed some database related glitches.
        
    - 1.21alpha25 :
    
        - The job's size is now always precomputed, even when
          'enforcement' is not set to 'strict'.
          
        - Introduction of the new 'trustjobsize' configuration directive
          to workaround some problematic printers. See the sample
          configuration file for details.
          
    - 1.21alpha24 :
    
        - Big database structure changes. An upgrade script
          is now included for PostgreSQL. LDAP users may just
          want to copy the new schema in place.
          
        - edpykota now recognizes the -o | --overcharge command line
          option to set an overcharging or undercharging factor
          on each print job's cost for a particular user.
          
        - repykota's output format has changed an is not limited to  
          80 characters wide anymore. People who used to parse 
          repykota's output in their own tools are strongly advised
          to parse dumpykota's output instead.
          
        - Introduction of the new 'maxdenybanners' configuration directive
          to allow up to N (the directive's value) deny banners to be
          printed even if the print account is forbidden to print.
          
        - Better detection of the number of copies in the PostScript parser.
        
    - 1.21alpha23 :
    
        - Traditional Chinese translation added.
    
        - Several fixes and performance improvements with the LDAP backend.
        
    - 1.21alpha22 :
    
        - edpykota's --delete command line option doesn't consider
          '*' to be the default argument anymore :-)
          
    - 1.21alpha21 :
    
        - dumpykota.cgi can now be run as a normal user as opposed to 
          a PyKota administrator. No read access to pykotadmin.conf is
          necessary anymore. This CGI scripts now honors the REMOTE_USER
          environment variable just like printquota.cgi does, see 
          the file cgi-bin/README for details.
          The command line tool dumpykota must still be run by a PyKota 
          administrator.
          
        - Fix exception's name error due to a change in recent versions 
          of PygreSQL.   
          
    - 1.21alpha20 :
          
        - Big bug fix wrt the datelimit field which was never reset.
          
    - 1.21alpha19 :
    
        - Introduction of the new dumpykota.cgi CGI script, which
          allows third party tools like phpPrintAnalyzer to be
          interfaced in real time with PyKota. To use it you
          may have to change your web server's configuration,
          please see pykota/cgi-bin/README
          
    - 1.21alpha18 :
    
        - Now dumpykota can dump the job history into CUPS' page_log
          format (even when the printing system is LPRng). This allows
          people to use tools like PrintAnalyzer or phpPrintAnalyzer 
          (two completely different beasts) with PyKota.
          
    - 1.21alpha17 :
     
        - Fixed field type's name in XML dumps.
        
        - Now drops priviledges to user 'pykota' when launched as root
          (directly or as a CUPS backend), if such an user exists.
          
        - Fixed PCL3/4/5 parser for some drivers' output.
        
    - 1.21alpha16 :
    
        - fixed several recently introduced bugs in LDAP backend code,
          thanks to Matt Hyclak.
          
    - 1.21alpha15 :
      
        - Normalized field names between LDAP and PostgreSQL backends
          in dumpykota's output.
          
        - dumpykota filtering capabilities are now supported within
          the LDAP backend too.
          
    - 1.21alpha14 :
    
        - dumpykota now supports extended filtering possibilities
          with the PostgreSQL backend (LDAP support will come soon).
          
    - 1.21alpha13 :
    
        - Added a --debug command line option to pkpgcounter, so
          that people may see what really happens, and what will
          be available in the future :-)
          
    - 1.21alpha12 :
    
        - Fixed a bug with non standard pkbanner command lines.
        
        - Now logs as errors the differences between precomputed and 
          computed job sizes. **CUPS only for now**
          
    - 1.21alpha11 :
    
        - Now user and group print quota entries can be stored directly
          below the user or group with the LDAP backend, using them
          as branches. Thanks to Stefan Wold for the patch.
          See the sample configuration file for details.
          
        - TLS is now supported with the LDAP backend. Thanks to Stefan
          Wold for the patch.
          See the sample configuration file for details.
          
        - edpkota now accepts the -U | --used value command line argument
          to preset the page counters to an initial value. Thanks to
          Wilson Roberto Afonso and Matt Hyclak for the patch.
          
    - 1.21alpha10 :
    
        - Now automatically reconnects up to three times if the LDAP server
          returned an error, possibly due to a timeout.
          
    - 1.21alpha9 :
    
        - Now searches the configuration files in 'pykota' system user's
          home directory. The 'pykota' system user and group have
          to be created by the system administrator, and given 
          sufficient permissions to PyKota's configuration files.
          
    - 1.21alpha8 :
    
        - Added the new reject_unknown directive to reject the
          creation of user or groups which are unknown to the system.
          
    - 1.21alpha7 :
    
        - Introduction of the pkmail email gateway.
        
    - 1.21alpha6 :
    
        - Fix for LPRng filenames.
        
        - Added Turkish translation thanks to Baris Cicek.
        
    - 1.21alpha5 :
    
        - Improved ESC/P2 miniparser for Stylus Photo printers.
        
        - Implemented the --savetoner command line option to pkbanner.
        
        - Improved pkbanner's help and messages.
        
    - 1.21alpha4 :
    
        - Fixed job's data file detection code for LPRng.
        
        - Fixed a problem with PJL and PowerSave mode, 
          thanks to Wilson Roberto Afonso.
        
    - 1.21alpha3 :
    
        - PyKota can now generate its own banners, and either account
          for them or not, depending on newly introduced directives
          in pykota.conf
          Thanks to Matt Hyclak for the patch.
          
    - 1.21alpha2 :
    
        - Installation script is now non interactive again, as it should
          have stayed from the beginning. Dependencies checking can
          be done by launching the checkdeps.py command.
          
    - 1.21alpha1 :
    
        - added an IPP messages miniparser. Extraction of the client
          hostname is now considered to be fiable, at least until
          the CUPS developpers change something...
        
    - 1.20 :
    
        - Version number changed.
        
    - 1.20beta :
    
        - Norwegian (Bokm� preliminary translation added.
        
        - German translation was finalized.
        
        - Greek translation was finalized.
        
        - For urgent legal reasons, a "privacy" directive has been added 
          to pykota.conf, to hide jobs' title, filename and options
          in the history.
          
