[web2py] Re: IS_IN_DB(multiple=True) problems on GAE

2012-09-15 Thread Alan Etkin
El viernes, 14 de septiembre de 2012 20:08:23 UTC-3, howesc escribió: > > i noticed that in 2.0.8 and trunk that IS_IN_DB when multiple=True tries > to do a belongs query with the string values posted by the user rather than > int values needed for the query (at least on GAE). bug report + propo

[web2py] Re: IS_IN_DB(multiple=True) problems on GAE

2012-09-15 Thread howesc
your patch does the correction in the DAL - i almost suggested the same thing (and since i was working off of 2.0.8 at the time that's why i missed your change). though i think it is possible to have key fields in GAE that are not ints so i was hesitant to make the change in the DAL myself. ei

[web2py] Re: IS_IN_DB(multiple=True) problems on GAE

2012-09-15 Thread Alan Etkin
> > your patch does the correction in the DAL - i almost suggested the same > thing (and since i was working off of 2.0.8 at the time that's why i missed > your change). though i think it is possible to have key fields in GAE that > are not ints so i was hesitant to make the change in the DAL

[web2py] Re: IS_IN_DB(multiple=True) problems on GAE

2012-09-15 Thread howesc
web2py default IDs are integers. we have some GAE users who are using named keysor at least i added support for named keys on GAE based on some user's requests. i don't have a generic solution for the DAL that chunks belongs queries in groups of 30, so it is still left as something the use

[web2py] Re: IS_IN_DB(multiple=True) problems on GAE

2012-09-15 Thread Alan Etkin
>i'm undecided anyway if this should be automatic for users or if they should be forced to realize that large belongs queries are >implemented as multiple datastore calls on GAE. For me, the automatic breaking should be the default, with a documented warning that surpassing the 30 item limit wi

[web2py] Re: IS_IN_DB(multiple=True) problems on GAE

2012-09-17 Thread howesc
that's fair. i don't know how to get the DAL to handle that based on how filters are applied when a select is executed On Saturday, September 15, 2012 3:45:30 PM UTC-7, Alan Etkin wrote: > > >i'm undecided anyway if this should be automatic for users or if they > should be forced to realize