Thanks Anthony.
Yes, actually I am looking now and I see that it was late and Yes, it
started working when I set lazy_tables to True, not False.
And omitting requires= was a mistake. I will test it without the entire
IS_IN_DB string now. I expect it will work just fine. Sometimes it's hard
to see the forest for the trees. Thanks for pointing that out.
On Friday, January 29, 2016 at 11:15:13 AM UTC-5, Anthony wrote:
>
> On Thursday, January 28, 2016 at 11:24:16 PM UTC-5, Dave wrote:
>>
>> Actually, maybe this is a bug. (or maybe just something that is not
>> supposed to work the way I am trying to use it...)
>>
>> If I turn off lazy_table support, everything works. I get a multi-select
>> box with the file_name populated (text area that I can multi-select).
>>
>
> Actually, it is the reverse -- you get the multi-select box with
> lazy_tables=True, but not with lazy_tables=False.
>
> Here is how it works. The multi-select widget is generated automatically
> whenever a field has an IS_IN_DB(..., multiple=True) validator.
> Furthermore, list:reference fields automatically get an IS_IN_DB validator,
> but only when lazy_tables=True. The reason is that the IS_IN_DB validator
> is created when the auto_validators() function is called, and that function
> only assigns validators to reference fields if the referenced table name is
> in db.tables. Here is the order of operations in each case:
>
> lazy_tables=True:
>
> 1. At the end of db.define_table(), the new "tablename" is added to
> db.tables.
> 2. Sometime later, when db.tablename is accessed, the table is fully
> defined via db.lazy_define_table.
> 3. db.lazy_define_table calls auto_validators, which adds the IS_IN_DB
> validator because "tablename" is in db.tables (see #1).
>
> lazy_tables=False:
>
> 1. Before the end of db.define_table(), db.lazy_define_table is called
> (prior to the new "tablename" being added to db.tables).
> 2. db.lazy_define_table calls auto_validators, which does *not* add
> the IS_IN_DB validator because "tablename" is *not *yet in db.tables
> (see #1).
> 3. The new "tablename" is added to db.tables *after*
> db.lazy_define_table and auto_validators have already run.
>
> I'm not sure I would call it a bug, but I think we may be able to adjust
> the code to make it work in either case (possibly by simply adding the new
> tablename to db.tables *before* db.lazy_define_table is called within
> db.define_table -- though I'm not sure if that will introduce some other
> problem).
>
> Maybe open a Github issue and link to this thread.
>
> If you do not want to use lazy tables, there is a simple workaround --
> just explicitly define the validator yourself:
>
> Field('children', 'list:reference file_entry',
> requires=IS_IN_DB(db, 'file_entry.id', '%(file_name)s', multiple=
> True))
>
> It appears you tried to do that, but you did not use the keyword argument
> "requires" and instead passed the validator as the third positional
> argument (which is actually the "length" argument).
>
> Note, the same issue exists with the default "represent" attribute of the
> list:reference field -- you will need to use lazy tables or otherwise
> explicitly specify that attribute as well.
>
> Anthony
>
>
--
Resources:
- http://web2py.com
- http://web2py.com/book (Documentation)
- http://github.com/web2py/web2py (Source code)
- https://code.google.com/p/web2py/issues/list (Report Issues)
---
You received this message because you are subscribed to the Google Groups
"web2py-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to web2py+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
