Redcloth upwards - Jan 2007 - <BR> and <P> in RedCloth v. Textism

Hi,

I''m using RedCloth right now for some simple markup in a Rails project.
It''s simple and easy - thanks for a great tool!

I have a question about an apparent discrepancy between RedCloth and 
Textile, relating to newline characters. I''ve played with the 
"hard_breaks" feature and that seems to generate it''s own
problems..

If I go onto the Textile test page:

http://www.textism.com/tools/textile/index.php

And enter the following text:

test
text

other

I get back:

         <p>test<br />
text</p>

         <p>other</p>

In IRB if I try to feed an equivalent string to RedCloth I get 
(accounting for hard_breaks on and off here):

irb(main):006:0> val = RedCloth.new("test\ntext\n\nother")
=> "test\ntext\n\nother"
irb(main):007:0> val.hard_breaks = true
=> true
irb(main):008:0> val.to_html
=> "<p>test<br />text<br />\nother</p>"
irb(main):009:0> val.hard_breaks = false
=> false
irb(main):010:0> val.to_html
=> "<p>test\ntext</p>\n\n\n\t<p>other</p> "

Basically, RedCloth either passes the single newline between "test"
and
"text" unparsed, or it adds only <br /> tags and fails to create
the
correct <p> tag for "other"..

Textile seems to convert the single newline into a <br /> character and 
the double newline into a <p> tag. The Textile behavior is greatly 
preferred for me.

I snooped around in the redcloth.rb source file and found this on line 
597:

     def hard_break( text )
         text.gsub!( /(.)\n(?!\Z| *([#*=]+(\s|$)|[{|]))/, "\\1<br
/>" )
if hard_breaks
     end

I believe that the problem is that this gsub is not accounting for the 
idea that there might be two newlines consecutively (or separated only 
by whitespace). In the case where there are two such newlines, I 
believe this pattern should not match and therefore the double newlines 
will be handed off for additional (correct) processing by the paragraph 
generator, rather than only one of them being sub''ed with a <br
/> tag.
I tried just changing "(.)" to "([^\n])" but that
didn''t have the
effect I was looking for..

I''d be happy to do a little work if someone can give me direction on 
patching this issue - this regex is pretty complicated and seems loaded 
with assumptions that I don''t understand, so I''m hesitant to
monkey
with it without help.

Or am I barking up a non-existent tree? Is there a way already to do 
what I want without a code change?

Thanks for any insight and assistance and especially for such a great 
tool!

Steve

p.s. I saw these postings in the archives which are related:
http://rubyforge.org/pipermail/redcloth-upwards/2006-June/000041.html
http://rubyforge.org/pipermail/redcloth-upwards/2006-August/000077.html

June 2006 claims that the exact problem I have is fixed! But I''m still 
having it.. Any patch around that can help?!

p.p.s. (I''m on version 3.0.4, at least according the folder that
it''s
installed - I did a gems install but just compared that to the tar.gz 
file from the website, which is the same..)
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://rubyforge.org/pipermail/redcloth-upwards/attachments/20070116/240e05b0/attachment.html

Hi,

First, I support any move to fix RedCloth! So SuperRedCloth as v4 seems 
great to me!

FYI, for anyone who cares, I''m working around my problems with RedCloth
and hard breaks by postprocessing the resulting html like so, which 
converts any <br /> followed by a newline into <p /> followed by a 
newline. It''s not perfect of course, but seems to fit my most painful 
issues:

     retval = textilize(h(value))
     retval.gsub(%r{<br />\s*\n}, "<p />\n")

Also, I see some mention of v3.03 or tip revision as being preferred to 
v3.04?

Can anyone provide a short description of what''s better about 3.03 or 
tip? And how I might go about deciding to change to one of these 
versions?

Thanks,

Steve

At 02:35 PM 1/23/2007, redcloth-upwards-request at rubyforge.org
wrote:
>Message: 6
>Date: Tue, 16 Jan 2007 12:41:32 -0800
>From: Steve Midgley <public at misuse.org>
>Subject: <BR> and <P> in RedCloth v. Textism
>To: redcloth-upwards at rubyforge.org.
>Message-ID: <20070116204855.26B8D524159A at rubyforge.org>
>Content-Type: text/plain; charset="us-ascii"
>
>Hi,
>
>I''m using RedCloth right now for some simple markup in a Rails 
>project.
>It''s simple and easy - thanks for a great tool!
>
>I have a question about an apparent discrepancy between RedCloth and
>Textile, relating to newline characters. I''ve played with the
>"hard_breaks" feature and that seems to generate it''s own
problems..
>
>If I go onto the Textile test page:
>
>http://www.textism.com/tools/textile/index.php
>
>And enter the following text:
>
>test
>text
>
>other
>
>I get back:
>
>          <p>test<br />
>text</p>
>
>          <p>other</p>
>
>In IRB if I try to feed an equivalent string to RedCloth I get
>(accounting for hard_breaks on and off here):
>
>irb(main):006:0> val = RedCloth.new("test\ntext\n\nother")
>=> "test\ntext\n\nother"
>irb(main):007:0> val.hard_breaks = true
>=> true
>irb(main):008:0> val.to_html
>=> "<p>test<br />text<br />\nother</p>"
>irb(main):009:0> val.hard_breaks = false
>=> false
>irb(main):010:0> val.to_html
>=> "<p>test\ntext</p>\n\n\n\t<p>other</p>
"
>
>Basically, RedCloth either passes the single newline between
"test"
>and
>"text" unparsed, or it adds only <br /> tags and fails to
create the
>correct <p> tag for "other"..
>
>Textile seems to convert the single newline into a <br /> character 
>and
>the double newline into a <p> tag. The Textile behavior is greatly
>preferred for me.
>
>I snooped around in the redcloth.rb source file and found this on line 
>
>597:
>
>      def hard_break( text )
>          text.gsub!( /(.)\n(?!\Z| *([#*=]+(\s|$)|[{|]))/, "\\1<br
/>"
> )
>if hard_breaks
>      end
>
>I believe that the problem is that this gsub is not accounting for the 
>
>idea that there might be two newlines consecutively (or separated only 
>
>by whitespace). In the case where there are two such newlines, I
>believe this pattern should not match and therefore the double 
>newlines
>will be handed off for additional (correct) processing by the 
>paragraph
>generator, rather than only one of them being sub''ed with a <br
/>
>tag.
>I tried just changing "(.)" to "([^\n])" but that
didn''t have the
>effect I was looking for..
>
>I''d be happy to do a little work if someone can give me direction
on
>patching this issue - this regex is pretty complicated and seems 
>loaded
>with assumptions that I don''t understand, so I''m hesitant
to monkey
>with it without help.
>
>Or am I barking up a non-existent tree? Is there a way already to do
>what I want without a code change?
>
>Thanks for any insight and assistance and especially for such a great
>tool!
>
>Steve
>
>p.s. I saw these postings in the archives which are related:
>http://rubyforge.org/pipermail/redcloth-upwards/2006-June/000041.html
>http://rubyforge.org/pipermail/redcloth-upwards/2006-August/000077.html
>
>June 2006 claims that the exact problem I have is fixed! But I''m
still
>
>having it.. Any patch around that can help?!
>
>p.p.s. (I''m on version 3.0.4, at least according the folder that
it''s
>installed - I did a gems install but just compared that to the tar.gz
>file from the website, which is the same..)
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://rubyforge.org/pipermail/redcloth-upwards/attachments/20070124/8c126af1/attachment.html