Re: sqlite3

[Eli Zaretskii]

> From: Lars Ingebrigtsen <larsi@gnus.org>
> Cc: emacs-devel@gnu.org
> Date: Sat, 18 Dec 2021 09:28:07 +0100
> 
> Lars Ingebrigtsen <larsi@gnus.org> writes:
> 
> > Right.  I'll do some further testing, but the real fix is to add some
> > backoff logic when the file is locked, which I'll do now.  (And in that
> > case, I think the `fsync' should be unnecessary.)
> 
> Now done, but since I'm no longer seeing the errors, I can't test
> whether this fixes the issue, so can you check?

It still fails from time to time here.  I have an SSD disk, maybe
that's a factor, somehow.

> Also -- it would be better to catch just the "Permission denied" error
> instead of all file errors in the `condition-case' -- but I guess we
> don't signal a more specific error for that?

Yes, we do: write-region calls report_file_errno, which includes the
specific error in the error data.