Version
v24.14.0
Platform
Darwin telsha 25.4.0 Darwin Kernel Version 25.4.0: Thu Mar 19 19:31:09 PDT 2026; root:xnu-12377.101.15~1/RELEASE_ARM64_T8132 arm64
Subsystem
No response
What steps will reproduce the bug?
- Create two files:
.env with KEY=original and .env.test with KEY=override.
- Run the following in the Node.js REPL:
// Load the first file
process.loadEnvFile(".env");
console.log(process.env.KEY); // 'original'
// Attempt to load the second file to override the value
process.loadEnvFile(".env.test");
console.log(process.env.KEY); // Still 'original'
Mitigation:
import { parseEnv } from 'node:util';
import { readFileSync } from 'node:fs';
// This forces the override
const integConfigs = parseEnv(readFileSync("apps/server/.env.integ", "utf8"));
Object.assign(process.env, integConfigs);
How often does it reproduce? Is there a required condition?
It reproduces 100% of the time. The condition is that a variable must already exist in process.env (either from the shell or a previous loadEnvFile call) for the "bug" to be visible.
What is the expected behavior? Why is that the expected behavior?
While the default behavior of not overriding is standard for dotenv, there is currently no programmatic way (via options) to force an override using process.loadEnvFile().
At the very least, subsequent calls to process.loadEnvFile() should support an { override: true } option to allow developers to switch environment contexts (e.g., switching to an integration/test config) without manually parsing files via util.parseEnv.
What do you see instead?
The variable remains locked to the first value it was assigned. Subsequent calls to loadEnvFile for the same key are silently ignored for that specific key.
Additional information
Most modern environment loaders (like the dotenv npm package) provide an override flag. Without this, process.loadEnvFile is significantly less useful for complex test runners or mono-repos where multiple .env files must be composed or swapped during a single process lifetime.
Version
v24.14.0
Platform
Subsystem
No response
What steps will reproduce the bug?
.envwithKEY=originaland.env.testwithKEY=override.Mitigation:
How often does it reproduce? Is there a required condition?
It reproduces 100% of the time. The condition is that a variable must already exist in
process.env(either from the shell or a previousloadEnvFilecall) for the "bug" to be visible.What is the expected behavior? Why is that the expected behavior?
While the default behavior of not overriding is standard for
dotenv, there is currently no programmatic way (via options) to force an override usingprocess.loadEnvFile().At the very least, subsequent calls to
process.loadEnvFile()should support an{ override: true }option to allow developers to switch environment contexts (e.g., switching to an integration/test config) without manually parsing files viautil.parseEnv.What do you see instead?
The variable remains locked to the first value it was assigned. Subsequent calls to
loadEnvFilefor the same key are silently ignored for that specific key.Additional information
Most modern environment loaders (like the
dotenvnpm package) provide anoverrideflag. Without this,process.loadEnvFileis significantly less useful for complex test runners or mono-repos where multiple.envfiles must be composed or swapped during a single process lifetime.