'g' freezes

'g' freezes

[Rémi Letot]

Hello again,

I have nntp and nnimap groups in my list. nnimap-stream is starttls.

When I hit 'g', the process freezes at some point. I only have "nntp
read: nnk" (replace nn by numbers) in the minibuffer, and I have to C-g
to get my gnus back. Or wait until it times out, which takes forever. Or
'killall openssl' twice in a row.

'C-g' doesn't get me a backtrace.

Is there anything I can do to help track this ?

Thanks,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> I have nntp and nnimap groups in my list. nnimap-stream is starttls.
>
> When I hit 'g', the process freezes at some point. I only have "nntp
> read: nnk" (replace nn by numbers) in the minibuffer, and I have to C-g
> to get my gnus back. Or wait until it times out, which takes forever. Or
> 'killall openssl' twice in a row.
>
> 'C-g' doesn't get me a backtrace.

What's in your "*nnimap <imap server name>...*" buffer(s) after you `C-g'?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> I have nntp and nnimap groups in my list. nnimap-stream is starttls.
>>
>> When I hit 'g', the process freezes at some point. I only have "nntp
>> read: nnk" (replace nn by numbers) in the minibuffer, and I have to C-g
>> to get my gnus back. Or wait until it times out, which takes forever. Or
>> 'killall openssl' twice in a row.
>>
>> 'C-g' doesn't get me a backtrace.
>
> What's in your "*nnimap <imap server name>...*" buffer(s) after you `C-g'?

Thousands of lines, here are the last ones, tell me if you need them all :

* 1753 FETCH (UID 1951 FLAGS (\Seen)).
* 1754 FETCH (UID 1952 FLAGS (\Answered \Seen)).
* 1755 FETCH (UID 1953 FLAGS (\Answered \Seen $MDNSent)).
* 1756 FETCH (UID 1954 FLAGS (\Answered \Seen $MDNSent)).
* 1757 FETCH (UID 1955 FLAGS (\Answered \Seen $MDNSent)).
* 1758 FETCH (UID 1956 FLAGS (\Seen $MDNSent)).
* 1759 FETCH (UID 1957 FLAGS (\Seen $MDNSent)).
* 1760 FETCH (UID 1958 FLAGS (\Seen $MDNSent)).
* 1761 FETCH (UID 1960 FLAGS (\Answered \Seen)).
* 1762 FETCH (UID 1961 FLAGS (\Answered \Seen)).
* 1763 FETCH (UID 1962 FLAGS (\Answered \Seen)).
* 1764 FETCH (UID 1963 FLAGS (\Answered \Seen)).
* 1765 FETCH (UID 1964 FLAGS (\Answered \Seen)).
* 1766 FETCH (UID 1965 FLAGS (\Seen)).
1150 OK FETCH completed..

Thanks,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> Thousands of lines, here are the last ones, tell me if you need them all :
>
> * 1753 FETCH (UID 1951 FLAGS (\Seen)).
> * 1754 FETCH (UID 1952 FLAGS (\Answered \Seen)).

Looks fine.

nnimap will use a lot of time the very first time you use it.  Just let
it run for a while.

I guess it should be giving progress reports when it's doing this...
I'll add that.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> Thousands of lines, here are the last ones, tell me if you need them all :
>>
>> * 1753 FETCH (UID 1951 FLAGS (\Seen)).
>> * 1754 FETCH (UID 1952 FLAGS (\Answered \Seen)).
>
> Looks fine.
>
> nnimap will use a lot of time the very first time you use it.  Just let
> it run for a while.

M-g on all groups (even unsubscribed ones) works fine and fast, is 'g'
different ?

Thanks,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> M-g on all groups (even unsubscribed ones) works fine and fast, is 'g'
> different ?

No, `g' should be faster than `M-g' on all the groups.

Perhaps your server doesn't support streaming?  Try setting
`nnimap-streaming' to nil in the nniamp method.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> M-g on all groups (even unsubscribed ones) works fine and fast, is 'g'
>> different ?
>
> No, `g' should be faster than `M-g' on all the groups.
>
> Perhaps your server doesn't support streaming?  Try setting
> `nnimap-streaming' to nil in the nniamp method.

That fixed it, thanks.

My server is indeed quite old: courier-imap from debian etch. Will be
upgraded when squeeze is out.

Thanks,
-- 
Rémi

Re: 'g' freezes

[Rémi Letot]

hobbes@poukram.net (Rémi Letot) writes:

> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
>
>> hobbes@poukram.net (Rémi Letot) writes:
>>
>>> M-g on all groups (even unsubscribed ones) works fine and fast, is 'g'
>>> different ?
>>
>> No, `g' should be faster than `M-g' on all the groups.
>>
>> Perhaps your server doesn't support streaming?  Try setting
>> `nnimap-streaming' to nil in the nniamp method.
>
> That fixed it, thanks.

Urgh, spoke too fast: now it doesn't hang, but it doesn't detect new
messages either. Only M-g seems to detect changes in the number of
messages. 

Thanks,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> Urgh, spoke too fast: now it doesn't hang, but it doesn't detect new
> messages either. Only M-g seems to detect changes in the number of
> messages. 

Does `g' do anything?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> Urgh, spoke too fast: now it doesn't hang, but it doesn't detect new
>> messages either. Only M-g seems to detect changes in the number of
>> messages. 
>
> Does `g' do anything?

yes, nntp groups work fine, only nnimap are affected. 'g' doesn't update
them, M-g does.

