I have a .env file with

# .env
...
REDIS_PORT=12345
PORT=6789
...

but when i run gen-env-types, I get

// env.d.ts
declare namespace NodeJS {
  interface ProcessEnv {
    ...
    REDIS_PORT: string;
    // no PORT
    ...
  }
}

and .env.example

# .env.example
...
REDIS_PORT=
# no PORT=
...

keep-comments stopped working in this version.

.env.example is overwritten by empty values.

return statement was removed while it was necessary. 😕

I know this can be arguable but if there was an option on whether or not comment should be kept would be nice.

My use case is following:

In my projects, I already have env.example file as a placeholder for how the actual env file is supposed to look. Which contains comments as well to give more clarity on what it is about. When I run the type generator it basically overwrites itself. However, blank lines and comments are removed 😑

For example:

Command

npx gen-env-types .env.example -o src/env.d.ts -e .

env.example

# Application port
PORT=8110

# Keycloak configuration
KC_URL="http://127.0.0.1:8200/auth"
KC_REALM="master"
KC_CLIENT_ID="sse-local"
KC_CLIENT_SECRET=

# RabbitMQ configuration
# Fully Qualified URL - when fully qualified URL is provided other options get ignored
RABBITMQ_URL="amqp://127.0.0.1:5672"
# Other options
RABBITMQ_HOST="127.0.0.1"
RABBITMQ_VHOST="/"
RABBITMQ_PORT=5672
RABBITMQ_USER=
RABBITMQ_PW=

Result

PORT=8110
KC_URL="http://127.0.0.1:8200/auth"
KC_REALM="master"
KC_CLIENT_ID="sse-local"
KC_CLIENT_SECRET=
RABBITMQ_URL="amqp://127.0.0.1:5672"
RABBITMQ_HOST="127.0.0.1"
RABBITMQ_VHOST="/"
RABBITMQ_PORT=5672
RABBITMQ_USER=
RABBITMQ_PW=

Possible flag could be
--keep-comments, -k
Defaulted to false

npx gen-env-types .env.example -o src/env.d.ts -k -e .

Currently I have some variables in my local dev env that I do not use in prod. Currently I manually change after running gen-env-types, If there was a way to set whether or not the requirement that a variable exist on a variable by variable basis, thatd be pretty sweet.

Please either add a flag to omit generating the .env.example file, or better yet read the current example file and don't change variables that already exist.

For example in my .env.example file I have

DB_TYPE=postgres

but after running gen-env, it contains

DB_TYPE=

and so is no longer an accurate example.

And while this doesn't affect me personally, dotenv and dotenv-save allow you to use a custom .env filename, so you may want to make this configurable in this tool as well (unless -e already does this?)

Hi,
Thanks for this usefull package 👌
I think if you just skip lines that starts with # gonna do the trick
Maybe I'll work on a PR
Peace

It would be great if the tool ignored blank lines.

If you have a blank line in your .env, or more commonly just end the file with a newline, then the generated .example file has an extra = line.

.env

DB_TYPE=postgres
DB_USER=myapp

generated .example file

DB_TYPE=
DB_USER=
=

notice the last line with just a =