Well, I attempted to rewrite @. But somehow the empty changeset got published before that push completed. So it is forever part of history now :/

It looks like hg phabread | hg import - produced an empty changeset, which I accidentally pushed. I have since rewritten @ on hg-committed to remove the empty changeset. This differential revision should be reopened (which I cannot do since I don't own it).

This patch currently fails to apply.

merge.py has been way too long for years and I welcome splitting it.

I'm going to queue this for stable, as it is a Python 3 compatibility issue affecting a distinguished Australian end-user :p

In D8478#126778, @martinvonz wrote:

Why do we even give the choice between 3.7 and 3.8? Just because it was easy to do so anyway? Since the Python runtime is embedded, there's very little reason to try to match a system Python or something, right? To be clear, I don't have an issue with allowing it; I was just surprised by it and wanted to make sure I understand it correctly.

This patch is buggy. I'll upload a new version hopefully in the next few hours.

In D8477#126516, @durin42 wrote:

Clarifying: we can still build hg-with-2.x as a traditional install, and can also do PyOxidizer 3.x? or do we lose 2.x support when we land this series?

I do not understand what the intended side-effect of running this command is supposed to be.