        - Minor changes in the debian directory.
        
    - 1.20alpha27 : 
    
        - Fixed many problems with translation files. In particular
          the Brazilian Portuguese translation file was incorrectly
          installed into the french "breton" locale directory.
          
        - Finished French translation.  
          
    - 1.20alpha26 :
    
        - LPRng support code should now better detect network
          printers' hostnames or IP addresses.
          
        - Fix for recently introduced bug in LPRng support.
        
        - Fix for incorrect locale setting when starting.
        
    - 1.20alpha25 :
    
        - pkprinters allows --list to normal users, but require user
          to be a PyKota administrator when another command line 
          option is used.
          
        - pkhint doesn't use absolute path to search for helper commands
          anymore.
          
        - Spanish translation ready for 1.20.   
        
    - 1.20alpha24 :
    
        - Now warnpykota only warns users limited by balance once.
        
        - Now warnpykota only warns users who have ever printed,
          to not cause confusion to users who have just opened
          their account.
        
        - Prepared help messages for translations.
        
    - 1.20alpha23 :
    
        - Add workaround for locale problem.
        
    - 1.20alpha22 :
    
        - Hopefully final fix for data encoding charset.
        
        - edpykota now refuse to work if the user is not a PyKota admin.
        
        - Fixes a recently introduced bug with users groups and PostgreSQL.
        
    - 1.20alpha21 :
    
        - Now dumpykota refuses to launch if the current user is not
          a PyKota administrator (a PyKota administrator is an user
          allowed to read the pykotadmin.conf file).
          
        - Now any PyKota administrator can use repykota and warnpykota
          with additionnal arguments, not only the root user.
          
        - Improved documentation a bit.
        
    - 1.20alpha20 :
    
        - Reduced delay in internal SNMP and PJL_over_TCP handlers
          from 2 to 1 second.
        
        - dumpykota can now output XML.
        
    - 1.20alpha19 :
    
        - Improved user groups and user groups quotas dumps in dumpykota.
        
        - Finished implementation of LDAP support in dumpykota.
        
        - The md5 sum of the job's datas is now available as an 
          hexadecimal digest in the PYKOTAMD5SUM environment variable.
          
        - The description, filename, title, and options fields
          are now stored as UTF-8 strings in both LDAP and PostgreSQL.
          Proper charset is used during extraction from the database.
          
    - 1.20alpha18 :
    
        - Experimental autopykota command to initialize user
          account and quota entries on first print while not 
          resetting the balance value if not needed.
          
        - Added the new "winbind_separator" directive to strip
          out the Samba/Winbind domain name while printing if
          needed (Off by default).
          
    - 1.20alpha17 :
    
        - Added --hardreset command line option to edpykota.
        
        - Fixed problem with accented chars in title, filename
          or options and LDAP backend.
          
    - 1.20alpha16 :
    
        - Internal handling of socket based PJL queries over
          port tcp/9100 is now available with the magic
          'pjl' value in the hardware accounter.
          
    - 1.20alpha15 :
    
        - LDAP pykotaAccountBalance objects were always created
          attached to the pykotaUser objects. This is now
          fixed and the value of balancebase in pykota.conf
          is now correctly used.
          
    - 1.20alpha14 :
    
        - Belgian translations added (copy of the french ones).
        
        - External hardware accounter now loops until
          a result can be read correctly.
          
        - pkhint now suggest the internal snmp handling instead
          of the external one.
          
    - 1.20alpha13 :
     
        - Hardware accounter now accepts the special
          'snmp' value, meaning ALL is done from main code
          without launching any subprocess.
          You need the http://pysnmp.sf.net Python module
          for this to work (tested with v3.4.2)
          
        - Increased delay between two loops iterations  
          to 2 seconds in waitprinter.sh to diminish
          CPU load.
          
    - 1.20alpha12 :
    
        - waitprinter.sh rewritten from scratch to wait
          for printer status stabilization.
        
    - 1.20alpha11 : 
    
        - Generic data dumper 'dumpykota' introduced.
          Works only with PostgreSQL backend for now.
          
    - 1.20alpha10 :
    
        - Added fix for incorrect job's size computation when
          hardware accounting fails.
          
    - 1.20alpha9 :
    
        - Small fixes wrt subprocesses.
        
    - 1.20alpha8 :
    
        - Added full support of PCL3 : HP Deskjet printers are now 
          supported.
        
    - 1.20alpha7 :
    
        - First draft of PCL3GUI (Deskjet 640C for example) analyzer.
          Needs more work, but no documentation is available.
        
    - 1.20alpha6 :
      
        - Improved mailandpopup.sh script.
        
        - Small fixes for LPRng.
        
        - Now exports PYKOTAPRINTERHOSTNAME for network printers'
          hostname or IP address. Set to 'localhost' when not
          meaningful.
          
        - Fixed issue with how mod_auth_ldap Apache modules
          fills the REMOTE_USER environment variable.
          
        - Fixed a bug in LDAP user deletion code.
        
    - 1.20alpha5 :
      
        - Introduces the new configuration directive 'onaccountererror'
          to specify what must be done whenever the software or
          hardware accounter fails. Allowed values are 'continue' and
          'stop' (default). 
          
        - Minor fix in software accounter for a recently introduced
          problem.
          
    - 1.20alpha4 :
    
        - Added support for more ghostscript devices in ESC/P2
          analyzer : stcolor, st800, escp, escpc are now supported.
          
    - 1.20alpha3 :
       
        - Small improvement in PostScript parser to avoid being
          fooled by a change in number of copies : 
          The user can't ask for 200 copies of page 1 and 1 copy 
          of page 2 to fool PyKota anymore : 
          this would have been accounted for as 1 copy only, but
          now the maximum value seen is used.
          
        - Relax checks for PCL5 header to accomodate some strange
          printer drivers.
          
        - Now hardware accounting method works like software
          accounting method wrt the subprocess' output.
          
        - Hardware and Software accounting stop the print queue  
          when the accounting subprocess didn't work. This will
          be configurable in the near future.
          
