Hunchentoot/cl-who Page Composition
I'm trying to put together a few pages in hunchentoot as an experiment, and I'm running into an unexpected wall. As an example, I have the following template macro.
(defmacro page-template ((&key title) &body body) `(with-html-output-to-string (*standard-output* nil :prologue t :indent t) (:html :xmlns "http://www.w3.org/1999/xhtml" :xml\:lang "en" :lang "en" (:head (:meta :http-equiv "Content-Type" :content "text/html;charset=utf-8") (:title ,(format nil "~@[~A - ~]Test Site" title))) (:body ,@body))))
Now when I have a pure text page, or one filled with html literals like
(define-easy-handler (test-page :uri "/") () (page-template (:title "Splash Page") (:p "Testing testing")))
everything is a-ok. The page outputs properly and I can see te efforts of my code instantly. However, when I have a page which is made up of redundant elements, it's not as simple. For example, lets say I have a page on which for whatever reason I want to display three RSS newsfeeds. This is a sufficiently complex component that I want to abstract it out, so to my minnd, I should be able to do something like
(define-easy-handler (test-feed :uri "/feeds") () (page-template (:title "Splash Page") (publish-newsfeed "http://nf-one.html") (publish-newsfeed "http://nf-two.html") (publish-newsfeed "http://nf-three.html"))) (defmacro publish-newsfeed (url &optional (item-limit 5)) (flet ((get-text (s-tree node-path) (car (last (xmls-tools:find-subtree s-tree node-path))))) (let ((rss-feed (xmls:parse (drakma:http-request url)))) `(:div :class "rss-feed" (:a :href ,(get-text rss-feed '("channel" "link")) :target "_top" (:h1 ,(get-text rss-feed '("channel" "title")))) (:ul ,@(mapcar #'(lambda (item) `(:li (:a :href ,(get-text item '("link")) :target "_top" (:h2 ,(get-text item '("title")))) (:p :class "date" ,(get-text item '("pubDate"))) (:p ,(get-text item '("description"))))) (let ((items (xmls-tools:find-all-children (xmls-tools:find-subtree rss-feed '("channel")) "item"))) (if (> (length items) item-limit) (subseq items 0 item-limit) items))))))))
But the result of the above is a "Server Error" page. I'm not quire sure why; page-template
is a macro so the calls to publish-newsfeed
shouldn't be expanded until they're in the context of with-html-output-to-string
. Can anyone tell me what I'm doing wrong?
Also, on closer inspection of the various Hunchentoot/cl-who tutorials, none of them seems to do this kind of page composition. Can anyone with some Hunchentoot experience tell me what the correct/canonical way of decomposing a page into components is?
EDIT:
Correct response by Ramarren below; the with-html-output
macros work under different evaluation rules. The version of publish-newsfeed that would actually work in this situation is actually
(defun publish-newsfeed (url &optional (item-limit 5)) (flet ((get-text (s-tree node-path) (car (last (xmls-tools:find-subtree s-tree node-path))))) (let* ((rss-feed (xmls:parse (drakma:http-request url))) (items (xmls-tools:find-all-children (xmls-tools:find-subtree rss-feed '("channel")) "item")) (ltd-items (if (> (length items) item-limit) (subseq items 0 item-limit) items))) (with-html-output (*standard-output* nil :indent t) (:div :class "rss-feed" (:a :href (get-text rss-feed '("channel" "link")) :target "_top" (:h1 (str (get-text rss-feed '("channel" "title"))))) (:ul (dolist (item ltd-items) (htm (:li (:h2 (:a :href (get-text item '("link")) :target "_top" (str (get-text item '("title"))))) (:p :class "date" (str (get-text item '("pubDate")))) (:p (str (get-text item '("description")))))))))))))
Note the removal of mapcar
for dolist
(I'm a Schemer, don't give me too much of a hard time about liking lambdas, but they weren't the right choice here), and the use of htm
to escape blocks of html s-exps (h-exps?) that wouldn't otherwise be in context for with-html-output
. Finally, I had to wrap text but NOT :href
properties in (str )
to get them to expand dynamically.