thread referral and nnir

thread referral and nnir

[Andrew Cohen]

I've just checked in a small change so that thread referral
(`gnus-summary-refer-thread', or `A T') can use nnir to search across
multiple groups. By default everything should work as before. If you use
`A T' with a (non-numeric) prefix arg then nnir will search for thread
articles across all groups on the server and take you to an ephemeral
group composed of these articles.

If you want to make this change the default without having to use a
prefix arg just set `gnus-refer-thread-use-nnir' to t. This also
reverses the meaning of a (non-numeric) prefix arg (i.e. with
`gnus-refer-thread-use-nnir' set to t a prefix arg will cause
`gnus-summary-refer-thread' to only search in the current group) so you
can easily get back to the old behavior on servers that might be slow.

At the moment this only works for nnimap servers using the imap search
engine, but should be easy to add for other engines as well.

Re: thread referral and nnir

[Dave Abrahams]

on Sat Jul 23 2011, Andrew Cohen <cohen-AT-andy.bu.edu> wrote:

> I've just checked in a small change so that thread referral
> (`gnus-summary-refer-thread', or `A T') can use nnir to search across
> multiple groups. By default everything should work as before. If you use
> `A T' with a (non-numeric) prefix arg then nnir will search for thread
> articles across all groups on the server and take you to an ephemeral
> group composed of these articles.
>
> If you want to make this change the default without having to use a
> prefix arg just set `gnus-refer-thread-use-nnir' to t. This also
> reverses the meaning of a (non-numeric) prefix arg (i.e. with
> `gnus-refer-thread-use-nnir' set to t a prefix arg will cause
> `gnus-summary-refer-thread' to only search in the current group) so you
> can easily get back to the old behavior on servers that might be slow.
>
> At the moment this only works for nnimap servers using the imap search
> engine, but should be easy to add for other engines as well.

Nice; it works well!

However, I do have a message for which `^' finds the parent article via
nnir but where `C-u A T' fails, saying the search produced no results.
On the other hand, when I execute the nnir IMAP search I prescribed in
the other thread

  or (header message-id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) (header references <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>)

it works perfectly.  So something is not quite sufficient in the query.

-- 
Dave Abrahams
BoostPro Computing
http://www.boostpro.com

Re: thread referral and nnir

[Andrew Cohen]

>>>>> "Dave" == Dave Abrahams <dave@boostpro.com> writes:

    Dave> on Sat Jul 23 2011, Andrew Cohen <cohen-AT-andy.bu.edu> wrote:

[...]


    Dave> Nice; it works well!

    Dave> However, I do have a message for which `^' finds the parent
    Dave> article via nnir but where `C-u A T' fails, saying the search
    Dave> produced no results.  On the other hand, when I execute the
    Dave> nnir IMAP search I prescribed in the other thread

    Dave>   or (header message-id
    Dave> <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>)
    Dave> (header references
    Dave> <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>)

    Dave> it works perfectly.  So something is not quite sufficient in
    Dave> the query.

OK, lets try to debug this. First, can you verify that `C-u A T'
searched the groups you think it should be searching? You can check the
*Messages* buffer to see. 

Assuming that it did, try to evaluate `nnimap-make-thread-query' on the
header of the parent message and see what it returns.

Re: thread referral and nnir

[Dave Abrahams]

on Sat Jul 23 2011, Andrew Cohen <cohen-AT-andy.bu.edu> wrote:

>>>>>> "Dave" == Dave Abrahams <dave@boostpro.com> writes:
>
>     Dave> on Sat Jul 23 2011, Andrew Cohen <cohen-AT-andy.bu.edu> wrote:
>
> [...]
>
>
>     Dave> Nice; it works well!
>
>     Dave> However, I do have a message for which `^' finds the parent
>     Dave> article via nnir but where `C-u A T' fails, saying the search
>     Dave> produced no results.  On the other hand, when I execute the
>     Dave> nnir IMAP search I prescribed in the other thread
>
>     Dave>   or (header message-id
>     Dave> <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>)
>     Dave> (header references
>     Dave> <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>)
>
>     Dave> it works perfectly.  So something is not quite sufficient in
>     Dave> the query.
>
> OK, lets try to debug this. First, can you verify that `C-u A T'
> searched the groups you think it should be searching? You can check the
> *Messages* buffer to see. 

