For me, the biggest issue was lack of syntax highlight on pull requests. I've created this Chrome extension to fix it: https://chrome.google.com/webstore/detail/gh-diff-highlight/...

Contributions[1] are welcomed :)

[1] https://github.com/danielribeiro/github-diff-highlight-exten...

Hey cool, my friend made the same extension only it uses github's server highlighted html instead of a javascript highlighter.

https://github.com/msolomon/github-diff-syntax-highlighter

This is excellent - congratulate your friend!

Honest question: What is this good for? I guess I don't use github (or any other visual vcs) enough.. but I really can't see any point in browsing file structure of the repo in a web browser.

I browse repos on github pretty frequently. Some use cases:

- It's someone else's repo (eg, a gem I'm using), and I just want to see one line real quick, rather than cloning the whole project locally and finding it there.

- It's my repo, but I my local working copy is on a different branch or has uncommitted changes, so I don't want to bother stashing + changing branches just to look up one line

- I'm poking around several repos that I have no intention of using- eg, looking at an interview candidate's public repos

For the "my repo, wrong branch" use case, you have some options:

  git show other-branch:path/to/file

or:

  git grep -e pattern other-branch

git-grep is indescribably awesome. To be fair though, I search github pretty frequently because it means I can search issues also. If issues were git repos themselves it would be a different ballgame.

It's good for when you want to browse a code base written in Java due to the deep folder hierarchies. And the reason somebody might do that is that they might be choosing between several different libraries, and want to evaluate the quality of their code a little before downloading them.

I am not a Java hater, but any Java or Scala repo has a very deep tree. Clicking down and refresh page is not fun. Also consider when you want to open multiple files which are on different levels of the tree. Without the tree ready you would go back and forth between pages or having multiple new tabs opened to open each file.

It seems like you are getting down-voted but I'm very curios about this as well. It has always seemed to me that the tree browser was very inefficient way of browsing a file structure especially if you get into the realm of deep directories and long file names.

A directory tree may not be the most optimal solution, but it's easier to browse than the default github view...

Personally I use it when trying to view past changes and understand what the state of the repo was at that particular moment.

Frankly, right now I just have a hard time remembering what path I was going down with the current process of traversing the file structure.

protip: you can press 't' when you're in the repo for Sublime-like search for files -- which I find more useful than a tree browser

I find that feature useful when I am already familiar with a code base. However, when I am spelunking unfamiliar code bases, a good tree browser would make the effort quicker (e.g. speeding conventions directory hierarchies to get to the meat of the implementation). Since I find myself performing such discovery fairly often, I see this extension being a real time saver.

do you know that you can press T on a repo page and get a nice fuzzy search? i find it much more efficient than clicking through a tree.

Search will never be a substitute for hierarchical organization of information in all cases. Imagine if the taxonomic tree of life was abandoned in favor of a fuzzy search on organism attributes. Sometimes the relationships between conceptual elements, whether files or organism, are more important than a fuzzy key word.

For this reason I detest the Windows 7 start menu for its anti-organization UI, for example.