Via nntp I only see a paragraph. On Jim reply I see several more paragraphs separated by lone dots.
I understood the dot issue had been already solved…
More. There is also a post from wa6mbz with only a “Hello All,” line. Where is the rest?
> Hello SUSE,
>
>
> Suse 12.2 still has a bug in the wireless networking aspect of things.
> If you sign up for wireless access to your router and then reboot your
> system, the computer forgets about the settings, the passwords and
> routers, just like 12.1 did.
> .
> To fix this mess, you still have to remove the old setting, take down
> the computer, ie, not just reboot, and then re-install the network info.
> Then it works until you reboot or shut down the next time.
> .
> This getting rather annoying and someone who knows what they are doing
> should get on it. I have noticed this on five of my older laptops,
> three toshibas and 2 lenovos. That means that it is not just me with a
> screwy network card.
> .
> Thank you in advance for your consideration.
Like caf, I’ve not observed this in 12.1 - have you reported it in
bugzilla? If so, what’s the bug number? If you haven’t, you might want
to after some steps are taken troubleshooting it here.
On Fri, 07 Sep 2012 00:40:41 +0000, Carlos E. R. wrote:
> Via nntp I only see a paragraph. On Jim reply I see several more
> paragraphs separated by lone dots.
> I understood the dot issue had been already solved…
>
> More. There is also a post from wa6mbz with only a “Hello All,” line.
> Where is the rest?
Both posts come back from the NNTP server properly, it’s probably your
leafnode cache that’s not handling it properly. Maybe that’s why you see
it when those of us who access the NNTP server directly don’t.
On 2012-09-07 02:58, Jim Henderson wrote:
> On Fri, 07 Sep 2012 00:40:41 +0000, Carlos E. R. wrote:
> Both posts come back from the NNTP server properly, it’s probably your
> leafnode cache that’s not handling it properly. Maybe that’s why you see
> it when those of us who access the NNTP server directly don’t.
No, I see the same issue when connected with pan directly without leafnode.
I deleted the group in pan, unsusbscribed it, told it to delete the articles, then subscribed again,
loaded the headers, and still those two articles are incomplete.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” (Minas Tirith))
On 2012-09-07 03:45, Carlos E. R. wrote:
> On 2012-09-07 02:58, Jim Henderson wrote:
> I deleted the group in pan, unsusbscribed it, told it to delete the articles, then subscribed again,
> loaded the headers, and still those two articles are incomplete.
>
I deleted all the groups in pan, exited the program (twice). On the second time, after readding the
server, now I can see the full post.
I don’t understand it.
–
Cheers / Saludos,
Carlos E. R.
(from 11.4 x86_64 “Celadon” (Minas Tirith))
If it works with knode, but not with claws or thunderbird, then it is probably a client problem rather than a server problem.
Since you already have knode installed, if you also have either claws or thunderbird installed you might try that. That will show wheter it varies between clients on the same computer with the same network connections.
On Fri, 07 Sep 2012 01:45:32 +0000, Carlos E. R. wrote:
> On 2012-09-07 02:58, Jim Henderson wrote:
>> On Fri, 07 Sep 2012 00:40:41 +0000, Carlos E. R. wrote:
>
>> Both posts come back from the NNTP server properly, it’s probably your
>> leafnode cache that’s not handling it properly. Maybe that’s why you
>> see it when those of us who access the NNTP server directly don’t.
>
> No, I see the same issue when connected with pan directly without
> leafnode.
What version of pan?
Pan is the newsreader I use, and I’m seeing the messages properly. I run
Pan 0.135.
> I deleted the group in pan, unsusbscribed it, told it to delete the
> articles, then subscribed again,
> loaded the headers, and still those two articles are incomplete.
Each time I load the headers, I see that they are complete.
On Fri, 07 Sep 2012 01:48:23 +0000, Carlos E. R. wrote:
> On 2012-09-07 03:45, Carlos E. R. wrote:
>> On 2012-09-07 02:58, Jim Henderson wrote:
>
>
>> I deleted the group in pan, unsusbscribed it, told it to delete the
>> articles, then subscribed again,
>> loaded the headers, and still those two articles are incomplete.
>>
>>
> I deleted all the groups in pan, exited the program (twice). On the
> second time, after readding the server, now I can see the full post.
>
> I don’t understand it.