Yeah, it's doing what I expect:

Retrieving newsgroup: nnir:((query . "(or (header message-id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) (header references <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>))") (unique-id . "m2aac4cxyd.fsf"))...
Fetching headers for nnir:((query . "(or (header message-id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) (header references <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>))") (unique-id . "m2aac4cxyd.fsf"))...done
Opening nnir server on nnimap:BoostPro-ephemeral...done
Opening server BoostPro
Searching nnimap+BoostPro:[Gmail]/All Mail...done
Search produced empty results.

It may be that Gmail just doesn't like the size of that query.
 
>
> Assuming that it did, try to evaluate `nnimap-make-thread-query' on the
> header of the parent message and see what it returns.

Assuming I know what you mean by "the header of the parent message," I
get:

  "(OR (OR HEADER Message-Id <m2vcvem90x.fsf@pluto.luannocracy.com> HEADER REFERENCES <m2vcvem90x.fsf@pluto.luannocracy.com>) (OR (OR HEADER Message-Id <CAHka5EVNxMNuVHNG5Judxewxf9ibhf=P0k18V0-XxbaaxpdySA@mail.gmail.com> HEADER REFERENCES <CAHka5EVNxMNuVHNG5Judxewxf9ibhf=P0k18V0-XxbaaxpdySA@mail.gmail.com>) (OR (OR HEADER Message-Id <m2fwminr94.fsf@pluto.luannocracy.com> HEADER REFERENCES <m2fwminr94.fsf@pluto.luannocracy.com>) (OR (OR HEADER Message-Id <BANLkTimwECBRcqK6Joi8Scbcdk8AhD2=VQ@mail.gmail.com> HEADER REFERENCES <BANLkTimwECBRcqK6Joi8Scbcdk8AhD2=VQ@mail.gmail.com>) (OR (OR HEADER Message-Id <m2pqp0hf43.wl%dave@boostpro.com> HEADER REFERENCES <m2pqp0hf43.wl%dave@boostpro.com>) (OR (OR HEADER Message-Id <AANLkTi=fmvEEZVK97cCHCJy3g4YPFY2pPWPqOmZM+N1t@mail.gmail.com> HEADER REFERENCES <AANLkTi=fmvEEZVK97cCHCJy3g4YPFY2pPWPqOmZM+N1t@mail.gmail.com>) (OR (OR HEADER Message-Id <m2mxkmq74b.wl%dave@boostpro.com> HEADER REFERENCES <m2mxkmq74b.wl%dave@boostpro.com>) (OR (OR HEADER Message-Id <AANLkTimQf5HnfrpiKfZY6Wgh6wjkDtdyZYmDDEcvux+6@mail.gmail.com> HEADER REFERENCES <AANLkTimQf5HnfrpiKfZY6Wgh6wjkDtdyZYmDDEcvux+6@mail.gmail.com>) (OR (OR HEADER Message-Id <m2wrjqqpm8.wl%dave@boostpro.com> HEADER REFERENCES <m2wrjqqpm8.wl%dave@boostpro.com>) (OR (OR HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) (OR HEADER REFERENCES <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com> HEADER Message-Id <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>)))))))))))"

It seems like an awful lot of "OR" nesting.

-- 
Dave Abrahams
BoostPro Computing
http://www.boostpro.com

Re: thread referral and nnir

[Andrew Cohen]

>>>>> "Dave" == Dave Abrahams <dave@boostpro.com> writes:


[...]

    >> 
    >> OK, lets try to debug this. First, can you verify that `C-u A T'
    >> searched the groups you think it should be searching? You can
    >> check the *Messages* buffer to see.

    Dave> Yeah, it's doing what I expect:

    Dave> Retrieving newsgroup: nnir:((query . "(or (header message-id
    Dave> <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>)
    Dave> (header references
    Dave> <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>))")
    Dave> (unique-id . "m2aac4cxyd.fsf"))...  Fetching headers for
    Dave> nnir:((query . "(or (header message-id
    Dave> <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>)
    Dave> (header references
    Dave> <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>))")
    Dave> (unique-id . "m2aac4cxyd.fsf"))...done Opening nnir server on
    Dave> nnimap:BoostPro-ephemeral...done Opening server BoostPro
    Dave> Searching nnimap+BoostPro:[Gmail]/All Mail...done Search
    Dave> produced empty results.

