Re: Working on SIM-82

On Tue, 2004-03-09 at 18:51, Mike Cannon-Brookes wrote:
> I'm a +1 on this _as long as you're willing to put in the testing  
> effort_.
> 
> Sorry to be a bore on this (we'd all like SiteMesh to be better), but...
> 
> Trust me when I say that Scott has put in a s**tload of time testing  
> SiteMesh on all of the different application servers we support (Orion,  
> Resin, Jetty, JBoss, Tomcat, Weblogic, Websphere, SunOne, JRun) and  
> we've found that developing SiteMesh is like playing whack-a-mole. Fix  
> something (like i18n) on one server and it then breaks on another.

Sure. That's why I was bringing it up to the list instead of just
hacking away until it works for *my* needs. I don't have every server,
not by a long shot - I'd need some kind of group support, at the very
least, participating in testing and fixing.

> The existing test suite is pretty good, but testing is laborious  
> because it really requires setting up all of these servers and  
> hammering at them one by one.
> 
> At the moment the only real SiteMesh issues that I know of are the HTML  
> file problems on some Tomcat versions, and the i18n problem with Resin  
> + Velocity decorators (I believe this one is a Resin bug but will check  
> with Scott).

Well, the Tomcat issues are pretty severe. It makes sense once you know
what the problems are, but the solution is unsatisfactory - and remember
that Tomcat is *probably* the most popular container out there. Being
unable to decorate static content in the most popular container out
there (again, I don't have substantive numbers, but I think it's a safe
guess) is Not A Good Thing.

Luckily, it's SIM-83 that gave me a good idea about how to fix it,
although I'm quite sure there are issues I won't anticipate offhand.
Again, my goal was to get some sense of community support for fixing the
thing once and for all. If I'm doing this as an island effort there's no
point as Tomcat isn't a "normal" container for me.

> Can we maybe do this on a branch if it's going to hack at the core?  
> (which I agree is a complete ball of wax at the moment - but most of  
> that is because all the container specific stuff are bug fixes gained  
> through much pain in the past).

Sure, I had no intention of putting into HEAD until it was ready.





-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click