Re: [css21] What's the used value of left/right when overconstrained in relpos?

On Wed, May 1, 2013 at 1:36 PM, Tab Atkins Jr. <jackalmage@gmail.com> wrote:

> On Wed, May 1, 2013 at 1:27 PM, Glenn Adams <glenn@skynav.com> wrote:
> > On Wed, May 1, 2013 at 11:24 AM, Tab Atkins Jr. <jackalmage@gmail.com>
> > wrote:
> >> On Tue, Apr 30, 2013 at 9:29 AM, Glenn Adams <glenn@skynav.com> wrote:
> >> > On Mon, Apr 29, 2013 at 2:02 PM, Tab Atkins Jr. <jackalmage@gmail.com
> >
> >> > wrote:
> >> >> On Mon, Apr 29, 2013 at 1:59 PM, Glenn Adams <glenn@skynav.com>
> wrote:
> >> >> > My general position is that no style value returned via
> >> >> > getComputedStyle()
> >> >> > should be 'auto' if there is a legal return value that denotes the
> >> >> > "resolved" or "as used" interpretation of 'auto'.
> >> >>
> >> >> I believe the current common position among spec is that "auto"
> >> >> resolves at the earliest common moment it can resolve, across all
> >> >> possibiliites.  That is, if it can resolve at computed-value time in
> >> >> some cases, but used-value time in others, it should just resolve at
> >> >> computed-value time.
> >> >
> >> > sounds reasonable
> >>
> >> Unfortunately, it looks like I said the exact opposite of what I meant
> >> to say.  My actual words are nonsensical, because they imply that we
> >> should resolve "auto" away at computed-value time even when it can't
> >> be done until used-value time.
> >>
> >> What I *meant* to say is "It should just resolve at used-value time".
> >> That is, a given keyword should always resolve at a single time; if
> >> it's possible to resolve it at several different times, resolve it at
> >> the earliest time that satisfies all cases.
> >
> > Thanks for clarifying. That is how I decoded your intention in your
> earlier
> > response.
>
> Cool, just making sure.  ^_^  This is somewhat contrary to your stated
> preference earlier in this thread, because it means that properties
> which only resolve "auto" at used-value time would still report "auto"
> from getComputedValue(), since only a small blacklist of properties
> return used values.
>

Hmm, I though we were on the same page but perhaps not.

So, are you saying that for the set of properties { height, ..., width }
and { bottom, ..., top } listed in [1], that 'auto' should be returned for
those cases where computed value is 'auto' but used value is a specific
dimension *and* the resolved value is the used value (according to the
algorithm defined in [1])?

[1] file:///Users/glenn/work/w3c/csswg/specs/cssom/Overview.html#resolved-values