widget status
Kevin Dangoor
able to do anything else on widgets (other than commit a couple of
patches). The API still has a couple more tweaks to go (see the
separate discussion about cleaned up widget APIs). These should be
minor tweaks, so if you want to start coding against the new API just
run turbogears.widgets.use_new_widgets() and have at it!
I hope that on Monday we can finalize those last details. At that
point, I'll flip-flop the old and new as I did earlier in the week
with controllers.py.
Kevin
--
Kevin Dangoor
Author of the Zesty News RSS newsreader
email: k...@blazingthings.com
company: http://www.BlazingThings.com
blog: http://www.BlueSkyOnMars.com
Randall
I know this is alpha software so things break, but I want to let you
know, if you didn't already, that the last patches broke lots of stuff.
You'll see if you run a DataController. First, the Form class
requires a non-keyword arg for the widgets(fields), so line 209 in
datacontroller.py needs to have the kwarg removed. After fixing that
I'm getting this error, which I don't know how to fix yet.
svn/turbogears/turbogears/fastdata/templates/datagrid.py", line 31, in
_pull
NameError: name 'name' is not defined
Randall
Randall
yet.
Alberto Valverde
Alberto Valverde
>
> The problem you're suffering is that DataGrid has been upgraded
> (since r637) to the new widget's API. I think Kevin has tried his
> best to keep everything backwards compatible with the old API (by
> creating newforms.py and newbase.py, etc) but missed it on this one
> (I wrote the patch and didn't think this could happen either, so I
> guess I'm the one to blame here)
Just clarifying myself... I wrote the DataGrid updater patch (#502),
not the new widgets API (which is Kevin's, Michele's and David's
great work ;)
Alberto
Randall
commented #use_new_widgets() function. Should I just uncomment that?
Don't mean to complain. Just trying to let you know what's broken if
you didn't already. I appreciate your work.
Randall
Kevin Dangoor
> I know this is alpha software so things break, but I want to let you
> know, if you didn't already, that the last patches broke lots of stuff.
I hadn't considered that aspect when I applied the patch. Here's what
I've done as a corrective measure: I've put the old fastdata package
in place and moved the new one into "newfastdata". In my current
mental state (it's a bit late for me), I couldn't come up with a good
way around circular imports, so if you want to use the new widgets
with the updated grid widget, you'll need to
import turbogears
from turbogears import newfastdata
turbogears.fastdata = newfastdata
I'm not even 100% certain this will work, because of the template.
Hopefully on Monday I'll be able to switch in the new widgets (thus
breaking everything for people with custom widgets, but hopefully it's
a quick migration).
Kevin
Randall
you had a good weekend. Just want to get this down while its on my
mind. I switched to the new API and I'm just using the DataController
against a simple two text field SQLObject class. There is a problem
with kid (I think) rendering for an edit and the output is nearly
useless.
TypeError: dispatch_error() got multiple values for keyword argument
'tg_errors'
I've stepped thru code trying to find out where the exception is thrown
but no luck yet. I have noted that in the kid template form.kid, when
obj is replaced with None in ${form.insert(obj, action=action)}, the
form renders fine. When obj is included it throws the exception shown
above. I checked that obj has both attributes (permissionId,
description) required by the form.
Randall
this is interesting:
File
"/home/randall/downloads/svn/turbogears/thirdparty/kid/kid/pull.py",
line 206, in _coalesce
for ev, item in stream:
File "<string>", line 66, in _pull
File
"/usr/lib/python2.4/site-packages/SQLObject-0.7.1dev_r1516-py2.4.egg/sqlobject/main.py",
line 898, in get
id = cls.sqlmeta.idType(id)
ValueError: invalid literal for int(): permissionId
It looks like the SO get method is sent the wrong value. That seems
weird to me though becuase by the time form.insert is called, the SO
instance has already been fetched.
Alberto Valverde
py:replace="field.display(value=value.get(field.name),
options=options.get(field.name))"
A you can see, the form template tries too fetch the object's value
via "get" (instead of normal attribute lookup).
I fixed, err, worked-around it, by sending the SO object as a
dictionary via database.so_to_dict.
I'k not sure wether this is a bug or new behaviour, though....
Alberto
Kevin Dangoor
needs to wrap incoming dictionaries in a util.DictWrapper.
Kevin
Michele Cella
> I ran into this problem too... The porblem is at neforms.py:415 in:
> py:replace="field.display(value=value.get(field.name),
> options=options.get(field.name))"
>
> A you can see, the form template tries too fetch the object's value
> via "get" (instead of normal attribute lookup).
>
> I fixed, err, worked-around it, by sending the SO object as a
> dictionary via database.so_to_dict.
> I'k not sure wether this is a bug or new behaviour, though....
>
I talked with Kevin about this two days ago, it's not a new behavior
but my fault. :-/
This will be fixed with an AttrDict wrapper.
Ciao
Michele
Michele Cella
>
> I talked with Kevin about this two days ago, it's not a new behavior
> but my fault. :-/
> This will be fixed with an AttrDict wrapper.
>
Wow, I'm really slow.
Sorry, didn't notice Kevin already replied.
Ciao
Michele