Using this file with Pegdown 1.1.0.

See nicoulaj/idea-markdown#42.

I recently started using Pegdown in a project I'm working on and have been very impressed. One thing I have noticed however is that it seems to produce incorrect HTML for Markdown images. For example, the following Markdown:

![Cascading Style Sheets](http://jigsaw.w3.org/css-validator/images/vcss)

Becomes:

<img src="http://jigsaw.w3.org/css-validator/images/vcss">Cascading Style Sheets</img>

When, according to http://daringfireball.net/projects/markdown/syntax it should be setting the alt attribute, as such:

<img src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Cascading Style Sheets" />

I noticed that the test clase test-cases/docs-maruku/images.html shows this working correctly, so perhaps a recent change as broken it?

To tried to get around this by just omitting the alt text in the markdown, however the parser then seems to not recognize it as an image. For example:

![](http://jigsaw.w3.org/css-validator/images/vcss)

Using other markdown parsers, this seems to be legal syntax.

Automatic links, as described by John Gruber's link above, seem to also have a slight issue. When given something as such:

<http://example.com/>

The result is:

<a href="http://example.com>">http://example.com></a>

The image problem seems to occur with or without extensions, while the automatic linking syntax problem seems to occur only when I have the extensions on (perhaps a conflict with GitHub style automatic linking?)

This implementation of Markdown is one of the best I've seen, thank you for sharing all your hard work!

NOTE: All instances of & lt; and & gt; above should actually be < and >. Not sure why Github Markdown is doing that.

It appears that putting a strong (bold) inside an emphasis (italic) using the same character renders incorrectly.

Example:

*This is a **test***

Expected Output:

<p><em>This is a <strong>test</strong></em></p>

_This is a _test**

Actual Output:

<p><em>This is a </em><em>test</em>**</p>

This is a test**

The bold tag doesn't need to be at the end: *This **is a ** test* is also rendered incorrectly.

The inverse isn't a problem (ie: **This is a *test*** renders fine).

It looks like a single asterisk is getting priority over a double asterisk. I haven't had time to dig into the source to see what's going on yet.

The Parser class has some inner classes the prevent it being subtyped. Parboiled generates some methods that can't be seen from a subtype.

See the fork stickycode/pegdown for a solution....

I'm implementing includes for source code samples and might look at folding it back into the core if it makes sense but it would probably be too many dependencies so thought it best as an external extension.

Parboiled is in the scala-tools repository, so I guess it might make sense to put pegdown there as well. Maven Central would be better though.

The ability to disable HTML via an option would be very useful for cases where people would like to use only markdown syntax. For example, markdown could be allowed in a blog's comments but not HTML so that commenters could safely include things such as code blocks and emphasis (this is the case I would like to use it in). PHP Markdown provides this through the use of the no_markup option.

This could theoretically be implemented either before or after using the markdown parser, but I think it would be much more beneficial and efficient to have this included in the parser itself.

Definition lists are an useful semantical markup means for many documents. While the original markdown does not include a syntax for this there are other implementations doing so.

So, what about making a pegdown extension for parsing a definition list markup?

The blog entry http://www.justatheory.com/computers/markup/modest-markdown-proposal.html is pretty good in explaining the appropriate syntax from markdown extra and furthermore proposes a slight improvement for it.

Here are some examples of links that should work but don't:

[ex@mple](http://example)
[example://](http://example)

This must have something to do with autolinks detection, I guess ?

When transforming some text to Markdown I observed that backslashes were disappearing:

  def processor = new PegDownProcessor(ALL & ~QUOTES & ~SMARTS)
  processor.markdownToText("hello\\ world")

  => <p>hello world</p> // instead of <p>hello\ world</p>

I had the same issue earlier with specs2 when using "regex".r.replaceAllIn(someString, groupMatches) so maybe that's something you're using internally somewhere?

In standard markdown, code blocks must be indented with 4 spaces or a tab, but it's not convenient if we copy some code from somewhere else, that we usually have to correct the format line by line manually, sometimes it's a hard work.

I wonder if pegdown support code blocks wrapped in ``` and ```, such as:

```
class A
class B
```

This will make things easier.

Pegdown is usually really fast, but something goes wrong when there's a bunch of links with bold _and_ italics inside the label.

For example, this basically never completes on my machine. Yes, it's odd, but I'm trying to write something to convert backward from HTML to Markdown with Pegdown extras (mostly for fun), and I was using a Wikipedia page. This is just a snippet:

 1. ^ [***a***][link143][***b***][link144][***c***][link145][***d***][link146][***e***][link147][***f***][link148][***g***][link149][***h***][link150][***i***][link151][***j***][link152][***k***][link153][***l***][link154][***m***][link155][***n***][link156][***o***][link157][***p***][link158][***q***][link159] Scheina, "Argentina," 401.

[link143]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-0
[link144]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-1
[link145]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-2
[link146]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-3
[link147]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-4
[link148]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-5
[link149]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-6
[link150]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-7
[link151]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-8
[link152]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-9
[link153]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-10
[link154]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-11
[link155]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-12
[link156]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-13
[link157]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-14
[link158]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-15
[link159]: http://en.wikipedia.org/wiki/ARA_Moreno#cite_ref-Conways_0-16

Now, if you replace all the *** with **, *, or nothing, it runs just fine. None of the other parsers I've been playing with seem to have this issue. (TextMate, Mac OS X quicklook, and a program called Readown.)

Reducing the number of links helps, too, but it still runs rather slow.

This is another low-priority oddball. When a fenced codeblock is immediately followed by a "<dd>", the codeblock is not processed.

