Vincent,

Vincent Legoll 写道：
> yesterday's today is in fact today's today...

I've tried to stop that but to no avail.

> I was too pessimistic, I'm not getting any merge conflict.

Happy to hear it.  Git can be remarkably clever with 3-way merges 
sometimes (and disappointingly dense at others).

> Your patch seems to be working nicely, I tested on a range
> of VMs:
>
> x86_64: alpine devuan fedora debian_stretch debian_buster
> i686: void

Thank you for being so thorough.  You've given me the confidence 
to push this to master as 
1a1faa78b0498fbb71f1533beb4b65817c1d3f2a.  Guess I'll leave this 
bug open since it only solves it for systemd.

> I'll try to come with something for the other init systems
> and add that to my series...

I hope we can avoid touching users' fstab but don't know enough 
about these other systems to say.

Thanks!

T G-R