The difference between patch queues and git branches is that git branches don't have any temporal context other than the base patch in git branches. There's no ordering or connection between git checkout -b b1 and git checkout -b b2 off a single patch.
This is not the case in hg, although I can't claim to fully understand it. Patch queues are really quite different from git's local cheap branches, and I think that trying to counter them with the pq's is a tough row to hoe.
Provided you're using qguards, mq allows you to ditch temporal connections between patches as well. An early version of this article left that out specifically because I modified the default behavior of qnew on my system to create guards so that I can use them that way; I've updated the article to use qguards explicitly to achieve git-like temporal independence.
EDIT: And today I learn that Hacker News doesn't support full Markdown.
The difference between patch queues and git branches is that git branches don't have any temporal context other than the base patch in git branches. There's no ordering or connection between git checkout -b b1 and git checkout -b b2 off a single patch.
This is not the case in hg, although I can't claim to fully understand it. Patch queues are really quite different from git's local cheap branches, and I think that trying to counter them with the pq's is a tough row to hoe.