How can I programmatically add script or stylesheet tag to a page specified in page's YAML front matter (meta)?
Assuming there is src/documents/posts/a.html.eco with following contents:
---
layout: default
scripts: ['a.js']
---
Blog post that requires a special javascript
and layout src/layouts/default.html.eco with following contents:
...
#getBlock('scripts').toHTML()
</body>
...
The final result for posts/a.html should be:
...
<!-- some extra stuff that was added when processing script tag -->
<script scr="/scripts/a.js"></script>
</body>
...
..while other pages shouldn't have a reference to /scripts/a.js
The comment above tag is just to show that there may be some processing envolved before injecting the tag.
I tried many approaches using different events in docpad.coffee file (including approach taken from docpad-plugin-livereload plugin) but every time I was facing the same problem - script tag was applied to all pages instead of being applied to a.html only. Here is one of my tries:
renderDocument: (opts) ->
{extension,templateData,file,content} = opts
if extension == 'html' and scripts = file.get('scripts')
if typeof scripts != 'undefined'
scripts.forEach (scriptName) ->
#docpad.getBlock('scripts').add('<!-- custom script tag here -->')
I've also tried render event, populateCollections (which is not documented however I found it in docpad-plugin-livereload plugin) and even extendTemplateData events and no luck so far.
I know there is a method of doing this right inside a layout:
#getBlock('scripts').add(#document.scripts or [])
..which is totally fine and it really works as expected however it doesn't seem to provide enough freedom for me to manipulate the content before it's injected to a page.. And even if it's possible I won't like the idea of having some heavy logic inside layout template, I want it to be in a plugin/docpad.coffee
Hopefully that makes sense
Try templateData.getBlock('scripts').add instead of docpad.getBlock('scripts').add
Related
I need to get the headline and the text separately out of a text content element. The reason is, to give the editor a simple way to add a content for a complicated section in my html theme.
I am new to TYPO3 an we run V11.5.16! I read and watched some tutorials and I got most of my site already working! Contents are dynamic and multilinguale so far.
To get contents from backend, I use Backend Layouts and copy the content from styles.content.get inside my setup.typoscript. I think this is the common way to do it, and as I said, it works. I output them using {contentXY->f:transform.html()} or {contentXY->f:format.raw()}.
For a text content element, I get something like:
<div id="c270" class="frame frame-default frame-type-text frame-layout-0">
<header>
<h2 class="">Headline</h2>
</header>
<p>Some Text</p>
</div>
Is it possible to get only "Headline"? And if so, it hopefully works also for getting out separately "Some Text"
Something like: {contentXY->f:transform.html(filterBy('h2'))}
Thanks in Advance!
EDIT:
According to Peter Krause's answer: I know, there is an extra content element for headers. But I need the text content element, because for the places in the html, I need header AND text. And the editors are technically not savy enough to fill in different content elements. Please don't ask in more detail. ):
You can handle header and body of an CE seperately, but not in a page context.
In page context you get the result from rendering the CEs, which is a string (with HTML).
For each CE there is a rendering information, which nowadays is also FLUID.
Depending on your installation it probably is FSC (ext:fluid_styled_content) or a Bootstrap extension.
This means: there are FLUID templates which can be overriden and modified.
In these templates you have access to each field of a CE separately.
Look for the templates stored in the defined paths (in TSOB) and add your own path for overides:
lib.contentElement {
templateRootPaths {
1 = ...
2 = ...
3 = ...your path...
}
partialRootPaths {
1 = ...
2 = ...
3 = ...your path...
}
layoutRootPaths {
1 = ...
2 = ...
3 = ...your path...
}
}
Thanks for all hints! I think, for my requirement, there is no solution out of the box. So i made a custom CE with Mask and edited the template html. For non-technical editors, it is the best solution in terms of data input. I hope this stands for future upgrades...
