Le Lundi 6 Février 2006 18:31, Tobias Toedter a écrit :
> On Monday 06 February 2006 18:00, Mathieu Roy wrote:
> > Le Lundi 6 Février 2006 17:48, Tobias Toedter a écrit :
> > > This commit just changes indentation and adds a placeholder condition
> > > around the formatting code. The condition will be modified to be able
> > > to selectively bypass the markup code when special markup tags are
> > > encountered (<nowiki>, <pre>, <code>).
> >
> > ??
> >
> > what means nowiki, what means pre?
>
> <nowiki> is a tag used by MediaWiki, to disable the wiki formatting. I
> think this is something we need, too.
>
> <pre> is just the plain old HTML <pre>, meaning to output the following
> text without markup modification.
>
> > And code is way to much restrictive IHMO. Code is meaningless for
> > installation that are no software related (and it exists). Verbatim is
> > neutral.
>
> Yes, I know that code has a context meaning, but I thought that quite a lot
> of projects might make use of it, because the large majority is about
> software. If you think that it's not necessary, we can just drop it.


What you mention as pre, nowiki and code refers to what I called "verbatim", 
see the related task comment.

Verbatim is a neutral word that exactly means what we mean, pre is 
un-understandable if you ignore HTML, nowiki is confusing (there's no wiki 
involved here), code is too specific.



> > Also, do we really want to use <tag>? Seems to me that we'll definitely
> > get into trouble as soon as someone will post in a comment bits of HTML,
> > which is very likely to happen. Our tags should be someone unique enough.
>
> Hm, that brings up another topic I'd like to discuss. Currently, the user
> submitted data is stored in the DB as HTML, i.e. with HTML tags being
> interpreted. The output of DB items gets no filtering at all, so that HTML
> attacks would be possible.
>
> So, our current scheme is this:
> input: htmlspecialchars(user input) -> database
> output: database -> browser
>
> It would be much cleaner to do it the other way round:
> input: user input -> database
> output: htmlspecialchars(database) -> browser

Do we really convert things to htmlspecialchars before inserting things in the 
database?

I don't remember but that indeed seems awkward. What happen then if we grab 
the database content to put it in a plain text mail, the plain text get html 
entities? Weird.

Are you sure we don't do the later?

> We're planning to perform some markup in the cookbook table during the next
> upgrade of Savane. Shouldn't we also convert the HTML data into normal text
> data?

We should.


-- 
Mathieu Roy

  +---------------------------------------------------------------------+
  | General Homepage:           http://yeupou.coleumes.org/             |
  | Computing Homepage:         http://alberich.coleumes.org/           |
  | Not a native english speaker:                                       |
  |     http://stock.coleumes.org/doc.php?i=/misc-files/flawed-english  |
  +---------------------------------------------------------------------+

Reply via email to