Re: Tree Iter: Text out of pango markup

[muppet <scott asofyet org>]

On Feb 4, 2005, at 2:51 AM, A. Pagaltzis wrote:

> * muppet <scott asofyet org> [2005-02-04 03:45]:
> > All you really want to do, though, is strip tags, so even
> > pango_parse_markup() would be overkill for you.  Things like
> > $markup =~ s/<[^>]*>//g are what perl does best.  :-)
>
> Is that sufficient in the case of Pango markup? Are there no edge
> cases like with HTML where this could break? How does Pango deal
> with angle brackets that are supposed to be displayed, and aren't
> part of markup?

Of course there are edge cases, nothing can ever be truly simple in 
computing.

To find out how Pango markup gets interpreted i usually just create a 
dummy project in glade, a window with a label, set the label to "Use 
Markup", and then type in the Label box and see what shows up.

If there's *anything* that causes the markup to be invalid, e.g. one of 
ten tags is unknown, or there's a single stray "<", the whole thing is 
considered to be not markup.

Luckily, the known tag list is small:
http://developer.gnome.org/doc/API/2.0/pango/PangoMarkupFormat.html


I also think that in the general case, if you control the text that 
goes in, you can ensure that it's valid, so simply stripping tags and 
converting entities would be a quick and easy solution.

--
If I lived in Teletubby Land, the homicide rate would be four.
  -- elysse