Alberto Valverde wrote: > Ok, refactored meta stuff to widgets.meta. I had to add a compound > attr to Widgets so meta can know when a field is compound without > circular imports. >
See below... > On 10/03/2006, at 19:40, Michele Cella wrote: > > That was not exactly the question I wanted to make, sorry. What I > meant is that unpacked_error dicts have a key for compound widgets > too when they have an error. I think it contains a list of all errors > for all inner widgets, this is just how FE works. I noticed this on > the browser test so cannot say for sure what format it's in (has > __str__ for sure :) You can try to reproduce it by commenting those > two lines beneath the comment (sic). Right now they are ignored. > I'm not quite tuned on this thing will take a look when I have time... :D > > > > I think that without ignore_key_missing = True disabled_fields will > > not > > work right, am I wrong? this wil also fix #659 I guess... not sure. > > Mmmm, I'm not sure but I think they will not :( However, I think we > can add it back safely as ignoring missing keys shouldn't be as > dangerous as accepting spurious keys. Anyway, I think that maybe we > shouldn't add because you can always override that behavior by > passing a custom schema to forms which you expect to disable some > fields. > Isn't disabled_fields intended to be used at display time? for example to redisplay a form without some fields? in that case the validator is already ready, so I guess we need the ignore_missing can you give it a shot while committing? I'm pretty sure it will fix also Jason bug. > What I'd really like to fix when this get's in (I have a sort of > cumbersome poor-man's svn with this patch numbering stuff :) Is the > allow_extra_fields = True (setting it to false) and possibly adding a > filter_extra_fields=True too. Needs some more testing though as some > tests where failing when I tried it (probably because some input- > expecting widget's didn't have a validator) but didn't look into it > too closely... There's a post in this thread where I explain better > the reasoning behind this... > We may ask Kevin what's opinion, should we expect a validator for every field? > My goal here is too make the default Schema @ validators as strict as > possible (secure by default) but with the possibility of loosening it > up by subclassing. > That's a good goal. ;-) > Well, I'm ready to commit :) , comments on validation9.patch @ #613? > Yep, :P - why return getattr(widget, 'full_qualified_form', True) shouldn't this be False for default? just wondering... - can we remove compound from the base widget class? and instead use getattr(widget, "compound", False)? it was there before but I've removed it since it semmed silly from a parent point of view... :D Other than that, wonderful work. Ciao Michele --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "TurboGears Trunk" group. To post to this group, send email to [EMAIL PROTECTED] To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/turbogears-trunk -~----------~----~----~----~------~----~------~--~---
