IS_MATCH is using the default strict=False parameter in this case, so
it returns the part that matches.
Using strict=True it won't match:
>>> IS_MATCH('[\w\.\-]+', strict=True)('my name')
('my name', 'invalid expression')
Anyways, userids are not supposed to have blanks.
On Dec 23, 10:46 am, Anthony <[email protected]> wrote:
> By default, the 'username' field gets the following validators:
>
> (IS_MATCH('[\w\.\-]+'), IS_NOT_IN_DB(db, table.username))
>
> The IS_MATCH is doing the truncating. Actually, I think that's a bug --
> IS_MATCH isn't supposed to transform the input, only check the regex
> pattern. For now, though, I suppose you could overwrite that tuple with
> your own validators:
>
> auth.settings.table_user.username.requires = [IS_MATCH('[\w\.\- ]+'),
> IS_NOT_IN_DB(db, auth.settings.table_user.username)]
>
> Note, I added a space to the IS_MATCH regex.
>
> Anthony
>
>
>
>
>
>
>
> On Friday, December 23, 2011 10:11:44 AM UTC-5, greenguerilla wrote:
>
> > Hi,
>
> > I noticed something strange when testing the registration of users.
>
> > If I attempt to register as user: 'My Name' (note the space), the form
> > validation passes and I get an 'email sent' message.
> > What actually happens though, is the the username gets inserted into
> > the db as 'my'.
> > I'm fine with the lowercasing but I would rather strip the spaces and
> > insert the user as 'myname'
> > I have been searching for a hook to do this and I have tried the two
> > methods:
>
> > auth.settings.register_onvalidation = clean_username
> > auth.settings.register_onaccept = clean_username
>
> > Where 'clean_username' is a function to strip all the spaces from the
> > entered text.
> > My problem is that the username has already been stripped by the time
> > it gets this far. Strangely enough this doesn't happen for the
> > username field in the login form.
>
> > Does anywhere know where this might be happening and how I can
> > override it?
>
> > Thanks,
>
> > John
