[Spambayes-checkins] spambayes CHANGELOG.txt, 1.29,
1.30 WHAT_IS_NEW.txt, 1.22, 1.23
Tony Meyer
anadelonbrin at users.sourceforge.net
Sun Dec 28 23:46:33 EST 2003
Update of /cvsroot/spambayes/spambayes
In directory sc8-pr-cvs1:/tmp/cvs-serv8306
Modified Files:
CHANGELOG.txt WHAT_IS_NEW.txt
Log Message:
Bring these closer to the actual current status.
Index: CHANGELOG.txt
===================================================================
RCS file: /cvsroot/spambayes/spambayes/CHANGELOG.txt,v
retrieving revision 1.29
retrieving revision 1.30
diff -C2 -d -r1.29 -r1.30
*** CHANGELOG.txt 29 Dec 2003 01:46:43 -0000 1.29
--- CHANGELOG.txt 29 Dec 2003 04:46:31 -0000 1.30
***************
*** 3,6 ****
--- 3,7 ----
Alpha Release 8
===============
+ Tony Meyer 29/12/2003 Web interface: Improve the 'online' help message for the review page, and add messages for the stats and home pages.
Tim Peters 29/12/2003 Many improvements to the mksets.py testtools script.
Tim Peters 28/12/2003 Outlook: export.py - the -n option now gives the number of Set subdirectories desired, instead of a number of msgs per Set subdir "to shoot for".
***************
*** 9,18 ****
--- 10,26 ----
Tim Peters 26/12/2003 Many improvements to the export.py script for Outlook.
Skip Montanaro 24/12/2003 storage: make state key a manifest constant
+ Mark Hammond 23/12/2003 Along with checking Outlook isn't running, check Outlook's mail transport also isn't running, and that an existing sb_server isn't.
+ Mark Hammond 23/12/2003 Fix uninstall problem - uninstall should be 100% clean, assuming Outlook isn't running.
Mark Hammond 23/12/2003 Tray app: Binary version failed to check for most recent version.
Skip Montanaro 23/12/2003 Sendmail annotates the Received: header with "(may be forged)" if it thinks the sender is forging its identity. Generate a token for this, if we are mining received headers.
Tony Meyer 22/12/2003 Move OE specific stuff out from UserInterface.py to oe_mailbox.py.
Mark Hammond 22/12/2003 Outlook: Default to background filtering being on for new versions.
+ Tony Meyer 22/12/2003 Web interface: A beginning at a more userfriendly interface to the testing setup.
+ Mark Hammond 22/12/2003 Outlook: When a frozen executable, addin.py becomes a mini-installer EXE for the DLL.
Mark Hammond 21/12/2003 Outlook: DWhen doing a "batch train" (eg, selecting multiple messages and saying "Delete as" or "Recover from",) the DB was saved in between each and every message. Now only saved at the end (which was always the intent)
Mark Hammond 21/12/2003 Outlook: DAs part of checking our configuration is invalid, make sure the user hasn't set us up such that either Spam/Unsure folders isn't also being watched for new messages
+ Mark Hammond 20/12/2003 dump_props.py: Add -c option, which writes output to the Windows clipboard.
+ Mark Hammond 20/12/2003 Outlook: Include the foldername in many messages, so help track down wierd bugs from user logs. Say what we are watching a folder for.
+ Mark Hammond 20/12/2003 Outlook: Fix [ 860410 ] SpamBayes allows top-level folders to be selected, and also prevent a single-select dialog from closing with no selection
Skip Montanaro 20/12/2003 Tokenizer: Solved the "backwards breakdown" problem with ip addresses in Received: headers.
Skip Montanaro 20/12/2003 Tokenizer: Tightened up recognition of hostnames and accepted bracketed or parenthesized ip addresses without requiring a leading space.
***************
*** 22,25 ****
--- 30,34 ----
Mark Hammond 19/12/2003 Outlook: Move some of our init code from OnConnection to OnStartupComplete
Tony Meyer 18/12/2003 Bring pspam into the modern SpamBayes world.
+ Mark Hammond 17/12/2003 Tray app: Better icons and icon loading code.
Tony Meyer 17/12/2003 Add the basis of a new experimental (and highly debatable) option to 'slurp' URLs.
Tim Peters 17/12/2003 Implemented the intended "tiling" version of x-use_bigrams.