Example:

~~~
this is code
~~~

:   this is a definition

If the definition has a term preceding it, it renders as expected:

~~~
this is code
~~~

Term
:   this is a definition

Now for the really weird stuff. If I add two linebreaks, the code renders correctly, but the definition now doesn't:

~~~
this is code
~~~


:   this is a definition

The same thing happens if we change the fenced code block to a normal code block, but leave the single space:

    this is code

:   this is a definition

The order matters. A fenced code block coming after a single "dd" always works.

I have just recently begun to use pegdown in a project and so far I am finding it to be a nice library. One thing I have noticed though, even with 0.9.x series, I am unable to escape special characters correctly when they are not placed within a markdown tag that generates a link.

For example if I do this:
example

The & does correctly become &.

If on the other hand I do something like what is found in the Special Chars.text, the special characters are not correctly replaced with the escaped counterparts.

I have attempted to use different combinations of the Extensions ranging from none to all and combination of the extensions seem to change how the special characters are handled.

Thank you for sharing this easy to use implementation of markdown.

the first character of the definition list seems to have a problem

I'm using pegdown, I want to know if it's possible to get the generated html for a part of markdown with pegdown, thanks!

The full question is here:

http://stackoverflow.com/questions/9655653/is-it-able-to-get-the-generated-html-for-a-part-of-markdown

See this gist for a test case: https://gist.github.com/1163474

Pegdown needs too long (hours) for parsing the simple string A<hr>B_C_D_E_F_G_H_I_J_K_L_M_N.

• You could also replace <hr> with any other html tag to get a too slow result.
• If I remove the underscores, everything is fast.
• If I remove the beginning A-text, everything is fast.
• If I remove the tag <hr> everything is fast.
• If I add an empty line between A and <hr> everything is fast.

I have this simple utility class with ThreadLocal static field:

public class MarkDownProcessor {
    private static final ThreadLocal<PegDownProcessor> PROCESSOR = 
        new ThreadLocal<PegDownProcessor>() {
            protected PegDownProcessor initialValue() {
                return new PegDownProcessor(Extensions.WIKILINKS + Extensions.AUTOLINKS);
            }
        };

    public static PegDownProcessor get() {
        return PROCESSOR.get();
    }
}

When I want to render markdown text to page, I can use this:

MarkDownProcessor.get().markdownToHtml(....);

But from time to time first call to markdownToHtml() renders same output as I passed into. Next call is OK. Maybe, it is a timing stuff. Can I use ThreadLocal for managing instances of PegDownProcessor or it is a bug?

(This has been discussed in #28, but the solution from commit 07dda00 is incomplete.)

Subclassing Parser does only work if the subclass resides in the same package as Parser, that is, org.pegdown. The MyParser subclass introduced in commit 07dda00 is inside that package. Relocating it to any other package results in an IllegalAccessError.

PegDownProcessor with enabled extension TABLES crashes with org.parboiled.errors.ParserRuntimeException for text:
"$c_3 \cdots k_3 + c_4 \cdots k_4 + c_5 \cdots k_5 = s |c_3 \cdots k_3 --- c_4 \cdots k_4| + |c_4 \cdots k_4 – c_5 \cdots k_5|$" (quotes added for clarity). PegDownProcessor version is 1.1.0.

This appears to trigger exponential parsing time:

Blaa

**SELECT @default_late_fees =
     CASE @default_days_out % @default_product_price_days_out
               <br>WHEN 0 <br>THEN (@default_days_out / @default_product_price_days_out) * @default_amount 
               <br>ELSE (FLOOR(@default_days_out / @default_product_price_days_out) + 1) * @default_amount 
          <br> END**

Thanks,
<br>S

Maruku supports CSS header declarations like this:

CSS: style.css

# Normal Markdown starts here

Etc. Etc.

Which will insert the appropriate <link /> tag into the rendered HTML.

This is super handy and would love to see it in pegdown.

Whenever an attribute name of an HTML tag contains underscores, the tag is not recognized as a tag. Instead, the characters are escaped.

For example, given input:

Some text

<div attr_with_underscore="the value">
then a div
</div>

Then some other text

We expect the block-level tag to be recognized and left as it is:

<p>Some text</p>

<div attr_with_underscore="the value">
then a div
</div>

<p>Then some other text</p>

But, instead, the opening tag gets escaped:

<p>Some text</p>

<p>&lt;div attr_with_underscore=&ldquo;the value&rdquo;&gt;
then a div
</div></p>

<p>Then some other text</p>

Hi Matthias,

I have a regression in specs2 with the latest 1.0.0.

A code block like this:

• this is an item

   // showing some code here
   val x = y + 1
  

• this is another item

used to be rendered with just one

 tag for the code block.

Now I have several 
 tags for each line of the block. For example, in my user guide, I used to have:

  
 
    
using logical operators
 
    
def beBetween(i: Int, j: Int) = be_>=(i) and be_<=(j)
// create a Seq Matcher from a Matcher

def allBeGreaterThan2: Matcher[Seq[Int]]   = be_>=(2).forall     // fail after the first failure

def allBeGreaterThan3: Matcher[Seq[Int]]   = be_>=(2).foreach    // like forall but execute all matchers and collect the results

def haveOneGreaterThan2: Matcher[Seq[Int]] = be_>=(2).atLeastOnce


   
Now I have:
 
  
using logical operators
 
  
def beBetween(i: Int, j: Int) = be_>=(i) and be_<=(j)
// create a Seq Matcher from a Matcher


  
def allBeGreaterThan2: Matcher[Seq[Int]]   = be_>=(2).forall     // fail after the first failure