Gnus development mailing list
 help / color / mirror / Atom feed
* Totally unrelated message shown in thread, sometimes
@ 2020-11-25  7:45 Pankaj Jangid
  2020-11-25  9:21 ` Adam Sjøgren
  0 siblings, 1 reply; 11+ messages in thread
From: Pankaj Jangid @ 2020-11-25  7:45 UTC (permalink / raw)
  To: Gnus Users

I am trying to reproduce but till now could not do so for all cases.

Message-1:

Subject: pgBadger v11.4 released
To: PostgreSQL Announce <pgsql-announce@lists.postgresql.org>
From: Gilles Darold via PostgreSQL Announce <announce-nore...@postgresql.org>
Reply-To: gi....@darold.net
Date: Tue, 24 Nov 2020 22:04:45 +0000
Message-ID: <160625548542.25817.12618241037859978217@wrigleys.postgresql.org>
X-Auto-Response-Suppress: All
Auto-Submitted: auto-generated
X-pglister-tags: related
X-pglister-tagsig: e6b21f5d742e4d0d4b3124e9270ec500e6e109bae4851f84fdd9df2b24dc218b

Message-2:

From: arthur miller <arthur.......@live.com>
To: Eli Zaretskii <el..@gnu.org>
Subject: RE: Emacs memory usage
Thread-Topic: Emacs memory usage
Thread-Index: AQHWwm+sdCjjswmvHUisXU6jXxlhlKnXdFdHgACNfmk=
Date: Wed, 25 Nov 2020 00:38:55 +0000
Message-ID: <AM0PR06MB6577CB2F2949742C4770579096FA0@AM0PR06MB6577.eurprd06.prod.outlook.com>
References: <AM0PR06MB6577612B793D344B85B3B0FA96FB0@AM0PR06MB6577.eurprd06.prod.outlook.com>, <83mtz6u42h.fsf@gnu.org>
In-Reply-To: <83mtz6u42h.fsf@gnu.org>

When I invoke `gnus-summary-insert-new-artiles' or perform some limiting
operation, the threading becomes inconsistent. The above two messages are
samples. Message-2 was shown as next message in thread with Message-1 as
false root.

When I go out to group-buffer and come back, it is all fine.

At some other instances also, I have observed inconsistencies. I'll try
to note it down next time.


^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Totally unrelated message shown in thread, sometimes
  2020-11-25  7:45 Totally unrelated message shown in thread, sometimes Pankaj Jangid
@ 2020-11-25  9:21 ` Adam Sjøgren
  2020-11-25 10:08   ` Pankaj Jangid
  2020-11-25 17:07   ` Eric Abrahamsen
  0 siblings, 2 replies; 11+ messages in thread
From: Adam Sjøgren @ 2020-11-25  9:21 UTC (permalink / raw)
  To: ding

Pankaj writes:

> When I go out to group-buffer and come back, it is all fine.
>
> At some other instances also, I have observed inconsistencies. I'll try
> to note it down next time.

I see this too once in a while. I don't have a recipe to reproduce it,
unfortunately.

I mostly get headers from a message from nndrafts:delayed mixed into
headers of an article in an unrelated newsgroup, when it happens.

It "feels" like it started happening around the time nnselect was
introduced, but I update emacs from git on an arbitrary schedule, so
that might be completely wrong.


  Best regards,

    Adam

-- 
 "Miawmiaw, what's miawmiaw?"                               Adam Sjøgren
                                                       asjo@koldfront.dk



^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Totally unrelated message shown in thread, sometimes
  2020-11-25  9:21 ` Adam Sjøgren
@ 2020-11-25 10:08   ` Pankaj Jangid
  2020-11-25 17:07   ` Eric Abrahamsen
  1 sibling, 0 replies; 11+ messages in thread
From: Pankaj Jangid @ 2020-11-25 10:08 UTC (permalink / raw)
  To: Adam Sjøgren; +Cc: ding

[-- Attachment #1: Type: text/plain, Size: 641 bytes --]

Adam Sjøgren <asjo@koldfront.dk> writes:

>> When I go out to group-buffer and come back, it is all fine.
>>
>> At some other instances also, I have observed inconsistencies. I'll try
>> to note it down next time.
>
> I see this too once in a while. I don't have a recipe to reproduce it,
> unfortunately.
>
> I mostly get headers from a message from nndrafts:delayed mixed into
> headers of an article in an unrelated newsgroup, when it happens.
>
> It "feels" like it started happening around the time nnselect was
> introduced, but I update emacs from git on an arbitrary schedule, so
> that might be completely wrong.


