On Sun, Nov 03, 2019 at 06:49:41PM +0100, Oswald Buddenhagen wrote: > > The problem I encounter is that after connecting to the server and > > apparently asking it about the list of messages in a mailbox, mbsync > > generates tons of warnings reading > > IMAP warning: unknown system flag \Unseen > > > > which appear to be generated for each messages ostensibly marked by the > > server with that flag. > > > i don't think what the server is doing is specification-conformant. > see https://tools.ietf.org/html/rfc3501#section-2.3.2 > you should report it to them.
The problem is that it is highly unlikely they will do anything about that: I had a similar problem with using mbsync to access IMAP servers of another (the largest) Internet giant in Russia—yandex.ru—and they had some other IMAP violation in their servers. I have reported the issue to them like 3 years ago, and they are still "working on it" (I pinged them; honestly, I did). I think the problem there may be two-fold: 1) most their users access their mail boxes using the stock webmail interface provided by that same service provider; 2) the overwhelming majority of the rest uses either the mobile app provided by the service or are using one of the "go-to" IMAP clients such as MS Outlook and Thunderbird. And I'm pretty confident that in the case of webmail and the app, it's unlikely IMAP even gets used: quite possibly it's some custom HTTP-based protocol over a websocket or something like this. So while I have so say in these matters, my opinion is that such bogosities are better worked around somehow, if possible. Still, regarding the flag in question — the RFC 3501 says: | Servers MAY permit the client to define new keywords | in the mailbox (see the description of the PERMANENTFLAGS response | code for more information). | A flag can be permanent or session-only on a per-flag basis. May it be that this flag somehow gets defined as permanent by _some_ client? I have no idea - which exactly, but I do occasionally use their web interface so, let's suppose, it's that client. > > The problem is exacerbated by the fact mbsync does not hold to its > > promise of not showing warnings is the "--quiet" (or "-q") command-line > > option is passed to it twice, > > > yeah, it's actually calling the error() function. it probably shouldn't, as > it just continues anyway. Should I open a bug? What is the procedure for reporting issues, anyway? _______________________________________________ isync-devel mailing list isync-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/isync-devel