    - 1.20alpha2 :
    
        - Now software accounting method reads accounter's
          answer until a line with a single integer is 
          obtained or until all of its output has been read.
          This should fix the problem some people have
          when strange locale related messages are sent
          when running subprocesses (Solaris only ?).
          
        - First version of ESC/P2 analyzer. Seems to work
          ok at least with 'epson', 'epsonc', 'eps9mid' 
          and 'esp9high' ghostscript devices. Needs more
          work to correctly analyze other devices.
        
    - 1.20alpha1 : 
    
        - Fixed problem with number of copies in PCLXL parser.
        
        - Fixed problem with number of copies in PCL5 parser,
          when number of copies was set outside of and before first
          page.
        
        - Added workaround for number of copies directly set in
          PostScript code.
          
        - Added german translation.
        
        - Fixed problem on first print for user group quota when
          "enforcement: strict" is used.
        
        - Fixed interpolation problem in french translation.
        
    - 1.19 :
    
        - Updated existing translations.    
        
        - Added final Thai and Greek translations.
        
        - Minor last minute bug fixes.
        
    - 1.19beta :
    
        - LPRng support for both software and hardware accounting seems
          to be OK.
          
        - Translation files fixed because of fuzzy matching problem.  
        
        - Fixed path to Swedish translation: 'sv' instead of 'se'.
        
        - Added first draft of Thai translation.
        
        - Fixed some minor bugs with Pychecker.
        
    - 1.19alpha35 :
    
        - Generic PDL analyzer now supports Binary PostScript documents
          in addition to DSC compliant ones.
        
    - 1.19alpha34 :
    
        - LPRng support : software accounting seems to work fine,
          but hardware accounting doesn't work yet.
   
    - 1.19alpha33 :
    
        - pykosd now accepts command line arguments. Enjoy !
          NB : due to a bug in xosd, you need xosd version 2.2.8
          or higher if you don't use the default C locale.
          
    - 1.19alpha32 :
    
        - Early (not working) version of the LPRng support rewrite.
        
        - Saner default configuration file.
        
    - 1.19alpha31 :
    
        - Fixed a double Python2.1 incompatibility with newer versions,
          which prevented the PCL5 and the PCLXL parsers to work with
          this older version of Python.
          
    - 1.19alpha30 :
    
        - Introduction of preliminary version of the pykosd command
        
        - Usernames, Groupnames and Printernames can now contain
          ANY character, expected '/' and '@'.
        
        - Integrated most of the Debian work from Sergio 
          Gonz�z Gonz�z
          
    - 1.19alpha29 : 
    
        - PCLXL parser should now correctly handle number of
          copies set for each page.
          
        - Added testsuite for generic PDL analyzer.
        
        - Typo fixed in PCLXL parser.
        
        - Correctly dispatches error messages depending on the
          situation : tools which can be run interactively will
          ALWAYS output their logs to stderr, excepted for
          their DEBUG messages, for which the "logger:"
          directive as set in pykota.conf will be honored.
          Tools which can't be run interactively will ALWAYS
          honor the "logger:" directive as set in pykota.conf
          
    - 1.19alpha28 :
    
        - The description field for printers can now be set,
          and is displayed in reports.
          
        - Fixed a small problem in SQL backend code.
          
        - Improved pkhint
        
        - First working PCL6 (aka PCLXL) parser !
          Doesn't handle copies yet.
        
    - 1.19alpha27 :
    
        - Additionnal 10% speed increase in PCL5 parser for GhostScript
          generated files (untested yet with other PCL files).
          
        - Reintroduced the use of mmap in PCL5 parser : this speedups
          PCL5 parsing by almost 4 times.
          This 4x with the 3x below give us in fact a twelvefold increase
          in performance for PCL5 parsing !
          
        - The Python accelerator Psyco is now used if available
          in the generic PDL parser : it speedups PCL5 parsing by
          almost three times automagically. 
          
    - 1.19alpha26 :
    
        - Fixed PCL5 parser according to the sources of rastertohp.
        
        - Fixed number of copies handling in PCL5 parser : the number
          of copies could vary from page to page.
          
        - Added more debug information to PCLXL parser.  
        
    - 1.19alpha25 :
    
        - Added new "description" field to printers in both
          PostgreSQL and LDAP. No code to handle this yet.
          
        - Reversed the fix for power save mode in printers :  
          no paper or tray open would have caused people to
          be undercharged.
          
        - Preliminary version of MySQL initialization script.
          
    - 1.19alpha24 :
    
        - Native PDF parsing code now works perfectly with
          Python 2.1 and above, and is way faster than before.
          GhostScript is not used anymore to convert PDF to PS,
          because PDF is parsed directly.
          
    - 1.19alpha23 :
    
        - Fixed waitprinter.sh wrt PowerSave mode and some
          printers. Thanks to Ryan Suarez.
        
        - Smart PDL analyzer now recognizes PDF too : two
          PDF parsing methods are used : a native one which
          is very fast but requires Python 2.3 or above, and
          a slower one which uses Ghostscript. PyKota detects
          automatically which method to use depending on your
          version of Python.
        
        - All tracebacks now include PyKota's version number.
          
    - 1.19alpha22 :
    
        - Ensures that the file pointer for job's data is 
          reset to the start of the job's datas after the
          job's size pre-computation phase.
          
        - Improved PostScript detection code for bad drivers
          which don't correctly switch languages between PJL
          and PS.
       
        - Even better debugging code.
        
    - 1.19alpha21 :
    
        - Additionnal check.
        
    - 1.19alpha20 :
    
        - More complete logging messages wrt accounters.
        
        - Now catches early failures occuring when the configuration
          files are parsed.
        
    - 1.19alpha19 :
    
        - Fixed over-verbose exits when displaying help
          or version.
          
        - Extended some PATH in the shell scripts  
        
        - Deleting an user now also deletes his payments history.
        
        - Doesn't explicitely ignore SIGCHLD anymore. This seems to
          cause problems.

    - 1.19alpha18 :
    
        - History of payments is now stored in the database, and
          updated each time an user's balance is modified with
          edpykota --balance. This history represents each time
          the user was given some positive or negative credit,
          but doesn't reflect price paid for jobs, which appear
          in the jobs history instead.
          
