progit | project aims to generate html pages

Either merging or rebasing in your above example can result in merge conflicts. It is possible that only one or the other would have conflicts too. To understand what rebasing does, consider the following diagram which shows what Topic would look like after rebasing on Master:

Well, you're getting into plumbing details; even if you have to explain Git to a team of coworkers, I'm surprised by the idea that this level of detail would be needed...

Anyway, the info/refs file would only exist on a remote meant to be accessed by HTTP with a dumb server. You probably won't find it (and don't need it) in your local repo. (The remote in this scenario is probably a bare repo, btw, so info would be at the repo root, as bare repos don't have a work tree and place the files you're used to seeing in .git at the root instead.)

If our remote is in something like github, tfs, etc... then you just don't need to worry about any of this as the server will manage things just fine. I guess if you served the repo as static content from a plain old web server then this would matter, and you'd have to set up the hook.

Most users will never use or see the update-server-info command; as its name suggests, it's for repos on the server side - remotes - to compensate for the lack of a git-aware HTTP server.

The post-receive hook is invoked after receiving a push; so on a dumb server scenario, you set this hook on the remote so that when you push to it, it responds by updating certain information (like the refs file).

The GET command you're looking at is an HTTP command, run when necessary by the git client when you do a fetch.

As far as I know, Git checkout will carry over modified files in your working directory when switching branches. The documentation supports this claim:

git checkout
To prepare for working on , switch to it by updating the index and the files in the working tree, and by pointing HEAD at the branch. Local modifications to the files in the working tree are kept, so that they can be committed to the .

As to why Git might have this behavior, this prevents accidentally wiping out changes in the working directory by changing branches.

I'd recommend learning the basics of git (how a branch works, how to merge, how to pull and push code) from the GUI. Once you know the basics, learning which commands correspond to which GUI actions should be easy.

Purists will tell you that you will always have more control in the terminal. This is true, but no need to throw yourself into the deep end as a beginner. Get functional, then get fancy.

Most anything you can do in the command line, you'd be able to do in the GUI, but sometimes the command line is the only way to do whatever it is you need to do. This is especially true when you get into the more advanced parts of git: rebase, cherry-pick. It's a great skill to have, especially in cases where the GUI is not available, like when you are remote accessing a server.

The general question of GUI vs CLI is explored at length in In the Beginning... Was the Command Line.

You're calling the executable with the cd command as an argument. That's the equivalent of running the following on the command line:

This seems to be a synthetic example, so it could just be a mistake. However, it could also be a side effect of a particular configuration, where local branch markdown-strip has no upstream set.

Remember, the upstream of any branch is a two-part setting that you can set with one command:

/pattern1/,/pattern2/ defines a range of lines between two patterns.

/^}/ only means a line that starts with a }

When I tried the Perl from MinGW, there is no error. The Cygwin based perl doesn't seem to have the problem (or bypass this problem).

Also refer to: Multi-line regular expression issue in Perl on Windows 10.