***************
*** 31,41 ****
--- 40,55 ----
Tony Meyer 15/12/2003 Add the missing code for the Habeas headers tokenizing (and deprecate).
Mark Hammond 15/12/2003 Fix [ 833439 ] default_bayes_customize.ini is confusing.
+ Mark Hammond 14/12/2003 Move the option loading code to a function, then call this function as the module loads.
Tim Peters 14/12/2003 Removed support code for the defunct experimental_ham_spam_imbalance_adjustment option
Mark Hammond 14/12/2003 Fix [ 856628 ] reload(Options) fails in windows binary
Mark Hammond 14/12/2003 Fix [ 859215 ] "Restore Defaults" causes assertion error at exit.
Tony Meyer 14/12/2003 ImapUI: When logging in was done by the UI (to show available folders) we assigned the imap_session object to the wrong name
+ Skip Montanaro 10/12/2003 Loosen constraints on HEADER_VALUE regular expression.
+ Skip Montanaro 10/12/2003 Add ability for "x-" options (deprecated, or experimental).
Mark Hammond 10/12/2003 Outlook: Try and add the Spam field to the 'Unsure' folder in the same way we do for the Spam and watch folders.
Tony Meyer 04/12/2003 Tray app: Change the default (double-click) behaviour of the tray to "review messages" rather than "display information".
Tony Meyer 04/12/2003 Tray app: use SetDefaultItem (so the default action is in bold in the menu).
+ Mark Hammond 03/12/2003 For the unittest scritps avoid sys.path munging.
+ Mark Hammond 03/12/2003 Add new test_programs unittest script and support file for unittest scripts.
Mark Hammond 02/12/2003 sb_server was ignoring command-line options; fix.
Richie Hindle 27/11/2003 Sjoerd's improved version of patch 831388.
***************
*** 52,55 ****
--- 66,70 ----
Barry A. Warsaw 25/11/2003 New script: sb_evoscore.py - A shim script between sb_xmlrpcserver.py and Ximian Evolution.
Skip Montanaro 25/11/2003 Added a makefile to the testtools directory to make using timcv easier.
+ Neale Pickett 18/11/2003 Cleanup sb_filter and sb_mboxtrain.
Richie Hindle 16/11/2003 Patch [ 842464 ] Correct installation instructions from "setup.py install" to "python setup.py install"
Skip Montanaro 13/11/2003 sb_filter: add -o/--option command line arg that allows user to set any options value from the command line
Index: WHAT_IS_NEW.txt
===================================================================
RCS file: /cvsroot/spambayes/spambayes/WHAT_IS_NEW.txt,v
retrieving revision 1.22
retrieving revision 1.23
diff -C2 -d -r1.22 -r1.23
*** WHAT_IS_NEW.txt 29 Dec 2003 01:46:43 -0000 1.22
--- WHAT_IS_NEW.txt 29 Dec 2003 04:46:31 -0000 1.23
***************
*** 28,53 ****
Outlook Plugin
--------------
! o
POP3 Proxy / SMTP Proxy
-----------------------
! o
Web Interface
-------------
! o
POP3 Proxy Service / POP3 Proxy Tray Application
------------------------------------------------
!
! o
IMAP Filter
-----------
! o
General
-------
! o
Transition
--- 28,144 ----
Outlook Plugin
--------------
! o Add slightly better statistics.
! o When installing, along with checking Outlook isn't running, check
! Outlook's mail transport also isn't running.
! o Fix uninstall problem - uninstall should be 100% clean, assuming Outlook
! isn't running.
! o Default to background filtering being on.
! o When doing a "batch train" (eg, selecting multiple messages and saying
! "Delete as" or "Recover from") the database was saved in between each
! and every message. Now it is only saved at the end.
! o Ensure that the Spam and Unsure folders aren't also being watched for
! new messages.
! o Include the folder name in many messages, to help track down wierd bugs
! from user logs, and also say what we are watching a folder for.
! o Don't allow top-level folders to be selected.
! o Don't allow a single-select dialog to be closed without a selection.
! o Add the Spam field to the 'Unsure' folder as is done with the Spam and
! watched folders.
! o Improve matters when the default (Outlook message) store is offline.
POP3 Proxy / SMTP Proxy
-----------------------
! o smtpproxy is now only a module, not a script. Use sb_server instead.
! o sb_server was ignoring command-line options; fix.
