Commits

Author Commit Message Labels Comments Date
Alexandre Bourget
Deserialization cache: don't add overhead for all Fields, only modify Field if decorator is used.
Alexandre Bourget
Implement an optional tiny deserialization cache. Documented. Tests pass, though no new tests were added to make sure this doesn't break in the future. The reason to implement this is that the deserialize() function gets called twice in a normal controller. Once with validate() and a second time with sync(). For custom Renderers that have complex deserialization (fetching to a database, creating temporary objects, appending to a DB relation), you don't wa…
Alexandre Bourget
Renamed ._params to .params everywhere. Lots of documentation fixes. Added `.errors`. - `.errors` point to the FieldSet's errors. This can be used to indicate we're re-displaying a form which has errors in it. - `.params` should be useable by extended classes - deprecation warning
Alexandre Bourget
#84 Make sure we can call .configure() multiple times, with consistent and predictable results. Before this patch, the outcome of calling configure() twice was unpredictable, thus no old code should be broken. Removed some documentation to ModelRenderer for params it didn't have. NOTE: there is duplication with documentation in docs/forms.txt from ModelRenderer. It is now possible to call a fs.configure(options=[fs.field1.required()]) and it will simply replace the `field1` in the _render_…
Alexandre Bourget
Revenu sur les changements : b980b3a13a81
Alexandre Bourget
Implemented the __repr__or a fieldset. Tests included. Will output: >>> fs = FieldSet(Two) >>> fs <FieldSet with ['id', 'foo']> >>> fs.configure(pk=True) >>> fs <FieldSet (configured) with ['id', 'foo']>
Alexandre Bourget
Update documentation about default Renderers, based on code. In the documentation, it was previously copied, and not in sync. Using a literalinclude will make sure the doc is in sync with the code.
Gael Pasgrimaud
fix 131
Alexandre Bourget
Merge main branch
Alexandre Bourget
Added tests for the new set(required=True) introduced two commits earlier.
Alexandre Bourget
Added insert_after() on the FieldSet object. Fixed display bug. Tweaked docs. Tests included. The display bug was just showing off the wrong object when raising the TypeError. The reason for adding insert_after(), is that in some cases, you want to add a field related to the previous field and not the next field. For example, you have a `password` field. If most of the time, you just want a single field for the password to appear, then you let it that way. If at some point, you want to…
Gael Pasgrimaud
add 1.3.3 entry
Gael Pasgrimaud
run test with SA > 0.6
Gael Pasgrimaud
Added tag 1.3.2 for changeset 916d1a7403f0
Gael Pasgrimaud
bump version
Tags
1.3.2
Gael Pasgrimaud
add exceptions module with PkError. add key value to Exception message
Gael Pasgrimaud
fix find links
Gael Pasgrimaud
fix find links
Gael Pasgrimaud
merge
Tarek Ziadé
added more information on why the model could not be instanciated. This can also happen if the model is unsynced with the DB so displaying the error helps
Tarek Ziadé
removed trailing space before my actual change
Gael Pasgrimaud
merge 756
Alexandre Bourget
Implement Field.set(required=True), to change the required-ness in-place, instead of creating a new object. Once you use the in-place mode, you don't want to switch back to those long chainings of .required().with_metadata(), etc that don't give easy line-wrapping (you must escape lines with \, and it's cumbersome).
Alexandre Bourget
Implemented value_objects on Fields and Renderers, to have access to objects rather than list of PKs.
Alexandre Bourget
Added support for value_objects
Gael Pasgrimaud
assume ShemaListProperty work
Gael Pasgrimaud
better test
Gael Pasgrimaud
add support for Schema*Property
Gael Pasgrimaud
changes
Gael Pasgrimaud
oops, broken test
  1. Prev
  2. Next