Links failing when there is class, text and title, all in in parentheses

[gocom]

The following:

"(class)(text)(title)":http://example.com/

"(class) (text)(title)":http://example.com/

"(class)(text) (title)":http://example.com/

"(class) (text) (title)":http://example.com/

Generates:

<p><a class="class" href="http://example.com/" style="padding-left:1em;padding-right:1em;">(title)</a></p>

<p><a href="http://example.com/">(class) (text)(title)</a></p>

<p><a class="class" href="http://example.com/" style="padding-left:1em;padding-right:1em;">(title)</a></p>

<p><a href="http://example.com/">(class) (text) (title)</a></p>

There seems to be no way to have class, title and content in parentheses at the same time.

[gocom]

The text seems to get eaten and parentheses are interpreted as padding.

[netcarver]

I had a quick look at this last night and it seems to be the $this->c definition that's causing it.

        $this->c = "(?:$this->clas|$this->styl|$this->lnge|$this->hlgn)*";

This regex allows repeated classes/styles etc so it's matching the class and text from (class)(text)(title) via the $this->clas part (the title is pulled off by the link parsing regex and used as the link text as the attribute gobbled it.) Then the attribute parsing code takes the (class)(text) that it got passed as the attributes and correctly pulls the class from it leaving (text) which then gets parsed as left and right padding. What a mess.

I'm going to do more work on the link parsing code anyway over the Christmas break so I'll take a closer look at fixing this as I do what I already had planned.

[netcarver]

Pushed a fix for this as part of the new branch (now merged into master and removed: see db84396 in particular with regard to this issue). Uses a new pattern that cuts out the * from the attributes regular expression but still allows matching of language/style/class#id in any order.

[netcarver]

Closed by a45f808.