On Wednesday, January 29, 2020 7:40:50 PM MST Han-Wen Nienhuys wrote:

On Wed, Jan 29, 2020 at 5:49 PM Martin Fick [off-list ref] wrote:

quoted

quoted

If you're actually doing the correct locking and packed-refs read (which
"real" implementations like libgit2 do) then no, I don't think that's
dangerous. And I think libgit2 properly complains when it sees a
repositoryformatversion above 0. I don't know offhand about JGit, or any
of the lesser-used ones like dulwich or go-git.

Today, some of these sound like shortcuts that are very likely taken quite
a bit by cleanup and other maintenance scripts (not necessarily formal
git tools), and the impact of these shortcuts is likely low with the
current model. However, I suspect these tools/scripts could be seriously
disruptive if we leave the refs dir around when using reftable,

Maybe we can leave the refs dir, but have no heads/ directory inside,
and make the whole thing read-only?

That might be a good enough safety. I guess the next question would be,  would 
it be OK for reftable to ignore and entries under the refs/ dir if they happen 
to appear there somehow?

-Martin

-- 
The Qualcomm Innovation Center, Inc. is a member of Code 
Aurora Forum, hosted by The Linux Foundation