    Dave> It may be that Gmail just doesn't like the size of that query.

    >> 
    >> Assuming that it did, try to evaluate `nnimap-make-thread-query'
    >> on the header of the parent message and see what it returns.

    Dave> Assuming I know what you mean by "the header of the parent
    Dave> message," I get:

    Dave>   "(OR (OR HEADER Message-Id
    Dave> <m2vcvem90x.fsf@pluto.luannocracy.com> HEADER REFERENCES
    Dave> <m2vcvem90x.fsf@pluto.luannocracy.com>) (OR (OR HEADER
    Dave> Message-Id
    Dave> <CAHka5EVNxMNuVHNG5Judxewxf9ibhf=P0k18V0-XxbaaxpdySA@mail.gmail.com>
    Dave> HEADER REFERENCES
    Dave> <CAHka5EVNxMNuVHNG5Judxewxf9ibhf=P0k18V0-XxbaaxpdySA@mail.gmail.com>)
    Dave> (OR (OR HEADER Message-Id
    Dave> <m2fwminr94.fsf@pluto.luannocracy.com> HEADER REFERENCES
    Dave> <m2fwminr94.fsf@pluto.luannocracy.com>) (OR (OR HEADER
    Dave> Message-Id
    Dave> <BANLkTimwECBRcqK6Joi8Scbcdk8AhD2=VQ@mail.gmail.com> HEADER
    Dave> REFERENCES
    Dave> <BANLkTimwECBRcqK6Joi8Scbcdk8AhD2=VQ@mail.gmail.com>) (OR (OR
    Dave> HEADER Message-Id <m2pqp0hf43.wl%dave@boostpro.com> HEADER
    Dave> REFERENCES <m2pqp0hf43.wl%dave@boostpro.com>) (OR (OR HEADER
    Dave> Message-Id
    Dave> <AANLkTi=fmvEEZVK97cCHCJy3g4YPFY2pPWPqOmZM+N1t@mail.gmail.com>
    Dave> HEADER REFERENCES
    Dave> <AANLkTi=fmvEEZVK97cCHCJy3g4YPFY2pPWPqOmZM+N1t@mail.gmail.com>)
    Dave> (OR (OR HEADER Message-Id <m2mxkmq74b.wl%dave@boostpro.com>
    Dave> HEADER REFERENCES <m2mxkmq74b.wl%dave@boostpro.com>) (OR (OR
    Dave> HEADER Message-Id
    Dave> <AANLkTimQf5HnfrpiKfZY6Wgh6wjkDtdyZYmDDEcvux+6@mail.gmail.com>
    Dave> HEADER REFERENCES
    Dave> <AANLkTimQf5HnfrpiKfZY6Wgh6wjkDtdyZYmDDEcvux+6@mail.gmail.com>)
    Dave> (OR (OR HEADER Message-Id <m2wrjqqpm8.wl%dave@boostpro.com>
    Dave> HEADER REFERENCES <m2wrjqqpm8.wl%dave@boostpro.com>) (OR (OR
    Dave> HEADER Message-Id
    Dave> <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>
    Dave> HEADER REFERENCES
    Dave> <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>)
    Dave> (OR HEADER REFERENCES
    Dave> <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>
    Dave> HEADER Message-Id
    Dave> <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>)))))))))))"

    Dave> It seems like an awful lot of "OR" nesting.

That looks right. I don't use gmail, but searches with dovecot have no
trouble with lengthy queries. Notice that the query that worked for you
is present in this lengthier query and properly OR'd with other clauses,
so it should have produced a superset of the articles. If its a gmail
problem we can try to work around it. Can you try a few more things?

Try entering the long query directly as an nnir search (as you did for
the search that succeeded). It calls all the same routines so this
should not make any difference, but I want to make sure.