Thanks,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> yes, nntp groups work fine, only nnimap are affected. 'g' doesn't update
> them, M-g does.

What's in the "*nnimap <server> ...*" buffer after you press `g'?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> yes, nntp groups work fine, only nnimap are affected. 'g' doesn't update
>> them, M-g does.
>
> What's in the "*nnimap <server> ...*" buffer after you press `g'?

plenty of lines like these:

* 1754 FETCH (UID 1952 FLAGS (%Answered %Seen)).
* 1755 FETCH (UID 1953 FLAGS (%Answered %Seen $MDNSent)).
* 1756 FETCH (UID 1954 FLAGS (%Answered %Seen $MDNSent)).
* 1757 FETCH (UID 1955 FLAGS (%Answered %Seen $MDNSent)).
* 1758 FETCH (UID 1956 FLAGS (%Seen $MDNSent)).
* 1759 FETCH (UID 1957 FLAGS (%Seen $MDNSent)).
* 1760 FETCH (UID 1958 FLAGS (%Seen $MDNSent)).
* 1761 FETCH (UID 1960 FLAGS (%Answered %Seen)).
* 1762 FETCH (UID 1961 FLAGS (%Answered %Seen)).
* 1763 FETCH (UID 1962 FLAGS (%Answered %Seen)).
* 1764 FETCH (UID 1963 FLAGS (%Answered %Seen)).
* 1765 FETCH (UID 1964 FLAGS (%Answered %Seen)).
* 1766 FETCH (UID 1965 FLAGS (%Seen)).
360 OK FETCH completed..

(this is the last one)

When I start gnus, my nnimap groups stay empty, even if gnus knew the
number of messages last time I stopped it. 

The only way to get the number of unread messages is to M-g them.

HTH,
-- 
Rémi

Re: 'g' freezes

[Rémi Letot]

hobbes@poukram.net (Rémi Letot) writes:

> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
>
>> hobbes@poukram.net (Rémi Letot) writes:
>>
>>> yes, nntp groups work fine, only nnimap are affected. 'g' doesn't update
>>> them, M-g does.
>>
>> What's in the "*nnimap <server> ...*" buffer after you press `g'?

And this is what gets in that buffer when I M-g a group :

* FLAGS (gnus-expire %Draft %Answered %Flagged %Deleted %Seen %Recent).
* OK [PERMANENTFLAGS (gnus-expire %* %Draft %Answered %Flagged %Deleted %Seen)] Limited.
* 35 EXISTS.
* 0 RECENT.
* OK [UIDVALIDITY 1059246001] Ok.
* OK [MYRIGHTS "acdilrsw"] ACL.
364 OK [READ-WRITE] Ok.
* 1 FETCH (UID 41083 FLAGS (%Flagged %Seen)).
* 2 FETCH (UID 41146 FLAGS (%Seen gnus-expire)).
* 3 FETCH (UID 41147 FLAGS (%Seen gnus-expire)).
* 4 FETCH (UID 41148 FLAGS (%Seen gnus-expire)).
* 5 FETCH (UID 41149 FLAGS (%Seen gnus-expire)).
* 6 FETCH (UID 41150 FLAGS (%Seen gnus-expire)).
* 7 FETCH (UID 41151 FLAGS (%Seen gnus-expire)).
* 8 FETCH (UID 41152 FLAGS (%Seen gnus-expire)).
* 9 FETCH (UID 41153 FLAGS (%Seen gnus-expire)).
* 10 FETCH (UID 41154 FLAGS (%Seen gnus-expire)).
* 11 FETCH (UID 41155 FLAGS (%Seen gnus-expire)).
* 12 FETCH (UID 41156 FLAGS (%Seen gnus-expire)).
* 13 FETCH (UID 41157 FLAGS (%Seen gnus-expire)).
* 14 FETCH (UID 41158 FLAGS (%Seen gnus-expire)).
* 15 FETCH (UID 41159 FLAGS (%Seen gnus-expire)).
* 16 FETCH (UID 41160 FLAGS (%Seen gnus-expire)).
* 17 FETCH (UID 41161 FLAGS (%Seen gnus-expire)).
* 18 FETCH (UID 41162 FLAGS (%Seen gnus-expire)).
* 19 FETCH (UID 41163 FLAGS (%Seen gnus-expire)).
* 20 FETCH (UID 41164 FLAGS (%Seen gnus-expire)).
* 21 FETCH (UID 41165 FLAGS (%Seen gnus-expire)).
* 22 FETCH (UID 41166 FLAGS (%Seen gnus-expire)).
* 23 FETCH (UID 41167 FLAGS (%Seen gnus-expire)).
* 24 FETCH (UID 41168 FLAGS (%Seen gnus-expire)).
* 25 FETCH (UID 41169 FLAGS (%Seen gnus-expire)).
* 26 FETCH (UID 41170 FLAGS (%Seen gnus-expire)).
* 27 FETCH (UID 41171 FLAGS (%Seen gnus-expire)).
* 28 FETCH (UID 41172 FLAGS (%Seen gnus-expire)).
* 29 FETCH (UID 41173 FLAGS (%Seen gnus-expire)).
* 30 FETCH (UID 41174 FLAGS (%Seen gnus-expire)).
* 31 FETCH (UID 41175 FLAGS ()).
* 32 FETCH (UID 41176 FLAGS ()).
* 33 FETCH (UID 41177 FLAGS ()).
* 34 FETCH (UID 41178 FLAGS ()).
* 35 FETCH (UID 41179 FLAGS ()).
365 OK FETCH completed..