I'm trying to optimalize web for speed and wanna ask about Eliminating render-blocking CSS and JS.
by JS 'm only using async attr. - lets say, throwin' it at plugins like flexslider, lightbox.. but should I also use this with the base scripts like ?:
<script src="https://cdnjs.cloudflare.com/.../4.5.3/js/bootstrap.min.js" async></script>
<script src="js/script.js" async></script>
Whenever i add async on some script and test it, that .js script won't just operate - as if not linked. What am I doing wrong? And is this enough ... ?
-by CSS - tring to impove it like this :
<link rel="preload" href="https://use.fontawesome.com/releases/v5.2.0/css/all.css" media="print" as="style" onload="this.onload=null;this.rel='stylesheet'" integrity="sha384-hWVjflwFxL6sNzntih27bfxkr27PmbbK/iSvJ+a4+0owXq79v+lsFkW54bOGbiDQ" crossorigin="anonymous"><noscript><link rel="stylesheet" href="https://use.fontawesome.com/releases/v5.2.0/css/all.css"></noscript>
This cuts down the time of rendering CSS resources, but again when using it on e.g fontawesome - icons are not loaded as if link was't there...So, is this a good way of rendering CSS please ?
JavaScript
When using async you have to ensure that load order does not affect your implementation.
Async basically says "load everything as quickly as you can, I don't care about load order".
If you haven't accounted for this then the easiest fix is to use defer on your JavaScript instead of async. This basically says "load everything after the HTML has loaded but please keep the order as some scripts depend on others".
This will be slightly slower overall but still fix the JavaScript being render blocking.
You should defer all scripts, except any scripts that are essential for above the fold operations (and then you should inline those scripts in a <script> tag in the <header>, obviously keep this to a minimum).
CSS
Render blocking CSS is anything sitting in an external file that relates to content "above the fold".
To understand this fully you need to understand how the browser render things but in essence anything that is visible without scrolling ("above the fold" content) is delayed if your CSS is in an external file as it needs that information to know how to present and lay things out.
What you need to do is find all the styles that apply to your above the fold content and inline them in a <style> tag in the page <header>. Yet again this needs to be kept to a minimum so you may need to make the above the fold CSS custom rather than using bootstrap....including the whole of bootstrap inline would not be good!
Then all other styles can sit in external style sheets.
This way the second the page's HTML is downloaded it has everything it needs to layout the page without waiting for any other requests.
Font Awesome
Ah fonts for icons. I won't go into why that is a bad practice from an accessibility and performance perspective as I have covered that numerous times before.
Instead I will simply say that for any "above the fold" icons you should instead swap them for inline SVGs. This is for the same reason as inlining your CSS, inline SVGs do not need a network request to render so the second the HTML is loaded your page can be displayed.
just a suggestion, have no way of testing atm but try putting 'async' before the source attribute. also, try adding a copied line with the attribute defer instead of async for wider browser support.
Unfortunately there's no details/summary element in the default CKEditor configuration of TYPO3 and I'm looking for a way to add it.
What I've been trying to do:
I searched and found a widget on https://ckeditor.com/cke4/addon/detail , but it's repository on GitHub has been archived and the widget does not work as expected. It requires 'api,widget' and this generates a JavaScript error:
[CKEDITOR.resourceManager.load] Resource name "api" was not found at "/typo3/sysext/rte_ckeditor/Resources/Public/JavaScript/Contrib/plugins/api/plugin.js?t=K24B-4e13cc129f".
When removing this requirement for "api", there's an error regarding line 72
CKEDITOR.api.parser.add(item, el);.
Then I found a similar widget at GitHub , which looks like an older version of the former without requirement for "api".
It already looks quite good, but is still a bit buggy: the HTML structure is changed when saving and the summary is duplicated. When switching to the source code, the HTML structure specified in the template ...
<details><summary>Summary</summary><div class="details-content"></div></details>
… get's partially lost.
I'm not sure if the widgets are buggy or if the editor is limited by the integration into TYPO3 and I was also not able to combine the two in a way that would lead to a working solution.
Update (Jul 22):
I successfully modified the Creating a Simple CKEditor Widget (Part 1) example to create
a widget with the following HTML structure:
<div class="expander">
<p class="expander-title">Title</p>
<div class="expander-content"><p>Content...</p></div>
</div>
With the help of a small JavaScript snippet and some CSS it now behaves almost like a details-summary element, but is not quite as good in terms of SEO and accessibility.
