On 2009-07-10, at 10:04EDT, P T Withington wrote: > On 2009-07-10, at 09:51EDT, André Bargull wrote: > >> On 7/10/2009 3:20 PM, P T Withington wrote: >>> Not approved. >>> >>> When Max submitted his change r13009 for review, I asked him: >>> >>> >>>> 5. Isn't making resize default to false for input text an API >>>> change? >>>> >>> >>> So maybe he took that out, in response to my comment? >>> >>> My question still remains. This is going to fix Amazon, but it >>> will be an API change that will potentially break other >>> applications that were depending on the existing default. >>> >>> [...] >>> >>> I think we need to ask: >>> >>> 1) Should input text resize default to true? If so, what does >>> that mean? (Since, as we see, you can't enter anything if it is >>> initially empty). >>> >> >> Hmm, you wanted it to default to true: >> "http://www.openlaszlo.org/pipermail/laszlo-dev/2008-June/015718.html >> " > > Thanks for reminding me! > > LPP-5678 was complaining that resizing inputtext _did_ work in LPS3, > it was just not the default. But in LPS4, we change the default for > <text> resize from false to true. And since <inputtext> is a > subclass of <text>, it inherited that default. Given that was an > explicit API change, and has been in LPS4 since forever, I still > think it would be a (potentially breaking) API change to change the > default for <inputtext> back. Although, I can see that it doesn't > make sense to resize an empty input text, LPP-5678 definitely wanted > <inputtext> to size to its initial content. > > How do we reconcile this?
Wait. I am officially confused. Despite what I said in the email that you referenced, if you look at the change set it says: > =================================================================== > --- WEB-INF/lps/lfc/kernel/swf/LzInputTextSprite.as (revision 9905) > +++ WEB-INF/lps/lfc/kernel/swf/LzInputTextSprite.as (working copy) > @@ -29,7 +29,10 @@ > this.yscroll = 0; > this.xscroll = 0; > > - this.resize = false; > + //@field Boolean resize: text width automatically resizes when > text is set. > + // default: false > + // > + this.resize = (args.resize == true); So, I didn't make inputtext default to resize=true, I just made it pay attention to the resize argument. But I only did it for the swf kernel. If you look at the current code base, there is still this vestigial code in the swf kernel, but it has no effect (that I can tell, there are no references to the sprite copy of resize). Not sure how to untangle this. It seems that in evolving the sprite's we moved the resizing behavior out of the kernel, up to generic code, and up to the <text> superclass. In the interim, it seems that the defaulting of resize to `false` for inputtext got lost. And the change by Max that Henry's binsearch found just made it finally take effect. --- So, I withdraw my objection to Henry's proposed change, since it seems to be restoring original intent (not that I am a strict constructionist, just that I can now see that his change is _not_ an API change). _______________________________________________ Laszlo-reviews mailing list [email protected] http://www.openlaszlo.org/mailman/listinfo/laszlo-reviews
