Force publish package version?

[Christopher Hendrix]

Trying to cut a new release of one of my packages, elm-matrix

Error: It looks like you are trying to bump from version 2.0.1 to 3.0.0.
This implies you are making a MAJOR change, but when we compare
the 2.0.1 API to the API you have now it seems that it should
really be a MINOR change (2.1.0).

I know that there are breaking changes, as internally it uses 0-indexing instead of 1-indexing now. This can’t be determined just by API changes.

Is there a way to force publish? Or has anyone else run into this and thinks it’s worth talking about?
The auto-differ/bumper is wonderful, but it still can’t hold a candle to someone who understands the package version implications.

​

[Max Goldstein]

I'm assuming the "internal" change in indexing is visible to the client? Then yes, I agree with you there should be a way to force a more breaking version. To my knowledge there is not.

I ran into this issue when adding a deprecation notice, which by the semvar spec must be a minor release, but the package manager took it as patch. I made the choice to not do a deprecation notice (it was a part of the API that was never meant to be exposed in the first place), meaning that  the package manager actually prevented me from following semvar (in this case).

[Jeff Smits]

This is surprising. I thought it was part of the idea of this semvar enforcement, that you would be able to change the version to something larger than strictly necessary according to analysis. I would call this a bug and report it to elm-package. At most this message should be a warning.

On 31 Jul 2015 06:57, "Max Goldstein" <maxgol...@gmail.com> wrote:

I'm assuming the "internal" change in indexing is visible to the client? Then yes, I agree with you there should be a way to force a more breaking version. To my knowledge there is not.

I ran into this issue when adding a deprecation notice, which by the semvar spec must be a minor release, but the package manager took it as patch. I made the choice to not do a deprecation notice (it was a part of the API that was never meant to be exposed in the first place), meaning that  the package manager actually prevented me from following semvar (in this case).

--
You received this message because you are subscribed to the Google Groups "Elm Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elm-discuss...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Evan Czaplicki]

Yeah, plan to support this case, but haven't had time yet.

Here's a hack you can do. Go to the released version, expose a value (maybe unit) and do a minor bump, add all your changes removing that constant.

It's silly and I apologize for that, but it'll do the job for now. Longer term we'll let people do exactly what you want to do as long as they give an explanation of why it really is a major change.

--
Sent from Gmail Mobile