nnimap-get-groups: %NoSelect vs. %Noselect

nnimap-get-groups: %NoSelect vs. %Noselect

[dick]

Dovecot capitalizes the no-select flag as "%Noselect" when nnimap-get-groups
expects "%NoSelect".  My internal fix was to lowercase-normalize all the
flags and then match on "%noselect".

Re: nnimap-get-groups: %NoSelect vs. %Noselect

[Eric Abrahamsen]

dick <dick.r.chiang@gmail.com> writes:

> Dovecot capitalizes the no-select flag as "%Noselect" when nnimap-get-groups
> expects "%NoSelect".  My internal fix was to lowercase-normalize all the
> flags and then match on "%noselect".

Looks like the RFCs also have %Noselect, so I expect we're just doing it
wrong here. I never noticed as all my Noselect groups don't actually
exist (they only hold children), so they were getting filtered out
later.

Does anyone know of any servers that use %NoSelect? Looks like this
should be a straight fix, otherwise.

Re: nnimap-get-groups: %NoSelect vs. %Noselect

[Lars Ingebrigtsen]

Eric Abrahamsen <eric@ericabrahamsen.net> writes:

> dick <dick.r.chiang@gmail.com> writes:
>
>> Dovecot capitalizes the no-select flag as "%Noselect" when nnimap-get-groups
>> expects "%NoSelect".  My internal fix was to lowercase-normalize all the
>> flags and then match on "%noselect".
>
> Looks like the RFCs also have %Noselect, so I expect we're just doing it
> wrong here. I never noticed as all my Noselect groups don't actually
> exist (they only hold children), so they were getting filtered out
> later.
>
> Does anyone know of any servers that use %NoSelect? Looks like this
> should be a straight fix, otherwise.

Yup.  I have no recollection of \NoSelect, and the server I was using
for testing at that time was Dovecot...  is it possible that they
released a version that had \NoSelect at some point?

But it's probably just a typo on my part.

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

Re: nnimap-get-groups: %NoSelect vs. %Noselect

[Eric Abrahamsen]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Eric Abrahamsen <eric@ericabrahamsen.net> writes:
>
>> dick <dick.r.chiang@gmail.com> writes:
>>
>>> Dovecot capitalizes the no-select flag as "%Noselect" when nnimap-get-groups
>>> expects "%NoSelect".  My internal fix was to lowercase-normalize all the
>>> flags and then match on "%noselect".
>>
>> Looks like the RFCs also have %Noselect, so I expect we're just doing it
>> wrong here. I never noticed as all my Noselect groups don't actually
>> exist (they only hold children), so they were getting filtered out
>> later.
>>
>> Does anyone know of any servers that use %NoSelect? Looks like this
>> should be a straight fix, otherwise.
>
> Yup.  I have no recollection of \NoSelect, and the server I was using
> for testing at that time was Dovecot...  is it possible that they
> released a version that had \NoSelect at some point?

Beats me. Maybe someone else will chime in.

> But it's probably just a typo on my part.

I'm inclined to just switch it, but we could also check for both, just
to be safe.

Re: nnimap-get-groups: %NoSelect vs. %Noselect

[Eric Abrahamsen]

On 07/19/19 15:45 PM, Lars Ingebrigtsen wrote:
> Eric Abrahamsen <eric@ericabrahamsen.net> writes:
>
>> dick <dick.r.chiang@gmail.com> writes:
>>
>>> Dovecot capitalizes the no-select flag as "%Noselect" when nnimap-get-groups
>>> expects "%NoSelect".  My internal fix was to lowercase-normalize all the
>>> flags and then match on "%noselect".
>>
>> Looks like the RFCs also have %Noselect, so I expect we're just doing it
>> wrong here. I never noticed as all my Noselect groups don't actually
>> exist (they only hold children), so they were getting filtered out
>> later.
>>
>> Does anyone know of any servers that use %NoSelect? Looks like this
>> should be a straight fix, otherwise.
>
> Yup.  I have no recollection of \NoSelect, and the server I was using
> for testing at that time was Dovecot...  is it possible that they
> released a version that had \NoSelect at some point?
>
> But it's probably just a typo on my part.

I'm just going to switch this to %Noselect, I think. I'll keep an eye on
bug reports, and we can switch to a check for both, if necessary.