Install the specified python version found in pyproject.toml

[paxcodes]

Since poetry currently does not export runtime.txt Heroku installs a different Python version than what is specified in pyproject.toml.

[zyv]

@paxcodes that’s correct, thank you for the heads up. We have historically committed runtime.txt into our repos, that’s why we are not having this problem.

However, now that we have written a buildpack, we could just export runtime.txt in the buildpack and agree internally, that we are going to specify minors in pyproject.toml.

@marns93 , @mm-matthias what do you think?

[mm-matthias]

@zyv There are several options:

1. Always write runtime.txt in the buildpack, possibly overwriting an existing runtime.txt. Requires pyproject.toml to contain the minor version number.
2. Always write runtime.txt in the buildpack. Raise an error if there is already an existing runtime.txt. Requires pyproject.toml to contain the minor version number.
3. If there is an existing runtime.txt, do not overwrite it in the generator. Optionally check if the version number in pyproject.toml matches the version in runtime.txt.

For our own purposes I'd go with option 2) as this makes sure the heroku and pyproject.toml versions of python always match.

What are your preferences @paxcodes ?

[paxcodes]

I prefer #2 for the same reason.

…

On Tue, Mar 3, 2020 at 5:38 AM mm-matthias ***@***.***> wrote: @zyv <https://github.com/zyv> There are several options: 1. Always write runtime.txt in the buildpack, possibly overwriting an existing runtime.txt. Requires pyproject.toml to contain the minor version number. 2. Always write runtime.txt in the buildpack. Raise an error if there is already an existing runtime.txt. Requires pyproject.toml to contain the minor version number. 3. If there is an existing runtime.txt, do not overwrite it in the generator. Optionally check if the version number in pyproject.toml matches the version in runtime.txt. For our own purposes I'd go with option 2) as this makes sure the heroku and pyproject.toml versions of python always match. What are your preferences @paxcodes <https://github.com/paxcodes> ? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2?email_source=notifications&email_token=ADIDWNUJPT3GWMY3MHHZSTTRFUB6NA5CNFSM4K772K62YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOENTQL6Q#issuecomment-593954298>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADIDWNUAN7B6DBL6MZIUV2LRFUB6NANCNFSM4K772K6Q> .

-- Rosalyn Margret “Pax” Williams

[zyv]

Let's go for raising an exception, otherwise we might end up with discrepancies we'll find out about only too late @mm-matthias & @marns93.