HTH,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> When I start gnus, my nnimap groups stay empty, even if gnus knew the
> number of messages last time I stopped it. 

What do you mean by "stay empty"?

If a group had (say) 25 unread articles, and you quit Gnus, and then
start Gnus, does the group buffer then say that there's 0 unread
articles in the group?

If so, then `G E' the group after the `g', and then `G E' the same group
after a `M-g', so that we can look at the differences.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> When I start gnus, my nnimap groups stay empty, even if gnus knew the
>> number of messages last time I stopped it. 
>
> What do you mean by "stay empty"?
>
> If a group had (say) 25 unread articles, and you quit Gnus, and then
> start Gnus, does the group buffer then say that there's 0 unread
> articles in the group?

yep

> If so, then `G E' the group after the `g', and then `G E' the same group
> after a `M-g', so that we can look at the differences.

at startup:

("nnimap+swoog:INBOX.linux.bxlug" 3
 ((1 . 41174))
 ((seen
   (40803 . 41174))
  (expire
   (41146 . 41174))
  (tick 41083))
 "nnimap:swoog"
 ((imap-status . "41161 41083 41160")
  (uidvalidity . "1059246001")
  (to-list . "linux-bruxelles@lists.bxlug.be")
  (auto-expire . t)))


after g:

("nnimap+swoog:INBOX.linux.bxlug" 3
 ((1 . 41174))
 ((seen
   (40803 . 41174))
  (expire
   (41146 . 41174))
  (tick 41083))
 "nnimap:swoog"
 ((imap-status . "41161 41083 41160")
  (uidvalidity . "1059246001")
  (to-list . "linux-bruxelles@lists.bxlug.be")
  (auto-expire . t)))

after M-g:

("nnimap+swoog:INBOX.linux.bxlug" 3
 ((1 . 41174))
 ((seen
   (40803 . 41174))
  (expire
   (41146 . 41174))
  (tick 41083))
 "nnimap:swoog"
 ((imap-status . "41161 41083 41160")
  (uidvalidity . "1059246001")
  (to-list . "linux-bruxelles@lists.bxlug.be")
  (auto-expire . t)))

HTH,
-- 
Rémi

Re: 'g' freezes

[Rémi Letot]

hobbes@poukram.net (Rémi Letot) writes:

> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
>>
>> If a group had (say) 25 unread articles, and you quit Gnus, and then
>> start Gnus, does the group buffer then say that there's 0 unread
>> articles in the group?
>
> yep
>
>> If so, then `G E' the group after the `g', and then `G E' the same group
>> after a `M-g', so that we can look at the differences.

Hello,

tried again with latest gnus, pulled minutes ago, and same
result. nnimap groups show zero unread count at startup, g doesn't
change anything, and M-g gets an unread count.

One more strange thing: when I G-E a group and C-c C-c the buffer, the
group gets an unread count. Not the real one, but a number. Then I M-g,
and it gets another number, still not the real one. I have to enter and
exit the group to get the best unread count. However this count is a
little bit above the real number. I think it sees messages which are
deleted but not expunged, but I don't know how to validate this.

Tell me if you need more info 

Thanks,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> after g:
>
> ("nnimap+swoog:INBOX.linux.bxlug" 3
>  ((1 . 41174))
>  ((seen
>    (40803 . 41174))
>   (expire
>    (41146 . 41174))
>   (tick 41083))
>  "nnimap:swoog"
>  ((imap-status . "41161 41083 41160")
>   (uidvalidity . "1059246001")
>   (to-list . "linux-bruxelles@lists.bxlug.be")
>   (auto-expire . t)))
>
> after M-g:
>
> ("nnimap+swoog:INBOX.linux.bxlug" 3
>  ((1 . 41174))
>  ((seen
>    (40803 . 41174))
>   (expire
>    (41146 . 41174))
>   (tick 41083))
>  "nnimap:swoog"
>  ((imap-status . "41161 41083 41160")
>   (uidvalidity . "1059246001")
>   (to-list . "linux-bruxelles@lists.bxlug.be")
>   (auto-expire . t)))

Those are totally identical, aren't they?  The other puzzling thing is
that they lack any `active' section in the marks lists.  Do you have any
third-party code that alters your group infos?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> after g:
>>
>> ("nnimap+swoog:INBOX.linux.bxlug" 3
>>  ((1 . 41174))
>>  ((seen
>>    (40803 . 41174))
>>   (expire
>>    (41146 . 41174))
>>   (tick 41083))
>>  "nnimap:swoog"
>>  ((imap-status . "41161 41083 41160")
>>   (uidvalidity . "1059246001")
>>   (to-list . "linux-bruxelles@lists.bxlug.be")
>>   (auto-expire . t)))
>>
>> after M-g:
>>
>> ("nnimap+swoog:INBOX.linux.bxlug" 3
>>  ((1 . 41174))
>>  ((seen
>>    (40803 . 41174))
>>   (expire
>>    (41146 . 41174))
>>   (tick 41083))
>>  "nnimap:swoog"
>>  ((imap-status . "41161 41083 41160")
>>   (uidvalidity . "1059246001")
>>   (to-list . "linux-bruxelles@lists.bxlug.be")
>>   (auto-expire . t)))
>
> Those are totally identical, aren't they?  

I think so, but I don't know what I could be searching for, so I pasted
them all anyway :-)

> The other puzzling thing is that they lack any `active' section in the
> marks lists.  Do you have any third-party code that alters your group
> infos?