Web Interface
-------------
! o Add the start of a user-friendly interface to the testing tools that are
! part of SpamBayes (for testing new features against corpora of ham and
! spam).
! o Fix a bug in the web interface where the probability would be
! incorrectly calculated on 'show clues'.
! o More robust code for parsing score headers - copes with the presence of
! logarithms.
! o More robust code for parsing evidence headers. Copes with ';' and ': '
! being part of a clue.
! o Increased the auth-digest login timeout from one minute to twenty.
! o Improve the 'Find Message' query on the front page of the web UI.
! o Add an 'advanced' word query (off by default).
! o Make the review messages page more customizable.
! o Provide a partially filled-in bug report message (please use it!).
! o Add basic statistics information.
! o Add a basic help system.
! o Add warning information, for example if the user has imbalanced
! training, or insufficient training.
! o Fixed an infinite loop when you break the browser connection to
! sb_server when sb_server is busy training.
POP3 Proxy Service / POP3 Proxy Tray Application
------------------------------------------------
! o Fix checking for most recent version.
! o Better icons and icon loading code.
! o Change the double-click (default) action to "review messages" and
! display the default in bold.
! o If a proxy is already running, don't start the service.
IMAP Filter
-----------
! o When a message is marked as deleted, it is now marked as read (seen) as
! well.
! o If sb_imapfilter.py is run without any switches, just serve the web
! interface (but don't launch a browser).
!
! sb_filter
! ---------
! o Print each message once, not once per argument.
! o Now obeys the notate_to and notate_subject options.
! o Added -o/--option command line argument, which allows setting any
! option via the command line.
! o Expanded the documentation.
! o Allow multiple types of mailboxes to be processed using mboxutils.getmbox.
! If any mailbox files are given on the command line, the output is always
! a Unix-style mailbox containing From_ lines.
! o If the -n switch was before the -d/-p switch, then the name wouldn't be
! used; this has been changed so that the -d/-p name is used wherever the
! -n switch is.
!
! Testing Suite
! -------------
! o Many improvements to the mksets.py script.
! o Many improvements to the rebal.py script.
! o Many improvements to the export.py script (for Outlook).
! o Added a makefile to the testtools directory to make using timcv.py easier.
!
! Tokenizer
! ---------
! o Sendmail annotates the Received: header with "(may be forged)" if it
! thinks the sender is forging its identity. Generate a token for this,
! if we are mining received headers.
! o Solved the "backwards breakdown" problem with IP addresses in Received:
! headers.
! o Tightened up recognition of hostnames and accepted bracketed or
! parenthesized IP addresses without requiring a leading space.
! o Add the missing code for the Habeas headers tokenizing (and deprecated it).
! o Removed support code for the defunct experimental_ham_spam_imbalance_adjustment
! option.
General
-------
! o Option names are always case insensitive, no matter what.
! o Moved the option loading code to a function.
! o Generalized the DirOfTxtFileMailbox class in mboxutils to assume all
! non-directory files contain a single message and to recursively descend
! into subdirectories of the argument directory.
! o Loosened constraints on HEADER_VALUE regular expression.
! o Import/Export data as utf-8 with sb_expimp.py.
! o Fixes to the which_database.py utility script.
! o Fix bug where if one was using Python 2.2, Windows and bsddb
! the database would never open correctly.
! o New script: sb_evoscore.py - A shim script between sb_xmlrpcserver.py
! and Ximian Evolution.
! o Fix the pspam scripts, muttrc and spambayes.el so that they work with
! the current SpamBayes package.
!
Transition
***************
*** 57,62 ****
<http://sourceforge.net/project/showfiles.php?group_id=61702>).
! Other than the changes listed in "Incompatible changes" at the top of
! this document, there are no known transition issues.
--- 148,160 ----
<http://sourceforge.net/project/showfiles.php?group_id=61702>).
! o To take advantage of improvements in the tokenizing engine at the core
! of SpamBayes, you may wish to do a full retrain; this is not necessary,
! however.
!
! o If you use the scripts in the testtools directory, you should examine
! how these scripts have changed (some command line switches are altered).
!
! See also the the changes listed in "Incompatible changes" at the top of
! this document.
***************
*** 64,68 ****
===================
The following bugs tracked via the Sourceforge system were fixed:
!