Assuming that this fails to find all the articles, can you try removing
some of the clauses and see when it starts to work? For example can you
try: 

"(OR (OR HEADER Message-Id <m2vcvem90x.fsf@pluto.luannocracy.com> HEADER
REFERENCES <m2vcvem90x.fsf@pluto.luannocracy.com>) (OR (OR HEADER
Message-Id
<CAHka5EVNxMNuVHNG5Judxewxf9ibhf=P0k18V0-XxbaaxpdySA@mail.gmail.com>
HEADER REFERENCES
<CAHka5EVNxMNuVHNG5Judxewxf9ibhf=P0k18V0-XxbaaxpdySA@mail.gmail.com>)
(OR (OR HEADER Message-Id <m2fwminr94.fsf@pluto.luannocracy.com> HEADER
REFERENCES <m2fwminr94.fsf@pluto.luannocracy.com>) (OR (OR HEADER
Message-Id <BANLkTimwECBRcqK6Joi8Scbcdk8AhD2=VQ@mail.gmail.com> HEADER
REFERENCES <BANLkTimwECBRcqK6Joi8Scbcdk8AhD2=VQ@mail.gmail.com>) 
(OR (OR HEADER Message-Id
<AANLkTi=fmvEEZVK97cCHCJy3g4YPFY2pPWPqOmZM+N1t@mail.gmail.com> HEADER
REFERENCES
<AANLkTi=fmvEEZVK97cCHCJy3g4YPFY2pPWPqOmZM+N1t@mail.gmail.com>) 
(OR (OR HEADER Message-Id
<AANLkTimQf5HnfrpiKfZY6Wgh6wjkDtdyZYmDDEcvux+6@mail.gmail.com> HEADER
REFERENCES
<AANLkTimQf5HnfrpiKfZY6Wgh6wjkDtdyZYmDDEcvux+6@mail.gmail.com>) 
(OR (OR HEADER Message-Id
<AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER
REFERENCES
<AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) (OR
HEADER REFERENCES
<CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>
HEADER Message-Id
<CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>))))))))"

Re: thread referral and nnir

[Andrew Cohen]

OK, I think I tracked this down. Gmail needs the "%" character to be
quoted (none of the other servers I tested needed this, but it doesn't
hurt them. Go figure).

I'm pushing this in a moment. Let me know if it fixes the problem. 

Re: thread referral and nnir

[Dave Abrahams]

on Sat Jul 23 2011, Andrew Cohen <cohen-AT-andy.bu.edu> wrote:

> Try entering the long query directly as an nnir search (as you did for
> the search that succeeded). It calls all the same routines so this
> should not make any difference, but I want to make sure.

Fails in the same way as expected.

> Assuming that this fails to find all the articles, can you try removing
> some of the clauses and see when it starts to work? For example can you
> try: 
>

<snip>

also fails.  I can't find any rhyme or reason in it.  You might want to
get yourself a Gmail account so you can experiment, but here's what I
I've discovered:

works:

--8<---------------cut here---------------start------------->8---
OR HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>
--8<---------------cut here---------------end--------------->8---

works:

--8<---------------cut here---------------start------------->8---
OR (OR HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) HEADER Message-Id <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>
--8<---------------cut here---------------end--------------->8---

works:

--8<---------------cut here---------------start------------->8---
(or (OR HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) HEADER Message-Id <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>)
--8<---------------cut here---------------end--------------->8---

works:

--8<---------------cut here---------------start------------->8---
(or (subject interview (OR HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>)) HEADER Message-Id <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>)
--8<---------------cut here---------------end--------------->8---

works:

--8<---------------cut here---------------start------------->8---
or HEADER References <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com> (or (OR HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) HEADER Message-Id <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>)
--8<---------------cut here---------------end--------------->8---

works:

--8<---------------cut here---------------start------------->8---
(OR (OR HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) (OR HEADER REFERENCES <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com> HEADER Message-Id <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>))
--8<---------------cut here---------------end--------------->8---


fails:

--8<---------------cut here---------------start------------->8---
(OR (OR HEADER Message-Id <m2wrjqqpm8.wl%dave@boostpro.com> HEADER REFERENCES <m2wrjqqpm8.wl%dave@boostpro.com>) (OR (OR HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) (OR HEADER REFERENCES <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com> HEADER Message-Id <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>)))
--8<---------------cut here---------------end--------------->8---

works:

--8<---------------cut here---------------start------------->8---
(OR HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>)
--8<---------------cut here---------------end--------------->8---

works:

--8<---------------cut here---------------start------------->8---
(OR (OR HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) HEADER REFERENCES <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>)
--8<---------------cut here---------------end--------------->8---

fails:

--8<---------------cut here---------------start------------->8---
(OR HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com> HEADER Message-Id <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>)
--8<---------------cut here---------------end--------------->8---

works:

