Setting up azure pipeline #107
Conversation
…nto azure_pipeline
| ne(variables['Build.SourceBranch'], 'refs/heads/master')) | ||
| pool: | ||
| vmImage: ubuntu-16.04 | ||
| timeoutInMinutes: 90 |
There was a problem hiding this comment.
Can we parameterize the time? Or is this a hard requirement?
azure-pipelines.yml
Outdated
| echo "dev.pandas.io" > CNAME | ||
| printf "User-agent: *\nDisallow: /" > robots.txt | ||
| git add --all . | ||
| git config user.email "[email protected]" |
azure-pipelines.yml
Outdated
| printf "User-agent: *\nDisallow: /" > robots.txt | ||
| git add --all . | ||
| git config user.email "[email protected]" | ||
| git config user.name "bhavaniravi" |
| # 6. The public key specified in "sshPublicKey" is the pair of the uploaded private key, and needs to be set as a deploy key of the repo where the docs will be pushed (with write access): https://github.com/pandas-dev/pandas-dev.github.io/settings/keys | ||
| - task: InstallSSHKey@0 | ||
| inputs: | ||
| hostName: 'github.com,192.30.252.128 ssh-rsa AAAAB3NzaC1yc2EAAAABIwAAAQEAq2A7hRGmdnm9tUDbO9IDSwBK6TbQa+PXYPCPy6rbTrTtw7PHkccKrpp0yVhp5HdEIcKr6pLlVDBfOLX9QUsyCOV0wzfjIJNlGEYsdlLJizHhbn2mUjvSAHQqZETYP81eFzLQNnPHt4EVVUh7VfDESU84KezmD5QlWpXLmvU31/yMf+Se8xhHTvKSCZIFImWwoG6mbUoWf9nzpIoaSjB+weqqUUmpaaasXVal72J+UX2B+2RPW3RcT0eOzQgqlJL3RKrTJvdsjE3JEAvGq3lGHSZXy28G3skua2SmVi/w4yCE6gbODqnTWlg7+wC604ydGXA8VJiS5ap43JXiUFFAaQ==' |
There was a problem hiding this comment.
Again, can this be an env parameter?
| - task: InstallSSHKey@0 | ||
| inputs: | ||
| hostName: 'github.com,192.30.252.128 ssh-rsa AAAAB3NzaC1yc2EAAAABIwAAAQEAq2A7hRGmdnm9tUDbO9IDSwBK6TbQa+PXYPCPy6rbTrTtw7PHkccKrpp0yVhp5HdEIcKr6pLlVDBfOLX9QUsyCOV0wzfjIJNlGEYsdlLJizHhbn2mUjvSAHQqZETYP81eFzLQNnPHt4EVVUh7VfDESU84KezmD5QlWpXLmvU31/yMf+Se8xhHTvKSCZIFImWwoG6mbUoWf9nzpIoaSjB+weqqUUmpaaasXVal72J+UX2B+2RPW3RcT0eOzQgqlJL3RKrTJvdsjE3JEAvGq3lGHSZXy28G3skua2SmVi/w4yCE6gbODqnTWlg7+wC604ydGXA8VJiS5ap43JXiUFFAaQ==' | ||
| sshPublicKey: 'ssh-rsa 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 [email protected]' |
|
@sara-02 Thanks for the suggestions. I am kind of done with the first iteration. I am wondering why the build is not triggered in this PR because when I tried doing it for #106 it triggered for that branch alone. The hard-coded values are from the pandas repo. I just replicated the method followed there |
|
Also, one more thing, more of a question about how does pandas-dev perform adding commits, and what strategy we follow. Do we retain all the commit history or squash it into a single commit before merging? |
|
@sara-02 When I worked on this issue for pandas they didn't ask me to squash the commits. Not sure if that's the standard |
Hey not an expert here, but in my last org, unless you were a public member of the group, the CI would not trigger. If you want to give it a shot, maybe try making your membership in python-sprint public instead of private. Go to |
|
Great job, looks really good. Couple of comments: The condition in the job is saying that this build will just run for commits to master, and not PRs. I don't think the CI is not running because of this, but the build wouldn't happen. The reason we have this in pandas is that PR builds are not able to access the secret variables for security reasons, so we can't push to the repo since we can't read the token. Unrelated, my preference would be to set up the CI with a hello world build as you did, and then add the build to the docs. Gradually if possible. Obviously you learned a lot about how it works @bhavaniravi, but would be nice if the rest of the people is able to follow step by step. And all this in a single PR is too much I think. About the squashing, GitHub has the option to do the squashing when merging a PR. For me there is no value in the local commits people do, so I've been squashing here too, so in the history of the repo we just have one commit per PR. Happy to have a discussion about this (this PR doesn't seem to be the best place, but an issue can be created). |
|
Forgot about the emails, usernames... All the ones in the azure pipelines config file are mostly useless. When I set up that in pandas I used the project dev list, just to have something nice. but My preference would be to have something hardcoded that is not the actual email of anyone, but it doesn't make a difference. |
1. Setup azure pipeline
2. Wrote a minimalistic
azure-pipeline.ymlby taking inspiration from pandas repo3. used
Makefilefrom #51 to add doc building pipeline