D
David Dorward
Looks like it might be a bug then.
Looks like it might be a bug then.
E
Els
Michael said:
Yes, I was entirely mistaken.
[snip explanation]
I saw in another message of his, that he got that from
http://www.htmlhelp.com/tools/csscheck/ which actually does give that
error (erroneously).
Thanks - that's one of the W3C pages I didn't come across before
E
Els
Els wrote:
[ .subsubsection#sverige ]
Note: I found this link easier to read:
http://www.w3.org/TR/REC-CSS2/selector.html#simple-selector
[ .subsubsection#sverige ]
Note: I found this link easier to read:
http://www.w3.org/TR/REC-CSS2/selector.html#simple-selector
E
Els
David said:
Or maybe it's because it's a CSS1 checker:
<quote>
Note: CSSCheck is primarily a CSS1 checker. Many CSS2 properties are
supported, but new CSS2 selectors will generate errors.
</quote>
(If I understood appendix B of the CSS1 specs correctly (of which I'm
not sure), only 1 class/id/pseudo-class is allowed per selector in
CSS1)
M
Michael Winter
It seems that my news server was sluggish last night; I was missing
about half of the thread when I replied.
[snip]
Mike
M
Michael Winter
On 26/02/2006 09:11, Els wrote:
[snip]
Yes, but I think I ended up paraphrasing the applicable paragraph.
Not intentionally, mind you.
Section 5 (Selectors) doesn't contain any examples of of mixed type
attribute selectors as far as I can see. However, the (normative)
grammar makes it unequivocally clear that they are valid for simple
selectors[1].
Mike
[1] I should point out that the simple_selector production in the
CSS 2 grammar is in error. It was corrected in CSS 2.1 to:
simple_selector
: element_name [ HASH | class | attrib | pseudo ]*
| [ HASH | class | attrib | pseudo ]+
;
The previous production made the selector entirely optional.
That is,
{
/* Declarations */
}
would be valid.
[snip]
Yes, but I think I ended up paraphrasing the applicable paragraph.
Not intentionally, mind you.
Section 5 (Selectors) doesn't contain any examples of of mixed type
attribute selectors as far as I can see. However, the (normative)
grammar makes it unequivocally clear that they are valid for simple
selectors[1].
Mike
[1] I should point out that the simple_selector production in the
CSS 2 grammar is in error. It was corrected in CSS 2.1 to:
simple_selector
: element_name [ HASH | class | attrib | pseudo ]*
| [ HASH | class | attrib | pseudo ]+
;
The previous production made the selector entirely optional.
That is,
{
/* Declarations */
}
would be valid.
E
Els
Michael said:On 26/02/2006 09:11, Els wrote:
[snip]
Yes, but I think I ended up paraphrasing the applicable paragraph.
Not intentionally, mind you.
You did pretty well though - when I first read it, I though you had
literally quoted from the specs. Only you hadn't <g>
[snip, did read it]
M
Michael Winter
On 26/02/2006 09:21, Els wrote:
[snip]
I'm not quite sure what you're trying to express, here.
.className or #id
but not
#id.className
An id, class, or pseudo-class selector can appear by itself [optionally
combined with an element], but more than one type of attribute selector
cannot be combined.
Or perhaps:
#id.className
but not
.class1.class2
Multiple attribute selector types can be used together, but only one of
each type can be used at once.
As I read it, the grammar permits the latter, but it also places very
tight constraints on the /order/ of simple selector components: element,
id, class, pseudo-class. A simple selector can contain any of these, but
if any are present, they must appear in that order and can only occur
once at most.
The reason why
.subsubsection#sverige a
fails is because the id selector follows the class name.
#sverige.subsubsection a
will be accepted.
Note that the W3C validator doesn't care much for the order, either
because it wasn't meant to be significant (but that fact couldn't be
expressed concisely in the grammar), or due to a bug.
Mike
[snip]
I'm not quite sure what you're trying to express, here.
.className or #id
but not
#id.className
An id, class, or pseudo-class selector can appear by itself [optionally
combined with an element], but more than one type of attribute selector
cannot be combined.
Or perhaps:
#id.className
but not
.class1.class2
Multiple attribute selector types can be used together, but only one of
each type can be used at once.
As I read it, the grammar permits the latter, but it also places very
tight constraints on the /order/ of simple selector components: element,
id, class, pseudo-class. A simple selector can contain any of these, but
if any are present, they must appear in that order and can only occur
once at most.
The reason why
.subsubsection#sverige a
fails is because the id selector follows the class name.
#sverige.subsubsection a
will be accepted.
Note that the W3C validator doesn't care much for the order, either
because it wasn't meant to be significant (but that fact couldn't be
expressed concisely in the grammar), or due to a bug.
Mike
E
Els
Michael said:
That's what I figured appendix B is saying.
Apparently it isn't though:
Yes, that CSS1 checker does accept that. I didn't expect that.
Well, I'll just stick to CSS2.1, which I mostly understand, and
combine that with what works in the browsers. Life can be so easy ;-)
J
Jonathan N. Little
Luigi said:
Okay I think I know what is going on here, when Luigi is trying to
specify a A element that is a child of and ELEMENT that is both
CLASS="subssubsection" AND ID="saverige" fails because
..subsubsection#sverige A
^^^^^^^^^^^^^^^^^^^^^^
this is the wrong way to specify the condition as it is being
interpreted as class "subsubsection#sverige" containing an illegal
character '#' thus giving you the "class error", to what you wish Luigi
I think it should be written:
#sverige[class="subsubsection"] a {...}
E
Els
Jonathan said:
..subsubsection#sverige is actually correct.
Proof:
<http://validator.w3.org/check?verbose=1&uri=http://here.locusmeus.com/temp/idclass.html>
E
Els
Els said:
Duh, how about this one:
<http://jigsaw.w3.org/css-validator/...i=http://here.locusmeus.com/temp/idclass.html>
J
Jonathan N. Little
Yep I noticed that too ;-) but yes it also validates. Hmmm I do not knowEls said:
why sir Luigi had an error, most likely a duplicate identifier somewhere.
E
Els
L
Luigi Donatello Asero
Michael Winter said:
So is the latter the correct form although the former is tolerated?
Is
#sverige.subsection a=
<a class="subsection" id="sverige"> ?
S
Steve Pugh
Luigi Donatello Asero said:
No. And surely you've seen enough CSS by now to understand something
as basic as a descendent selector? Ignore the order of the #sverige
and .subsection, would you ask whether "#sverige a" matched <a
id="sverige"> or whether ".subsection a" matched <a
class="subsection">?
#sverige.subsection a
is the selector for
<foo class="subsection" id="sverige"><a>
The selector for <a class="subsection" id="sverige">
would be
a#sverige.subsection
Steve
M
Michael Winter
As far as the grammar for CSS 1 is concerned, yes. For CSS 2 (and
later), the order is not important and the selectors are equivalent.
No, see Steve's reply.
Mike