If I replace the elements <div class="expander"> and <p class="expander-title"> with <details> and <summary> in the widget, it unfortunately doesn't work properly anymore and changes the structure when saving. For some reason the RTE treats them differently.
I have already manually added the following to the RTE configuration:
processing:
allowTags:
- details
- summary
allowTagsOutside:
- details
- summary
I use org.asciidoctor.convert plugin for gradle to generate API documentation for my team. I include files:
include::{snippets}/index/curl-request.adoc[]
and want to place it's content into spoiler or anything like that. Is there any way to somehow hide dynamicaly asciidoc elements? I try to use
pass:[<details open>
include::{snippets}/index/curl-request.adoc[]
</details>]
but it is not processed include inside it. Any ideas will be higly appreciated. Without hiding snippets my documentation have almost unlimited scrol :). If no such way with ascii doc, suggestions of other documentation formats, where i can include files content and place it into the spoiler is also appreciated.
As this was so helpful for me - I added here the solution from Robin's comment.
No CSS and Javascript needed!
Here is an example:
+++ <details><summary> +++
Check `reference.conf`:
+++ </summary><div> +++
----
play {
http {
secret.key = asdf
secret.key = ${?SECRET_KEY}
...
}
...
}
----
+++ </div></details> +++
This creates a collapsed element:
..and this expanded image:
Update
As Foad's answer looks more elegant, I tried his proposed solution for a ReDoc Markup with no success.
As Guillaume Grossetie has mentioned here
Using passthrough to include raw HTML is not recommended because now your document is tightly coupled with the output.
and it is deprecated. The new syntax for collapsible/foldable sections is
.some description
[%collapsible]
====
this
is
going
to be
folded
====
This is a late answer but maybe it will help you/others.
The following asciidoc features are the key for implementing dynamic showing/hiding of blocks:
If your output is only HTML then you can embed any HTML in the document using
++++ any HTML contents ++++
This includes
<style> tags containing custom CSS classes
custom HTML tags for show/hide functionality like <input type="checkbox" />
<script> tags containing Javascript code to hide/show some blocks, e.g. by adding event listeners to checkboxes.
As #LightGuard has already mentioned, role attributes are converted to CSS class references. E.g.
[source,role="someCssClass"]
----
...
----
is converted to something like
<div class="someCssClass">
</div>
So you can reference this CSS class from Javascript code and implement show/hide by modifying the display CSS attribute.
For a simple example implementation see https://raw.githubusercontent.com/NorbertSandor/jsinterop.xyz/master/jsinterop-project/jsinterop-website/src/main/asciidoc/index.asciidoc (near the top of the file).
I hava an editor template for, let's say, date:
#model DateTime
#section AdditionalJavaScript2
{
/* some js code */
}
#Html.TextBox("", Model.ToString("d.M.yyyy"), new { #class = "date" })
Now, I would like to put some js code into the HEAD section, but this doesn't work.
Of course, I have a this section in my layout.cshtml:
<head>
...
#RenderSection("AdditionalJavaScript2", required: false)
</head>
It works from the plain view, but not from partial view (editor template).
Why?
And, is there a workaround?
Thanks,
Igor
A partial-view does not use a template, it returns "raw" html to be included in your page (by Javascript). It does not have access to anything but the stream it returns itself.
Think of it like this: You typically call a partial view from Javascript/AJAX to get some new html. You get the return, and replace some DIV-tag. How can the system (FireFox, Chrome, ...) know, that there is some extra section of data that needs to replace something in the HEAD tag.
There are some workarounds:
Don't put the script in the HEAD
Add a parameter switch betweed the html and the script. You need to client-side calls, one to get the html, and one for the script. You include the calls to the partial-view on two locations on your page.
Separate the script and the html using some pre-defined tag like <!-- SEPERATOR -->, and let the calling code split the result, and put it in the correct position.