        - Jobs history web reports now include the job's size in
          bytes. Also columns were reordered to render the report
          more easily readable.
          
    - 1.19alpha17 :
    
        - Mysterious bug (program exited unexpectedly without leaving
          any log message) which occured once in a while when
          cancelling jobs should be fixed now.
          
        - New 'crashrecipient' directive. Disabled by  
          default. See sample configuration file to see how
          to activate it. It is suggested that you activate it.
          
        - New PYKOTAJOBSIZEBYTES environment variable exported,  
          which contains the job's size in bytes. Empty jobs
          are now denied. This value is now stored in the job
          history (but not yet displayed from the web interface)
          
        - Preliminary work on payments storage : the PostgreSQL  
          and LDAP schemas should be OK now, but no code is written
          yet to support this functionnality.
          
    - 1.19alpha16 :
    
        - the newuser and newgroup LDAP specific directives now accept
          a second parameter to attach(), "fail" or "warn". If no entry
          is found to attach PyKota attributes to, and "fail" was used,
          then action is aborted and program exits in error. If "warn"
          is used, a new entry is created but a warning message is logged.
          
        - The job's originating hostname is now stored in the database,
          and exported as PYKOTAJOBORIGINATINGHOSTNAME in the environment.
        
    - 1.19alpha15 :
    
        - The old pykota filter is now removed. This means that
          currently PyKota doesn't support LPRng anymore, but 
          LPRng support will be re-added, improved, in the near future.
          
    - 1.19alpha14 :
    
        - PYKOTAPRECOMPUTEDJOBSIZE and PYKOTAPRECOMPUTEDJOBPRICE 
          are now exported as environment variables.
          
        - Heavy CPU usage is now completely fixed it seems, by 
          introducing some 0.01 second sleeping when nothing to do.
          
        - Pre-computation of the job's size is now only
          done when enforcement is "STRICT", this saves
          CPU for people who don't want to use this new
          feature.
          
        - New 'enforcement' directive which accepts either
          STRICT or LAXIST. See sample configuration file
          for details.
          
    - 1.19alpha13 :
    
        - Get back to an old version of the polling loop which
          worked fine, but non-optimally. Will be optimized
          as time permits.
          
    - 1.19alpha12 :
    
        - Job's size is now precomputed unconditionnally using 
          the generic PDL analyzer.
          NB : This value is not yet used to pre-allow/deny the
          job.
        
    - 1.19alpha11 :
    
        - The pykota filter is now completely deprecated and
          can't be used anymore. LPRng users MUST download 
          an older version of PyKota. CUPS users MUST use
          cupspykota. Better support for LPRng will come
          in the near future.
          
    - 1.19alpha10 :
    
        - pykotme now uses generic PDL Analyzer directly instead
          of spawning a new process.
          
        - Improvements to generic PDL Analyzer.
        
        - 25% Speedup in PCL parser. Getting more of the Python version
          will probably be difficult, since the completely table driven 
          parser I wrote is slower than the original algorithm. Then,
          the original algorithm is still used for now.
          Probable that more speed could be achieved using mmap, at the
          expense of severe memory consumption.
        
        - configuration simplifications : the "accounter" directive's
          'hardware' arguments now accepts what was in the "requester"
          directive as it's content. The "requester" directive is not
          supported anymore. See sample configuration file for details.
        
    - 1.19alpha9 :
       
        - pkpgcounter is now just a wrapper around the PDLAnalyzer class
        
        - pykotaOptions LDAP attribute now accepts non-ascii characters
          too.
          
        - First try at cupspykota main loop rewrite
        
    - 1.19alpha8 :
    
        - 'querying' accounting method is now called 'hardware'
        
        - 'external' accounting method is now called 'software'
        
        - 'stupid' accounting method doesn't exist anymore
        
        - 'snmp' requester doesn't exist anymore
        
        - code simplifications in external requester
        
    - 1.19alpha7 :
     
        - Database schema change (both LDAP and PostgreSQL) to
          allow storage of the job-originating-hostname attribute.
          An SQL upgrade script is present. The code is not yet
          capable of setting/retrieving this attribute. Support
          for this functionnality will be added in the future.
          
    - 1.19alpha6 :
    
        - problems connecting to SMTP servers are now catched.
        
        - pykotme now uses pkpgcounter to compute the size of the
          job. This modification adds PCL5 and soon PCLXL (PCL6)
          support. You can now get quotes for PCL in addition to 
          PostScript.
          
        - pkpgcounter now includes first try at PCLXL parser.  
          It is currently disabled until it's finished.
          
    - 1.19alpha5 :
    
        - cupspykota polling loop fix for MacOSX (at least)
        
        - logging subsystem now logs PID too, to ease diagnosing
          problems under heavy load.
          
        - pkpgcounter enhancements.
        
        - pkprinters --skipexisting now tells pkprinters to not
          modify existing printers when --add is used.
          pkprinters logs a warning message whenever --add is
          used and a printer already exists.
          
        - LDAP schema now allows accented chars in pykotaFileName
          and pykotaTitle attributes.
        
    - 1.19alpha4 :
    
        - a bad copy&paste prevented pkpgcounter to work at all, this
          is now fixed.
          
        - improved PCL support in pkpgcounter, with the use of
          HP PCL/PJL Reference Set
          PCL5 Printer Language Technical Quick Reference Guide
        
        - pkpgcounter now uses mmap to speed things up : duration
          is halved with the sample 87 pages PCL document I use for
          testing.
          
    - 1.19alpha3 :
    
        - pkprinters now accepts the --list command line argument
          to list informations about printers.
          
    - 1.19alpha2 :
    
        - pkpgcounter is now included as a smart external
          accounter which can handle both PostScript and PCL.
          
    - 1.19alpha1 : 
    
        - PYKOTALIMITBY environment variable is now exported too.
        
        - Inclusion of Michele Baldessari's work on autotools.
        
