https://bugzilla.wikimedia.org/show_bug.cgi?id=64901

--- Comment #7 from [email protected] ---
(In reply to Gabriel Wicke from comment #6)
> (In reply to ssastry from comment #4)
> > (In reply to Gabriel Wicke from comment #2)
> > 
> > > The lost line sounds like a somewhat orthogonal issue, likely with selser.
> > > It might be helped by paragraph wrapping, but ideally we would not depend 
> > > on
> > > that to make selser work.
> > 
> > No, this is not a selser issue. It is a p-wrapping issue. Since _rtselser
> > modifies the top-level, all the top-level node separators go through regular
> > wts and without a p-tag, the text-node following the p-tag doesn't get 2
> > newlines, just one.
> 
> Let me rephrase: There's no rule that selser has to work this way. We could
> also use our dsr information to substring the original wikitext between
> elements if the HTML text content there was not modified.
> 
> So selser is working the way it's currently designed, but there might still
> be room to improve the way our selser implementation solves the wider selser
> problem.

Actually, I misspoke. I remembered adding this selser functionality to
separators already. So, I looked into the code again. 

So, let us use a test snippet:
----------
<blockquote>a</blockquote> b

c <blockquote>d</blockquote>
----------

Save it, generate html, edit it with a comment on the end and run selser on it
as follows:

[subbu@earth lib] node parse --trace selser --selser --oldtextfile /tmp/foobar
--oldhtmlfile /tmp/foobar.html < /tmp/foobar.edited.html

and see this in the output:
---------
[WTS]          | TEXT: " b\n\nc " ; SOL: false
[WTS]          | ---> SEP: ""
[WTS]          | sol false  b
c 
[WTS]          | ---> OUT: " b\nc "
---------

The absence of the P-wrapping tag leads to the "b\n\nc" to be a single
text-node. Since there is no diff-markers on text-nodes (something we can fix
by augmenting our dom-diff maybe), the text nodes go through regular WTS and
looks like our regular WTS is normalizing newlines there.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
Wikibugs-l mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/wikibugs-l

Reply via email to