On Wed, 06 May 2026 at 10:24:08 +0200, Oswald Buddenhagen wrote:
> On Tue, May 05, 2026 at 09:25:37PM +1000, Seth McDonald wrote:
> > So by instead telling
> > maildir_validate() to never create regardless of the value of
> > ctx->is_inbox, we steer back towards the expected behaviour.
> >
> yeah, except that this seems too easy.
> it must still report that INBOX exists, while not actually creating it.
So assuming the maildir is accessible (though perhaps not yet existent),
you want maildir_validate() to still return DRV_OK here? I don't see
how that can work, given that returning DRV_OK
(a) permits the rest of maildir_open_box() to create a .uidvalidity file
in the maildir (which may not exist in a dry run); and more
importantly,
(b) if that goes okay, the following call to box_confirmed() will add
ST_PRESENT to the state, which is what's preventing create_box()
from being called and messing up the proxy logic.
Related, I noticed the assertion is hit when the near-side store exists
prior to dry running mbsync, not just when it is INBOX. So the fix
likely shouldn't just be a change that accounts for the special case of
INBOX.
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
