On Fri, Jul 25, 2003 at 05:40:59PM +0200, Anders Engström wrote: > Hi, > > a recent thread discussed WW2 and Sitemesh. I've got some > questions/thoughts on WW1.3 and Sitemesh. > [snip] > > What was the initial motivation to put the last executed action in the > request? > > After seeing this I thought that perhaps the taglibs would use this as a > "last resort" when looking up properties - but this is not the case. >
I'm still interested to know what the intention of this design was. So - I dug through CVS and found the following log entry for ServletDispather revision 1.29: === Cut from CVS === Revision 1.29 / (view) - annotate - [select for diffs] , Wed Jan 23 01:58:48 2002 UTC (18 months ago) by salaman Branch: MAIN Changes since 1.28: +10 -3 lines Diff to previous 1.28 Made the stack's head value be stored in the request's attribute map upon it being popped from the stack. This is to ease in integration of other programs which might want to interact with Webwork's action after it's done processing. APPLICABILITY: This will allow sitemesh's filter to push this value back in the stack before decorating,and popping it when done, therefore giving the decorator access to anything that was available from the decorated page. =================== Any comments on this? Should SiteMesh handle this stuff (pushing the value on the stack before decorating), or should webwork allow tag-libs to look at the STACK_HEAD request attribute if everything else fails while looking for a value? //Anders -- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anders Engström [EMAIL PROTECTED] . http://www.gnejs.net PGP-Key: ED010E7F . [Your mind is like an umbrella. It doesn't work unless you open it.]
pgp00000.pgp
Description: PGP signature
