Quotes issues

[jamienk]

• Quotes after parentheses:

("Stop!" he said) <--opening quote is wrong 

(this is also messed up when a quote comes after a dash,
but your dash handling forces a space, see #6)

• Closing big quote after nested small quote

"I told her 'This is crazy'" <--closing quote incorrect

[danyaPostfactum]

closing quote incorrect

Why? I get: “I told her ‘This is crazy’”. What's wrong?

[jamienk]

I get “I said ‘This is nuts’“ <--final closing quote is backwards

[danyaPostfactum]

@jamienk autocorrect does not handle text across different text nodes at present. Text can be separated to a text node by formatting elements (for example, you make 'This is nuts' italic), or in some cases while editing (sometimes browser splits whole text into a sequence of small text nodes).
This can cause mentioned quote issue.

[jamienk]

I see what you're saying about the text node being split. How and why and when does this happen? I see it in Firebug. Why does CKEDITOR add line breaks in text nodes? (I understand that the line breaks are part of the HTML formatting and can be controlled, but why inside text nodes?)

[danyaPostfactum]

I see it in Firebug

Yes, this is Firefox issue. Anyway, I need to write CharacterWalker, that would jump between nodes and skeep bookmarks (they keep cursor position).
What line breaks do you mean? \n character? I don't see it. <br> element? But it's impossible. Text nodes contain character data only.

[jamienk]

When I inspect the text node that I write it has line breaks in firebug so that the text inside the paragraph tags is broken into multiple lines. I can't see the line breaks get created because I have to hide and show the text in order for firebug to refresh the contents.

[jamienk]

Example, when inspected in Firebug:

<p>
She’
s a big fan. A guy she met and doesn’
t like
wrote to her that he’
s thinking about her all the time
, and she texted back “
I’
m thinking about you all the time too!”
</p>

Where is that whitespace coming from? Not seeing this in Chrome. Can't find anything on it.

[danyaPostfactum]

There is no line breaks. This is just a presentation of a sequence of text nodes.
Other browsers normalize text by merging sibling text nodes while editing.

By the way, an element has a normalize() method, that invokes text normalization.

[jamienk]

Hm, so when you wrote: "autocorrect does not handle text across different text nodes at present" << does that mean that if we first normalize() then the nodes will be merged and it'll work?

[jamienk]

(and BTW, than you for this great work!)

[danyaPostfactum]

does that mean that if we first normalize() then the nodes will be merged and it'll work?

Yes. But I don't think this is good idea. Moreover, formatting element nodes (<b>, <i> etc) break autocorrect also. So I created CharacterIterator, that abstracts me from nodes, just giving me a sequence of characters.

You can try current master branch. It mostly done, but I did not tested yet.

[danyaPostfactum]

(and BTW, than you for this great work!)

Thank you for testing and your bug reports!

[jamienk]

From my small set of test cases it works good!
Nice and fast work. I'm gonna check it out to learn from you! Will report back any issues...