callback receives the submited value.
submit-form-action handles a form submit for some form button, making sure that no field values are lost in the process.
(defun get-request-action (action-name &key (ignore-missing-actions *ignore-missing-actions*))
"Gets an action from the request. If the request contains
*action-string* parameter, the action is looked up in the session and
appropriate function is returned. If no action is in the parameter,
returns nil. If the action isn't in the session (somehow invalid),
raises an assertion."
(when action-name
(let* ((app-wide-action (weblocks/app-actions:get-action action-name))
(session-action (get-session-action action-name))
(request-action (or app-wide-action session-action)))
;; TODO: rethink this form. May be throw a special condition instead of string
(unless ignore-missing-actions
(assert request-action (request-action)
(concatenate 'string "Cannot find action: " action-name)))
request-action)))
what if I have to redefine a route which already exists? I have to restart the entire lisp system
reset-routes(https://github.com/40ants/weblocks/blob/reblocks/src/routes.lisp#L55) and reload your system.
2 replies
Hello! I'm using Weblocks to make a webcomic website for a friend and I'm infinitely grateful for a way to create nice modern websites while completely avoiding Javascript. :)
Now, to the problem I'm having. The big picture: toasts.
The currently smaller picture: Making them disappear on timer.
I'll send a link to what I have now in a few minutes. It doesn't remove toasts on timer, and it asks the server to render them all every time one is created or closed (which is not a big deal really since it's the server code which creates the toasts anyway).
3 replies
Then click again on "render many" and without delay click on "render few"
You will see that AFTER "render few" (the last action from the UI) has finished doing its work,
the results will be replaced by an older action from the UI ...
How to avoid this? In a filtering widget for example, typing a word fast will result in the wrong filtration of a long list.
(after if finishes rendering the newest results according to the last input word, it will then render the previous results from the previous UI action when
the word was only partially typed into the box)"
https://justpaste.it/77h4x
2 replies
calling (update mywidget) from a new thread
How to make a new thread aware of the *session'* it was created from ?
executable code to illustrate the problem: https://gist.github.com/bamboospirit/e6ad95503ec40d2c7649f4aa14ad4979#file-thr-update-lisp
github issue 40ants/weblocks#54
(defmacro define-attic (attic-name routes &key (prefix "/"))
(with-gensyms (app route-name)
`(progn
(defapp ,attic-name :prefix ,prefix)
(defroutes ,route-name
,@routes)
(defmethod weblocks/session:init ((,app ,attic-name))
(declare (ignorable ,app))
(,(alexandria:symbolicate :make- route-name))))))
(defun start-attic (port)
(weblocks/server:start :port port))
(define-attic attic
(("/" (with-html (:h1 "Hello")))))
(start-attic 40002)
@C-Entropy I suggest you to create an app's prototype without all this macrology first. When it will work, start creating an abstractions.
Also, please, provide a full reproducable piece of code with the package definition. This way it will be easier to help you.
And use triple backticks ` to post formatted code like this:
(defmacro define-attic (attic-name routes &key (prefix "/"))
(with-gensyms (app route-name)
`(progn
(defapp ,attic-name :prefix ,prefix)
(defroutes ,route-name
,@routes)
(defmethod weblocks/session:init ((,app ,attic-name))
(declare (ignorable ,app))
(,(alexandria:symbolicate :make- route-name))))))
(defun start-attic (port)
(weblocks/server:start :port port))
(define-attic attic
(("/" (with-html (:h1 "Hello")))))
(start-attic 40002)("/" (with-html ...)) is wrong
("/" (make-instance 'hello))
(defmethod render ((hello hello))
(with-html (:h1 "Hello hello"))), am I right?
Yes, this is how it should be used.
Here is the real world example: https://github.com/ultralisp/ultralisp/blob/f2cb634fb302d3fec98c1ad98e6759e4f6060af3/src/widgets/main.lisp#L43-L69
Do you mean you want to make some page have a special header?
There is no easy way, but you might do the following:
- define a special variable like
(defvar *additional-headers*) - define an
render :aroundmethod to wrap this one: https://github.com/40ants/weblocks/blob/reblocks/src/page.lisp#L103-L129. In this method you need to bind*additional-headers*toNIL. - define a custom
render-headers :aftermethod for https://github.com/40ants/weblocks/blob/reblocks/src/page.lisp#L59-L85 and render content of*additional-headers*. - during rendering of the widget, add new pieces of header to
*additional-headers*.
This way you will be able to extend header from widgets, rendered to the page.
But this will work only for the initial loading of the page. When some widgets will be updated o replaced/created by AJAX, header will not change.
If you want to modify a visible header, then the best way will be to make it as a widget and to update it as any other widget.
(with-html-form (:POST
(lambda (&key file text &allow-other-keys)
(format t "~a ~a~%" file text))
:enctype "multipart/form-data")
(:p "Please choose a file:")
(:input :type "file"
:name "file")
(:input :type "text"
:name "text")
(:input :type "submit"
:value "Upload"))
Sooo … I'm on https://40ants.com/weblocks/quickstart.html
(render task-1) gives the error message
<DEBUG> [17:25:01] weblocks/widgets/render-methods render-methods.lisp (render :around) -
debugger invoked on a SIMPLE-ERROR in thread
#<THREAD "main thread" RUNNING {70052141F3}>:
Please, use push-dependency in code, wrapped with with-collected-dependencies macro.and the web app is not interactive. The debugger says
<ERROR> [17:27:03] weblocks/request-handler request-handler.lisp (handle-request :around app h0) -
Unhandled exception
<WARN> [17:27:03] weblocks/request-handler request-handler.lisp (handle-request :around app h0) -
Invoking interactive debugger because Weblocks is in the debug modeAny help with this, please?
I'm on sbcl 2.1.5 on an m1 MacBook Pro.
3 replies