I don't think so, I only require nnir, sigbegone (to make signatures
less visible), and spam. I have some functions to customize gnus, but
nothing that serious (I don't even know what that could do)

Note that this particular group may have been "all read" when I tested
that... Do you need that info for a group where some messages are
unread?

I started to have this behaviour when I put (nnimap-streaming nil) in my
config, but of course I can't tell what would happen without that since
g would simply freezes :-)

Thanks,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> I started to have this behaviour when I put (nnimap-streaming nil) in my
> config, but of course I can't tell what would happen without that since
> g would simply freezes :-)

If you're using a Courier IMAP server, this may be fixed now.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> I started to have this behaviour when I put (nnimap-streaming nil) in my
>> config, but of course I can't tell what would happen without that since
>> g would simply freezes :-)
>
> If you're using a Courier IMAP server, this may be fixed now.

I do use an old Courier IMAP.

Just tried, still hangs at startup and with 'g'. Killing openssl (two
times) allows gnus to start, and M-g works.

HTH,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> Just tried, still hangs at startup and with 'g'. Killing openssl (two
> times) allows gnus to start, and M-g works.

It hangs?  With nnimap-streaming set to nil?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> Just tried, still hangs at startup and with 'g'. Killing openssl (two
>> times) allows gnus to start, and M-g works.
>
> It hangs?  With nnimap-streaming set to nil?

no no, with nnimap-streaming set to nil it doesn't freeze, I
misunderstood what you meant.

Thanks,
-- 
Rémi

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> When I start gnus, my nnimap groups stay empty, even if gnus knew the
>> number of messages last time I stopped it. 
>
> What do you mean by "stay empty"?
>
> If a group had (say) 25 unread articles, and you quit Gnus, and then
> start Gnus, does the group buffer then say that there's 0 unread
> articles in the group?
>
> If so, then `G E' the group after the `g', and then `G E' the same group
> after a `M-g', so that we can look at the differences.

Hello,

as already stated, there is no difference in the G-E buffer before and
after the M-g.

I did some more tests, and I can't relate the number of unread articles
shown after the G-E to anything. So at startup, all nnimap groups
display a 0 unread count. After I G-E my inbox (for exaple), it displays
a number, which is close to the reality but not right. That number is in
no way related to any previous unread count. At that point I can't
select the group. Gnus tells me 'Can't select group' when I try to enter
it.

After I M-g, I get a more realistic unread count, and I can select the
group. G-E then doesn't change the article count anymore, and g is still
not working. But I do see activity in the imap log when I g, so I guess
it's doing its things, but doesn't reach the point where it should
update the unread count in the group buffer.

Do you want a test mailbox on my server to investigate ?

Thanks,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> So at startup, all nnimap groups display a 0 unread count.

Like I asked before (I think), do you have anything odd in the way Gnus
displays group lines?  Because "0" is just an incredibly unlikely thing
to have displayed in the group buffer.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> So at startup, all nnimap groups display a 0 unread count.
>
> Like I asked before (I think), do you have anything odd in the way Gnus
> displays group lines?  Because "0" is just an incredibly unlikely thing
> to have displayed in the group buffer.

Sorry, missed that. Here are two of my groups at startup :

*        0: nnimap+swoog:INBOX
*        0: nnimap+swoog:INBOX.linux.bxlug

The bxlug one is really "all read", but my inbox contains more than 200
unread messages.

HTH,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

>> Like I asked before (I think), do you have anything odd in the way Gnus
>> displays group lines?  Because "0" is just an incredibly unlikely thing
>> to have displayed in the group buffer.
>
> Sorry, missed that. Here are two of my groups at startup :
>
> *        0: nnimap+swoog:INBOX
> *        0: nnimap+swoog:INBOX.linux.bxlug

Oh, groups with only ticked articles...

What does

(gnus-active "nnimap+swoog:INBOX")

say?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>>> Like I asked before (I think), do you have anything odd in the way Gnus
>>> displays group lines?  Because "0" is just an incredibly unlikely thing
>>> to have displayed in the group buffer.
>>
>> Sorry, missed that. Here are two of my groups at startup :
>>
>> *        0: nnimap+swoog:INBOX
>> *        0: nnimap+swoog:INBOX.linux.bxlug
>
> Oh, groups with only ticked articles...

That's right for bxlug (although probably just a coincidence), but
certainly not right for my inbox.

>
> What does
>
> (gnus-active "nnimap+swoog:INBOX")

(1 . 0)

HTH,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

>> What does
>>
>> (gnus-active "nnimap+swoog:INBOX")
>
> (1 . 0)

And this is with the latest Gnus?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>>> What does
>>>
>>> (gnus-active "nnimap+swoog:INBOX")
>>
>> (1 . 0)
>
> And this is with the latest Gnus?

pulled again minutes ago just to be sure.

Note that this is before M-g. After M-g it becomes 

(272 . 81803)

HTH,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> pulled again minutes ago just to be sure.
>
> Note that this is before M-g. After M-g it becomes 
>
> (272 . 81803)

Could you post new `G E' outputs before and after the `M-g'?  And are
you 100% sure that you don't have any local code that interferes with
nnimap behaviour in any way?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> pulled again minutes ago just to be sure.
>>
>> Note that this is before M-g. After M-g it becomes 
>>
>> (272 . 81803)
>
> Could you post new `G E' outputs before and after the `M-g'?  And are
> you 100% sure that you don't have any local code that interferes with
> nnimap behaviour in any way?

