How to stop TYPO3 from stripping <p> tags from RTE content (or touching it at all) - typo3

Question

I have a TYPO3 which strips any <p> tags from content I create when the RTE editor is enabled for the field in question before it saves it in the DB. And it seems I cannot find a way to disable this behavior with a TypoScript. As soon as I disable the RTE editor I can save <p> tags and they get correctly rendered in the frontend. They also get correctly rendered in the frontend when I simply add them directly in the database in the tt_content table in the bodytext field.
When I switch to the edit source mode of the RTE I see all <p> tags in place. Before and after a save (also they never make it to the DB) so it looks like they get converted to (linux) line breaks or something and get converted back to <p> tags when the editor loads them in the backend. But those line breaks of course have no effect to in the frontend.
I thought this behavior would be controlled by RTE.default.proc but everything there looks good to me (p is already in the allowed tags and there is no clue why it could be stripped). I've also tried to disable the RTE.default.proc.entryHTMLparser_db and RTE.default.proc.exitHTMLparser_db as I wouldn't mind it if the HTML content as you see it in the edit source mode of the RTE (so with RTE still enabled!) would not be touched at all - in fact I would prefer it - but this had no effect. On the other hand when I add tags to the allowed tags which weren't there before (like <button>) this works so the things I try to add to RTE.default.proc aren't ignored in general.
So how can I stop TYPO3 from stripping my <p> tags from RTE content or touching it at all? I'd prefer a solution with TypoScript but meanwhile I would also be happy about an ugly hack in a sys extension as long as it works...

Answer 1

I'm not sure if this is a bug or not but the solution to my problem lies in p.rmTagIfNoAttrib = 1. At least in my TYPO3 version (v6.1.7 and nearly only built in extensions) I cannot find this setting in the preset TypoScript of the page or the RTE editor so I'm guessing it defaults to 0. In my logic 0 means false so I'd say the default would read as "remove tag if there is no attribute?: no!".
However TYPO3 seems to work after its own logic. Adding the following statement to the page TS sloves my problem and <p> tags are preserved:
RTE.default.proc {
entryHTMLparser_db {
tags {
p.rmTagIfNoAttrib = 1
}
}
}

Answer 2

The reasons for this behavior are explained in the TYPO3 manual:
Many of the transformations performed back and forth in the TYPO3 backend date back to when it was a challenge to incorporate a RTE in a browser. It was then sometimes needed to fall back an a simple <textarea> where rich text had to be presented in a simple enough way so that editors could work with it with no visual help.