[css-backgrounds-3] Disallow double `//` in `border-image`

[cdoublev]

0 // 1 is valid according to the border-image grammar, but I am not sure the original intent was to allow //. The grammar has been modified (10 years ago) to prevent a trailing /. I think the following value definition would prevent //:

     <'border-image-source'>
- || <'border-image-slice'> [ / <'border-image-width'> | / <'border-image-width'>? / <'border-image-outset'> ]?
+ || <'border-image-slice'> [ / <'border-image-width'> | [ / <'border-image-width'> ]? / <'border-image-outset'> ]?
  || <'border-image-repeat'>

There is no WPTs for a similar input. There is an existing WPT and I cannot say if this change would represent a back-compat issue.

Similarly, I am noting that you cannot set <bg-size> without defining <bg-position>: <bg-position> [ / <bg-size> ]?. Alternatively, it could be simplified in a consistent way with:

     <'border-image-source'>
- || <'border-image-slice'> [ / <'border-image-width'> | / <'border-image-width'>? / <'border-image-outset'> ]?
+ || <'border-image-slice'> [ / <'border-image-width'> [ / <'border-image-outset'> ]? ]?
  || <'border-image-repeat'>

[faceless2]

It's used in https://github.com/web-platform-tests/wpt/blob/master/css/css-backgrounds/border-image-shorthand-003.htm, which is passing in all browsers: https://wpt.live/css/css-backgrounds/border-image-shorthand-003.htm

[cdoublev]

Oh, sorry, I searched for //, but not / /.

[Loirooriol]

Also tested in https://wpt.fyi/results/css/css-backgrounds/parsing/border-image-valid.html

[cdoublev]

Ok, with some hindsight, considering the fact it is tested, that I probably ignore valid reasons to support this syntax, I prefer to not hold this request any longer. Feel free to re-open it if you think it is worth it. Thanks to both of you for your answers!

[Loirooriol]

The 1st time that I saw / / it also seem surprising to me, but disallowing it could break sites. May be better to forbid in linters?

[tabatkins]

Allowing a // was indeed intentional, fwiw. I'd have to go dig around in mail archives to find the conversation, as I think this was designed pre-GitHub, but if my memory is correct we explicitly discussed this.