I can't be 100% sure (this is emacs :-), but to my knowledge there is
nothing. That's an emacs-snapshot package from
http://emacs.orebokech.com on debian sid, there are a bunch of emacs
extensions (debian specific or not), but nothing related to gnus.

I git pulled gnus a few minutes ago, and I have that in my .emacs to
make it use git gnus instead of the bundled one:

---------
;; using git gnus
(setq load-path (cons (expand-file-name "~/.emacs.d/site-lisp/gnus/lisp") load-path))

(require 'info)
(if (featurep 'xemacs)
    (add-to-list 'Info-directory-list "~/.emacs.d/site-lisp/gnus/texi/")
  (add-to-list 'Info-default-directory-list "~/.emacs.d/site-lisp/gnus/texi/"))

(require 'gnus-load)
----------

BTW, I have nnimap-need-unselect-to-notice-new-mail set to t in my
select method. It seems that was needed with the old nnimap, but I don't
know if that might be related.

G E are still 100% identical before and after M-g (I ediffed them to be
sure). They are quite long, so I removed long ranges of numbers :

("nnimap+swoog:INBOX" 3
 ((1 . 44957)
  (44959 . 46017)
  (46019 . 46305)
  46307
  (46309 . 48802)
  (48804 . 49217)
  (49219 . 49299)

...

  (81580 . 81587)
  (81589 . 81601)
  (81603 . 81660)
  (81662 . 81803))
 ((reply 278 302 331 337 385 395
	 (412 . 413)
	 (416 . 417)
	 420 443 456 480 508 510 537 574 579 612 620 626 638 656 658 663 10938 12138 12542 12577
	 (12663 . 12664)
	 12688 12883 13204 14279 14282 15218 15394 16217 16299 16482 18597 19150 19676 34263 34271 34318 34669 34824 35530 35612 35635 35673 35705 44853 44926 44942 44950 47976 48926 48988 49151 49154 49966 50663 50815 51537 52388 52741 52920 56596 57648 62564 65439 65614 65752 66347 66375 67399 68407 68520 69088 69157 69163 69212 69500 69604 69615 69623 69658 69677 69989
	 (70222 . 70223)
	 70716 71004 72337 75255 75390
	 (76010 . 76011)
	 76691 76705 77210 77238 77358 80119 80230 80233 80532 81479)
  (tick 273 275)
  (seen
   (272 . 276)
   (278 . 284)
   (287 . 294)
   (296 . 305)
   (307 . 313)
   (316 . 325)
   (328 . 344)
   (347 . 349)
   351
   (353 . 354)
   356
   (358 . 361)
   (363 . 375)
   (377 . 382)

...

   (643 . 644)
   (650 . 653)
   (655 . 668)
   (671 . 676)
   679 747 2184 10108 10266
   (10808 . 10809)
   10888 10938 10953 11021 11131 11387 11472 11584
   (11670 . 11672)
   11812 11842 12050 12108 12130 12138 12166 12281 12302 12337 12401 12501 12530 12542 12558 12564 12570 12577 12581
   (12663 . 12664)
   12683 12688 12693 12696 12714 12786 12883 12971 12974
   (12998 . 12999)
   13064
   (13156 . 13157)
   13204 13208 13239 13375 13697 14125 14198 14279 14282 14299 14304 14409 14421 14436 14452
   (14463 . 14465)
   14480 14565 14603 14707 14990 15082 15084 15136 15215 15218 15291 15294 15329 15369 15394 15452 15600 15716 15760 15775 15800 15955 15971 16129 16217 16299 16373 16472 16482 16493 16510 16638 16662 16675 16695 16708 16711 16735 16747 16753 16786 16804 16977 17707 17771 17781 17838 17906 18070 18113 18258 18313 18467 18597 18616 18815 18873 18929 18940
   (19150 . 19151)

...

   81682 81685 81687
   (81689 . 81704)
   (81716 . 81717)
   81783
   (81790 . 81803)))
 "nnimap:swoog"
 ((modseq)
  (uidvalidity . "1220472259")
  (active 272 . 81808)
  (permanent-flags . t)
  (imap-status . "81002 272 81001")))



HTH,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> G E are still 100% identical before and after M-g (I ediffed them to be
> sure). They are quite long, so I removed long ranges of numbers :

This may be fixed now.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> G E are still 100% identical before and after M-g (I ediffed them to be
>> sure). They are quite long, so I removed long ranges of numbers :
>
> This may be fixed now.

Sorry, same symptom...

Thanks,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> Sorry, same symptom...

(gnus-active "group.name") still returns `(1 . 0)' after `g', but not
after `M-g'?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

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

> (gnus-active "group.name") still returns `(1 . 0)' after `g', but not
> after `M-g'?

edebug though `nnimap-update-info' in nnimap.el and see what goes wrong.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
>
>> (gnus-active "group.name") still returns `(1 . 0)' after `g', but not
>> after `M-g'?
>
> edebug though `nnimap-update-info' in nnimap.el and see what goes wrong.

wow, I'm afraid this is beyond my abilities, I don't know elisp, edebug,
nor imap... I'm not a programmer, just a happy user.

Now if there is some humanely undestandable documentation, or if you (or
someone else) are able to guide me through jabber or irc, I can follow
instructions and interpret results. Or I can get you a test mailbox on
my server to try and debug this issue.

I'll try to find some edebug documentation, but I'm afraid this won't
help if I don't know what I'm searching for...