A URL containing the details of these bugs can be made by appending the
--- 162,166 ----
===================
The following bugs tracked via the Sourceforge system were fixed:
! 818871, 833439, 803798, 787676, 860410, 856628, 859215
A URL containing the details of these bugs can be made by appending the
***************
*** 73,90 ****
Feature Requests Added
======================
! No feature requests tracked via the Sourceforge system were added for this
! release.
!
Patches integrated
===================
! No patches tracked via the Sourceforge system were integrated for this
! release.
!
Deprecated Options
==================
! [add explanation of deprecated options here]
The following options have been deprecated in this release:
--- 171,203 ----
Feature Requests Added
======================
! The following feature requests tracked via the Sourceforge system were
! added for this release:
! 827138
Patches integrated
===================
! The following patches tracked via the Sourceforge system were integrated
! for this release.
! 842464, 831388, 809008, 831388
Deprecated Options
==================
! SpamBayes now has a method of noting options that are deprecated and which
! will not be available in future releases (it is likely that options will
! only be deprecated for one release before being removed). Deprecated
! options will not be offered in the graphical interfaces (Outlook plugin
! and web interface), and will be listed in the "What's New" file (this
! file) for each release.
!
! Deprecated options have the same name as previously, but now begin with
! "x-" (so "extract_dow" is now "x-extract_dow"). You can continue to use
! the original name (eg "extract_dow") in your configuration file, but will
! receive warnings in your log file or console window. We recommend that you
! examine this output every time you upgrade SpamBayes to ensure that you are
! not using any newly deprecated options.
!
! Discussion of why any particular option has been deprecated can be found in
! the spambayes-dev archives (at <http://mail.python.org/pipermail/spambayes-dev>).
The following options have been deprecated in this release:
***************
*** 97,110 ****
========================
! [add explanation of experimental options and a pointer to the testing
! setup here]
The following experimental options have been added in this release:
! o [Tokenizer] search_for_habeas_headers
! o [Tokenizer] reduce_habeas_headers
! o [Classifier] use_bigrams
! o [URLRetriever] slurp_urls
! o [URLRetriever] cache_expiry_days
! o [URLRetriever] cache_directory
! o [URLRetriever] only_slurp_base
! o [URLRetriever] web_prefix
--- 210,258 ----
========================
! SpamBayes now has a method of noting options that are experimental and
! which may be removed or made permanent in future releases (many experimental
! options will only be experimental for one release before being removed or
! fully integrated). Experimental options will not be offered in the
! graphical interfaces (Outlook plugin and web interface), and will be
! listed in the "What's New" file (this file) for each release.
!
! Experimental options begin with "x-" (as do deprecated options). If you
! start using an experimental option and it later becomes permanent you can
! continue to use the "x-" name in your configuration file, but will
! receive warnings in your log file or console window. We recommend that you
! examine this output every time you upgrade SpamBayes to ensure that you are
! using the correct name for all options.
!
! Discussion of why experimental options and results from using them can be
! found in the spambayes-dev archives (at
! <http://mail.python.org/pipermail/spambayes-dev>). Ideally, we would like
! users to test these options out on their mail and let us know the results.
! This can be as simple as turning on the option and emailing
! spambayes at python.org with anacdotal results after a period of time, or the
! full testtools scripts can be used. For details about using these, please
! read the "README-DEVEL.txt" file that comes with the SpamBayes archive.
!
! Experimental options are always turned off by default.
The following experimental options have been added in this release:
! o [Tokenizer] x-search_for_habeas_headers
! o [Tokenizer] x-reduce_habeas_headers
! These generate tokens based on the Habeas headers (see
! <http://habeas.com> for more details).
!
! o [Classifier] x-use_bigrams
! By default, SpamBayes uses unigrams tokens that are basically
! single words (split on whitespace). This option enables both unigrams
! and bigrams (pairs of words), but uses a 'tiling' scheme, where only
! the set of unigrams and bigrams that have the strongest effect on
! the message are used.
!
! o [URLRetriever] x-slurp_urls
! o [URLRetriever] x-cache_expiry_days
! o [URLRetriever] x-cache_directory
! o [URLRetriever] x-only_slurp_base
! o [URLRetriever] x-web_prefix
! If these are used, if a message is scored as 'unsure', and could use
! more tokens in it's classification, then text from any URLs in the
! message is retrieved and used, if it makes a difference to the
! classification.
More information about the Spambayes-checkins
mailing list