We want to avoid possible scenarios where assets are built wrong and consequently cached wrong.
Test 1. test how assets are generated
Assumption:
Generate in production mode (in test?)
Load the page and check an element has correct style
Load the rev-manifest.json and check it has been injected into the page
Q: what happens if we are not running this in production mode - added tag for production test
Test 2. test assets are cached when no change occurred
Assumption:
We load a page and check assets are loaded for the first time
We load the same page and check assets have been cached
Q: Can we actually check status, and is status 304 or 200? - as though the server returns 200
We can use the page event to track the response end.
When cached this is usually about 0.100 rather then vs 20.
Test 3. test fresh copies of assets are served when they've changed
Assumption:
We'd need to spin up a server, check it, and make a change, is this any different to check 2?
Suggestion: I don't think this is an end2end we could look to add uint tests against our build scripts, but this would probably be more a general thing?
Test 4. performance testing cached vs fresh assets fetching
Assumption:
We need to investigate web prefs metrics in playwright, we can do this in web page speed test, but wouldn't be part of the build
Suggestion: this isn't part of our e2e tests since the numbers will change over time, all we could really say is that call 2 is faster then call 1
Example:
Largest Contentful Paint: .503S vs .435S
Page Weight: 44kb vs 5k
We want to avoid possible scenarios where assets are built wrong and consequently cached wrong.
Test 1. test how assets are generated
Assumption:
Generate in production mode (in test?)
Q: what happens if we are not running this in production mode - added tag for production test
Test 2. test assets are cached when no change occurred
Assumption:
Q: Can we actually check status, and is status 304 or 200? - as though the server returns 200
We can use the page event to track the response end.
When cached this is usually about 0.100 rather then vs 20.
Test 3. test fresh copies of assets are served when they've changed
Assumption:
We'd need to spin up a server, check it, and make a change, is this any different to check 2?
Suggestion: I don't think this is an end2end we could look to add uint tests against our build scripts, but this would probably be more a general thing?
Test 4. performance testing cached vs fresh assets fetching
Assumption:
We need to investigate web prefs metrics in playwright, we can do this in web page speed test, but wouldn't be part of the build
Suggestion: this isn't part of our e2e tests since the numbers will change over time, all we could really say is that call 2 is faster then call 1
Example:
Largest Contentful Paint: .503S vs .435S Page Weight: 44kb vs 5k