Closed mibollma closed 1 year ago
I have the same problem actually. However, if following the suggestions (https://github.com/mathieudutour/github-tag-action/issues/141, https://github.com/mathieudutour/github-tag-action/issues/85, https://github.com/mathieudutour/github-tag-action/issues/26) to set default_bump
to false, the values new_version
and new_tag
will always be undefined even if there is commit (is that because the commit messages are not in semantic format?):
> Run mathieudutour/github-tag-action@v6.1
with:
github_token: ***
default_bump: false
release_branches: .*
dry_run: true
default_prerelease_bump: prerelease
tag_prefix: v
create_annotated_tag: false
fetch_all_tags: false
Previous tag was 401.104.5, previous version was 401.104.5.
Analyzing commit: Test nightly build and release
The commit should not trigger a release
Analyzing commit: Test nightly build and release
The commit should not trigger a release
Analyzing commit: Test nightly build and release
The commit should not trigger a release
Analyzing commit: Test nightly build and release
The commit should not trigger a release
Analyzing commit: Test nightly build and release
The commit should not trigger a release
Analyzing commit: Test nightly build and release
The commit should not trigger a release
Analyzing commit: Test nightly build and release
The commit should not trigger a release
Analysis of 7 commits complete: no release
This is what I used in the yaml:
- name: Get next version
id: next_tag_version
uses: mathieudutour/github-tag-action@v6.1
with:
github_token: ${{ secrets.GITHUB_TOKEN }}
default_bump: false
release_branches: .*
dry_run: true
At last, I switched to https://github.com/anothrNick/github-tag-action as alternative solution...
We wrote our own action to bump the version, which for our use-case was quite simple with a bit of git and regex.
The documentation states
However, to me it seems that a new version is always calculated even if there is no new commit and already a version on the current commit. This is especially a problem when re-executing jobs that failed in a later stage.
The output looks like this: