When I build my VueJs application, it automatically imports the app.js and chunk-vendor.js files with the preload attribute. This is great as it speeds up the page load time of my application.
I've looked at #vue/preload-webpack-plugin and I can see that I can preload specific webchunks or assets. This has the effect of preloading those files on all routes.
The thing I would really like to do is preload webchunks based on the initial route that is loaded (the first route the user visits).
Lets say I have two routes; home and accounts. Both of these routes are lazy loaded. When a user opens /home as their first page, I would like to preload the js and css webchunks related to the homepage. If the user initially opens the /accounts page, I would like to preload the webchunks related to the accounts page.
Its not possible to use wildcards in preload statements, so I know I can't do this statically.
Any ideas of how this could work? Has anyone heard of such a project being suggested elsewhere?
EDIT: Something I tried as an experiment was injecting preload headers into my index.html file using the beforeRouteEnter method. Whilst I could see the preload header in my DOM, I found that browsers did not observe the header in time, so the image I was experimenting with was not pre-loaded. In any case, this wouldn't have worked for a dynamically named file, but useful to know.
With SSR it is possible and framework like Nuxt does it automatically, because it builds separate html file for each route. So this html can be "tailored" for this specific route and include/preload all the code route needs...
Without SSR it much harder. #vue/preload-webpack-plugin works by injecting preload links into the index.html at build time and since there is only one index.html for a whole app, you can't make it route dependent (with this plugin). So what Vue CLI does is prefetching all the async chunks by default (clearly preferring speed over bandwidth usage)
I can imagine a solution in the form of Webpack plugin, which replaces preload-webpack-plugin and instead of generating preload/prefetch links at build time just generates some inline script with the map of "route name => chunk name" (some well defined naming convention would be needed) that would inject the links dynamically to the DOM base on the current URL. But even with my "googling skills" I wasn't able to find anything like that...
Related
I am building a react blog app, I am using a functional component where I will be using some internal and external links both. For the internal links I am using Link from react-router-domwhich is working fine but for the external link I am not able to decide will an Link from react-router-dom works, which directs to the path of the external URL or an <a> tag should be fine.
The purpose of using react-router-dom is to navigate to application routes by making changes in the DOM and not reloading the whole page. This scenario is applicable to internal links.
When coming towards external links. It is something that is not the part of our application. We cannot render it our application context. So, a solution to that is using an a tag for external links.
Link is basically a wrapper of an tag with a lot of upside functionalities like,
A can know when the route it links to is active and
automatically apply an activeClassName and/or activeStyle when given
either prop.
The will be active if the current route is either the linked
route or any descendant of the linked route.
To have the link be active only on the exact linked route, use
instead or set the onlyActiveOnIndex prop.
Read the rest at https://knowbody.github.io/react-router-docs/api/Link.html
You can use the anchor tag if you plain something plain. use Link for ease of use.
I've seen in some sample code that a route template ("{id:int}") on top of razor page causes the links to that page to use another pattern:
https://localhost/Movies/Edit/6
instead of
https://localhost/Movies/Details?id=6
My question is how asp.net manages to change all the links to that pattern, does it know about that page before rendering it?
Does it collaborate with other pages when processing a page?
When the application first starts, a collection of attribute routes are built. The routes are built for any Razor file with an #page directive in the root Pages folder, and for any other routes that have been defined via PageRouteConventions.
When you use the Url helper to generate links, or the anchor tag helper (which uses the Url helper behind the scenes), the link that gets generated is based on the attribute route that was built for the page that you pass to the helper.
In attribute routing, route parameters are added as segments in the URL, which is why the values are not appended as query string values. If you prefer query strings, don't declare route values as part of the #page directive.
Run the dotnet publish -c Release command and take a look inside the bin/Release folder.
You will not find your .cshtml files with html in them. What happaned where did all the html go? And how does this relate to the question?
You gotta remember that cshtml will endup being your regular ol' c# and all that fancy razor templating syntax end's up being c#. This process has many names and transpilation is one of them performed by transpilers.
Okey so now that we can safely assume that when you have a Index.cshtml file it will get populated in to some sort of an object, let's call it RazorPage.cs this will just store all the configuration for this page. Now let's say this index page is living in a folder called Home now we can have a dictionary Dictionary<string, RazorPage> and let's say that the key will be "/Home/Index". Following along based on transpiled #page "{id:int}" syntax, it might generate a template string for the route and store that in the RazorPage in a RouteTemplate parameter.
So when you use asp-page tag helper it will find the correct RazorPage and it can know the template for the url, populating it with the values you provided.
I haven't seen the actual implementation this is just my guess.
My question is how asp.net manages to change all the links to that pattern, does it know about that page before rendering it?
Yes it knows everything about the page at run time. Most likely the services.AddMvc() service takes care of loading in all the razor pages / views / controllers, at startup.
Does it collaborate with other pages when processing a page?
Highly likely no, unless you mean components/layouts/partials. It will however struggle to resolve a page if you have identical route for 2 pages.
I'm using Next.js with Polaris (from following their guide here). Using Link component imported from either Polaris or Next.js is not updating the url in the embedded app.
While the url path doesn't change, the view does change correctly (i.e. the component for the new path does render). Same result with breadcrumbs and url prop for ResourceList.
Is this expected behavior?
Ended up figuring this out after many variations of trying to make links work correctly.
Answer: no, that's not expected behavior and links should change url path for Shopify embedded apps.
For Polaris, in order to make links change the URL you will need to install this library (#shopify/react-shopify-app-route-propagator). There is enough instructions on the page to figure out how to install it.
