Reputation: 2146
In one of our branches we're doing a major refactor to the codebase, which involves a lot of files moving to new locations, renaming and editing/reordering a lot of files.
When I look at the final diff overview in a pull-request I see a lot of files added or deleted, even though they were renamed (in the relevant diff-commit they're recognized as renamed, but not in the total overview).
I tried changing the rename-threshold from 50 to 25, and it helped in part of the times (some files are renamed and more than 75% different).
When all the files seems to be added and deleted instead of renamed it's much harder doing code review.
If I change the similarity threshold to 1% - how it will affect the diff process? It might mistake about files and think they are renamed / copied even though they weren't?
How are you handling refactoring like that?
How do you keep track of a lot of files that were changed and renamed in a pull-request?
Thanks!
Command line example. At the right side the similarity threshold is reduced to 25%
P.s we use Bitbucket, so that's also where we're looking at pr's and doing code review. But basically Bitbucket behave the same as the git cli in that part.
Upvotes: 2
Views: 405
Reputation: 1806
I use Beyond Compare 4 for review. It provides a satisfying diff of all files changed/deleted/renamed/added etc. After installing it change your .gitconfig to look something like this:
[diff]
tool = bc4
[difftool "bc4"]
path = "C:/Program Files (x86)/Beyond Compare 4/BCompare.exe"
[merge]
tool = bc4
[mergetool "bc4"]
keepTemporaries = false
trustExitCode = true
keepBackup = false
path = "C:/Program Files (x86)/Beyond Compare 4/BCompare.exe"
Then simply run: git difftool --dir-diff master mybranch
Upvotes: 1