With pan do you have both your local leafnode cache and the remote server
configured?
The reason I ask is that confronted with multiple servers hosting the
same groups, Pan picks one or the other when an article is present on
multiple servers. It was initially designed to pull multipart binaries
from alt.binaries groups (as a primary purpose), and in that setup when
there are multiple parts spread across multiple servers, but no server
has all the parts, and it transparently reconstructs the original group
of posts with multiparts.
This also works on forums.novell.com (since the groups are all on both
servers).
If you reproduce this and get anything other than the full article, let
me know. If you can’t, then it’s a client-side issue and a bug needs to
be logged against the client.
> If you reproduce this and get anything other than the full article, let
> me know. If you can’t, then it’s a client-side issue and a bug needs to
> be logged against the client.
i get it all via telnet, but did not via nntp (into TBird)–are you
saying i should log a bug against TB?
but but, i can see the lone dotted message in this thread ??
> On 09/07/2012 07:46 PM, Jim Henderson wrote:
>
>> If you reproduce this and get anything other than the full article, let
>> me know. If you can’t, then it’s a client-side issue and a bug needs
>> to be logged against the client.
>
> i get it all via telnet, but did not via nntp (into TBird)–are you
> saying i should log a bug against TB?
Yes. If you pull it via telnet the way I described, you’re imitating an
NNTP client (the NNTP protocol is a clear text/plain text protocol).
So if you get the message executing the same commands that Thunderbird is
using (which you are doing), then it’s a bug against thunderbird.
> but but, i can see the lone dotted message in this thread ??
If the telnet commands return the full message, then the server is doing
what it is supposed to be doing and the client you’re using is
interpreting the response incorrectly.
On Fri, 07 Sep 2012 21:36:01 +0000, nrickert wrote:
> hendersj;2484654 Wrote:
>> If the telnet commands return the full message, then the server is
>> doing what it is supposed to be doing and the client you’re using is
>> interpreting the response incorrectly.
>
> I’m not sure that’s completely correct.
>
> If the server is doing it correctly, then telnet should show those lines
> as beginning with “…” (two dots).
If that’s the case (though ISTR that the NNTP RFCs say that is optional
behaviour, but I could be wrong), then the bug is the NNTP software
itself - and I can pass that along but don’t hold much hope that that bug
would be addressed by Highwinds.
On Fri, 07 Sep 2012 21:53:17 +0000, Jim Henderson wrote:
> If that’s the case (though ISTR that the NNTP RFCs say that is optional
> behaviour, but I could be wrong), then the bug is the NNTP software
> itself - and I can pass that along but don’t hold much hope that that
> bug would be addressed by Highwinds.
OK, reading the RFC again, I see that it should be “dot-stuffed” (ie,
“…”) if the line is “.” CRLF. For some reason the NNTP server itself
doesn’t recognize this when it comes from the gateway “unstuffed”.
The server erroneously doesn’t recognize this and stuff it. In theory
the server should never return that format, but I think I can probably
work around it in the gateway.
I’m a little surprised the server doesn’t automatically convert it when
it sees it inbound from the gateway - because the termination string if
not converted would terminate the message getting into the server in the
first place.
Perhaps the best option is to convert ^.
to ^
rather than trying to
dot-stuff it since the server seems to get confused by that.
If I have time this weekend, I’ll see if there’s something further that I
can do or if I missed something in the fix that I provided before.
On Fri, 07 Sep 2012 22:18:05 +0000, Jim Henderson wrote:
> If I have time this weekend, I’ll see if there’s something further that
> I can do or if I missed something in the fix that I provided before.
Actually, had a quick look at it this evening, and it seems that perhaps
the solution I’d used might be confusing some newsreaders.
The fix I put in place replaced “.” on a line by itself with ". ".
It seems some newsreaders don’t cope with that well as they strip the
trailing space. Technically a newsreader should be able to cope with
this properly - moreover, I should be able to duplicate the issue using
NNTP only for those newsreaders that are perhaps broken, if my reader
will do what I ask it to (I’ll post a test message for those having
issues to look at).
Based on my recent read of the RFC, it should be replaced with
“…” (which is what my newsreader actually does).
Here’s a message that in theory (I’ll check the post myself from telnet
to see if it got posted the way I intended) should break the NNTP readers
that have problems.
…
This text is below a line with ". " as the only characters on the line.
