From: Stuart Donaldson <stuartd@al...> - 2003-01-15 22:23:10
> -----Original Message-----
> From: Ian Bicking [mailto:ianb@...]
> Sent: Wednesday, January 15, 2003 1:08 PM
> To: Webware devel
> Subject: [Webware-devel] ExtraURLPath
> I'm going to review the New algorithm, as used by ExtraURLPath. These
> are the things I want to do:
> * Handle the infinite loop on a 404
> * Fallback to Main.py/index.py like 0.7 behavior
> * Handle ExtraURLPath=0 (i.e., signal 404) -- this won't be triggered
> in the current version, but hopefully after 0.8 we can
> ditch all the
> old path handling code.
> Is there anything else I need to look at? I'm just going to be
> reviewing that entire piece of code, so if there's any issues I should
> be able to look at them.
> I'm still not clear on the includeURL stuff -- is there
> anything I need
> to look at with the path handling? It didn't seem like it from my
> current understanding, but I'm still confused about the plan for that.
Unless I hear much concern to the contrary, I plan to roll the includeURL
fix into CVS tonight. I posted a patch to try and get some preliminary
review before potentially destabilizing the tree, but with public CVS access
down, that may be more hassle than it is worth now.
It would be great if you could test in that environment, or at least take a
look at it.
Check out the extraurl patch I did as well, it is an attempt to fix some of
the problems you are planning on addressing.
The plan on the includeURL handling stuff is as follows:
includeURL() forward() and callMethodOfServlet() take a URL
The URL is considered relative to the current servlet in the current
/Testing/IncludeURLTest includes "Dir/IncludeURLTest2" to actually
Absolute references are absolute within the current context.
"/Testing/Dir/IncludeURLTest2" referencing "/Main" will reference
It is not currently possible to access outside of the current
A second patch, not currently planned for the 0.8 release, will introduce an
optional argument where the Context can be specified.
There was also a psp:include bug related to all of this, and I need to
double-check to make sure that it works there as well.
Does this make sense and sound reasonable?