parse the html

[Reynaldo]

i've parse the html to $html

Code: Select all

    regsub -all "\n" $html "" html
    set nopage [string first "<div id=date>" $html 0]
    set news [string range $html $nopage [expr [string first "<ul>" $html $nopage] - 1]]

the output $html:

Code: Select all

<div id=tanggal> 
Tuesday              , 
27/09/2005 09:11              EST</div>
Today news is bla bla bla bla.              </a></div>
<div id=summary> 
Bla bla bla bla bla news today.            </div>
<div id=titlebiru>Read also :</div>
<ul>

How to regexp the date, news topic, and the news? into $variable, so the output will be:
Tuesday, 27/09/2005 09:11 EST, Today news is bla bla bla. Bla bla bla bla news today.

[De Kus]

I'd suggest to run a regexp on the complete html without string range and regsub changes.
if you are too unfamilar with regexpand dont want to link to the complete page I'd recommned you to try out this one:
http://forum.egghelp.org/viewtopic.php?t=9972

[demond]

suppose you want to extract what is contained between some opening and closing tags:

Code: Select all

[demond@whitepine demond]$ tclsh8.4
% set str "<tag attr=foo>some text</tag>"
<tag attr=foo>some text</tag>
% regexp {<tag.*?>(.*?)</tag>} $str -> str
1
% set str
some text
%

[De Kus]

Tip: use '[^>]' or '[^<]' instead of '.' in cases where <tag> </tag> is not unique, because regex tends to match the widest match, not the shortest match.

[demond]

Tip: use '[^>]' or '[^<]' instead of '.' in cases where <tag> </tag> is not unique, because regex tends to match the widest match, not the shortest match.

there's no need of that, '*?' is a non-greedy quantifier, see the docs

[De Kus]

ah sorry. must have missed it, hmm, should try if it is actually faster than the other variant.