--8<---------------cut here---------------start------------->8---
(OR (OR HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) HEADER REFERENCES <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com> HEADER Message-Id <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>)
--8<---------------cut here---------------end--------------->8---

fails (note the doubled space before the last message-id):

--8<---------------cut here---------------start------------->8---
(OR (OR HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) HEADER REFERENCES <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com> HEADER Message-Id  <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>)
--8<---------------cut here---------------end--------------->8---

fails:

--8<---------------cut here---------------start------------->8---
OR HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> (OR HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com> HEADER REFERENCES <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com> HEADER Message-Id <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>)
--8<---------------cut here---------------end--------------->8---

fails:

--8<---------------cut here---------------start------------->8---
(OR (OR (OR (OR (OR HEADER Message-Id <m2wrjqqpm8.wl%dave@boostpro.com> HEADER REFERENCES <m2wrjqqpm8.wl%dave@boostpro.com>) HEADER Message-Id <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) HEADER REFERENCES <AANLkTimeXrbYJDTmXhBWsOMg-jEmfJ9T4TdcVLuuHCfy@mail.gmail.com>) HEADER REFERENCES <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>) HEADER Message-Id <CAHka5EXZTg-=UX4O+LdLt6YTvi7GFBmwLs7x2cFKRK9t+cx1zA@mail.gmail.com>)
--8<---------------cut here---------------end--------------->8---

-- 
Dave Abrahams
BoostPro Computing
http://www.boostpro.com

Re: thread referral and nnir

[Dave Abrahams]

on Sun Jul 24 2011, Andrew Cohen <cohen-AT-andy.bu.edu> wrote:

> OK, I think I tracked this down. Gmail needs the "%" character to be
> quoted (none of the other servers I tested needed this, but it doesn't
> hurt them. Go figure).
>
> I'm pushing this in a moment. Let me know if it fixes the problem. 

Success!  Now I can officially switch back to Gnus... once I get one
more feature:

  I want it to bring me to the message that I'm currently viewing within
  that thread

That is an important part of the corresponding Wanderlust feature I
pointed you at.  In a long thread, it can be really tough to re-orient
onesself.

[Also (I realize this is like asking for the moon, but) it would be
really great if the resulting ephemeral group name were something more
meaningful to humans.]

-- 
Dave Abrahams
BoostPro Computing
http://www.boostpro.com

Re: thread referral and nnir

[Andrew Cohen]

>>>>> "Dave" == Dave Abrahams <dave@boostpro.com> writes:


[...]


    Dave> Success!  Now I can officially switch back to Gnus... once I
    Dave> get one more feature:

    Dave>   I want it to bring me to the message that I'm currently
    Dave> viewing within that thread

Easy enough. I'll push something when I get a chance.

[...]


    Dave> [Also (I realize this is like asking for the moon, but) it
    Dave> would be really great if the resulting ephemeral group name
    Dave> were something more meaningful to humans.]

Sorry, the way nnir passes queries around is through the group name. 

Re: thread referral and nnir

[Dave Abrahams]

on Sun Jul 24 2011, Andrew Cohen <cohen-AT-andy.bu.edu> wrote:

>>>>>> "Dave" == Dave Abrahams <dave@boostpro.com> writes:
>
>
> [...]
>
>
>     Dave> Success!  Now I can officially switch back to Gnus... once I
>     Dave> get one more feature:
>
>     Dave>   I want it to bring me to the message that I'm currently
>     Dave> viewing within that thread
>
> Easy enough. I'll push something when I get a chance.
>
> [...]
>
>
>     Dave> [Also (I realize this is like asking for the moon, but) it
>     Dave> would be really great if the resulting ephemeral group name
>     Dave> were something more meaningful to humans.]
>
> Sorry, the way nnir passes queries around is through the group name. 

What about at least making the buffer name more reasonable?

-- 
Dave Abrahams
BoostPro Computing
http://www.boostpro.com

glitches [was: thread referral and nnir]

[Dave Abrahams]

on Sun Jul 24 2011, Dave Abrahams <dave-AT-boostpro.com> wrote:

> on Sun Jul 24 2011, Andrew Cohen <cohen-AT-andy.bu.edu> wrote:
>
>>>>>>> "Dave" == Dave Abrahams <dave@boostpro.com> writes:
>>
>>
>> [...]
>>
>>
>>     Dave> Success!  Now I can officially switch back to Gnus... once I
>>     Dave> get one more feature:
>>
>>     Dave>   I want it to bring me to the message that I'm currently
>>     Dave> viewing within that thread
>>
>> Easy enough. I'll push something when I get a chance.
>>
>> [...]
>>
>>
>>     Dave> [Also (I realize this is like asking for the moon, but) it
>>     Dave> would be really great if the resulting ephemeral group name
>>     Dave> were something more meaningful to humans.]
>>
>> Sorry, the way nnir passes queries around is through the group name. 
>
> What about at least making the buffer name more reasonable?