Thanks,
-- 
Rémi

Re: 'g' freezes

[Rémi Letot]

hobbes@poukram.net (Rémi Letot) writes:

> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
>
>> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
>>
>>> (gnus-active "group.name") still returns `(1 . 0)' after `g', but not
>>> after `M-g'?
>>
>> edebug though `nnimap-update-info' in nnimap.el and see what goes wrong.

<snip>

> I'll try to find some edebug documentation, but I'm afraid this won't
> help if I don't know what I'm searching for...

ok, actually it helped :-)

I read some doc, didn't understand most of it, but found how to edebug
that function. I still didn't understand what happened while edebugging,
but saw some unsubscribed group names passing by, which must probably be
abnormal.

To test more easily, I unsubscribed every group except my inbox, and
tadaaaa, it got the right unread count at startup, and g did update
it. I then subscribed groups one by one, untill I found the culprit.

So the problem is caused by one precise group. Everything works
perfectly when that group is not subscribed, and does not work anymore
when that group is subscribed. The only obvious particular thing that I
can see in that group is that its name has a "-" in it. If that can't be
the problem, just tell me what info you need to debug the problem.

I haven't yet renamed it or tested anything more, to keep that problem
"alive" if you need a testbed.

HTH,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> So the problem is caused by one precise group. Everything works
> perfectly when that group is not subscribed, and does not work anymore
> when that group is subscribed. The only obvious particular thing that I
> can see in that group is that its name has a "-" in it. If that can't be
> the problem, just tell me what info you need to debug the problem.

So if that single group exists, the nnimap update starts failing on all
the nnimap groups?  What's the name of the mailbox?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> So the problem is caused by one precise group. Everything works
>> perfectly when that group is not subscribed, and does not work anymore
>> when that group is subscribed. The only obvious particular thing that I
>> can see in that group is that its name has a "-" in it. If that can't be
>> the problem, just tell me what info you need to debug the problem.
>
> So if that single group exists, the nnimap update starts failing on all
> the nnimap groups?  What's the name of the mailbox?

Yes, the group is :

nnimap+swoog:INBOX.1_GTD.PROJETS.services-domicile

I haven't tried each and every group, so I can't be sure that this is
the only one giving such behaviour, but that's my only group name
containing a "-", which is suspect in itself :-)

HTH,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> nnimap+swoog:INBOX.1_GTD.PROJETS.services-domicile
>
> I haven't tried each and every group, so I can't be sure that this is
> the only one giving such behaviour, but that's my only group name
> containing a "-", which is suspect in itself :-)

Doesn't look like there's anything majorly weird about the name:

(utf7-encode "INBOX.1_GTD.PROJETS.services-domicile" t)
=> "INBOX.1_GTD.PROJETS.services-domicile"

(Speaking of which, there was a missing encode in the fetching code, but
it shouldn't really matter...)

Could you see in the *imap log* buffer for the EXAMINE and the following
FETCH command for this group?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> nnimap+swoog:INBOX.1_GTD.PROJETS.services-domicile
>>
>> I haven't tried each and every group, so I can't be sure that this is
>> the only one giving such behaviour, but that's my only group name
>> containing a "-", which is suspect in itself :-)
>
> Doesn't look like there's anything majorly weird about the name:
>
> (utf7-encode "INBOX.1_GTD.PROJETS.services-domicile" t)
> => "INBOX.1_GTD.PROJETS.services-domicile"
>
> (Speaking of which, there was a missing encode in the fetching code, but
> it shouldn't really matter...)
>
> Could you see in the *imap log* buffer for the EXAMINE and the following
> FETCH command for this group?

Did g several times, here are some results :

23:19:23 841 EXAMINE "INBOX.1_GTD.PROJETS.services-domicile".
23:19:23 842 UID FETCH 1:* FLAGS.

23:23:09 1171 EXAMINE "INBOX.1_GTD.PROJETS.services-domicile".
23:23:09 1172 UID FETCH 1:* FLAGS.


BTW, I have subscribed only this group and my inbox, all others are
unsubscribed. But gnus does an examine&fetch on a bunch of other
groups. Is this normal ?

HTH,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> 23:19:23 841 EXAMINE "INBOX.1_GTD.PROJETS.services-domicile".
> 23:19:23 842 UID FETCH 1:* FLAGS.

Can you find your (last) "*nnimap <server> ...*" buffer, clear it, and
eval the following there:

(progn (nnimap-send-command "EXAMINE \"INBOX.1_GTD.PROJETS.services-domicile\"") (nnimap-send-command "UID FETCH 1:* FLAGS"))

Post the output here.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> 23:19:23 841 EXAMINE "INBOX.1_GTD.PROJETS.services-domicile".
>> 23:19:23 842 UID FETCH 1:* FLAGS.
>
> Can you find your (last) "*nnimap <server> ...*" buffer, clear it, and
> eval the following there:
>
> (progn (nnimap-send-command "EXAMINE \"INBOX.1_GTD.PROJETS.services-domicile\"") (nnimap-send-command "UID FETCH 1:* FLAGS"))
>
> Post the output here.

There is no output. To be sure that I did the right thing, I also tried
with INBOX, which generated plenty of output. Does the group need to be
subscribed for this to work ?

I did some more testing with subscribing other groups. All groups with a
"-" in the name generate the problem. But some other groups, with
nothing fancy in the name (if a - can be called fancy) also generate the
problem. So it might not be the name after all...

