Hi Ville,

Sorry for the late reply because we are on vacation lately.

I appreciate your email and understand your concern about users potentially
encountering unexpected breaking changes in a minor release. However, given
that major releases occur over several years, it’s not always feasible to
strictly adhere to semantic versioning and delay breaking changes until a
major release. Therefore, our approach is to avoid introducing breaking
changes in patch releases and to clearly document any breaking changes in
the changelog when they do occur.

Thanks

*Ovilia*


On Thu, Sep 26, 2024 at 6:57 AM Ville Brofeldt <ville.v.brofe...@gmail.com>
wrote:

> Hello all,
>
> On Apache Superset, we recently encountered an issue when upgrading
> ECharts from 5.4.1 to 5.5.1. A breaking change introduced in #19513 <
> https://github.com/apache/echarts/pull/19513> caused a regression, which
> we identified after reviewing the changelog. While this wasn’t a major
> issue to resolve, we were under the impression that ECharts followed
> semantic versioning, where breaking changes would have resulted in a major
> version bump (in this case 6.0.0).
>
> I understand that it might be preferable to base major releases on the
> introduction of significant new features, but it seems that many users or
> organizations could encounter unexpected issues during upgrades, as they
> may assume a minor version bump won’t introduce breaking changes.
>
> To help clarify expectations, a few potential alternatives come to mind:
>
> Keep the current versioning scheme, but add documentation to clarify how
> major, minor, and patch releases are structured.
> Adopt Semantic Versioning, and always bump the major version when a
> breaking change is introduced.
> Adopt Semantic Versioning, but defer the introduction of breaking changes
> to major releases.
> Please note, this isn't a complaint, but rather a suggestion to make the
> versioning logic clearer so that the community can better understand what
> to expect during updates. I’d love to hear what others think about this.
>
> Best regards,
> Ville

Reply via email to