svu-next wrong version bump

[Mzack9999]

The action incorrectly bumps Y in X.Y.Z

See https://github.com/projectdiscovery/utils/releases

v0.4.23 -> v0.5.0 -> v0.6.0

[dwisiswant0]

From what I can tell, svu is behaving as expected. When it detects a feat (or similar - see this table), it bumps the minor (Y) version, which is why the progression goes like that. Patch bumps (Z) only occur with fixes or other non-feature changes.

• v0.4.23 -> v0.5.0 because: feat(structs): add FilterStructToMap utility function by @hanXen in https://github.com/projectdiscovery/utils/pull/674
• v0.5.0 -> v0.6.0 because: feat: adds html package by @dwisiswant0 in https://github.com/projectdiscovery/utils/pull/681

So from what I can see, the versioning here looks correct and in line with SemVer spec, where new features increment the minor version & patches increment only the patch number. Could you clarify what part seems wrong to you, @Mzack9999?

[dwisiswant0]

Before projectdiscovery/utils#662, utils repo was using the mathieudutour/github-tag-action action to handle version bumps. That action works in a very similar way to what we’re using now: it looks at commit messages and applies bumping rules based on what it finds (see their table).

The key difference with that action is how it behaves when no matching commit message is detected. If there’s no feat, fix, or other keyword that maps directly to an X/Y/Z bump, the action doesn’t just stop, it falls back to a default_bump. By default, that fallback is set to patch, so in those cases utils repo will always get at least a patch increment.

What that means in practice is, regardless of whether we use the projectdiscovery/actions/svu-next action or the older mathieudutour/github-tag-action, the resulting version progression would be effectively the same. Both actions are driven by commit messages, and both follow SemVer rules when deciding how to bump the version. The only nuance is the default fallback, which just ensures there’s always a bump even if no keyword is present.

TLDR: The current action is not introducing unexpected behavior, it’s consistent with the previous approach, and the version progression you’re seeing (v0.4.23 -> v0.5.0 -> v0.6.0) would have happened either way.