HTH,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

>> (progn (nnimap-send-command "EXAMINE \"INBOX.1_GTD.PROJETS.services-domicile\"") (nnimap-send-command "UID FETCH 1:* FLAGS"))
>>
>> Post the output here.
>
> There is no output. To be sure that I did the right thing, I also tried
> with INBOX, which generated plenty of output. Does the group need to be
> subscribed for this to work ?

Nope.

No output what-so-ever?  Not even an error message?  If you use a
mailbox name that you know not exists (e.g. "LALA"), do you get an error
message then?

Could the "." characters be the problem?  Do you have any groups with
"." that does work?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>>> (progn (nnimap-send-command "EXAMINE \"INBOX.1_GTD.PROJETS.services-domicile\"") (nnimap-send-command "UID FETCH 1:* FLAGS"))
>>>
>>> Post the output here.
>>
>> There is no output. To be sure that I did the right thing, I also tried
>> with INBOX, which generated plenty of output. Does the group need to be
>> subscribed for this to work ?
>
> Nope.
>
> No output what-so-ever?  Not even an error message?  If you use a
> mailbox name that you know not exists (e.g. "LALA"), do you get an error
> message then?

Oops, sorry, the output was above :

* FLAGS (\Draft \Answered \Flagged \Deleted \Seen \Recent).
* OK [PERMANENTFLAGS ()] No permanent flags permitted.
* 1 EXISTS.
* 0 RECENT.
* OK [UIDVALIDITY 1255709781] Ok.
* OK [MYRIGHTS "acdilrsw"] ACL.
3511 OK [READ-ONLY] Ok.
* 1 FETCH (UID 1 FLAGS (\Seen)).
3512 OK FETCH completed..

That group contains only one read message.

> Could the "." characters be the problem?  Do you have any groups with
> "." that does work?

"." is the mailbox separator, all my groups except INBOX have "." in the name.

HTH,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> Oops, sorry, the output was above :
>
> * FLAGS (\Draft \Answered \Flagged \Deleted \Seen \Recent).
> * OK [PERMANENTFLAGS ()] No permanent flags permitted.
> * 1 EXISTS.
> * 0 RECENT.
> * OK [UIDVALIDITY 1255709781] Ok.
> * OK [MYRIGHTS "acdilrsw"] ACL.
> 3511 OK [READ-ONLY] Ok.
> * 1 FETCH (UID 1 FLAGS (\Seen)).
> 3512 OK FETCH completed..

Looks very normal.  Is there anything that separates this output from
output from groups that work?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> Oops, sorry, the output was above :
>>
>> * FLAGS (\Draft \Answered \Flagged \Deleted \Seen \Recent).
>> * OK [PERMANENTFLAGS ()] No permanent flags permitted.
>> * 1 EXISTS.
>> * 0 RECENT.
>> * OK [UIDVALIDITY 1255709781] Ok.
>> * OK [MYRIGHTS "acdilrsw"] ACL.
>> 3511 OK [READ-ONLY] Ok.
>> * 1 FETCH (UID 1 FLAGS (\Seen)).
>> 3512 OK FETCH completed..
>
> Looks very normal.  Is there anything that separates this output from
> output from groups that work?

Tried some others, nothing that I can see.

Thanks,
-- 
Rémi

Re: 'g' freezes

[Rémi Letot]

hobbes@poukram.net (Rémi Letot) writes:

> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
>
>> Looks very normal.  Is there anything that separates this output from
>> output from groups that work?
>
> Tried some others, nothing that I can see.

One more info: I can't rename that particular group. 

I tried G r on another group, which worked, but G r on the problematic
group just beeps, and gnus says that it can't rename it.

HTH,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> I tried G r on another group, which worked, but G r on the problematic
> group just beeps, and gnus says that it can't rename it.

What does the "*nnimap <server> ..." buffer say after you try to `G r' it?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> I tried G r on another group, which worked, but G r on the problematic
>> group just beeps, and gnus says that it can't rename it.
>
> What does the "*nnimap <server> ..." buffer say after you try to `G r' it?

2096 NO Can't RENAME the currently-open folder

HTH,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> 2096 NO Can't RENAME the currently-open folder

Ok; fixed.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> 2096 NO Can't RENAME the currently-open folder
>
> Ok; fixed.

nope. Just pulled, "./configure && make", and same error.

HTH,
-- 
Rémi

nnimap rename (was: Re: 'g' freezes)

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

>>> 2096 NO Can't RENAME the currently-open folder
>>
>> Ok; fixed.
>
> nope. Just pulled, "./configure && make", and same error.

It works for me on the servers I've tested.  What IMAP server type do
you use?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: nnimap rename (was: Re: 'g' freezes)

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>>>> 2096 NO Can't RENAME the currently-open folder
>>>
>>> Ok; fixed.
>>
>> nope. Just pulled, "./configure && make", and same error.
>
> It works for me on the servers I've tested.  What IMAP server type do
> you use?

old version of courier, will be updated when debian squeeze is out.

It worked on most groups before commit
d173e96bece12e3b0d343c8c0d138289de56e111, only the groups responsible
for breaking g (for me) where not possible to rename. 

Now, I can't rename any group (haven't tried all of them, but quite a
bunch), problematic or not. The error message is exactly the same as it
was on the problematic groups before the commit.

I don't know git (or any other dvcs), but I guess it would be possible
for me to revert that specific commit and see if it restores the old
behaviour ? Would it be usefull ? Can anyone tell me how to do that ?

