Clone from GitHub confuses simpleminded users if the repository has Pages
Created originally on Bitbucket by twic (Tom Anderson)
GitHub's 'Pages' feature stores the text of the pages as files in the repository they are attached to, in a branch called gh-pages.
When you clone a repository with hg-git, it updates to the tip. If the last thing that happened in the repository is a change in the pages, then the tip will be on the gh-pages branch. Thus, a hg-git clone of this repository will update to the gh-pages branch.
That is, you can clone a repository which you expect to have code in it, look in the directory, and just find a few HTML pages and random bits of JavaScript.
If you are not intimately familiar with the repository you are cloning, and the way GitHub implements pages, and especially if you have just had a rather filling lunch, you may be quite confused by this.
I have no idea if this is something that hg-git has any influence over, but it would be very useful if cloning a GitHub repository always updated to the master branch, or at least some reasonably plausible true branch of the repository, rather than gh-pages.