Skip to content

Latest commit

 

History

History
54 lines (40 loc) · 2.69 KB

MAINTAINERS.md

File metadata and controls

54 lines (40 loc) · 2.69 KB
Raw

Creating a release

Travis CI deploy to PyPI on tags (2 in 1 ☕)

1. Change the version number in the __version__ variable at init.py file inside instapy package [folder] to the new version number (follow SemVer).

E.g. for a bug fix:

__version__ = "0.1.5" # <= from 0.1.4

2. add and commit the change made to init.py regarding version increment and push it..

git add .
git commit -m "new bug fix version: 0.1.5"
git push   # you don't have to push, but it would be better

3. Make a new git tag with the exact version number you have just set in the 1st step,

git tag 0.1.5

4. Push that new tag

git push --tags

5. Ta Da 🎉 Github will make a new release by that tag and send a request to Travis CI that "Hey bro, I've just received a new tag!" and Travis CI will immidiately check,

But if all those above are true, then Travis CI will start building the jobs at all of the three stages available- static analysis, test and then deploy to PyPI automatically with the version you have set in __version__ variable at init.py file included in the freshly tagged release.

Advice

Before incrementing the __version__ at init.py and tagging that same version with git to push tags, make sure your code works really good and the build with the same code has passed before.
I mean do anything you can merge PRs or git push locally and see that build has passed.
And only then, increment the version number and make a new tag with that new version and push that tag.
It will make sure your build will surely pass and deployment will succeed always 🎉

Bad Situation

You push a tag to Github which ends up failing at static analysis or test stage and obviously - not being deployed to PyPI.
At that point, you will end up with a new release at Github and not having that release deployed to PyPI 😥

a-) You can either push new commits to Github which will solve those problems and you build passes at Travis CI so that you are ready to deploy to PyPI.
Then make the distribution files manually and upload that to PyPI by hand.
Which will have your Github and PyPI stay at the same version in parallel.

b-) You can first delete the release at Github (it allows readding releases) and then push the commits which will solve problems and have Travis CI build pass and are ready to deploy.
Then make that release tag again and push it to Github once again.
It will make a new Github release and also, obviously, pass static analysis & test stages and deploy to PyPI successfully.