Left-flanking / right-flanking (again?)

[mildsunrise]

I'm kinda pissed off. I implemented emphasis according to the spec, and now my parser is rendering _foo_ as <p>_foo_</p>. I verified that, according to the rules in 0.17, the two delimiter runs are both right- and left- flanking. So my parser is actually right!

So I go checkout 0.18 and I find the definition has changed a bit, it's now even more permissive. So _foo_ is still <p>_foo_</p>. I also checked master before submitting this.

Then I find in #302 (and in CMark code) that the code treats start of line and end of line as whitespace, and @jgm attempted to correct the spec. However _foo_ is still having both left- and right- flanking delimiter runs.

I suppose this would be fixed by changing:

 A [left-flanking delimiter run](@left-flanking-delimiter-run) is
-a [delimiter run] that is (a) not followed by [unicode whitespace],
+a [delimiter run] that is (a) not followed by [unicode whitespace] or end of line,
 and (b) either not followed by a [punctuation character], or
 preceded by [unicode whitespace] or a [punctuation character] or
 the beginning of a line.

 A [right-flanking delimiter run](@right-flanking-delimiter-run) is
-a [delimiter run] that is (a) not preceded by [unicode whitespace],
+a [delimiter run] that is (a) not preceded by [unicode whitespace] or end of line,
 and (b) either not preceded by a [punctuation character], or
 followed by [unicode whitespace] or a [punctuation character] or
 the end of a line.

But those two definitions are getting very messy IMO...

[jgm]

+++ Xavier Mendez [Mar 10 15 13:59 ]:

I'm kinda pissed off. I implemented emphasis according to the spec, and
now my parser is rendering foo as

foo

. I verified that,
according to the rules in 0.17, the two delimiter runs are both right-
and left- flanking. So my parser is actually right!

So I go checkout 0.18 and I find the definition has changed a bit, it's
now even more permissive. So foo is still

foo

. I also
checked master before submitting this.

Then I find in [1]#302 (and in CMark code) that the code treats start
of line and end of line as whitespace, and [2]@jgm attempted to correct
the spec. However foo is still having both left- and right- flanking
delimiter runs.

My original instinct in adjusting the spec was just to add:
"for these purposes the beginning and end of the line count
as whitespace." I then decided to make it more explicit, but
as you've pointed out, I didn't get it quite right.

For the sake of simplicity, would it be better to say the
following?

A left-flanking delimiter run is
-a [delimiter run] that is (a) not followed by [unicode whitespace],
and (b) either not followed by a [punctuation character], or
preceded by [unicode whitespace] or a [punctuation character].
For these purposes the beginning and end of line count as
[unicode whitespace].

[mildsunrise]

Yeah.

[jgm]

See commonmark/commonmark.js#12 for a related issue.