At 2023-09-29 07:44+0300, Amir Goldstein <amir73il@xxxxxxxxx> sent:
That's a good guess.
It helps to CC the author of the patch in this case ;-)
Ha, oops! Sorry, I have been spoiled by web-based bug reporting interfaces
and actually e-mailing maintainers is weirdly a novel experience.
The question is whether we should fix it.
The rule of thumb is that if users complain than we need to fix it,
but it's a corner case and if the only users that complained are willing
to work around the problem (hint hint) then we may not need to fix it.
Gotcha. In my specific situation, unless there is an alternate escaping
mechanism or mount API that works identically in both <6.5 and >=6.5, I'm
going to have to do some version detection to work around the regression
in released 6.5 kernels regardless of whether the old functionality is
restored later. As long as there is *some* way of using overlayfs with
arbitrary path nemes (including commas), I'm fine with whatever you
decide. (If there isn't currently a way to mount a path with commas in it,
you may consider me complaining.)
But if the status quo remains, I would like an answer to:
Is there a new way to escape commas for overlayfs options?
Deferring the question to Christian.
BTW, in case there's a difference between them, the actual code I'm
working with uses the libc mount() function, not the mount CLI tool. I've
also experimented with the functions in libmount but they seems to have
the same problem with commas. Either is fine as long as I can support
mounting arbitrary path names.
Thanks,
Ryan
