forrest-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From David Crossley <>
Subject Re: locationmaps not working?
Date Mon, 03 Oct 2005 02:24:46 GMT
Ross Gardler wrote:
> Diwaker Gupta wrote:
> >Ross Gardler wrote:
> >
> >>Cool. Does this mean the issue you raised earlier in this thread about
> >>the LM verifying the existence of a file is not critical?
> >>
> >>I think you had a valid point about this when using the LM to rewrite
> >>links, but I'm not sure if this was all one of the strange effects you
> >>were getting.
> >
> >This issue still exists (try changing to or some 
> >random name in samples/locationmap/index.xml and see what happens). I 
> >certainly think its critical (and I think an implementation that does 
> >_not_ do this check would be simpler). What do other people think about it?

I became aware of that recently. We actually had
a bad URL for that demo, but we didn't know it was
broken. Suddenly there was a change in behaviour.
It didn't used to verify. See where i needed to
change the demo.
svn log main/fresh-site/src/documentation/content/locationmap.xml
Restore upgrade from excalibur-sourceresolve-1.1 to excalibur-sourceresolve-2.1
Use a successful result for the rewriteDemo/index locationmap
demo in seed-sample.
See "changed behaviour for locationmap"

> There are two distinct uses of the locationmap. One is for URL rewriting 
> (like you are doing). In this instance I think you have a good case for 
> not doing the check for existence, at least during development. However, 
>  at publication time I think this is a valid check as it will highlight 
> broken links.
> The other use of the LM is where we use it to resolve the location of a 
> source file, as in the sitemap stuff. I this instance the check has to 
> be made, otherwise a meaningless error is given (null pointer).
> One solution would be to add an attribute to the locator node that 
> allows us to set whether we should check for existence. This could have 
> three values: "always", "never" or "publication". The last option would 
> allow us (with the addition of a new CLI parameter) to check that the 
> file exists when we try and publish the site, but would allow us to 
> ignore errors during development.

However we do not want to ignore errors during development,
so the default is good.


> I don't have the time to implement this yet. If this proposal would 
> solve your problem then please add it to the issue tracker (along with 
> any other alternatives you can think of).
> Ross

View raw message