Some red flags with the current grammar for reserved-statement #635
Description
Reserved statements:
reserved-statement = reserved-keyword [s reserved-body] 1*([s] expression)
reserved-keyword = "." name
reserved-body = *([s] 1*(reserved-char / reserved-escape / quoted))
reserved-char = content-char / "."
If I read this correctly, then these strings are valid:
// reserved-keyword reserved-body expression
.foo
.foo . {$baz}
.foo .... {$baz}
.foo .bar {$baz}
Do we really want to allow '.' in the reserved body?
Escaped text:
text-escape = backslash ( backslash / "{" / "}" )
quoted-escape = backslash ( backslash / "|" )
reserved-escape = backslash ( backslash / "{" / "|" / "}" )
To me this looks like a red flag.
Not only for the implementers of a parser, but also for the users of the final syntax.
In most cases escaped strings can be handled pretty low level (tokenizer).
But in this case knowing that we can now accept reserved-escape we need a lot of context.
And that allows us to support strings like this:
.foo .\{\|\} {$baz}
So I find the reserved-statement to be too lax.
Or I am reading the grammar wrong?