nnimap blindly assumes that imap server can store arbitrary flags

nnimap blindly assumes that imap server can store arbitrary flags

[James Cloos]

I see this kind of error, lately, when I ?-tick an article:

,----
| 238728 UID STORE 4467 +FLAGS.SILENT (gnus-dormant)
| 238728 BAD invalid flag list to STORE command
`----

Gnus needs to store its own flags in .newsrc.eld.

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

Re: nnimap blindly assumes that imap server can store arbitrary flags

[Lars Magne Ingebrigtsen]

James Cloos <cloos@jhcloos.com> writes:

> I see this kind of error, lately, when I ?-tick an article:
>
> ,----
> | 238728 UID STORE 4467 +FLAGS.SILENT (gnus-dormant)
> | 238728 BAD invalid flag list to STORE command
> `----
>
> Gnus needs to store its own flags in .newsrc.eld.

It does store these flags in .newsrc.eld.  But I guess sending flags
that the server don't understand to the server is kinda rude?  But does
it matter?

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/

Re: nnimap blindly assumes that imap server can store arbitrary flags

[Simon Josefsson]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> James Cloos <cloos@jhcloos.com> writes:
>
>> I see this kind of error, lately, when I ?-tick an article:
>>
>> ,----
>> | 238728 UID STORE 4467 +FLAGS.SILENT (gnus-dormant)
>> | 238728 BAD invalid flag list to STORE command
>> `----
>>
>> Gnus needs to store its own flags in .newsrc.eld.
>
> It does store these flags in .newsrc.eld.  But I guess sending flags
> that the server don't understand to the server is kinda rude?  But does
> it matter?

Gnus should only send it to the server if either gnus-dormant is
mentioned as a PERMANENTFLAGS, or more likely, \* is part of the
server-provided PERMANENTFLAGS.  See its documentation:

      PERMANENTFLAGS

         Followed by a parenthesized list of flags, indicates which of
         the known flags the client can change permanently.  Any flags
         that are in the FLAGS untagged response, but not the
         PERMANENTFLAGS list, can not be set permanently.  If the client
         attempts to STORE a flag that is not in the PERMANENTFLAGS
         list, the server will either ignore the change or store the
         state change for the remainder of the current session only.
         The PERMANENTFLAGS list can also include the special flag \*,
         which indicates that it is possible to create new keywords by
         attempting to store those flags in the mailbox.

/Simon

Re: nnimap blindly assumes that imap server can store arbitrary flags

[Lars Magne Ingebrigtsen]

Simon Josefsson <simon@josefsson.org> writes:

> Gnus should only send it to the server if either gnus-dormant is
> mentioned as a PERMANENTFLAGS, or more likely, \* is part of the
> server-provided PERMANENTFLAGS.  See its documentation:

Yup.  But that function doesn't have access to that information at that
point, so it's easier to just send everything.

So -- does it matter?  :-)  If it does, it can be fixed, obviously...

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/

Re: nnimap blindly assumes that imap server can store arbitrary flags

[James Cloos]

>>>>> "LMI" == Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

>> Gnus needs to store its own flags in .newsrc.eld.

LMI> It does store these flags in .newsrc.eld.  But I guess sending
LMI> flags that the server don't understand to the server is kinda rude?
LMI> But does it matter?

The flags get lost the next time g or M-g is called on those groups,
so it seemed like it was no longer storing them locally.

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

Re: nnimap blindly assumes that imap server can store arbitrary flags

[Lars Magne Ingebrigtsen]

James Cloos <cloos@jhcloos.com> writes:

> The flags get lost the next time g or M-g is called on those groups,
> so it seemed like it was no longer storing them locally.

Hm...  it's not supposed to overwrite those flags unless the IMAP server
is announcing that it's storing \* flags.  Is it announcing that?

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/

Re: nnimap blindly assumes that imap server can store arbitrary flags

[James Cloos]

>>>>> "JC" == James Cloos <cloos@jhcloos.com> writes:
>>>>> "LMI" == Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

JC>>> Gnus needs to store its own flags in .newsrc.eld.

LMI>> It does store these flags in .newsrc.eld.  But I guess sending
LMI>> flags that the server don't understand to the server is kinda rude?
LMI>> But does it matter?

JC> The flags get lost the next time g or M-g is called on those groups,
JC> so it seemed like it was no longer storing them locally.

I looked at the .newrc.eld again.  The only messages from nnimap groups
which are marked dormant are /old/.  

(Side note:  is it useful to compact a sequence such as (... 55 56 ...)
into (... (55 . 56) ...)?  Is it faster than leaving a sequence of two
uncompacted?)

I like to mark my own posts dormant so that I can readily notice
replies, but I don't see anything recent in the dormant sexps.

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

Re: nnimap blindly assumes that imap server can store arbitrary flags

[James Cloos]

>>>>> "LMI" == Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

LMI> Hm...  it's not supposed to overwrite those flags unless the IMAP server
LMI> is announcing that it's storing \* flags.  Is it announcing that?

Permanent flags are: %Seen %Answered %Deleted %Flagged %Draft

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6

Re: nnimap blindly assumes that imap server can store arbitrary flags

[Lars Magne Ingebrigtsen]

James Cloos <cloos@jhcloos.com> writes:

> (Side note:  is it useful to compact a sequence such as (... 55 56 ...)
> into (... (55 . 56) ...)?  Is it faster than leaving a sequence of two
> uncompacted?)

They take the same number of cons cells, but the latter allows expanding
to (55 . 57) without generating any garbage or manipulating the
sequence, so it's (marginally) faster in the long run.  Possibly.

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/

Re: nnimap blindly assumes that imap server can store arbitrary flags

[Lars Magne Ingebrigtsen]

James Cloos <cloos@jhcloos.com> writes:

> Permanent flags are: %Seen %Answered %Deleted %Flagged %Draft

Right, so there's definitely a bug in there somewhere.

However, none of my IMAP groups seem to lack %*, so it's slightly tricky
to reproduce, but I'll try to synthesise... 

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/

Re: nnimap blindly assumes that imap server can store arbitrary flags

[James Cloos]

>>>>> "LMI" == Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

LMI> They take the same number of cons cells, but the latter allows expanding
LMI> to (55 . 57) without generating any garbage or manipulating the
LMI> sequence, so it's (marginally) faster in the long run.  Possibly.

Cool.  

-JimC
-- 
James Cloos <cloos@jhcloos.com>         OpenPGP: 1024D/ED7DAEA6