Skip to content

Conversation

@jeffomatic
Copy link
Contributor

No description provided.

Copy link
Contributor

@dominic dominic left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

👍 from me. I'd be interested in bikeshedding on the branch names. Something like:

release-1.0
release-1.1

1.0-stable
1.1-stable

chain-1.0
chain-1.1

@jeffomatic
Copy link
Contributor Author

Re: naming, I'm down for whatever. Among those choices, x.y-stable read the best to me.

@dominic
Copy link
Contributor

dominic commented Mar 2, 2017

I like that stable emphasizes that main isn't as well.

@kr
Copy link
Contributor

kr commented Mar 2, 2017

On the topic of naming things and "say what you mean", the word "minor" sounds weird to me for our versioning scheme. Our releases like 1.1 or 1.2 are a pretty big deal. But that's beyond the scope of this PR.

@jeffomatic
Copy link
Contributor Author

I just made an update recommending that we use chain-core-server-<major>.<minor>.0 as the branch point for these version branches.

There is a weird edge case if there are changes to other packages (such as the SDK) that precede the chain-core-server release tag and are slated for the next minor version. For example, if main somehow contains a Java SDK commit that is slated for 1.2 and precedes the chain-core-server-1.1.0 release tag. This is technically possible and would mess with our unified version branch, but I doubt it will ever happen.

@jeffomatic
Copy link
Contributor Author

Also, re-constructing the 1.0-stable branch will take some historical spelunking, since we will probably have to backport various SDK and installer updates that landed after the most recent 1.0.x Chain Core release.

@kr
Copy link
Contributor

kr commented Mar 2, 2017

There is a weird edge case …

Yeah we should make it a rule to avoid that situation. Like we should close the tree leading up to a (so-called) minor release 1.n and don't open the tree for n+1 development until everything is clear.

@iampogo iampogo force-pushed the historical-branches branch from 00385c3 to e8fb4e6 Compare March 2, 2017 20:13
@dominic
Copy link
Contributor

dominic commented Mar 2, 2017

we should close the tree leading up to a (so-called) minor release

I've seen other repos branch early into 1.x-development or 1.x-candidate when their root branch is approaching a release to avoid this issue.

@iampogo iampogo merged commit cff5a7d into main Mar 2, 2017
@iampogo iampogo deleted the historical-branches branch March 2, 2017 20:16
@kr
Copy link
Contributor

kr commented Mar 2, 2017

I'm not sure what branching early buys us. We'll just have a tone of commits on the candidate branch and almost none on main during that period. If possible, I'd rather we do the stabilization work leading up to a release directly on main, rather than splitting our attention between the upcoming release and future development.

In other words, it's something I'd hope we would do as a team anyway even aside from all the branching mechanics. It's less confusing if we make sure not to begin future dev work until the release is out the door.

iampogo pushed a commit that referenced this pull request Mar 2, 2017
This commit relaxes the Ruby version requirement to 2.0, and ticks the SDK
version to 1.1.1.

This is a backport of #658 and #670 for the 1.1-stable branch.

Closes #678
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants