RegEx to shrink text with HTML tags

>([^>]{0,400})[^<]*

(>([^<]{0,400})[^<]*

Ozo, your regex captures a string between the ending bracket of a tag, takes up to 400 chars and only makes sure there's no closing bracket in the capture. Then it throws the stringaway up to the next opening bracket.

The OP asks for no break inside the middle of a word and the text can have any number of tags, opening and closing, but must not break layout (i.e., must not end up with crippled tags).

This is hardly a trivial task. Consider bold text
 withimages. The regex needs to understand that
 and  do not need to be scanned for a matching ending tag,  does. So, the regex can "break in the middle" of
 and and  but not "in the middle" of  ... .

That's at least my understanding of the question. I wonder what would happen if a opening / ending tag contains more than 400 chars (i.e., a table, a div).

I'm sorry I am not posting a solution (yet) because I think it is not easily (or at all) possible with a simple regex. A parsing language like XSLT (available in PHP) might be an easier means to tame this beast :), but only when this is about XHTML and not just HTML.

Regards,
-- Abel --

The crippled part again (oh, I hate this new EE rich text formatter that cripples just about any post...)

This is hardly a trivial task. Consider <b>bold text with <img ...> images and <br> newlines</b>. The regex needs to understand that <br> and  do not need to be scanned for a matching ending tag,  but <b> does. So, the regex can "break in the middle" of <br> and <img>  but not "in the middle" of <b>...<b>

The crippled part again (oh, I hate this new EE rich text formatter that cripples just about any post...)

This is hardly a trivial task. Consider < b >bold text with  images and < br > newlines. The regex needs to understand that < br > and < img > do not need to be scanned for a matching ending tag,  but < b > does. So, the regex can "break in the middle" of < br > and < img >  but not "in the middle" of < b >...< / b >

*edited by Netminder to fix formatting issue. Note that there are now spaces after the < and before the > for tags that can possibly be converted by the Rich Text feature*

sorry for the clutter. I have no idea how to fix this... See the code snippet for uncrippled text... :S

This is hardly a trivial task. Consider <b>bold text with <img ...> 
images and <br> newlines</b>. The regex needs to understand 
that <br> and  do not need to be scanned for a matching ending tag, 
but <b> does. So, the regex can "break in the middle" of <br> 
and <img>  but not "in the middle" of <b>...<b>

Select allOpen in new window

Last time, then I give up...

This is hardly a trivial task. Consider 
 
<b>bold text with <img ...> images and <br> newlines</b>. 
 
The regex needs to understand that <br> and 
do not need to be scanned for a matching ending 
tag, but <b> does. So, the regex can "break in the middle"
 of <br> and <img> but not "in the middle" of <b>...<b>

Select allOpen in new window

Hello abel,
thanks for your answer. You pointed out a problem I did not think of before: what happens if the whole string starts with a tag that ends at the end?

So I think the best solution would be to "kill" all unclosed HTML tags and leave the rest as is... still the question, how to do it?

Can you elaborate a bit on the "kill" bit? How would that look like? Surely this may well end up onto a very crippled HTML bit (i.e., when a "tr" is killed, but the "td" is not, or when the "html" is killed and the header tags not, let alone lists and selectboxes...).

Btw: the bug with the RTE editor is being researched by the EE team.

Something that didn't work out for you, I see, considering the layout you made :)

@abel:
good point... I start to think that it might be better to leave only and  tags and strip the rest totally... will think about it a little.