TSC has been versioned as 0.x since its initial release, despite being a widely-used, production-grade library. Staying at 0.x implies instability and allows semver-breaking changes without a major version bump, which is confusing for users who depend on it.
Proposal: release the next version as 1.0, signalling stability and committing to semver conventions going forward. This means:
- Breaking changes require a major version bump (2.0, etc.)
- Non-breaking additions use minor bumps (1.1, 1.2, ...)
- Bug fixes use patch bumps (1.0.1, ...)
This aligns with the approach used by other Tableau SDKs (e.g. the Migrations SDK).
Open questions:
- Are there any known breaking changes we'd want to ship before locking 1.0?
- Should we do a deprecation pass first (e.g.
no_extract, old auth patterns)?
TSC has been versioned as
0.xsince its initial release, despite being a widely-used, production-grade library. Staying at0.ximplies instability and allows semver-breaking changes without a major version bump, which is confusing for users who depend on it.Proposal: release the next version as
1.0, signalling stability and committing to semver conventions going forward. This means:This aligns with the approach used by other Tableau SDKs (e.g. the Migrations SDK).
Open questions:
no_extract, old auth patterns)?