Important note on library usage: AppProvider had to be in the parent component for this.context.polaris.appBridge to work correctly. All the logic for context needed to be in a child component. This issue might've been unique to just my case, but maybe not.
How do I navigate to another page within my Elm application?
Specifically, I am trying to navigate from the Home page to the Contributor page.
The Contributor page is located under my Domain folder.
The file structure is below:
- Home.elm
- Domain
- Contributor.elm
I thought I could do something like this:
a [href "Domain/Contributor.elm"] [text "click me!"]
or this:
a [href "Domain/Contributor.html"] [text "click me!"]
However, neither of these paths work.
Note:
The app that I'm working on is NOT a SPA.
You are using elm-live, which is a development server. It targets a single Elm source file as its entry point, so unless your Elm code is built as a single page application, you won't be able to do any navigation to another file (though there is nothing wrong with hard-coding href links that link elsewhere).
elm-live is also only for development. You wouldn't want to run it on a production server.
If you are trying to avoid a SPA and would rather have each Elm file represent the complete functionality for a single page, perhaps you could go with the default functionality of elm make, which generates an HTML file that contains inline javascript compiled from Elm code. This is, in essence, what drives the elm-lang.org website. If you look at the source code, you'll see the html generated by the default elm make command, compiled against each Elm file "page" of the application.
On the other hand, if you are trying to build a SPA, #Bill's answer is a good starting point.
I don't believe you can do the sort of navigation you are trying to do within an Elm app without building a SPA. You are attempting to use the HTML href attribute to navigate. That attribute needs to be a real URL. Without using something like the Elm navigation package, you wont's have support for multiple routes.
Simple navigation in Elm is fairly straightforward. I wrote a blog post on this subject.
Also, here is the github repo that demonstrates the work in this post.
I have read this in many places "You will eventually want to create your own runtime directory and keep it in your own source repository...". Can anyone tell me how to do that? What if I don't want to lose some of the OOTB components?
Currently I am just planning to have a separate webapp for custom developed components. Let's say, I want to have "ootb" mount point for the OOTB components and blank "" mount point for custom developed components. How should I do that? This is what I have tried without success:
<webapp-list>
<webapp name="webroot" http-port="8080" https-enabled="false">
<root-screen host=".*/ootb" location="component://webroot/screen/webroot.xml"/>
</webapp>
<webapp name="customroot" http-port="8080" https-enabled="false">
<root-screen host=".*" location="component://customroot/screen/customroot.xml"/>
</webapp>
</webapp-list>
If this does not work then one other solution that I can think of is to just have the "customroot" entry, and add the "webroot" as SubScreenItem in it. The "customroot" screen will just be blank, and my custom decorator will be present in the "customapps" screen which will be a counter part of the "apps" screen. And all my screens will use the "customapps" screen.
Although I haven't tried what I wrote above, but that somehow feels like a hack. I believe there should be some better way to do this.
And yes, I have read the article, I want to use localhost and there should be some way to do it with localhost too.
As explained in the other StackOverflow question you linked to (on the word "article") the webapp element used at runtime is selected based on the "moqui-name" context-param from the web.xml file for the webapp (in or out of a WAR file). Unless you are deploying multiple WAR files or other forms of webapps this is not useful.
What you are describing would be handled by adding subscreens in the screen hierarchy at the desired points. The general idea with the screen hierarchy in Moqui is that you can have root screens of "applications" mounted through various means (see the annotations on the subscreens element or the Making Apps with Moqui book for details on the 3 ways of doing this). Part of the point of this is to AVOID multiple webapps mounted in the servlet container because that makes things more complicated, including: handling authc and sessions, configuration and deployment, and so on.
Generally for an application in a component you'll want to use a database record to add a subscreen to an existing screen in the hierarchy, mainly from the "webroot" component. Here is an example of that from the example app in Moqui (this adds an "example" path elements under the "apps" path element, where the apps.xml screen is mounted under the root screen, putting it at /apps/example):
<moqui.screen.SubscreensItem screenLocation="component://webroot/screen/webroot/apps.xml"
subscreenName="example" userGroupId="ALL_USERS" menuTitle="Example" menuIndex="8" menuInclude="Y"
subscreenLocation="component://example/screen/ExampleApp.xml"/>
Here is an example from PopCommerce to mount the root screen of the application under the root screen instead of the "apps" screen (i.e. making it located at /popc instead of /apps/popc; note that this means the decoration in the apps.xml screen will not be used because it's not in the render path):
<moqui.screen.SubscreensItem screenLocation="component://webroot/screen/webroot.xml"
subscreenName="popc" userGroupId="ALL_USERS" menuTitle="POP Commerce" menuIndex="9" menuInclude="N"
subscreenLocation="component://PopCommerce/screen/PopCommerceRoot.xml"/>
I think I might have asked a confusing question, but thanks for your time David. If I try to rephrase my question, it would be: "How to have a decorator screen which will not use any HTML from the webroot or apps screens?"
I think I found the answer. I just added my customroot screen as SubScreenItem under webroot screen, and mentioned the attribute standalone="true" in it. Now my URL: localhost:8080/customroot/foo does not use anything mentioned in webroot or apps screens.
Just that, now if I want to have all my components to be at root level in URL like: localhost:8080/foo
I think the only way to do that would be to shift the OOTB components to some other URL like: localhost:8080/ootb/apps/AppList
To do that I will have to add webroot as SubScreenItem of the customroot screen, and replace the webapp entry of webroot with that of customroot.
Damn, I tried so hard and it still is confusing.