It would be great if SmartGit could be configured so it doesn’t allow to commit until I fix all of the spelling problems in the commit message.

This way, if I have a typo, it will make sure I noticed it. This, of course, should be optional.

Rebase ("Rebase HEAD to", "Rebase Interactive From") fails because of not stashed changes. It would be great to autostash/unstash files as for Pull-rebase feature.

More info: http://www.syntevo.com/doc/display/SU/Change+Sets

I know there's not a Git feature exactly like SVN's "Changelist", so perhaps that is why SmartGit doesn't have this feature. However, IntelliJ provides its own changelist functionality for Git projects, so I'm hoping there's a way to do this within SmartGit.

I often have several tickets/tasks pending at the same time, so being able to group change files by some label (e.g., ticket number) makes managing commits much easier.

As a user I wish I could see the guessed charset encoding of each file in the File Compare tool of Smartgit. And eventually in the Conflict Solver tool of  Smartgit.

There is nothing more frustrating than having someone on your team screw up line endings. While obviously not a long-term solution to the problem, it would be nice if SmartGit could recognize a line ending error message from Git and provide an option to automatically fix the files in question.

Currently the submodule configuration disallows adding a submodule in a non-existent directory. If I want to add a new submodule in externals/foo_project, for example, it doesn't let me do that in one step.

I often find confusing to open repo log when I have a file selected. I end up opening file log, then I realize this, close it, unselect the file and reopen the log.

I think that two separate buttons would be good. The file one should be grayed out when no file is selected.

Our team would very much like the ability to customize the format of a cherrypick commit message, by specifying a format string in the smartgit.properties file. We are imagining a format like:

-Dsmartgit.cherryPick.customCommitMessage="(CP %2 from %4) %3 (Cherry picked from commit %1)"

%1 = full SHA
%2 = short SHA
%3 = original commit message text

%4 = branch original commit was cherrypicked from

That would really help us understand our development flow better when looking at complicated branch hierarchies in the log window, as we could make it immediately clear when a commit is a cherrypick, and where it came from.

It also allows us to add information we need to the commit message to enhance our build and code review automation.

Because opening large repositories is quite slow (~5-10s), I keep all my frequently used repositories open. But unfortunately all these repositories are closed every time I want to open one more repository. Selecting and opening all the already opened repositories again just to keep them open is quite annoying. And also when cloning or adding a new repository, it will be opened automatically while all other repositories are closed.

In my opinion, repositories shouldn't be closed automatically but only when the user triggers it manually.