I notice that whenever threads are moved around in these forums all
posts in a particular thread do get moved to new forum/sub forum in
the web frontend.
But when viewing threads through nntp only the latest post is seen. Is
this how it is supposed to work ?
–
GNOME 3.10.1
openSUSE 13.1 (Bottle) (x86_64) 64-bit
Kernel Linux 3.11.6-4-desktop
On 2013-12-01 14:42, vazhavandan wrote:
> I notice that whenever threads are moved around in these forums all
> posts in a particular thread do get moved to new forum/sub forum in
> the web frontend.
>
> But when viewing threads through nntp only the latest post is seen. Is
> this how it is supposed to work ?
Yes.
Only the web side posts are moved. The person doing the move has to make
a post saying that he is going to move it, stop posting, and wait till
that post has gone on the nntp side. Then he does the move, and posts a
new one saying that it is moved.
The entire thread is on the new location on the web side. On nntp side,
however, only the last post saying that it was moved appears on the new
side, and it is a courtesy so that we can find the thread on nntp -
otherwise we can not post at all.
The previous post, saying not post more, remains at the end of the
thread on the old location on nntp side - and when we see it we must
refrain from posting there, because if we do we create a new thread on
the web side at the old forum, with just a single post.
It is somewhat confusing, and those lone posts do happen some times. It
needs some effort on the nntp posters to notice those moves (because the
closing notice is at the end). In thunderbird I mark them with a
different colour.
It is also confusing if the last post does not say where it was moved
from, because we can not know unless we happen to have seen it before
(and we need to read those posts before answering). Some moderators even
go to the extra effort of posting a quote of the OP, which is very nice
for us 
–
Cheers / Saludos,
Carlos E. R.
(from 12.3 x86_64 “Dartmouth” at Telcontar)
On Sun, 01 Dec 2013 13:42:39 +0000, vazhavandan wrote:
> I notice that whenever threads are moved around in these forums all
> posts in a particular thread do get moved to new forum/sub forum in
> the web frontend.
>
> But when viewing threads through nntp only the latest post is seen. Is
> this how it is supposed to work ?
Yes, because NNTP doesn’t support moves or edits. There’s no way to
effectively move the entire thread that preserves the timestamps (in
particular).
That’s the way it has always worked.
Jim
–
Jim Henderson
openSUSE Forums Administrator
Forum Use Terms & Conditions at http://tinyurl.com/openSUSE-T-C
On 2013-12-01 19:50, Jim Henderson wrote:
> On Sun, 01 Dec 2013 13:42:39 +0000, vazhavandan wrote:
> Yes, because NNTP doesn’t support moves or edits. There’s no way to
> effectively move the entire thread that preserves the timestamps (in
> particular).
Mmm… in my private leafnode server I can edit a post if I want to, but
not using leafnode tools, but brute force file edit. Ie, to do that
nicely someone would have to create new tools.
–
Cheers / Saludos,
Carlos E. R.
(from 12.3 x86_64 “Dartmouth” at Telcontar)
On Sun, 01 Dec 2013 19:58:07 +0000, Carlos E. R. wrote:
> On 2013-12-01 19:50, Jim Henderson wrote:
>> On Sun, 01 Dec 2013 13:42:39 +0000, vazhavandan wrote:
>
>> Yes, because NNTP doesn’t support moves or edits. There’s no way to
>> effectively move the entire thread that preserves the timestamps (in
>> particular).
>
> Mmm… in my private leafnode server I can edit a post if I want to, but
> not using leafnode tools, but brute force file edit. Ie, to do that
> nicely someone would have to create new tools.
That’s because leafnode uses a simple file structure rather than a
database backend. Most NNTP servers don’t use a simple file store for
their spool, but rather use a fairly complex database structure.
Leafnode isn’t intended to be used as a high-volume NNTP server, so it
doesn’t matter if it uses straight file storage or not.
Jim
–
Jim Henderson
openSUSE Forums Administrator
Forum Use Terms & Conditions at http://tinyurl.com/openSUSE-T-C
On 2013-12-01 22:19, Jim Henderson wrote:
> On Sun, 01 Dec 2013 19:58:07 +0000, Carlos E. R. wrote:
>> Mmm… in my private leafnode server I can edit a post if I want to, but
>> not using leafnode tools, but brute force file edit. Ie, to do that
>> nicely someone would have to create new tools.
>
> That’s because leafnode uses a simple file structure rather than a
> database backend. Most NNTP servers don’t use a simple file store for
> their spool, but rather use a fairly complex database structure.
Ah. Makes sense.
They could use an mbox type archive, that’s usable for high volume,
specially as you normally only append. Same as imap servers on mbox.
> Leafnode isn’t intended to be used as a high-volume NNTP server, so it
> doesn’t matter if it uses straight file storage or not.
Ha. Good point.
However, I’d bet that a server using leafnode type structure on a
reiserfs would run as fast.
I have not experimented with “inn” to know what it uses, though.
–
Cheers / Saludos,
Carlos E. R.
(from 12.3 x86_64 “Dartmouth” at Telcontar)
On Mon, 02 Dec 2013 01:38:23 +0000, Carlos E. R. wrote:
> On 2013-12-01 22:19, Jim Henderson wrote:
>> On Sun, 01 Dec 2013 19:58:07 +0000, Carlos E. R. wrote:
>
>
>>> Mmm… in my private leafnode server I can edit a post if I want to,
>>> but not using leafnode tools, but brute force file edit. Ie, to do
>>> that nicely someone would have to create new tools.
