Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly
Appearance settings

Support namespaced references again#4154

Merged
ethomson merged 2 commits intomasterlibgit2/libgit2:masterfrom
ethomson/namespaceslibgit2/libgit2:ethomson/namespacesCopy head branch name to clipboard
Mar 6, 2017
Merged

Support namespaced references again#4154
ethomson merged 2 commits intomasterlibgit2/libgit2:masterfrom
ethomson/namespaceslibgit2/libgit2:ethomson/namespacesCopy head branch name to clipboard

Conversation

@ethomson
Copy link
Member

@ethomson ethomson commented Mar 6, 2017

Support git_repository_set_namespace again (the equivalent of GIT_NAMESPACE); this was subtly broken with worktree support.

This adds a simple namespace test (equivalent to rugged's) and simplifies the namespace handling by simply appending the namespace (if it exists) to the gitdir and commondir.

/cc @carlosmn

Edward Thomson added 2 commits March 6, 2017 14:11
refs::namespace: add namespace tests
d24ae06 
These simple tests only ensure that we enforce the existence of a
namespace; these mirror the rugged tests, they are not exhaustive.
refdb_fs: honor the namespace
8358056
@ethomson
Copy link
Member Author

ethomson commented Mar 6, 2017

So there are two things here that probably bear discussion:

  1. The interaction between namespaces and worktrees. If you set a namespace, this will now look for namespaces in both. It's not exactly clear that this is what git itself would do, because git's namespaces are very narrowly applied to "server" functionality. I don't know that there are commands that obey namespaces and worktrees both. So I think our implementation is okay, but if there exists a command in the future (or in the present, and I just don't know it) that obeys both then we may need to rethink this.

  2. You can create a refdb that is not backed by a filesystem path. (See refs::iterator::empty, for example.) We handle this (and have since the inception of the refdb) by simply creating paths like refs/heads/master (instead of prefixing them with the gitdir). This is obviously pretty wrong. I haven't fixed this here, since it's a) a longstanding bug not a regression and b) I want to unblock Update libgit2 to b31b236... rugged#687. But I will open a PR to deal with this when I have a few spare minutes.

carlosmn
carlosmn approved these changes Mar 6, 2017
View reviewed changes
Copy link
Member

@carlosmn carlosmn left a comment

Choose a reason for hiding this comment

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

Namespaces are such an underused feature that I'm happy with just proving the minimal support works. We can worry about their interaction with worktrees if someone ever actually cares.

@ethomson ethomson merged commit b31b236 into master Mar 6, 2017
@johnhaley81 johnhaley81 mentioned this pull request Apr 20, 2017
Merged
@ethomson ethomson deleted the ethomson/namespaces branch January 9, 2019 10:18
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@carlosmn carlosmn carlosmn approved these changes

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@ethomson @carlosmn
Morty Proxy This is a proxified and sanitized view of the page, visit original site.