> Guillaume Lerouge wrote:
> > Hi,
> >
> > On Wed, Nov 4, 2009 at 3:22 PM, Marius Dumitru Florea <
> >
[hidden email]> wrote:
> >
> >> Hi Guillaume,
> >>
> >> Guillaume Lerouge wrote:
> >>> Hi Marius,
> >>>
> >>> On Wed, Nov 4, 2009 at 2:20 PM, Marius Dumitru Florea <
> >>>
[hidden email]> wrote:
> >>>
> >>>> Embed HTML elements pose some problems:
> >>>>
> >>>> * they are loaded asynchronously and any attempt to access even the
> >>>> simplest DOM property like nodeType through JavaScript raises an
> >>>> exception if the object is not fully loaded (closed to
> >>>>
https://bugzilla.mozilla.org/show_bug.cgi?id=331672 ).
> >>>>
> >>>> * they can't be fully controlled from JavaScript. The specific plugin
> >>>> that handles each of them decides the display and behavior (e.g.
> >>>>
https://bugzilla.mozilla.org/show_bug.cgi?id=250855 a flash movie
> might
> >>>> auto start in edit mode and, unless you want to see a movie while
> >>>> editing, that's not nice).
> >>>>
> >>>> * they aren't reloaded after they are detached and reattached to the
> DOM
> >>>> tree using JavaScript which means that any DOM manipulation involving
> >>>> one of the ancestors of a embed node can lead to the disappearance of
> >>>> the embedded object. Sadly, appendChild, insertBefore and replaceChild
> >>>> detach the target node before attaching it to the new location. As a
> >>>> consequence there's no way to move an embed node inside the DOM tree
> >>>> without detaching it.
> >>>>
> >>>> So I propose we replace the embed HTML elements with static image
> >>>> placeholders just after the edited document finishes loading. The
> >>>> placeholder will have the same size as the embedded object in view
> mode.
> >>>>
> >>>> Note, ckeditor already does this. You can check by copying the HTML
> from
> >>>>
http://jira.xwiki.org/jira/browse/XWIKI-3975 to the source of
> >>>>
http://ckeditor.com/demo .
> >>>>
> >>>> I'm +1.
> >>>>
> >>> Sounds good. It will also make the page easier to edit since it will be
> >> less
> >>> heavy.
> >>>
> >>> Only question: where do we display text for that image? The information
> >> text
> >>> we put with the image has to be easy to internationalize.
> >> Indeed that's something to think about. One option is to use a generic
> >> icon that stands for "Embedded Object" in the center of the placeholder
> >> (which also has a border) and use a translatable text as the tool tip of
> >> the placeholder. WDYT?
> >>
> >
>
> > Maybe we could use a DIV + text in the DIV + the background-image CSS
> > property so that the text can be internationalized easily yet is
> diaplayed
> > above the image.
>
> But that text will be editable. Maybe not when the embedded object is
> generated by a macro, but the editor is not bound to the xwiki/2.0
> syntax. It should work the same with HTML syntax or a syntax that allows
> you to write "video:URL". Making the DIV container read-only is
> unfortunately not straightforward at the moment because the
> contentEditable=false attribute is not well supported by the current
> browsers. Using an image placeholder is the simplest and quickest
> solution right now.
>
