requirement must use official version in order to publish them on pypi

[sebastienbeau]

@hparfr to discuss We can no publish to pypi if we depend of a fork @PierrickBrun

[hparfr]

Yes but we can't use branch without the fork.

[rvalyi]

two possible workarounds:

1. publish our own forked copier package under a different name. Cons: would force us to maintain it.
2. monkey patch copier. Cons: monkey patching should be used wisely...

I didn't think further what would be the best...

[rvalyi]

@sebastienbeau I had to do this trivial fix to get it to work https://github.com/akretion/docky/pull/154/commits/a714674f92ec61ef047dbbd62dbabf8e57a77bca

[bealdav]

OK for me to not use a fork of copier. I really prefer to know that we use explicitely copier and not encapsulate it

[sebastienbeau]

@hparfr let's merge it ?

[bealdav]

yes maybe no docky init is an option if .env created by copier !? @sebastienbeau

[sebastienbeau]

Copier should not create the ".env" file as this file is not commited. It's not the logic of copier, copier is here to create, then update the versioned file. Moreover you want to generate an env file without updating your project, so copier is not the right tools for doing it.

So we have the following possibility for the .env

• no generation (you do a copy paste from other project, and we need to document it)
• having an example of .env called .env-example in the copier template
• adding a check in docky run and if the file is missing propose to create one with default value (that can be updated manually)

What do you think ?

[bealdav]

+ 1 for .env-sample with instruction inside

[PierrickBrun]

.env-sample seems good to me

It would be great if docky could refuse to start if there is no .env file also

[sebastienbeau]

Docky init removed, let's merge it

[hparfr]

please update the readme also ;-)

[sebastienbeau]

@hparfr readme is ok for you ?