- From: Xidorn Quan <quanxunzhen@gmail.com>
- Date: Sat, 10 May 2014 10:40:14 +1000
- To: "L. David Baron" <dbaron@dbaron.org>
- Cc: www-style list <www-style@w3.org>
- Message-ID: <CAMdq69_cyu3B7kXAFMckpfYAqMkPtrO19HP1Yr4um8pZmr_gMw@mail.gmail.com>
On Sat, May 10, 2014 at 10:30 AM, L. David Baron <dbaron@dbaron.org> wrote: > On Saturday 2014-05-10 10:23 +1000, Xidorn Quan wrote: > > On Thu, May 8, 2014 at 9:08 AM, L. David Baron <dbaron@dbaron.org> > wrote: > > > > > http://dev.w3.org/csswg/css-counter-styles/#counter-style-speak-as > > > says, when describing the 'auto' value: > > > # If the system is override, this value has the same effect as the > > > # overridden style’s speak-as. > > > > > > This has an unusual interaction with the definition of the override > > > system, which says in > > > http://dev.w3.org/csswg/css-counter-styles/#valuedef-override that: > > > # If a counter style uses the override system, any unspecified > > > # descriptors must be taken from the overridden counter style > > > # specified, rather than taking their initial values. > > > > > > It means that when the override system is used, speak-as: auto is > > > equivalent to having an omitted speak-as, and is *not* equivalent to > > > the overridden system having speak-as: auto. Both of these seem > > > wrong to me. > > > > > > Instead, I would propose that the 'auto' value say: > > > # If the system is override, this value has the same effect that > > > # 'auto' would have for the overridden counter style. > > > which seems more consistent with how the override system otherwise > > > works. > > > > > > > To my understanding, there is no difference between your proposal and the > > current spec, though your wording seems to be clearer. > > The difference is that given: > > @counter-style base { > system: cyclic; > symbols: ‣; > speak-as: numeric; > } > > @counter-style derived { > system: override base; > speak-as: auto; > } > > The current spec says that the counter style 'derived' is spoken as > numeric, whereas I propose that it be spoken as bullet (which is > what 'auto' means for the cyclic system). > > (This is what your patches in [1] implement for 'speak-as', although > you followed the spec for the equivalent wording for 'range'.) > You are right (though the wording is different for 'auto' between 'speak-as' and 'range'.) I misunderstood the spec, and agree with what you proposed. - Xidorn
Received on Saturday, 10 May 2014 00:41:22 UTC