On Mon, 04 May 2026 at 15:00:27 +0200, Oswald Buddenhagen wrote:
> On Mon, May 04, 2026 at 09:09:05PM +1000, Seth McDonald via isync-devel wrote:
> > Which gives the impression that is_fake is either intended to
> > indicate something other than "this is a dry/fake run",
> >
> it's indicating that the specific box is fake (has no actual representation
> in the store it's attached to).
Okay, so would it be correct to say that each fake box is expected to be
individually created and deleted? Since that would explain why is_fake
is toggled on and off in the proxy create_box() and delete_box(),
respectively.
If so, then that's partly the problem here. The is_fake logic appears
to expect the fake mailbox to be selected, created, opened, examined,
and deleted, in that order (it doesn't consider/impact the modification
logic). But after selecting it, the program actually attempts to open
the mailbox first. And this opening logic includes creating the mailbox
if it doesn't exist, thereby skipping the proxy's creation logic and
messing with the is_fake logic.
That isn't the whole story though. In fact, this problem isn't confined
to just dry runs. AFAICT, using the same configuration file, the
mailbox creation logic is just never run, even without the --dry-run
option. For example, if you set a breakpoint at box_created(), it will
never be hit. This is the case from at least mbsync v1.3.0 (and perhaps
earlier; I stopped checking here).
Am I correct in thinking this behaviour is problematic, or am I missing
something (which is very possible)?
Take care,
Seth McDonald.
--
E9D1 26A5 F0D4 9DF7 792B C2E2 B4BF 4530 D39B 2D51
_______________________________________________
isync-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/isync-devel