[-- Attachment #2: Screen Shot 2020-11-25 at 3.36.25 PM.png --]
[-- Type: image/png, Size: 29310 bytes --]

^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Totally unrelated message shown in thread, sometimes
  2020-11-25  9:21 ` Adam Sjøgren
  2020-11-25 10:08   ` Pankaj Jangid
@ 2020-11-25 17:07   ` Eric Abrahamsen
  2020-11-25 17:32     ` Adam Sjøgren
  1 sibling, 1 reply; 11+ messages in thread
From: Eric Abrahamsen @ 2020-11-25 17:07 UTC (permalink / raw)
  To: ding

Adam Sjøgren <asjo@koldfront.dk> writes:

> Pankaj writes:
>
>> When I go out to group-buffer and come back, it is all fine.
>>
>> At some other instances also, I have observed inconsistencies. I'll try
>> to note it down next time.
>
> I see this too once in a while. I don't have a recipe to reproduce it,
> unfortunately.
>
> I mostly get headers from a message from nndrafts:delayed mixed into
> headers of an article in an unrelated newsgroup, when it happens.

Wow, that's pretty weird. Are any of the groups involved search groups,
or otherwise created with nnselect? If not I don't see how it could be
involved. Pulling messages in from nndrafts is awfully weird, though.



^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Totally unrelated message shown in thread, sometimes
  2020-11-25 17:07   ` Eric Abrahamsen
@ 2020-11-25 17:32     ` Adam Sjøgren
  2020-11-25 17:40       ` Eric Abrahamsen
       [not found]       ` <87d001nenj.fsf@ust.hk>
  0 siblings, 2 replies; 11+ messages in thread
From: Adam Sjøgren @ 2020-11-25 17:32 UTC (permalink / raw)
  To: ding

Eric writes:

>> I mostly get headers from a message from nndrafts:delayed mixed into
>> headers of an article in an unrelated newsgroup, when it happens.
>
> Wow, that's pretty weird. Are any of the groups involved search groups,
> or otherwise created with nnselect?

No, they are nntp groups, so it might very well be completely unrelated.

It "feels" like a structure that doesn't get cleared or gets shared
somehow. I will try to find more details next time it happens. I
remember C-u g'ing and the "raw" display was still a mix.

> If not I don't see how it could be involved. Pulling messages in from
> nndrafts is awfully weird, though.

Indeed!


  Best regards,

    Adam

-- 
 "Let me forget about today until tomorrow"                 Adam Sjøgren
                                                       asjo@koldfront.dk



^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Totally unrelated message shown in thread, sometimes
  2020-11-25 17:32     ` Adam Sjøgren
@ 2020-11-25 17:40       ` Eric Abrahamsen
       [not found]       ` <87d001nenj.fsf@ust.hk>
  1 sibling, 0 replies; 11+ messages in thread
From: Eric Abrahamsen @ 2020-11-25 17:40 UTC (permalink / raw)
  To: ding

Adam Sjøgren <asjo@koldfront.dk> writes:

> Eric writes:
>
>>> I mostly get headers from a message from nndrafts:delayed mixed into
>>> headers of an article in an unrelated newsgroup, when it happens.
>>
>> Wow, that's pretty weird. Are any of the groups involved search groups,
>> or otherwise created with nnselect?
>
> No, they are nntp groups, so it might very well be completely unrelated.
>
> It "feels" like a structure that doesn't get cleared or gets shared
> somehow.

Huh, I agree this sounds like the most likely explanation, but with no
nnselect groups involved something really spooky would have to be
happening. Do let us know if you come up with more clues.



^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Totally unrelated message shown in thread, sometimes
       [not found]           ` <87pn3zkfos.fsf@ucl.ac.uk>
@ 2020-12-25 23:40             ` Adam Sjøgren
  2020-12-26  5:54               ` Eric Abrahamsen
  0 siblings, 1 reply; 11+ messages in thread
From: Adam Sjøgren @ 2020-12-25 23:40 UTC (permalink / raw)
  To: ding

Eric writes:

> A suggestion: I get strange behaviour whenever I have tried to use any
> of the daemons.  Usually, hitting M-g on the group before entering it
> fixes things.  Next time you see a strangely threaded summary buffer,
> leave with Q (gnus-summary-exit-no-update, not lower case q), hit M-g on
> the group, and enter it again.  See if you get different threading and
> report here?

This "two articles get mixed together"-thing just happened to me again.
It was just after my laptop had a network hiccup (losing connection to
the access point).

More specifically, an article in a newsgroup,
nntp+fb:feedbase.blog.risks, showed it's own headers mixed with headers
from an article in my nndraft:delayed group. It looked spooky:

 · https://koldfront.dk/misc/gnus/mixup.png

Duplicate From, Subject, Newsgroups lines, and the Organization, Gcc,
OpenPGP, X-Now-Playing and X-Gnus-Delayed all coming from the delayed
article rather than the f.b.risks article. 

