Yep, that would be good.
On 10/11/07, André Bargull <[EMAIL PROTECTED]> wrote: > But first you need to add a setter for password in LzInputTextSprite... > > And people try to use constraints for "password" quite often: > http://forum.openlaszlo.org/showthread.php?t=4684 > http://forum.openlaszlo.org/showthread.php?t=7394 > http://forum.openlaszlo.org/showthread.php?t=9862 > ... > > JIRA: http://www.openlaszlo.org/jira/browse/LPP-704 > > > Or, a klever hacker would realize that they should just define a > > setter for the parent password attribute that applies the setting > > where it is actually needed (saving the overhead of events that the > > person who wrote the $once constraint was trying to avoid). I.e., > > 'push' the password field to where it is needed. [cc-ing John. there > > is probably a place in the docs for the following:] Remember Adam's > > first principle of constraints: "They are really handy for rapid > > prototyping, but eventually you want to eliminate them." On > > 2007-10-10, at 18:43 EDT, Henry Minsky wrote: > >> > If you look in the edittext component, it says the internal text field > >> > is using a 'once' > >> > constraint to look at the parent.password > >> > > >> > lps/components/lz/edittext.lzx: > >> > > >> > <_internalinputtext name="field" > >> > password="$once{parent.password}" > >> > > >> > So the internal field is not looking at the parent.password value more > >> > than once at init time. > >> > That could be changed to an 'always' constraint, I imagine that would > >> > work better. > >> > > >> > > >> > > >> > On 10/10/07, James Robey <[EMAIL PROTECTED]> wrote: > >> > >>> >> Hi all, I think i may have found a bug in the edittext component, > >>> >> but i > >>> >> thought someone might show this to be correct behavior in a way I > >>> >> don't yet > >>> >> see. The issue is setting password="true" on an <edittext/> > >>> >> component. It > >>> >> doesn't seem like the password attribute can be constrained. This > >>> >> example > >>> >> might also reveal a bug with $immediately{} constraints wherein > >>> >> "classroot" > >>> >> and "parent" don't exist when the constraint is evaluated. > >>> >> > >>> >> Anyone have advice on how to make this work? > >>> >> > >>> >> Test case: none of these fields (save the last hardcoded editext) has > >>> >> password styling. > >>> >> > >>> >> <canvas proxied="false" debug="true"> > >>> >> <debug/> > >>> >> > >>> >> <class name="pwtest"> > >>> >> <attribute name="passwordB" type="boolean" value="true"/> > >>> >> <attribute name="passwordS" type="string" value="true"/> > >>> >> <attribute name="passwordE" type="expression" value="true"/> > >>> >> > >>> >> <simplelayout axis="y"/> > >>> >> > >>> >> <edittext password="${classroot.passwordB}"/> > >>> >> <edittext password="${classroot.passwordS}"/> > >>> >> <edittext password="${classroot.passwordE}"/> > >>> >> > >>> >> <edittext password="$immediately{classroot.passwordB}"/> > >>> >> <edittext password="$immediately{classroot.passwordS}"/> > >>> >> <edittext password="$immediately{classroot.passwordE}"/> > >>> >> > >>> >> <edittext password="$once{classroot.passwordB}"/> > >>> >> <edittext password="$once{classroot.passwordS}"/> > >>> >> <edittext password="$once{classroot.passwordE}"/> > >>> >> > >>> >> <edittext password="true"/> > >>> >> </class> > >>> >> > >>> >> <pwtest/> > >>> >> > >>> >> </canvas> > >>> > >> > > >> > > >> > -- > >> > Henry Minsky > >> > Software Architect > >> > [EMAIL PROTECTED] > >> > > -- Henry Minsky Software Architect [EMAIL PROTECTED]
