[NLP-er]

Originally Posted by James Birkett

I think it parses inward.
I.e. [img][b]text[/b][/img]
It'll parse the image tag, then the bold I think.

I come to this conclusion because you can wrap noparse tags around BBCodes, so if they went outwards then the tags would be parsed before the noparse being parsed.. if that makes sense.

I already checked it - it parses from deepest, but images are parsed separately. So there is several blocks of code to parse bbcode. I had to parse images myself in my bbcode, because I need whole content parsed before I do my own changes.