Exiting the group with Q, M-g'ing and entering the group again didn't
make any difference - the same mixed up article headers were shown
again. I tried M-g'ing on both groups as well, no difference.

Quitting Gnus and starting again made the inclusion of headers from the
delayed article stop, and the f.b.risks article looked as expected
afterwards.

I guess it's sort of reasonable that things get out of whack when the
network disappears, it's just a weird result.


  Merry christmas,

    Adam

-- 
 "Scare yourself all away from the light                    Adam Sjøgren
  When you look down - you never left the ground"      asjo@koldfront.dk



^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Totally unrelated message shown in thread, sometimes
  2020-12-25 23:40             ` Adam Sjøgren
@ 2020-12-26  5:54               ` Eric Abrahamsen
  2020-12-26  5:58                 ` Eric Abrahamsen
  2020-12-26 14:20                 ` Adam Sjøgren
  0 siblings, 2 replies; 11+ messages in thread
From: Eric Abrahamsen @ 2020-12-26  5:54 UTC (permalink / raw)
  To: ding

Adam Sjøgren <asjo@koldfront.dk> writes:

> Eric writes:
>
>> A suggestion: I get strange behaviour whenever I have tried to use any
>> of the daemons.  Usually, hitting M-g on the group before entering it
>> fixes things.  Next time you see a strangely threaded summary buffer,
>> leave with Q (gnus-summary-exit-no-update, not lower case q), hit M-g on
>> the group, and enter it again.  See if you get different threading and
>> report here?
>
> This "two articles get mixed together"-thing just happened to me again.
> It was just after my laptop had a network hiccup (losing connection to
> the access point).
>
> More specifically, an article in a newsgroup,
> nntp+fb:feedbase.blog.risks, showed it's own headers mixed with headers
> from an article in my nndraft:delayed group. It looked spooky:
>
>  · https://koldfront.dk/misc/gnus/mixup.png

Spooky, indeed. Does this only ever happen with delayed articles, or
with other draft articles, too? Was drafting and delaying the outgoing
article the last thing you did before opening this group and seeing the
fubar'ed headers? Or had you delayed it a while ago? And did you see the
doubled headers on only one message in this group, or for every message?



^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Totally unrelated message shown in thread, sometimes
  2020-12-26  5:54               ` Eric Abrahamsen
@ 2020-12-26  5:58                 ` Eric Abrahamsen
  2020-12-26 14:20                 ` Adam Sjøgren
  1 sibling, 0 replies; 11+ messages in thread
From: Eric Abrahamsen @ 2020-12-26  5:58 UTC (permalink / raw)
  To: ding

Eric Abrahamsen <eric@ericabrahamsen.net> writes:

> Adam Sjøgren <asjo@koldfront.dk> writes:
>
>> Eric writes:
>>
>>> A suggestion: I get strange behaviour whenever I have tried to use any
>>> of the daemons.  Usually, hitting M-g on the group before entering it
>>> fixes things.  Next time you see a strangely threaded summary buffer,
>>> leave with Q (gnus-summary-exit-no-update, not lower case q), hit M-g on
>>> the group, and enter it again.  See if you get different threading and
>>> report here?
>>
>> This "two articles get mixed together"-thing just happened to me again.
>> It was just after my laptop had a network hiccup (losing connection to
>> the access point).
>>
>> More specifically, an article in a newsgroup,
>> nntp+fb:feedbase.blog.risks, showed it's own headers mixed with headers
>> from an article in my nndraft:delayed group. It looked spooky:
>>
>>  · https://koldfront.dk/misc/gnus/mixup.png
>
> Spooky, indeed. Does this only ever happen with delayed articles, or
> with other draft articles, too? Was drafting and delaying the outgoing
> article the last thing you did before opening this group and seeing the
> fubar'ed headers? Or had you delayed it a while ago? And did you see the
> doubled headers on only one message in this group, or for every message?

I guess what I'm getting at with that last question is: is there any
chance that the article number of the delayed message in the delayed
group was the same as the article number of the article with doubled
headers in the feedbase group?



^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Totally unrelated message shown in thread, sometimes
  2020-12-26  5:54               ` Eric Abrahamsen
  2020-12-26  5:58                 ` Eric Abrahamsen
@ 2020-12-26 14:20                 ` Adam Sjøgren
  2020-12-26 19:30                   ` Eric Abrahamsen
  1 sibling, 1 reply; 11+ messages in thread
From: Adam Sjøgren @ 2020-12-26 14:20 UTC (permalink / raw)
  To: ding

Eric writes:

> Does this only ever happen with delayed articles, or with other draft
> articles, too?

