Am I not understanding something here? I've not committed these files. They shouldn't exist in a change-set anywhere. While I do appreciate Mercurial's efforts to preserve any work I've done that I might want to keep why isn't the sensible design choice on Mercurial's part:
a) Ignore them as per the general process where they effectively are if you remember to pull/update before committing changes in other files.
or
b) Ask user if s/he'd like them overwritten / if no, give directions on how to merge with the proper args that tell Mercurial to ignore files with uncommitted changes in them.
How did we get to:
c) Refuse to do anything forcing user to Google for some hack-ish solution to work around a popular version control system's complete inability to take into account that sometimes we forget to pull/update before commits and most of us (unfortunately) have a ton of config files we never want committed.
What is it I'm not getting here? Why is it all or nothing like this? And given that there's a reason for it, why is there no "duh" way to establish these files as off-limits in the first place? I'm getting semi-ranty here but I am legitimately trying to understand. What's the win here that hasn't sunk in yet that makes such awkwardness necessary?