When looking at these threads, Gnus tends to generate many buffers with
crazy-long names, which tends to make using ido un-fun.

Also, I recently got the following odd-looking summary buffer by using this
feature.  Note how there was apparently a reply to my message of today
last Wednesday:

--8<---------------cut here---------------start------------->8---
  .  > Tuesday 12: Daniel Pfeifer * [ryppl-dev] $<> expressions in include_directories command
  .  +-> Tuesday 12: Troy Straszheim * 
  .  | \-> Tuesday 12: Daniel Pfeifer * 
  .  \-> Tuesday 12: Dave Abrahams * 
  .  > Wednesday 13: Daniel Pfeifer * [ryppl-dev] Re: [CMake] $<> expressions in include_directories command
  .    \-> Wednesday 13: Daniel Pfeifer * 
  .      \-> Thursday 14: Dave Abrahams * 
  .        \-> Today, 04:12: Daniel Pfeifer * 
  .          \-> Today, 07:56: Dave Abrahams * 
  .            \-> Wednesday 13: Daniel Pfeifer * 
--8<---------------cut here---------------end--------------->8---


When I go to the last article and `^', the buffer changes as follows:

--8<---------------cut here---------------start------------->8---
  .  > Tuesday 12: Daniel Pfeifer * [ryppl-dev] $<> expressions in include_directories command
  .  +-> Tuesday 12: Troy Straszheim * 
  .  | \-> Tuesday 12: Daniel Pfeifer * 
  .  \-> Tuesday 12: Dave Abrahams * 
  .  > Wednesday 13: Daniel Pfeifer * [ryppl-dev] Re: [CMake] $<> expressions in include_directories command
  .    \-> Wednesday 13: Daniel Pfeifer * 
 R       +-> Wednesday 13: Andreas Pokorny * Re: $<> expressions in include_directories command
  .      | \-> Wednesday 13: Daniel Pfeifer * [ryppl-dev] Re: [CMake] $<> expressions in include_directories command
  .      \-> Thursday 14: Dave Abrahams * 
  .        \-> Today, 04:12: Daniel Pfeifer * 
  .          \-> Today, 07:56: Dave Abrahams * 
--8<---------------cut here---------------end--------------->8---

The new article comes in with a negative article number, which caused me
to see some `Can't mark negative article numbers' errors; no surprise I
suppose.

Why wasn't the new article found by the original search?  I don't know,
but I think it was posted via gmane and thus had a strange In-Reply-To field
in its header:

--8<---------------cut here---------------start------------->8---
References: <52baa07f-b976-40ae-9816-153bfdce1775@a7g2000vby.googlegroups.com>
	<CANLz8R6CPNY+YadXYdmxbX3o-iD0AAphczV1qjjmLBy0rwwo0w@mail.gmail.com>
	<CAFKLa4JvPLu4tEorhQH32+DTz+8aoPbqQSsV8jWaVC9wRVex1g@mail.gmail.com>
	<CANZMcXpGO23DwejZda5NguQuLdvKZ-57UC_PdjyPiYkE1-hxSQ@mail.gmail.com>
In-Reply-To: <CANZMcXpGO23DwejZda5NguQuLdvKZ-57UC_PdjyPiYkE1-hxSQ@mail.gmail.com>
--8<---------------cut here---------------end--------------->8---

Then, in trying to reproduce the issue I did `C-u A T' on one of
today's articles from the ephemeral buffer and got

--8<---------------cut here---------------start------------->8---
Debugger entered--Lisp error: (args-out-of-range [] 8)
  nnir-warp-to-article()
  gnus-warp-to-article()
  gnus-summary-refer-thread((4))
  call-interactively(gnus-summary-refer-thread nil nil)
--8<---------------cut here---------------end--------------->8---

which was a big surprise.

-- 
Dave Abrahams
BoostPro Computing
http://www.boostpro.com