Spec the @for at-rule #2986
Spec the @for at-rule #2986
Conversation
spec/variables.md
Outdated
| ## Syntax | ||
|
|
||
| <x><pre> | ||
| **Variable** ::= '$' Identifier | NamespacedVariable | ||
| **Variable** ::= VariableName | NamespacedVariable | ||
| **VariableName** ::= '$' Identifier |
There was a problem hiding this comment.
To make the difference clearer, I think this should probably be something like NamespacelessVariable or LocalVariable.
There was a problem hiding this comment.
After writing this, I saw that #2850 (comment) suggested PlainVariable.
LocalVariable seems wrong to me, as it could be the name of a global variable.
There was a problem hiding this comment.
I will let you decide which name I should use before updating my PR.
|
|
||
| * While `i` is not equal to `to`: | ||
|
|
||
| * Declare a variable with `rule`'s `VariableName` as name and value `i` in the `rule`'s `ForBlock`'s scope | ||
| * Let `scope` be a new [scope]. |
There was a problem hiding this comment.
do we actually create a new scope at each iteration ?
There was a problem hiding this comment.
@nex3 was there anything wrong with my previous proposal ? According to the spec for variables, the curly braces of the ForBlock are already introducing a new scope.
There was a problem hiding this comment.
I don't think it's actually detectable either way right now, but if we do ever add full closure semantics to Sass we'll want a closure to capture the current value of the iteration variable rather than the final value (that is, we'll want it to behave like for/let in JS and not like for/var). This text makes that explicit.
|
@nex3 by curiosity, what is this weird |
A |
This represents a non-namespaced variable named (as opposed to Variable which accepts both). Several parts of Sass don't allow namespaced names for the variables they need (loops, argument declaration, etc...)
|
I updated my markdown-link-check fork but it still failed to fetch GitHub URLs 😕. I'm re-running tests (again) to see if that clears it out. |
|
Oh, I missed #2994. |
As a basis for the fast-track proposal in #2931, this writes specs for the
@forat-rule.I intentionally skipped any mention of the unit conversion for
fromandto, because dart-sass, libsass and ruby-sass have 3 different behaviors for it (which is precisely what #2931 is meant to solve).Implementing strictly the spec as written here would probably produce the libsass behavior (relying on the fact that adding a unitless number to a number with unit is allowed in Sass and produces a number with that unit). Specifying the ruby-sass behavior is a matter of adding a step
Convert `to` to `from`'s unitbefore asserting that both values are integers (but currently, dart-sass does the conversion the other way, which is why I prefer keeping this line as part of the proposal instead). Spec'ing the existing dart-sass behavior would make the spec more complex for something that would have to be reverted by the proposal anyway (converting numbers to strip the unit for the loop variable).