Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Conversation

implausible
Copy link
Member

@implausible implausible commented Jan 28, 2019

This PR begins to address the many incoming deprecations from Libgit2. It also handles a lot of the additional features Libgit2 has exposed since 0.27.0

  • Libcurl dependency has been removed
  • IndexNameEntry is new and exposed.
  • IndexReucEntry is new and exposed.
  • Mailmap is new and exposed.
  • Fixed bug that caused segfault due to garbage collection in remote.prototype.getRefspec
  • Added pending feature, signing commits during rebase, into Nodegit's fork of libgit2. See Add sign capability to git_rebase_commit libgit2/libgit2#4913 for more details.

This release also improves some of our capacity to handle code generation for statically allocated arrays as needed for git_index_reuc_entry. Check out the field mode and oid on git_index_reuc_entry here https://libgit2.org/libgit2/#HEAD/type/git_index_reuc_entry.

This PR also changes GitBuf.prototype.set and GitBuf.prototype.grow to sync.

- Added git_index_name_entry
- Added git_index_reuc_entry
- Added git_mailmap
- Added git_path_is_gitfile
certificateCheck used to require passing 1 to ignore certificate failures. Now we need to pass 0 to inform libgit2 that the certificate was deemed valid.
@implausible
Copy link
Member Author

implausible commented Jan 28, 2019

There is something spooky in git_stash_save on OSX only.

@implausible implausible force-pushed the bump/libgit2-preview-v0.28 branch from 2172a65 to c363ca3 Compare January 29, 2019 15:16
@implausible
Copy link
Member Author

implausible commented Jan 29, 2019

Looks like it wasn't a segfault, but test pollution from an a couple async routines not being waited on in the stash test suite. 🤞

@implausible implausible force-pushed the bump/libgit2-preview-v0.28 branch from c363ca3 to f41c963 Compare January 29, 2019 15:43
@implausible
Copy link
Member Author

As a follow up to this:

  1. It would be goodness to have some automation for building the ownership connection between GitRebaseOptions and GitRebase on the return result of rebase init and rebase open.
  2. It would be nice to invalidate nodegit objects after callbacks have been completed. This would prevent segfaults.

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.

2 participants