>>
>> That’s because leafnode uses a simple file structure rather than a
>> database backend. Most NNTP servers don’t use a simple file store for
>> their spool, but rather use a fairly complex database structure.
>
> Ah. Makes sense.
>
> They could use an mbox type archive, that’s usable for high volume,
> specially as you normally only append. Same as imap servers on mbox.
That doesn’t work so well for xover queries and the like. Most NNTP
servers have different options for configuring the spool depending on the
message volume and intended use. High-volume NNTP servers host millions
and millions of messages, not merely tends or hundreds of thousands - and
they deal with a large number of communications.
>> Leafnode isn’t intended to be used as a high-volume NNTP server, so it
>> doesn’t matter if it uses straight file storage or not.
>
> Ha. Good point.
>
> However, I’d bet that a server using leafnode type structure on a
> reiserfs would run as fast.
I doubt it. Flat text file performance is never as good as a fully-
indexed system.
> I have not experimented with “inn” to know what it uses, though.
I have, but INN isn’t what’s used here - what’s used here are high-end
commercial servers from Highwinds Software.
Jim Henderson
openSUSE Forums Administrator
Forum Use Terms & Conditions at http://tinyurl.com/openSUSE-T-C
On 2013-12-02 02:41, Jim Henderson wrote:
> On Mon, 02 Dec 2013 01:38:23 +0000, Carlos E. R. wrote:
>> Ah. Makes sense.
>>
>> They could use an mbox type archive, that’s usable for high volume,
>> specially as you normally only append. Same as imap servers on mbox.
>
> That doesn’t work so well for xover queries and the like. Most NNTP
> servers have different options for configuring the spool depending on the
> message volume and intended use. High-volume NNTP servers host millions
> and millions of messages, not merely tends or hundreds of thousands - and
> they deal with a large number of communications.
I suppose. But so do large imap servers. I don’t know who is using or
not dovecot at large, but it can use mbox files. Not what they
recommend, though.
>>> Leafnode isn’t intended to be used as a high-volume NNTP server, so it
>>> doesn’t matter if it uses straight file storage or not.
>>
>> Ha. Good point.
>>
>> However, I’d bet that a server using leafnode type structure on a
>> reiserfs would run as fast.
>
> I doubt it. Flat text file performance is never as good as a fully-
> indexed system.
The designers of Reiserfs claim that you can store databases directly on
it, using the filesystem as the base-root of the database system, using
files directly as the records (not the tupla (is that the name?) but the
individual cells.
>> I have not experimented with “inn” to know what it uses, though.
>
> I have, but INN isn’t what’s used here - what’s used here are high-end
> commercial servers from Highwinds Software.
I know
> 7.7> 2013-12-02 03:13:05 Telcontar fetchnews 17441 - - <200 Welcome to forums.novell.com! (Typhoon v2.1.2.394)
> <7.6> 2013-12-02 03:13:05 Telcontar fetchnews 17441 - - nntp.opensuse.org: 0 articles posted.
It identifies itself as “Typhoon”.
–
Cheers / Saludos,
Carlos E. R.
(from 12.3 x86_64 “Dartmouth” at Telcontar)
On Mon, 02 Dec 2013 02:28:08 +0000, Carlos E. R. wrote:
> I suppose. But so do large imap servers. I don’t know who is using or
> not dovecot at large, but it can use mbox files. Not what they
> recommend, though.
Exactly my point.
>> I doubt it. Flat text file performance is never as good as a fully-
>> indexed system.
>
> The designers of Reiserfs claim that you can store databases directly on
> it, using the filesystem as the base-root of the database system, using
> files directly as the records (not the tupla (is that the name?) but the
> individual cells.
That’s a database abstraction, really, not the same as using the
filesystem as a database. But either way, not an option with the
configuration we use.
>>> I have not experimented with “inn” to know what it uses, though.
>>
>> I have, but INN isn’t what’s used here - what’s used here are high-end
>> commercial servers from Highwinds Software.
>
> I know
>
>
>> 7.7> 2013-12-02 03:13:05 Telcontar fetchnews 17441 - - <200 Welcome to
>> forums.novell.com! (Typhoon v2.1.2.394)
>> <7.6> 2013-12-02 03:13:05 Telcontar fetchnews 17441 - -
>> nntp.opensuse.org: 0 articles posted.
>
>
> It identifies itself as “Typhoon”.
Yes, that’s the product name.
Jim Henderson
openSUSE Forums Administrator
Forum Use Terms & Conditions at http://tinyurl.com/openSUSE-T-C
On 2013-12-02 03:34, Jim Henderson wrote:
> On Mon, 02 Dec 2013 02:28:08 +0000, Carlos E. R. wrote:
>> The designers of Reiserfs claim that you can store databases directly on
>> it, using the filesystem as the base-root of the database system, using
>> files directly as the records (not the tupla (is that the name?) but the
>> individual cells.
>
> That’s a database abstraction, really, not the same as using the
> filesystem as a database. But either way, not an option with the
> configuration we use.
I know, I know. I’m only speculating on theories
–
Cheers / Saludos,
Carlos E. R.
(from 12.3 x86_64 “Dartmouth” at Telcontar)