    - 1.18 :
    
        - Improved documentation
        
    - 1.18beta2 :
      
        - Fix for raw jobs with cupspykota
        
    - 1.18beta :
    
        - Small fix for new waitprinter.sh script
        
        - Better tracebacks on error conditions
        
        - Improved translations
        
        - Improved documentation
        
    - 1.18alpha15 :

        - waitprinter.sh replaced by a new and possibly faster
          and more powerful one.
    
        - Names can now begin with a digit.
        
        - Stability improvement for quotas set on printers groups.
        
    - 1.18alpha14 :
    
        - Catches more (all ?) problems with SMTP servers instead
          of breaking.
          
        - Italian translation added.
        
    - 1.18alpha13 :
    
        - Pre and Post hooks are now available in the pykota
          filter as well.
          
        - The pykota filter doesn't check last user's quota
          anymore when delayed hardware accounting is used : 
          this will be done anyway the next time the last 
          user will print again. This saves some database
          queries which are unneeded.
        
    - 1.18alpha12 :
    
        - Pre and Post hooks to external commands with many 
          many environment variables available are useable 
          from the cupspykota backend.
          See sample configuration file for details and help.
          
    - 1.18alpha11 : 
    
        - Hopefully final fix wrt groups (users and printers)
        
    - 1.18alpha10 :
    
        - Fixed nested printer groups accounting.
        
        - Fixed user groups + printer groups problem.
        
    - 1.18alpha9 :
    
        - Fix for LDAP problem when job price was 0.
        
        - Fix for empty user groups with LDAP.
        
        - Preliminary version of the snmpprinterstatus command added.
        
    - 1.18alpha8 :
    
        - Don't retrieve data when not needed.
          This avoids many database queries.
        
    - 1.18alpha7 :
    
        - New experimental ldapcache directive to improve
          LDAP caching.
          
    - 1.18alpha6 :
        
        - New configuration directive added : maildomain
          to specify the domain to send mail to when user 
          email address is not set in PyKota database.
          
        - More complete email message headers.  
          
    - 1.18alpha5 :
    
        - Debian packaging specific files, contributed by
          Peter Hawkins are now included in the project's tree.
          
        - Preliminary work on future low-level LDAP caching to speed
          up modifications (at least).
          
    - 1.18alpha4 :
    
        - Applied some of the modifications suggested by
          Peter Hawkins, for better Debian integration.
          Most notably cupspykota, pykota, and the shell
          scripts are now installed into /usr/share/pykota
          instead of into /usr/bin
          BEWARE : double check that your symbolic link
          to cupspykota is correct, and that your pykota.conf
          file use correct path for the shell scripts like
          waitprinter.sh and al.
          
        - pkhint command added, to help in configuring correct
          accounting method for PyKota.
          
    - 1.18alpha3 :
    
        - Fixed a rare bug in cupspykota backend which gave 
          incorrect "CUPS backend died abnormally" messages.
        
    - 1.18alpha2 :
    
        - Added the pkprinters command line tool to manage printers
          exclusively.
        
    - 1.18alpha1 :
    
        - Preliminary work on Relationnal Database Independance.
        
        - Sample report made from OpenOffice.org added as PDF and SXC.
        
    - 1.17 :
    
        - Finalized release.
        
    - 1.17alpha2 : 
    
        - Fix stupid software accounting bug in the CUPS backend.
        
    - 1.17alpha1 :
    
        - User/Group deletion code in edpykota rewritten.
        
    - 1.16 :
    
        - Preliminary documentation on how to use OpenOffice.org
          to create personnalized reports for PyKota + PostgreSQL.
          
        - Spanish installation guide for CUPS+PyKota+PostgreSQL
          added to the documentation. Doesn't yet document v1.16
          though.
        
        - Portuguese Portuguese translation replaces Brazilian 
          Portuguese Translation. Brazilian portuguese 
          translation is still available under the 'br' locale.
        
    - 1.16beta3 :
    
        - Swedish translation added.
        
        - Small fix in documentation.
        
    - 1.16beta2 :
    
        - Small fix for job cancelling code while job already
          printing in CUPS backend.
          
        - Big fix wrt printing policies for unknown users.  
        
    - 1.16beta1 :
    
        - Manpages upgraded.
        
    - 1.16alpha26 :
    
        - Job price is now computed like this :
        
          SUM((NbPages * PricePerPage) + PricePerJob)
          
          For current printer and all the printers groups it is 
          a member of, if any.
          
          This may be difficult to grasp, but offers unprecedented
          flexibility.
          
    - 1.16alpha25 :
    
        - Email address can be set at user creation time,
          using the following syntax : 
          
            $ edpykota --add username/email@yourdomain.com
        
    - 1.16alpha24 :
    
        - CGI script now can show the printing history.
        
        - Bug wrt number of copies should be fixed.
        
        - Major code refactoring => It is now possible to
          automatically add printers on first print.
        
    - 1.16alpha23 :
    
        - Fixed potential accuracy problem if a user printed on
          several printers at the very same time.
        
    - 1.16alpha22 :
    
        - edpykota now allows adding printers to printer groups
        
    - 1.16alpha21 :
    
        - Allows the dot in user and printer names.
        
        - Printer groups are now cached too.
        
    - 1.16alpha20 :
    
        - Fixed several small bugs with the help of PyChecker.
        
    - 1.16alpha19 :
     
        - Fixed a bug in PostgreSQL backend's code.
        
    - 1.16alpha18 :
    
        - Initial code for printer groups support. Printer groups have
          to be added with external tools like psql or gq for example.
          Printer groups seems to work fine with users. Untested with user 
          groups for now.

        - Several minor visual problems fixed in reports.
        
        - Fixed a bug in LDAP date handling.
        
        - Fix for LDAP problem with some combination of OpenLDAP + Python-LDAP
          + some particular settings.
          
        - mailandpopup.sh shell script added. See sample configuration file
          for details.
        
        - Several environment variables are now exported when executing 
          external commands (accounters, requesters, mailers) :
           
            PYKOTAUSERNAME => user who is printing
            PYKOTAPRINTERNAME => printer on which the job is being sent
            PYKOTATITLE => Job's title
            PYKOTAFILENAME => Job's filename
            PYKOTACOPIES => number of copies asked
            PYKOTAOPTIONS => print command options (-o with CUPS)
          
    - 1.16alpha17 :
    
        - Added portugues translation.
        
