History of TikiWiki Parser Page of Doom

some errors I have noticed with the wiki parser on my own site that also does not work here:
---
{CODE}{CODE}
---
*fee^foo
bar^
---
!heading
*[http://google.com|google] sometext:
**double bullet point 1 ending with colon (the addtional sub text below this is missing):
++text preceeded by double plus signs blahblahblahblahblahblah blah blahblah blah blahblahblahblahblahblah blah blahblah blah blahblahblahblahblahblah blah blahblah blah blahblahblahblahblahblah blah blahblah blah blahblahblahblahblahblah blah blahblah blah blahblahblahblahblahblah blah blahblah blah blahblahblahblahblahblah blah blahblah blah
++sub text of double bullet point 1

*[http://www.google.com|google] the word google should be a link preceded by a bullet, also:
**"double bullet point 2"
++"more text" this should be subtext of double bullet point 2

~pp~ this is text that should not be parsed ~/pp~
~np~ showing ~pp~ and ~/pp~ non-parsed ~/np~ does not work

however,

~pp~ showing ~np~ and ~/np~ non-parsed does work ~/pp~

---
I wonder why \nr rowends in tables are "recommended"? It disalows any nice formatting of multy-line cells (like using lists or divs etc. inside sells) and renders tables almost unusable...

||{div}r1c1{/div}|r1c2
{div}
*r2c1item1
*r2c1item2{/div}|r2c2||

---
{img src=http://bitweaver.org/ link='http://www.bitweaver.org/wiki/edit.php?page_id=452'}
Now it takes extra ' to make the link attribute with = char inside the URL work, when in 1.3 it worked fine without it. This could cause inconsistent results when upgrading existing bw.o site to HEAD
{img src=http://bitweaver.org/ link=http://www.bitweaver.org/wiki/edit.php?page_id=452}