Preparing Watermark for incorporation to core component suite (RF-12563)

[lfryc]

This issue was opened to track changes of Watermark component in order to incorporate it into core component suite.

I did following modifications:

• fully utilized CDK XML templates
• used UIOutput as base class instead of UIInput
• removed suffix attribute (now inputs are automatically determined)
• removed for attribute (unified usage through nesting into input components)
• removed support for non-JSF inputs

[buildhive]

RichFaces » sandbox #238 SUCCESS This pull request looks good (what's this?)

[blabno]

Oh, I cry. Why did you remove suffix,for and non-JSF inputs support? It was so great. When I was writing watermark the rich:calendar was broken and I could not nest the watermark inside rich:calendar. In future we may have same situation, i.e. with components not from RF. As to "suffix" I don't believe in automatic discovery of input with that jQuery trick. What if there is more then one input in the component? i.e. visible input and a hidden one?

[lfryc]

Hey Bernard,

thanks for a feedback! It is very appreciated!

Just note the changes above are just proposals to be discussed, let me explain my thoughts:

• at overall, I tried to made an API as much JSF-like as possible
• and made implementation as minimal as possible while staying feature complete

Here are detailed explanations:

• for attribute seems unnecessary
  • component has one-to-one mapping component-to-watermark
  • separation of watermark definition from input component may be confusing, thus I would prefer nesting
• I have removed suffix because it does not play a nice with a JSF components behavior
  • JSF user should not be forced to know what is particular ID (or its part) of the input inside component DOM
  • jQuery trick can be surely improved to be more accurate - having :visible selector is very good suggestion
• I have removed support for non-JSF usage because there is no other RF component which supports that
  • this is certainly subject for discussion
• watermark now works when nested in calendar
  • in RF 4.3 we will make sure watermark works with each component in RF where it does make sense
  • library inter-operability issues will be addressed as they come (to be honest, we don't have much inter-operability issues reported in issue tracker)

What do you think?

My analysis might miss something important and that's why I'm really glad you can review it.

[lfryc]

It would be very helpful if you could explain in detail where the component does not meet its use cases - what are additional requirements I missed.

We could discuss details on team meeting - I have added it to agenda: https://community.jboss.org/wiki/RichFacesTeamMeetingAgenda2012-10-30 By any chance would you be able to come today?

[lfryc]

We have agreed on following points on team meeting [1]

• remove @for attribute from original watermark implementation
• use CSV's approach for finding proper input element - generalize implementation as RichFaces script utility function (with clientId and callback as parameters)
  • https://github.com/richfaces/components/blob/develop/validator/ui/src/main/resources/META-INF/resources/org.richfaces/richfaces-csv.js#L110
• rename to placeholder
• replace watermark's @suffix and @for with @selector
• Chrome support for placeholders' classes might be forced in watermark widget

[1] https://community.jboss.org/wiki/RichFacesTeamMeetingAgenda2012-10-30

[buildhive]

RichFaces » sandbox #244 SUCCESS This pull request looks good (what's this?)

[lfryc]

Watermark is now prepared to be included in core suite - it just has to be renamed to Placeholder.

[lfryc]

The placeholder component was adopted to core suite in https://github.com/richfaces/components/commit/cdf39c4a7232e34a686ff70b50cf6ab93b8c4146