        - Major code refactoring in accounters.
        
    - 1.16alpha16 :
    
        - Code refactoring, and addition of a real HTML reporter, to 
          have nice looking and useful web based reports (in the future).
        
        - Added utolower configuration option to convert all usernames
          to lowercase during printing. 
          
    - 1.15alpha15 :
    
        - Now includes the papwaitprinter.sh shell script to wait for
          AppleTalk enabled printers.
          
    - 1.16alpha14 :
    
        - Now printing on a printer not defined in the Quota Storage,
          results in the job being cancelled or stopped, depending
          on the printing system.
          
    - 1.16alpha13 :
    
        - Small bugfix wrt Python v2.1
        
        - If username is an empty string, user root is assumed.
        
        - 'postgresql' is not recognized anymore as a storage backend,
          you have to use 'pgstorage' instead. 'postgresql' used to
          be supported for compatibility with very old PyKota versions,
          this is not the case anymore.
          
    - 1.16alpha12 :
    
        - Job price added to job history, to keep accounting correct
          if a printer price per page or per job is modified and the
          history is not reset : the history would have given a false
          amount of money charged before the printer's prices were
          modified, this addition solves the problem.
          
    - 1.16alpha11 :
    
        - Some preliminary modifications to the database structure for 
          PostgreSQL and OpenLDAP. A script to upgrade a PostgreSQL
          database is now provided. LDAP attributes now support
          eq, pres and sub indexes. 
          
        - Job history now contains additional informations :  
          number of copies, filename, title and print options.
          
    - 1.16alpha10 :
        
        - Some modifications done to external policy handling.
          See sample configuration file for details.
    
        - Spanish translation added.
        
    - 1.16alpha9 :
    
        - mailto can be set to external(...) to warn users above
          quota with any command of your choice (e.g. linpopup).
          Several parameters are available, see sample configuration
          file for details.
          
    - 1.16alpha8 :
    
        - Now cupspykota transmits SIGTERMS to the original  
          backend when the user deletes a job being processed.
          
        - The documentation is now licensed under the terms of the
          GNU GPL too.
          
    - 1.16alpha7 :
    
        - The CUPS backend is now fully functionnal (it seems).
        
        - From now on, the use of the pykota filter with CUPS 
          is deprecated. The pykota filter still has to be
          used with LPRng though.
        
    - 1.16alpha6 :
    
        - Bug fixed : when edpykota was used to add a printer
          with no user or group name, a bad user/group name "*" was
          added to the database.
          
    - 1.16alpha5 :
    
        - New CUPS backend supports device enumeration !
          This means that NO manual CUPS configuration has to
          be done. All can now be done from CUPS web interface.
        
    - 1.16alpha4 :
    
        - New CUPS backend added. 
        
    - 1.16alpha3 :
       
        - Minimal FAQ document added.
        
    - 1.16alpha2 :
    
        - CGI script now partially supports browser's language preference
        
    - 1.16alpha1 :
    
        - Ensures that the LDAP connection is really closed at end.
        
        - More complete error messages in case of LDAP operation failure.
        
    - 1.15 :
     
        - Documentation improved.
        
    - 1.15beta :
    
        - Command line tools now accept multiple printer
          names or wildcards, separated with commas.
          
    - 1.15alpha9 :
    
        - External policy can be used to auto-add unknown users 
          and any other interesting things you may want to do. 
          WARNING : You can't auto-add printers for now. 
          They have to be created manually before.
          Look at sample configuration file for examples.
          
    - 1.15alpha8 :
    
        - Job history can be disabled if needed.
          This feature may be useful for people who use the LDAP
          backend and don't want their LDAP tree to grow out of
          control.
        
    - 1.15alpha7 : 
    
        - More work done on caching wrt groups.
        
    - 1.15alpha6 :
    
        - Performance bottleneck with LDAP solved.
        
    - 1.15alpha5 :
    
        - Caching mechanism for all database queries.

    - 1.15alpha4 :
    
        - Problem with Python 2.3 fixed. Needs some definitive 
          modification.
          
    - 1.15alpha3 :
    
        - CGI script now tries to restrict view when the 
          REMOTE_USER environment variable is set. This allows
          people to protect its use with .htaccess files.
          When unprotected, or if REMOTE_USER is 'root', then 
          unrestricted access is granted.
          NB : REMOTE_USER is automatically set by the web server
          (e.g. Apache) when a page with restricted access is viewed.
          
    - 1.15alpha2 :
    
        - Better detection of LPRng.
        
        - Outputs a warning if printing system is unknown.
        
    - 1.15alpha1 :
    
        - Bug fix in SQL code.
        
    - 1.14 : 
    
        - The PostgreSQL database connection is now always closed before
          exit. The LDAP database connection too, but it didn't suffer
          from the problem anyway, AFAICT.
        
    - 1.14beta2 :
    
        - An LDAP-specific configuration field was introduced, usermail.
          Use it to specify which LDAP attribute contains a valid email
          address for the user. See sample configuration file for
          details. 
        
    - 1.14beta :
    
        - Two new LDAP-specific configuration fields were introduced,
          newuser and newgroup. Use them to specify if you want to add
          quota information to existing entries or elsewhere.
          See sample configuration file for details.
          
    - 1.14alpha9 :
    
        - Documentation improved.
        
        - pykotme now also prints the user's account balance value
          if the user is registered in the Quota Storage and his
          limiting factor is "balance".
          
    - 1.14alpha8 :
    
        - bin/waitprinter.sh is now included to wait for
          SNMP enabled printers to be idle again before
          asking for their internal page counter.
          See sample configuration file for details.
          
    - 1.14alpha7 :
    
        - edpykota now doesn't check if the user/group is above
          quota. The check is only done at print time. This 
          prevents the sending of email messages when creating
          user quotas if quota is already reached at creation
          time.
          
    - 1.14alpha6 :
    
        - Configuration file split and moved to /etc/pykota/pykota.conf
          and /etc/pykota/pykotadmin.conf to prevent simple users to
          have Read/Write access to the Quota Database.
          Don't forget to :
          
                $ chmod 640 /etc/pykota/pykotadmin.conf
                
        - storageuser and storageuserpw configuration fields reintroduced
          for the same reason.
          
