I've been working on this for several hours and have been trying out a number of suggestions in similar questions here.
I can run my gatsby site develop build just fine, but when I run "gatsby build" I'm running into this error.
Invalid plugin options for "gatsby-source-contentful":
- "accessToken" is required
- "spaceId" is required
not finished open and validate gatsby-configs, load plugins - 1.427s
I have one .env file named ".env" with both of the necessary env variables, and in my gatsby-config file, I have it configured like this...
resolve: 'gatsby-source-contentful',
options: {
spaceId: process.env.CONTENTFUL_SPACE_ID,
accessToken: process.env.CONTENTFUL_ACCESS_TOKEN
}
I've tested out the accessToken and spaceId by manually adding them into my gatsby-config plugins options and ran "gatsby build" and they are working fine. When I change the config back to the env variables though, it's no good and I get the error I mentioned previously.
One thing I tried just recently was to make another .env file and name it ".env.production" and it threw a different error like this...
“error Cannot access Contentful space “*********ab” on environment “master” with access token “***************************************1234”. Make sure to double check them! (value)”
This led me to look into the master environment on contentful but I'm pretty new to it and it's got me confused.
I can find a lot of questions here with the same type of problem but I'm yet to find a solution that works. Any gatsby experts out there who've had this problem?
According to Environment variables docs:
In development, Gatsby will load environment variables from a file
named .env.development. For builds, it will load from .env.production.
So, in your scenario, just duplicate and rename your .env file to .env.development and .env.production. Gatsby will take the needed data depending on the fired command.
Related
I'm working on two React projects.
In both projects, the manifest.json file is set up correctly and The serviceworker is also registered.
In one of the projects, the question of adding to the home screen is not asked.
From the research I've done, I've found that a project that doesn't work properly can't find a file in .license format, but in a project that works properly, it's a file with a .license.txt format.
What makes the difference between the two files when it is taken from the building project? How can this problem be solved?
Photo of the project that works properly and displays the install app prompt
The project photo does not work properly and does not show the install app prompt
error message in console
Finally I found the solution to the problem. If you have ejected the project, you can add a field called extractComments with a false value in the webpack.config.js file in the TerserPlugin section. If you have not ejected the project, you can solve this problem by creating a webpack.config.js file and adding the Terser module with the default settings and adding the extractComments field with a false value. The following links may be helpful in understanding and resolving the issue.
license-to-kill-your-PWA
terser-plugin
The build files generated through create react app have different names(hash code) every time.
I would like to give a custom names for the generated files.
Is there any possibility to do the same?
You can change the output filename by customizing the filename property in webpack config -- refer to https://webpack.js.org/guides/caching/
The default implementation is kept like this because, because every time you build an asset, it generates a new name and browsers won't be able to serve a cached response.
If you change the name to a constant you might need to clear the browser cache manually/ disable cache to see your changes immediately. (I think...Applicable only in prod mode as dev mode makes use of Hot module replacement)
Steps to change file name in CRA.
npm run eject This will unwind the hidden configs from CRA and show some additional config folders
Move to the config folder.
Edit file webpack.config.js (Somewhere around line 172 - 180 you can see filename: section where this is defined)
Following up to my comment, if you absolutely must change Webpack configuration you can also consider libraries such as:
Craco
Rescripts
So I have a file named .env with the following contents
NODE_PATH=./src
NODE_ENV=what
TEST=test
And I am calling that in my index.js in my react app.
require("dotenv").config();
console.log(process.env);
...
shows the following output
NODE_ENV: "development"
PUBLIC_URL: ""
I thought maybe I declared another .env file somewhere else, but thats not the case. I searched my project for the PUBLIC_URL and it's not located anywhere in my project. I don't even know what else to check at this point.
In react code you have to compile the environment variables in at, well, compile time because at run-time there it is only possible to access a fake process.env object. Unless you are using server side rendering.
See also: Passing environment-dependent variables in webpack
If you're using CRA then you'll need to do: REACT_APP_TEST=test and reload the dev server to have it show up as expected in your app.
If you have used create-react-app for bootstrap your project then you have to use environment variables like REACT_APP_NODE_ENV=development.
After adding any new environment variable, you have to restart the development server.
I used Sencha CMD to build an app and want to ensure that I'm only downloading what is required for the app to run. I have read some places that the way to do this is by using the Required [] clause to only include the required classes, however I don't think I understand this clearly. For example I have a page that has an {xtype: 'grid'} within it, but I haven't put Requires : [ 'Ext.grid.Panel' ], yet it still displays properly. In my mind if I haven't included the requires, and CMD is doing what I think it should, this page should fail to load because Ext.grid.Panel wasn't included. The fact that it's working is making me think that CMD is automatically including it for me, but what else is it automatically including and potentially bloating the download? How can I ensure that I'm only downloading what's required for the app to run?
I found the answer to my problem. I was developing my app on the same server that I'm serving it from, and when I was running "sencha app build production", it's putting all of the "build" files into a separate directory. I was assuming the microloader/bootstrap process was then going to look in that directory but I guess that's not how it works. I had to then copy those "build" files to a new directory and include my index.html and bootstrap.js in the new directory and then it appears to work, only loading the required files. It's certainly not clear anywhere that this is what's required. Hopefully this will help anyone else having the same issue.
I am currently looking into React, React-habitat, and Webpack.
My question is does Webpack also bundle the index.html file where I reference the bundle or is this kept separate? Is it possible to exclude this without specifically stating this in the Webpack config file?
this is the current structure am envisioning.
https://imgur.com/a/98or9
I know that all the dependencies found in my entry file for Webpack are bundled. The reason I would like to know this is because I am doing some research on how a CMS can be built around the three topics I mentioned above. I need to know if the index.html file is also bundled or not because I would like to edit the original index file(if it is not bundled) instead of repackaging everything for every change.
Hope someone understands what am looking for.