Revision history for Class-DBI-FormBuilder
0.43 Tues 9 Aug 2005, 1:20
- added as_multiform() method
- added create_from_multiform() method
- added post_processor architecture (e.g. for pretty-printing)
- added PrettyPrint post-processor
- added form_final(), which applies field processors in the '__FINAL__' slot
of process_fields
0.421 Wed 3 Aug 2005, 23:55
- skip annoying tests in 51.related.t
0.42 Wed 3 Aug 2005, 23:50
- changed to using column objects throughout, and stringifying them via $column->mutator
or $column->accessor whenever a field name is required (pointed out by Dave Howorth)
- PROBABLY STILL BROKEN THOUGH - works fine unless you are customising your CDBI accessors,
which was the point, but should be fixed in the next release
- added column_meta() and column_type() methods as public methods
- plugins now receive a column object instead of a field name
- processors now receive a column object as an extra argument (used by +VALUE)
- changed to retain values as objects wherever they are supplied as objects, and stringifying
just before sending to CGI::FB
- added pretty-printing feature
0.41 Tues 2 Aug 2005, 23:40
- added _load_meta() method to access lots of column metadata via DBI:
- removed requirement to load Class::DBI::Plugin::Type
- fixed _db_order_columns()
- support MySQL SET columns in _get_col_options_for_enumlike() without needing the
patch to CDBI::mysql
- removed the code that sets the 'required' field list to the list in a 'Required' column
group - this has been deprecated for a while
- added TIMESTAMP processor. form_timestamp now uses this instead of DISABLED, so it can be easily
replaced
- added +SET_$foo($bar) processor
- deprecated +SET_VALUE($value) processor - use +SET_value($value)
- altered the precedence rules for processors - a custom non '+' processor will have the effect
of dropping all auto-generated processors, so that only the custom processors run. '+'
processors are run in addition to automatically set processors.
- re-organised POD
0.4 Fri 22 July 2005, 23:00
- added +READONLY processor - note this also sets the class of the input field to 'ReadOnly'
- added +DISABLED processor - note this also sets the class of the input field to 'Disabled'
- added +SET_VALUE($value) processor
- added +ADD_FIELD processor
- added +NULL processor
- added 'timestamp' to @BASIC_FORM_MODIFIERS, and form_timestamp method
which defaults to the DISABLED preprocessor
- added 'process_extras' to @BASIC_FORM_MODIFIERS, with form_process_extras method
- fixed _get_type to detect column type for related fields
- much more sophisticated chaining of preprocessors
- the default is for fields not to be required (CGI::FB defaults to required, if a field has
a validation spec, but in CDBI::FB, *all* fields have a validation spec)
0.3511 Thur 21 Jul, 2005 09:00
- added 'int' entry to ValidMap (reported by Dave Kamholz)
0.351 Wed 20 Jul, 2005 01:00
- cleaned up code in _db_order_columns - still not understanding
the empty 'All' group issue though
- factored out calculation of field names for form_has_many() and form_might_have() into
_multiplicity_fields()
0.3502 Mon 18 Jul, 2005 23:15
- added test for _db_order_columns fatal error when looking up
implicitly defined 'All' column group. Still returns an empty list,
so haven't got this one figured out yet, but at least the $@ is not
being propagated further than expected.
0.3501 Sun 17 Jul, 2005 22:30
- POD: re-arranged and fixed a few bits and pieces
0.35 Sun 17 Jul, 2005 15:30
- implemented field pre-processing architecture
0.3453 Sun 17 Jul, 2005 13:30
- improved subclassability by storing the CDBI::FB (sub)class on the
caller, and then using that instead of __PACKAGE__
0.3452 Sat 16 Jul, 2005 22:30
- half fixed _db_order_columns bug. The internal lookup could die
for the 'All' group if this group was not set up explicitly.
Looks like Maypole (via CDBI::Loader?) normally does set this up
explicitly, but hand-rolled setups normally will not. The fix will
return the columns for the 'All' group (via CDBI->columns( 'All' )
in non-db order if the group is not defined explicitly.
Reported by dkamholz.
- highlighted Class::DBI::Plugin::Type issue in GOTCHAS (reported by dkamholz)
- add 'our $pager' declaration to pager template (reported by dkamholz)
0.3451 Thur 14 Jul, 2005 13:40
- removed restriction on setting keepextras to 1
0.345 Thur 14 Jul, 2005 10:00
- only add pk fields to forms built for objects, not for
classes or searches. Empty pk hidden fields were being submitted
as empty strings rather than undef in some situations.
- renamed form_hidden() to form_pks()
0.3441 Wed 13 Jul, 2005 23:10
- an extra test
0.344 Wed 13 Jul, 2005 23:00
- don't send undefs to create() for columns not represented in form data
0.343 Wed 13 Jul, 2005 13:24
- previous fix was broken if caller tries to set keepextras to 1. Now
we just croak, the proper fix is to remove pks from validation, or
figure out what fields are required and set that explicitly.
0.342 Wed 13 Jul, 2005 10:45
- fixed bug where keepextras set by caller were overridden in _get_args
(reported by David Kamholz)
- added big warning about as_form_with_related to pod.
0.341 Mon 11 Jul, 2005 09:45
- fixed rt link
- tests now correctly bail if DBD::SQLite2 is not available
0.34 Fri 1 Jul, 2005 16:30
- update_from_form_with_related() is beginning to work
0.33 Thur 30 Jun, 2005 11:00
- some pod formatting fixes
- spot of refactoring inside _splice_form
- basic untested brobably broken update_from_form_with_related() added
- multiple pk hidden fields bug still not fixed
0.32 Wed 29 Jun, 2005 15:00
- lots of refactoring in form_* methods
- fixed disappearance of has_many fields
- as_form_with_related seems to be rendering forms correctly now
- except for multiple pk hidden fields bug
(haven't tried processing submitted forms yet)
- added plugin architecture (by James Tolley)
- added dependency on UNIVERSAL::require
0.31 Tues 28 Jun, 2005 15:00
- as_form_with_related getting better
0.3_1 Tues 28 Jun, 2005 10:00
- first code for as_form_with_related
- new dependency on List::Util
0.3 Fri 24 Jun, 2005 14:20
- added automatic configuration of validation settings
- new dependency on Class::DBI::Plugin::Type
0.2 Tues 14 Jun, 2005 23:10
- added search_opt_cmp and search_opt_order_by options to search_form and
search_where_from_form
- has_a fields are now automatically added to the required list
0.121 Sun 12 Jun, 2005 13:20
- included 34.might_have.t in MANIFEST
0.12 Sat 11 Jun, 2005 23:50
- added might_have support
0.11 Fri 10 Jun, 2005 12:00
- POD fixes
0.1 Thur 9 Jun, 2005 22:45
First version, released on an unsuspecting world.