        - Code cleaning for future implementation of email addresses 
          support in the PostgreSQL and LDAP backends.
          
    - 1.14alpha5 :
        
        - Big bug fixed when changing the prices for a printer
        
    - 1.14alpha4 :
    
        - When adding an user, the user itself and its account balance
          are now added in the same LDAP entry. Before they were added
          as two different LDAP entries.
          
    - 1.14alpha3 :
    
        - Email field added to PostgreSQL database. An upgrade script
          is included in initscripts/postgresql.
        
    - 1.14alpha2 :
    
        - Typo fixed when using the --prototype option to edpykota
        
    - 1.14alpha1 :
    
        - Poor man's treshold included : users with a low account
          balance (below 1.0 by default, configurable) are sent
          a warning message.
          
        - All warning messages to users are now configurable.  
          NB : Warning messages to the administrator are not.
        
    - 1.13 :
    
        - Some small fixes wrt group handling. 
        
        - The whole module code was passed through PyChecker and
          several small code inconsistencies were fixed,
          as well as a bug in the groups code.
          
    - 1.13alpha1 :
    
        - Bug fix wrt LDAP.
        
    - 1.12 :
    
        - More complete test added in edpykota.
        
    - 1.12alpha2 :
    
        - The previous bug fix was incomplete. This is corrected.
        
    - 1.12alpha1 :
    
        - Small bug fix.
        
    - 1.11 :
    
        - Now includes the pykotme command line tool, to produce
          printing quotes for users.
          
    - 1.10 :
    
        - Bug fixed wrt quota reports when print limiting is by 
          account balance.
          
    - 1.09 :
    
        - Translations finalized.
        
        - Manpages updated.
        
        - Installation documentation improved wrt LDAP support.
        
    - 1.09beta2 :
    
        - The CGI script for quota reports works again, and
          is much more powerful.
        
    - 1.09beta :
    
        - edpykota finalized
        
    - 1.09alpha2 :
    
        - LDAP schema modified a bit. Please upgrade.
        
        - LDAP backend works !
        
        - repykota and warnpykota allow the root user to pass
          users or groups names on the command line. Users 
          and groups names may contain wildcards.
          
        - repykota and warnpykota only reports or warns the
          current user when launched by non-root users.
          
        - Minor bug fixes.  
    
    - 1.09alpha1 :
    
        - More work on LDAP storage backend. Many options
          were added to /etc/pykota.conf to give some
          hints to the LDAP storage backend.
          
        - Severe database access optimizations were done by
          rewriting most of the Quota Storage backends drivers.
      
    - 1.08 :
    
        - Major bug fix wrt LPRng support for remote jobs.
        
        - Major bug fix wrt increase/decrease account balances.
        
    - 1.08alpha8 :
    
        - Code refactoring.
        
        - A single user/password pair is used to connect
          to the database backend. The storageuser configuration
          field, and its associated passwords storageuserpw, are 
          not used anymore.
          
        - You can now set PyKota in debug mode, see sample configuration  
          file for details. 
          WARNING : only Quota Storage backend queries are logged in 
          debug mode, for now.
        
    - 1.08alpha7 :
    
        - Very latest LDAP schema.
        
        - Code enhancement wrt easy pluggability of PyKota attributes
          and object classes into an existing LDAP directory.
        
    - 1.08alpha6 : 
    
        - Minor bug corrections
        
        - More work on LDAP : new schema
        
    - 1.08alpha5 :
    
        - More good work on LDAP storage.
          repykota now works reasonably well with the sample LDAP directory.
        
    - 1.08alpha4 :
    
        - LDAP schema included. 
          THIS IS JUST FOR PEOPLE TO DISCUSS ABOUT THIS.
          LDAP SUPPORT DOESN'T WORK YET, AND THE SCHEMA
          MAY CHANGE IN THE FUTURE. PLEASE TELL US
          WHAT YOU THINK ABOUT THIS SCHEMA ON THE 
          MAILING LIST.
          
    - 1.08alpha3 :
    
        - External accounting methods were partly rewritten :
         
          - No more "broken pipe" should happen.
          
          - They now take care of the number of copies
            This may be unneeded though, if the postscript
            file already does this, because this would
            overcharge users (number of copies counted
            two times). NEEDS MORE TESTING.
            
          - The sample configuration file now contains  
            an external accounting method example which should
            work with all DSC compliant Postscript files.
            
        - Some small bugs were fixed.    
        
    - 1.08alpha2 :
    
        - Now works with net-snmp v5.0 and above.
          It already worked, but the sample configuration 
          file didn't contain appropriate values...
        
    - 1.07 : Release of the Shame !
    
        - The external accounting methods driver was left out of 
          the CVS tree. Nobody could use it since it wasn't included !
          
    - 1.06 :
    
        - Severe bug fixed : if you had a printer in 
          power saving state which couldn't answer immediately, 
          a bug was triggered.
          It was caused by a bad copy/paste which
          forgot to import a Python module at run time.
          
    - 1.05 :
    
        - External accounters are finally available !
          This means that you can plug any page accounting
          method you like by setting the appropriate
          'accounter' field in /etc/pykota.conf
          See the sample conf/pykota.conf.sample to 
          learn how to do.
          
          NB : Both 'external' and 'stupid' accounting methods
               account a job size just before *this* job is 
               sent to the printer. 
               The original 'querying' method accounts a job
               size just before *the next* job is sent to
               the printer.
        
    - 1.05alpha3 :
    
        - A 'stupid' and unreliable accounting method was
          implemented to serve as an example on how to
          do this sort of things. This method only counts
          the 'showpage' statements in the input data.
          See sample configuration file for details.
          Pluggable accounting methods work, but I advise
          you TO NOT USE THIS ONE WHICH IS JUST AN EXAMPLE.
          It is not reliable enough to be used.
          Use the 'querying' accounting method instead.
          
    - 1.05alpha2 :
    
        - Pluggable accounting methods.
        
        - Better error handling.
        
