Re: [css-multicol-1][css-sizing-3] definition of column-width (editorial)

> On 7 Jun 2018, at 20:08, Tab Atkins Jr. <jackalmage@gmail.com> wrote:
> 
> On Wed, Jun 6, 2018 at 2:04 PM, Yves Lafon <ylafon@w3.org> wrote:
>> On Wed, 6 Jun 2018, Chris Lilley wrote:
>> 
>>> Hi Yves
>>> 
>>> On 06-Jun-18 05:24, Yves Lafon wrote:
>>>> 
>>>> It would be good to either add an denote in [css-multicol-1] to point to
>>>> those new values (even stating that their stability is not granted yet), or
>>>> at least to have a central place where all possible definitions for a
>>>> property (or function, or anything else) can be found.
>>> 
>>> I thought that already existed and was automatically updated
>>> 
>>> https://drafts.csswg.org/indexes/
>>> 
>>> although the date there is from 2017, it does seem to be constantly
>>> updated (for example, a look for font-variant-alternates, which was recently
>>> moved to Fonts 4, does indeed point to Fonts 4 not 3).
>> 
>> 
>> Great, but it doesn't seem to be complete, as a starter, multicol-1 is
>> missing.
>> Also it show other "conflicts", like for 'column-gap' defined in multicol-1
>> [1], not showing in your index page, also defined in align-3 [2]
> 
> The index generation currently has issues with "diff specs",
> unfortunately.  It relies on the linking logic to find out where it
> should point to (because it doesn't want to point to *every possible
> place* something is defined in, if some of those places are
> superceded).  The linking model has some logic for how to handle
> definitions appearing/disappearing between TR/ED and across levels. In
> particular, if something was defined in level N of a spec, and is not
> in level N+1, Bikeshed assumes it was removed for a good reason, and
> silently removes the level N anchor from the possible results.
> 
> Diff specs that omit large sections of the previous level, thus, cause
> some problems here. That's the source of all the "missing definitions"
> problems that people have pointed out in the indexes, as far as i can
> tell.

Ok, so diff specs (from a multicol-1 pov), as they define some properties
of multicol-1 are basically invalidating it, and there is difference in that specific 
case between sizing-3 which adds new values and align-3 which redefines a property
(with the same definition, fortunately…).

> Ultimately this is a problem on my side; I need to figure out a way to
> communicate that a spec is a diff and should interact differently with
> the level-based logic.
Ok, good to see it’s a known issue!
Thanks,

-- 
Baroula que barouleras, au tiéu toujou t'entourneras.

        ~~Yves

Received on Friday, 8 June 2018 20:38:14 UTC