HTH,
-- 
Rémi

Re: nnimap rename

[Rupert Swarbrick]

[-- Attachment #1: Type: text/plain, Size: 1476 bytes --]

hobbes@poukram.net (Rémi Letot) writes:
> It worked on most groups before commit
> d173e96bece12e3b0d343c8c0d138289de56e111, only the groups responsible
> for breaking g (for me) where not possible to rename. 
>
> Now, I can't rename any group (haven't tried all of them, but quite a
> bunch), problematic or not. The error message is exactly the same as it
> was on the problematic groups before the commit.
>
> I don't know git (or any other dvcs), but I guess it would be possible
> for me to revert that specific commit and see if it restores the old
> behaviour ? Would it be usefull ? Can anyone tell me how to do that ?

You want to "checkout" the old version rather than
"reverting". Reverting refers to committing a change to undo an existing
change in the repository. What you need is time travel, which is much
simpler...

In git, you can say "the commit before" with a trailing caret, so I
think you probably want to use something like

  git checkout d173e96^

If you want to jump back and forwards easily, I recommend creating a
branch with a name, like

  git checkout -b old d173e96^

Now you can go to the new version with "git checkout master" and back
with "git checkout old". Moreover, you can cherry-pick commits into your
"old" branch to help work out what's going on.

When you finally want to get rid of the "old" branch, just checkout a
different branch (master, for example) and type "git branch -d old".

Rupert

[-- Attachment #2: Type: application/pgp-signature, Size: 315 bytes --]

Re: nnimap rename

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

>>>>> 2096 NO Can't RENAME the currently-open folder
>>>>
>>>> Ok; fixed.
>>>
>>> nope. Just pulled, "./configure && make", and same error.
>>
>> It works for me on the servers I've tested.  What IMAP server type do
>> you use?
>
> old version of courier, will be updated when debian squeeze is out.

All the other servers allow you to rename a group after doing an EXAMINE
on it.  Courier will complain that it's open.  Doing a CLOSE on the
group doesn't help, either.  (And it's the wrong thing to do,
apparently, since it has side effects for \Deleted articles.)  So the
only way to make Courier unselect a mailbox is to SELECT a different
one?  What if you only have one?

What am I missing here?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: nnimap rename

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>>>>>> 2096 NO Can't RENAME the currently-open folder
>>>>>
>>>>> Ok; fixed.
>>>>
>>>> nope. Just pulled, "./configure && make", and same error.
>>>
>>> It works for me on the servers I've tested.  What IMAP server type do
>>> you use?
>>
>> old version of courier, will be updated when debian squeeze is out.
>
> All the other servers allow you to rename a group after doing an EXAMINE
> on it.  Courier will complain that it's open.  

I checked out a pre d173e96bece12e3b0d343c8c0d138289de56e111 gnus, and I
can confirm that I had no problem renaming most groups before that, on
my courier server. Only the groups which hose 'g' refuse to be renamed,
and only if I M-g them (or subscribe them and use g) before renaming.

After that commit, I can't rename any group, subscribed or not.

HTH,
-- 
Rémi

Re: nnimap rename

[Lars Magne Ingebrigtsen]

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

>  So the only way to make Courier unselect a mailbox is to SELECT a
> different one?  What if you only have one?

SELECTing a mailbox that doesn't exist seems to do the trick on
Courier.  Hacked and pushed.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: nnimap rename

[Rémi Letot]

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

> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
>
>>  So the only way to make Courier unselect a mailbox is to SELECT a
>> different one?  What if you only have one?
>
> SELECTing a mailbox that doesn't exist seems to do the trick on
> Courier.  Hacked and pushed.

pulled, made, tested, and yesss! Now I can rename all groups, including
the 'g' killer, no matter which are subscribed :-)

I somehow hoped that this would make 'g' work right, but no :-(

Thanks,
-- 
Rémi

Re: 'g' freezes

[Rémi Letot]

hobbes@poukram.net (Rémi Letot) writes:

> hobbes@poukram.net (Rémi Letot) writes:
>
>> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
>>
>>> Looks very normal.  Is there anything that separates this output from
>>> output from groups that work?
>>
>> Tried some others, nothing that I can see.
>
> One more info: I can't rename that particular group. 
>
> I tried G r on another group, which worked, but G r on the problematic
> group just beeps, and gnus says that it can't rename it.

And if I rename it with icedove, then try g in gnus, I have no error
(nnimap groups do not update, but gnus doesn't realizes that the group
is gone). Of course gnus realizes that the group is not there anymore if
I M-g it, but g does not give any error.

HTH,
-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> Of course gnus realizes that the group is not there anymore if I M-g
> it, but g does not give any error.

It's not really supposed to.  Groups that go missing on the server will
appear with "*" in the article number slot if you restart Gnus.
Probably.  Perhaps not nnimap groups...

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: 'g' freezes

[Rémi Letot]

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

> hobbes@poukram.net (Rémi Letot) writes:
>
>> Sorry, same symptom...
>
> (gnus-active "group.name") still returns `(1 . 0)' after `g', but not
> after `M-g'?

yep

-- 
Rémi

Re: 'g' freezes

[Lars Magne Ingebrigtsen]

hobbes@poukram.net (Rémi Letot) writes:

> My server is indeed quite old: courier-imap from debian etch. Will be
> upgraded when squeeze is out.

Hm.  I wonder whether we can detect this situation in any reliable
way...  the streaming makes `g' so so so much faster, but this is the
second report I've had that it doesn't work, so it looks like it's not
that uncommon.

The waiting loop should probably be time-limited.  If no data has
arrived in a couple of seconds, then we could just break out of the
loop.  That way we'd at least not hang forever waiting for the data.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen