commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeff Turner <>
Subject Re: Latka notes
Date Wed, 06 Mar 2002 03:20:14 GMT
On Tue, Mar 05, 2002 at 09:54:55PM -0500, Serge Knystautas wrote:
> I don't think I explained my idea 1 very well based on comments.
> The issue I'm facing is making sure my latka suite stays current with the
> site as developers maintain the site.  For an initial import, I might have a
> webapp with 350 JSPs... I don't want to manually build the latka suite for
> that.  More importantly, as the team of developers add more JSPs, I want to
> know that some schmuck didn't add a JSP to CVS and not add it to the
> functional test.
> I would start with just a file system dir list, recursing through subdirs
> and listing all *.jsp files.  (no link checking or wget or anything
> complex... just pure dir.list() with recursion).  This would create a list
> like:
> /test.jsp
> /index.jsp
> /login.jsp
> /contacts/list.jsp
> /contacts/add.jsp
> /contacts/edit.jsp
> It would take these URLs and build the latka suite file.  If there already
> was a latka suite file, it would append the files not already in there.
> Really this is just a tool/script to maintain the latka suite file.

Perl, or shell would be easiest.

If you want to keep location-independence (list URLs, not file paths),
you could write a servlet that returns a list of directory contents as
XML. Then use Ant's <get> to retrieve the XML listing, <style> to
transform it into a Latka XML file, and off you go.

> I was getting somewhat ahead of myself with the "meta" decription, but I was
> thinking the filtering could get more advanced so that I included "*.jsp"
> and "*.html", but excluded everything in "/include" since those pages are
> not directly called.  Maybe I should just fashion this as an ant task so I
> can leverage it's include/exclude file pattern capabilities.

Except in Ant you can't iterate over items in a fileset. Maybe there's
some 3rd party tasks for that.

> On an unrelated note, I thought of two other validators... a) link checking,
> and b) xml validation (if your servlet is supposed to output XML instead of
> HTML).  Anyone else working on these kind of validators?

XML validation should be very easy (XPath does it for free). Link
checking might need the JTidy one's capabilities first. AFAIK, no-one is
working on them.


> Serge Knystautas
> Loki Technologies - Unstoppable Websites
> ----- Original Message -----
> From: "Jeff Turner" <>
> To: "Jakarta Commons Developers List" <>
> Sent: Tuesday, March 05, 2002 5:34 PM
> Subject: Re: Latka notes

To unsubscribe, e-mail:   <>
For additional commands, e-mail: <>

View raw message