headers
Lars Magne Ingebrigtsen | 1 Dec 2010 17:52
Face
Picon
Favicon
Gravatar

Re: API for adding gwene feeds?

Matt Lundin <mdl <at> imapmail.org> writes:

> I'm wondering (naively) if there's any undocumented api for adding rss
> groups to gwene. It would be quite lovely to be able to enter urls
> directly from Emacs/Gnus instead of the web form.

You often have to (or want to) fiddle a bit with the group names that
Gwene suggests, so I'm not sure what the interface for that would be
other than a web page...

--

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi <at> gnus.org * Lars Magne Ingebrigtsen
Permalink | Reply | 0 comments |
headers
Lars Magne Ingebrigtsen | 1 Dec 2010 17:52
Face
Picon
Favicon
Gravatar

http://prog21.dadgum.com/atom.xml

Ted Zlatanov <tzz <at> lifelogs.com> writes:

> This article has the full HTML in the original feed at
> http://prog21.dadgum.com/atom.xml.  I'm not sure if the full HTML was
> missing when Gwene got the article or if the parser was unhappy.

My guess would be that the feed was buggy when Gwene pulled it down.
Because that would be more convenient for me.  :-)

--

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi <at> gnus.org * Lars Magne Ingebrigtsen
Permalink | Reply | 0 comments |
headers
jidanni | 2 Dec 2010 03:55
Favicon
Gravatar

Re: gwene chewed up articles

>>>>> "LMI" == Lars Magne Ingebrigtsen <larsi <at> gnus.org> writes:
LMI> jidanni <at> jidanni.org writes:

LMI> I have no idea.  Gwene just uses curl to download the data.  I guess if
LMI> the host is very slow, curl might exit without reading the entire feed?

Does Gwene check the curl exit value etc.

Better yet, does Gwene even make sure it even has got the closing XML
tag at the end of the line with a link on it before processing it, not
to mention any of the higher level closing XML tags that should be all
seen before saying that we have all the information needed to make a
post. I mean those closing XML tags are there for a purpose.
Permalink | Reply | 2 comments |
headers
Steinar Bang | 2 Dec 2010 10:02
Picon
Picon
Favicon

Re: gwene chewed up articles

>>>>> jidanni <at> jidanni.org:

> ... I mean those closing XML tags are there for a purpose.

Yes.

To make XML an inappropriate format for streaming.
Permalink | Reply | 0 comments |
headers
sam tygier | 2 Dec 2010 13:20
Picon
Favicon

spam filtering

Hi

some gmane groups seem to get vast amounts of fairly obvious spam. eg gmane.comp.misc.oopic is the spam
filtering not picking up any of these?

also it would be great if there was a quicker way to mark spam (its gotten better, but not perfect yet), maybe a
way of selecting several messages and pressing a spam button once.

thanks

Sam
Permalink | Reply | 12 comments |
headers
Steinar Bang | 2 Dec 2010 13:33
Picon
Picon
Favicon

Re: spam filtering

>>>>> sam tygier <samtygier <at> yahoo.co.uk>:

> also it would be great if there was a quicker way to mark spam (its
> gotten better, but not perfect yet), maybe a way of selecting several
> messages and pressing a spam button once.

Since some use the spam reporting as a kind of walk-by-vandalism, I
think that having the reporting slow and cumbersome is actually a good
idea.
Permalink | Reply | 0 comments |
headers
Nicolas Alvarez | 3 Dec 2010 02:32
Picon
Gravatar

Re: More NNTP encryption on news.gmane.org

Lars Magne Ingebrigtsen wrote:
> Steinar Bang <sb <at> dod.no> writes:
> 
>> Ah... because STARTTLS uses good old port 119?
> 
> Yeah.  STARTTLS is a kinda gross hack, but it's a very, very nice gross
> hack.  I mean, you basically just connect to the old port, and then say
> STARTTLS (or a variant thereof -- it's a different command in every
> protocol, of course), and then you start doing the TLS stuff.  I think
> the people who came up with this hack probably feels very proud of
> themselves.
> 
> And they should!  :-)
> 

HTTP has a similar feature, that nobody in the world actually supports, but 
it's there in the spec...

--

-- 
Nicolas
Permalink | Reply | 0 comments |
headers
Olly Betts | 3 Dec 2010 12:21
Favicon
Gravatar

Re: gmane search results HITSPERPAGE

On 2010-11-28, Lars Magne Ingebrigtsen <larsi <at> gnus.org> wrote:
> Andrew Cohen <cohen <at> andy.bu.edu> writes:
>
>> Searching on gmane allows limiting the number of search results through
>> the HITSPERPAGE parameter. Going through the thread from 2006 it appears
>> that the max value for this is 1000. Any chance this could be increased?
>> It seems as if the limit was imposed for speed reasons. But maybe this
>> could go to 2000 without serious negative impact?

What's your intended use here?  Why is 2000 really any better than 1000?

> I guess so...  Olly?

Well, it could be raised to any arbitrary value, or removed entirely.
But we don't want people using search.gmane.org in an automated way (it's
just one machine), and it seems unlikely a human would want more than 1000
results per page.

There is a plan to set up a "search api" server - a higher (or perhaps
no) limit might make sense there.

Cheers,
    Olly
Permalink | Reply | 1 comment |
headers
Andrew Cohen | 3 Dec 2010 14:48

Re: gmane search results HITSPERPAGE

>>>>> "Olly" == Olly Betts <olly <at> survex.com> writes:

    Olly> On 2010-11-28, Lars Magne Ingebrigtsen <larsi <at> gnus.org> wrote:
    >> Andrew Cohen <cohen <at> andy.bu.edu> writes:
    >> 
    >>> Searching on gmane allows limiting the number of search results
    >>> through the HITSPERPAGE parameter. Going through the thread from
    >>> 2006 it appears that the max value for this is 1000. Any chance
    >>> this could be increased?  It seems as if the limit was imposed
    >>> for speed reasons. But maybe this could go to 2000 without
    >>> serious negative impact?

    Olly> What's your intended use here?  Why is 2000 really any better
    Olly> than 1000?

I picked 2000 more or less at random. I was doing some fairly generic
searches that were returning results in the few thousand range (it was
easier to pull down the several thousand results and then use gnus to
search through the summary buffer, since I didn't know precisely what I
was searching for).

    Olly> Well, it could be raised to any arbitrary value, or removed
    Olly> entirely.  But we don't want people using search.gmane.org in
    Olly> an automated way (it's just one machine), and it seems
    Olly> unlikely a human would want more than 1000 results per page.

That makes sense. I only suggested 2000 since my immediate "need" (which
really isn't hugely important) was in that range, but it wouldn't be a
big deal if the limit went even higher, or just stayed at 1000.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Lars Magne Ingebrigtsen | 3 Dec 2010 18:01
Face
Picon
Favicon
Gravatar

Re: gwene chewed up articles

jidanni <at> jidanni.org writes:

> Does Gwene check the curl exit value etc.

Gwene does not.  I've now added a check on $? (that's the right thing,
right?  I should learn Perl some day) and bails if it's not 0.

> I mean those closing XML tags are there for a purpose.

Yeah...  to take up space, and make you buy bigger hard disks!  It's a
conspiracy! 

--

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi <at> gnus.org * Lars Magne Ingebrigtsen
Permalink | Reply | 0 comments |

Gmane