Command
new, test
Is this a regression?
The previous version in which this bug was not present was
Likely 19.0.x
Description
Note: I maintain Jasmine.
jasmine-core is a plain old script/CommonJS hybrid, not an ES module, but the Angular dev tools load it via <script type="module">. That appears to be the cause of the global this problems that the angular CLI works around by monkey patching the window object to look like it's from an old version of the GJS runtime . It could cause other problems as well.
#29104 appears to have introduced both the bug and the GJS workaround.
jasmine-core 6.x will use globalThis to find the global object, which should do away with the need for the GJS workaround. But it will also emit a warning if it's loaded as an ES module in the browser , since that's an unsupported and un-tested configuration that's already known to have caused problems.
Minimal Reproduction
Run ng new and accept the defaults. (I used @angular/cli 20.3.5.)
In the created project, run npm t.
Click the debug button in Karma.
View source in the resulting tab.
Look for the script tag that loads jasmine.js.
If you want to see the warning:
Run ng new and select the zoneless option. This isn't strictly necessary but it will avoid errors related to Don't expose private APIs on the jasmine namespace jasmine/jasmine#2078 .
Patch package.json as shown below.
Remove package-lock.json and node_modules.
Run npm install.
Run npm t.
diff --git a/package.json b/package.json
index 0474f8f..2fff3a9 100644
--- a/package.json
+++ b/package.json
@@ -37,12 +37,20 @@
"@angular/cli": "^20.3.5",
"@angular/compiler-cli": "^20.3.0",
"@types/jasmine": "~5.1.0",
- "jasmine-core": "~5.9.0",
+ "jasmine-core": "github:jasmine/jasmine#6.0",
"karma": "~6.4.0",
"karma-chrome-launcher": "~3.2.0",
"karma-coverage": "~2.2.0",
"karma-jasmine": "~5.1.0",
"karma-jasmine-html-reporter": "~2.1.0",
"typescript": "~5.9.2"
+ },
+ "overrides": {
+ "karma-jasmine": {
+ "jasmine-core": "6.0.0-alpha.0"
+ },
+ "karma-jasmine-html-reporter": {
+ "jasmine-core": "6.0.0-alpha.0"
+ }
}
- }
\ No newline at end of file
+ }
Exception or Error
Your Environment
_ _ ____ _ ___
/ \ _ __ __ _ _ _| | __ _ _ __ / ___| | |_ _|
/ △ \ | '_ \ / _` | | | | |/ _` | '__| | | | | | |
/ ___ \| | | | (_| | |_| | | (_| | | | |___| |___ | |
/_/ \_\_| |_|\__, |\__,_|_|\__,_|_| \____|_____|___|
|___/
Angular CLI: 20.3.5
Node: 20.19.0
Package Manager: npm 10.8.2
OS: darwin arm64
Angular: 20.3.4
... common, compiler, compiler-cli, core, forms
... platform-browser, router
Package Version
------------------------------------
@angular-devkit/architect 0.2003.5
@angular-devkit/core 20.3.5
@angular-devkit/schematics 20.3.5
@angular/build 20.3.5
@angular/cli 20.3.5
@schematics/angular 20.3.5
rxjs 7.8.2
typescript 5.9.3
zone.js 0.15.1
Anything else relevant?
No response
Command
new, test
Is this a regression?
The previous version in which this bug was not present was
Likely 19.0.x
Description
Note: I maintain Jasmine.
jasmine-core is a plain old script/CommonJS hybrid, not an ES module, but the Angular dev tools load it via
<script type="module">. That appears to be the cause of the globalthisproblems that the angular CLI works around by monkey patching the window object to look like it's from an old version of the GJS runtime. It could cause other problems as well.#29104 appears to have introduced both the bug and the GJS workaround.
jasmine-core 6.x will use
globalThisto find the global object, which should do away with the need for the GJS workaround. But it will also emit a warning if it's loaded as an ES module in the browser, since that's an unsupported and un-tested configuration that's already known to have caused problems.Minimal Reproduction
ng newand accept the defaults. (I used @angular/cli 20.3.5.)npm t.If you want to see the warning:
ng newand select the zoneless option. This isn't strictly necessary but it will avoid errors related to Don't expose private APIs on thejasminenamespace jasmine/jasmine#2078.npm install.npm t.Exception or Error
Your Environment
Anything else relevant?
No response