Revisiting div class/id shorthand

[yhtez]

As of #137 (which as far as I can tell fixes the ambiguity issue outlined in #35 and #28) I don't see any reason why the shorthand syntax for classes and ids can't be expanded to allow the case with no element tag such as .awesome-class#unique-id instead of div.awesome-class#unique-id.
I'd be happy to look into making a PR to add this functionality if you wouldn't be apposed to this syntax now.

[lambda-fairy]

Good idea!

The only issue would be the interaction with named attributes. Currently class/id shorthands and named attributes can be interleaved in an arbitrary way, so e.g.

div contenteditable? #unique-id dir="rtl" .awesome-class {}

is valid syntax.

If we do the obvious thing and allow omitting the element name with no other changes, then

contenteditable? #unique-id dir="rtl" .awesome-class {}

would be valid syntax. That's not ambiguous but it could be confusing.

I think an easy solution would be to say that any classes/ids must come before named attributes. That eliminates the confusing example I gave above, and I think it was an accident that the interleaving was allowed in the first place.

What do you think @zth0? If you're happy with this tweak then I'm happy to take a PR 🙂

[yhtez]

The way I was planning to go about it would have been to only allow omitting element name if only if the first attribute of the element is a class or id shorthand such that

.awesome-class contenteditable? dir="rtl" #unique-id {}
#unique-id contenteditable? dir="rtl" .awesome-class {}

is valid syntax (still allowing any order of class and id attrs beyond the implicit div) but

contenteditable? #unique-id dir="rtl" .awesome-class {}

isn't (which prevents it from being confusing.)

I'm not sure if you meant allowing omitting element name even without any attrs but I'd prefer not to go about it this way as I think this causes more breakage than is necessary.
For example blocks with no element name would instead be wrapped in a div https://github.com/lfairy/maud/blob/f652784adbdd57d1510e83fd17ed2a888655eb8b/maud/tests/basic_syntax.rs#L30-L39 would generate the html hello<div> ducks geese</div> swans

I'm not sure if this matches what you were thinking but I think that going about it this way instead would prevent this from breaking any old code and should be simpler to implement.

[yhtez]

Opened PR #173 with my suggested implementation, let me know what you think and if I missed anything or you would rather your idea.

[lambda-fairy]

Oh I forgot about plain blocks. That makes sense to me!

[lambda-fairy]

BTW, issues should stay open until the PR has been merged.