Dan Miron wrote: > Hi guys! I'd like to know what you think about this matter, which has > been raised on http://jira.xwiki.org/jira/browse/XWIKI-3375. > As I posted there, this is what I think: > > -the parameters for the two macros are completely different, so i see no > point for having the box specific parameters (cssClass, title, image, > width and blockTitle) among the rss macro's ones, this will lead to > confusion for the user. The box parameters are deduced from the rss > feed's properties and then passed to the box macro. No need for them to > be exposed in the Rss Macro. Therefore, extending the RssMacroParameters > from the BoxMacroParameters is unreliable. > > -extending the RssMacro from the AbstractBoxMacro<RssMacroParameters> > doesn't mean simply implementing AbstractBoxMacro.parseContent instead > of Macro.execute. Currently, most of the code about the box around the > css macro is placed in the box macro's implementation, which is > DefaultBoxMacro, so we do make use of the existing implementations. > Giving up using the box macro internally means giving up using most of > the features already presented in the box macro and rewriting them > > -and finally, the most time costing disadvantage is that extending the > rss macro from the box basically means rewriting this macro from > scratch, because it involves redesigning it, task that will cost us time. > > So, therefore, I'm -1 for this.
Inheritance implies an "is-a" relationship. So, the important question is: is a RSS display a Box? Will a RSS display always be a box? IMHO, no. The fact that currently the RSS macro uses a box is a presentation detail that might change in the future. Hierarchies are harder to change, compositions are easier. +1 for composition (RSS not extending Box, but using one internally, as it is done now). -- Sergiu Dumitriu http://purl.org/net/sergiu/ _______________________________________________ devs mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/devs
