Tres Seaver wrote:
I'm pretty sure you're not allowed dots in ids of HTML elements.
The attached HTML and XHTML pages both pass the W3C validator (one HTMl
4.01 strict, the other XHTML 1.0 strict). Both use dotted IDs for form
elements. Neither renders properly in Firefox 1.5 (the selectors
based on the dotted ID, as well as those based on attribute, don't get
applied the form elements).
At
least not when they are prefixed by 'form'. A zope.schema.TextLine in a
formlib form, for example, generates HTML like this:
input id=form.title class=textType type=text value=Changed
title size=20 name=form.title/
In particular, we want to style the fields to look a bit more like
Plone, and we have special cases for title and description that we
want to style by ID, not class.
However, CSS like this has no effect in Firefox or Safari:
#form.title {
font-size: 160%;
font-family: dtml-var headingFontFamily;
font-weight: normal;
width: 99%;
}
#form.description {
font: 100% dtml-var fontFamily;
font-weight: bold;
height: 6em;
width: 99%;
}
Not all that surprising - it could look like I'm styling a form with
class=description.
Not according to the CSS specs. Note that we might be better off making
it easy to place a custom CSS class on the widget, rather than fighting
browser failures on various other selectors.
I had another look at the specs, and I agree that dots should be valid, but
it's pretty horrendous that neither Firefox nor Safari will let me style
that. If I've messed up the CSS syntax, let me know.
- -1 on changing the default behavior; to paraphrase, broken browsers
you will have always with you.
+1 on changing 'renderTag' to use some kind of local policy (e.g.,
look up an 'ITagElementCleaner' utility) to spply such transforms.
E.g.::
def renderTag(tag, **kw):
cleaner = querytUtility(IWidgetTagCleaner)
if cleaner is None:
kw = cleaner(kw)
...
Yeah, this is nicer. I'd be happy to supply a patch for this. It would also
suport the case where you wanted to other standard transformations on
attributes.
I'd like to understand what the chances are of Plone 3 (whch will run on
Zope 2.10.x, whichever is latest when Plone is released in mid-May or a bit
later) being able to depend on this. We may need some kind of monkey patch
in Plone if it can't go into a real Zope release in time,
Martin
--
View this message in context:
http://www.nabble.com/zope.app.form-generates-invalid-HTML-tf3276583.html#a9117875
Sent from the Zope3 - dev mailing list archive at Nabble.com.
___
Zope3-dev mailing list
Zope3-dev@zope.org
Unsub: http://mail.zope.org/mailman/options/zope3-dev/archive%40mail-archive.com