Now that you ask, I can only ever remember seeing my long-lived delayed
article being mixed in.

(I guess that points the suspicion towards the gnus-delay-send-queue
demon-handler.)

> Was drafting and delaying the outgoing article the last thing you did
> before opening this group and seeing the fubar'ed headers? Or had you
> delayed it a while ago?

It is an old delayed article that I hadn't touched recently - it was
last edited on October 4, 2020, and it's set to send on July 14, 2021.

> And did you see the doubled headers on only one message in this group,
> or for every message?

Only in the one specific message shown on the screenshot.


Eric writes:

> I guess what I'm getting at with that last question is: is there any
> chance that the article number of the delayed message in the delayed
> group was the same as the article number of the article with doubled
> headers in the feedbase group?

I suspected that too, but the delayed article is ~/News/drafts/delayed/2
while the article from f.b.risks has this Xref header:

  Xref: feedbase.org feedbase.blog.risks:6131

Anyway, given how rarely it happens and that it seems network
instability is "needed", I guess it isn't worth too much time to chase
down.


  Best regards,

    Adam

-- 
 "It started out that computer scientists were              Adam Sjøgren
  worried nobody was listening to us. Now I'm worried  asjo@koldfront.dk
  that too many people are listening."



^ permalink raw reply	[flat|nested] 11+ messages in thread

* Re: Totally unrelated message shown in thread, sometimes
  2020-12-26 14:20                 ` Adam Sjøgren
@ 2020-12-26 19:30                   ` Eric Abrahamsen
  0 siblings, 0 replies; 11+ messages in thread
From: Eric Abrahamsen @ 2020-12-26 19:30 UTC (permalink / raw)
  To: ding

Adam Sjøgren <asjo@koldfront.dk> writes:

> Eric writes:
>
>> Does this only ever happen with delayed articles, or with other draft
>> articles, too?
>
> Now that you ask, I can only ever remember seeing my long-lived delayed
> article being mixed in.
>
> (I guess that points the suspicion towards the gnus-delay-send-queue
> demon-handler.)
>
>> Was drafting and delaying the outgoing article the last thing you did
>> before opening this group and seeing the fubar'ed headers? Or had you
>> delayed it a while ago?
>
> It is an old delayed article that I hadn't touched recently - it was
> last edited on October 4, 2020, and it's set to send on July 14, 2021.
>
>> And did you see the doubled headers on only one message in this group,
>> or for every message?
>
> Only in the one specific message shown on the screenshot.
>
>
> Eric writes:
>
>> I guess what I'm getting at with that last question is: is there any
>> chance that the article number of the delayed message in the delayed
>> group was the same as the article number of the article with doubled
>> headers in the feedbase group?
>
> I suspected that too, but the delayed article is ~/News/drafts/delayed/2
> while the article from f.b.risks has this Xref header:
>
>   Xref: feedbase.org feedbase.blog.risks:6131

Okay, thanks for all data points!

> Anyway, given how rarely it happens and that it seems network
> instability is "needed", I guess it isn't worth too much time to chase
> down.

The bug itself isn't too bad, but understanding why it's happening might
lead to fixes for other stuff.



^ permalink raw reply	[flat|nested] 11+ messages in thread

end of thread, other threads:[~2020-12-26 19:31 UTC | newest]

Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-11-25  7:45 Totally unrelated message shown in thread, sometimes Pankaj Jangid
2020-11-25  9:21 ` Adam Sjøgren
2020-11-25 10:08   ` Pankaj Jangid
2020-11-25 17:07   ` Eric Abrahamsen
2020-11-25 17:32     ` Adam Sjøgren
2020-11-25 17:40       ` Eric Abrahamsen
     [not found]       ` <87d001nenj.fsf@ust.hk>
     [not found]         ` <87k0u8zdv8.fsf@tullinup.koldfront.dk>
     [not found]           ` <87pn3zkfos.fsf@ucl.ac.uk>
2020-12-25 23:40             ` Adam Sjøgren
2020-12-26  5:54               ` Eric Abrahamsen
2020-12-26  5:58                 ` Eric Abrahamsen
2020-12-26 14:20                 ` Adam Sjøgren
2020-12-26 19:30                   ` Eric Abrahamsen

Gnus development mailing list

This inbox may be cloned and mirrored by anyone:

	git clone --mirror http://inbox.vuxu.org/ding

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V1 ding ding/ http://inbox.vuxu.org/ding \
		ding@inbox.vuxu.org
	public-inbox-index ding

Example config snippet for mirrors.
Newsgroup available over NNTP:
	nntp://inbox.vuxu.org/vuxu.archive.emacs.gnus.general


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git