    - 1.05alpha1 :
    
        - SECURITY file added to help improve PyKota's security.
        
        - Extracting the printer's internal page counter is now
          tried several times, waiting several seconds between
          two tries. This lets the time to warm up for some printers 
          which don't answer when they are sleeping (my Apple 
          LaserWriter 16/600 PS is in this case, maybe others too)
          
        - Small display bug fixed in repykota
          
    - 1.04 :
    
        - Default print policy for users/groups unknown from 
          the print quota system is now DENY instead of ALLOW,
          since ALLOW can generate inaccurate results
          (incorrect job sizes charged to the wrong persons)
          
        - LPRng support works !
        
        - A bug was introduced some time ago wrt printers' default
          policy for unknown users. It is now corrected.
        
        - repykota now displays prices per job and per page for each
          printer, if they are defined.
          
        - Minor display bug fixed in repykota.
        
        - Problem when running repykota with an empty database was fixed.
        
    - 1.03 :
    
        - Upgrade script included for earlier versions.
          Please look inside the initscripts subdirectory.
          THE DATABASE SCHEMA HAS CHANGED, PLEASE UPGRADE.
          
        - repykota now reports account balances too.
        
        - PyKota now has a nice logo, see the logos
          subdirectory.
        
        - Manual pages were finally updated.
        
        - Group quotas seem to work now !
        
        - The new database schema allows to keep an history of all
          jobs as well as charge users per page and/or per job, 
          and track users' account balance.
          
        - edpykota now accepts much more command line options to  
          use the new functionnalities. See edpykota --help for
          details.
          
        - The installation script now allows to install the sample
          configuration file during first installation.
          
        - More group quota code works, but still not finished.  
        
        - The CGI script displays a link to PyKota's website.
        
        - The job history is now kept, this will allow per-period
          reports in the future.
          
        - After having modified the quota for an user with edpykota,
          a quota check is done to eventually warn the user/admin about
          a quota which is too low to print.
          
        - A workaround is provided for HP Printers : their internal
          page counter is only saved to NVRAM in a 10 increment, so
          if you switch them off and then on, the reported page counter
          may be lower than the real number of pages printed.
          See http://web.mit.edu/source/third/lprng/doc/LPRng-HOWTO-15.html
          We unconditionnally set the last job's page count to 
          abs(int((10 - abs(lastcounter(snmp) - lastcounter(storage)) / 2))
          in this case. 
          For a more accurate accounting, never switch your HP printers
          off.
          
        - A fix is provided for printers which only have a volatile  
          page counter (reset to 0 every time you switch the printer on)
          This should allow PyKota to work reasonably fine with HP Laserjet
          4L/5L/6L, not perfect, but better than nothing.
          See http://web.mit.edu/source/third/lprng/doc/LPRng-HOWTO-15.html
          For a more accurate accounting, never switch your HP printers
          off. This is untested, please report any problem.
          
        - A bug was fixed when edpykota --add was used with users who already
          had a quota on the specified printer.
          
        - A small display bug in repykota was introduced in preliminary 
          1.03 versions, and fixed later on.
          
        - Some minor bugs which happened in rare situations were fixed.
        
        - Support for AppleTalk printers was added, see sample configuration
          file for details.
          
        - Users and group printing can now be controlled (limited) either by  
          print quota or by account balance.
        
    - 1.02 :
    
        - The installation script now checks for software availability
          and in case a software is missing asks the user if he
          wants to continue with the installation or abort it.
          
        - The configuration file pykota.conf is now expected to be 
          found in /etc instead of in /etc/cups
          The installation script prompts the user to see if he
          wants to move an old configuration file to the new location
          if needed.
          
        - Improved documentation.
        
        - You can now set the tcp/ip port on which the Quota Storage
          Server is listening, see sample configuration file for details.
        
        - Better general error handling.
        
        - Upgrade script for pre 1.01 PostgreSQL database schema is now
          included as well.
        
    - 1.01 :
    
        - The configuration file now accepts an option   
          to choose the recipient(s) of the email messages :
          
            - DevNull means no one will receive them.
            - User means only the user will receive them.
            - Admin means only the admin will receive them.
            - Both means the User and the Admin will receive them.
            
        - The configuration file now uses hard-coded default values
          when an option is not set. See sample configuration file
          for details.
          
        - Manual pages are included since 1.00, but I forgot to    
          add this information to this file.
          
        - Redistribution terms for the official package have  
          softened and are now fully GPL compatible : 
          unrestricted modification is now allowed even for
          the version number.
          
    - 1.00 :
    
        - edpykota now accepts a --noquota option. This
          disable quota checking while still doing page
          accounting. This is really useful for people
          who don't want to limit their users but want
          to know how much pages they print.
          
        - Some untested scripts were added to retrieve
          the life time page counter of non-SNMP printers.
          
        - Every directory now has a specific README file.  
        
    - 0.99 :
    
        - Under some circumstances while the user wasn't allowed
          to print, he didn't receive any email message. It is
          now fixed.
          
        - When an user wasn't allowed to print, the quota for
          the previous user wasn't updated. This is now fixed.
          
    - 0.98 :
    
        - Correctly handle the case where the printer is switched off.
        
        - Small bug wrt syslog fixed.
        
    - 0.97 :
    
        - edpykota accepts wildcards on its command line for users/groups too
          if the --add option is not set, e.g. :
          
              $ edpykota --printer lp --softlimit 50 --hardlimit 100 "jer*"
              
        - If no user name is passed at all, then a default wildcard of "*"
          which means apply the command on ALL users for this printer is used.
          
        - Small bug fixes.
        
    - 0.96 :

        - Options requester, policy, admin, adminmail and gracedelay can now
          be set either globally or per printer. The printer option has
          priority if both are defined.

        - More powerful configuration parser.
        
        - If all options are defined globally, there's no need to 
          define a section for each printer in the configuration
          file anymore. Just define a [global] section and it's ok.

    - 0.95 :

        - External requesters for printers finally added.

        - Full internationalization (english and french are supported)
 
        - More complete quota usage report

        - CGI script to access to the quota usage report

        - Several bugs fixed.


    - 0.9 : 

        - First public version