Not receiving salmon slaps from identi.ca

[eschnou]

Hi all,

[Please let me know if you would prefer me to ask these questions on statusnet-dev.]

I'm testing salmon interop between status.net and my node-ostatus library and I'm not really going far: I never receive any slaps :-( This is weird, because it used to work a couple weeks ago and I have not changed much to my code. I recall seeing an email from Andrea about statusnet not being swat0 anymore, is this related ?

- I post from the identi.ca web UI a message with mention of @lau...@eschenauer.be

- I see in my logs a webfinger request followed by a request for the user atom stream

But I never see any incoming salmon request, and don't see other errors.

Can anyone confirm if this is a general issue or just a problem on my end ?

Thanks,

Laurent

[Tuomas Koski]

Hi Laurent,

I tried to check your replies from my "hack platform" and while
requesting the host-meta file (from
http://eschenauer.be/.well-known/host-meta) I got HTTP/1.1 503 Service
Temporarily Unavailable

Let me know when all is up again and I'll give it another run.

Cheers,
--
Tuomas

[Laurent Eschenauer]

Hello Tuomas,

 

Let me know when all is up again and I'll give it another run.

Back online, thanks for letting me know how it works for you. Can you already confirm me that you are receiving slaps from identi.ca ?

You can also ping me by IM: I'm eschnou on freenode, and laurent.eschenauer at gmail.com on Gtalk.

Cheers,

Laurent

[Tuomas Koski]

Hi

On 16 March 2011 10:56, Laurent Eschenauer <laurent.e...@gmail.com> wrote:
> Hello Tuomas,
>
>>
>> Let me know when all is up again and I'll give it another run.

Discovery seemed to work nicely and just "slaped" you (endpoint
http://eschenauer.be/salmon/user/laurent). Did you get it?

On 16 March 2011 10:56, Laurent Eschenauer <laurent.e...@gmail.com> wrote:
> Back online, thanks for letting me know how it works for you. Can you
> already confirm me that you are receiving slaps from identi.ca ?

Yes I am. (http://www.lobstermonster.org/profile/koski)

On 16 March 2011 10:56, Laurent Eschenauer <laurent.e...@gmail.com> wrote:
> You can also ping me by IM: I'm eschnou on freenode, and laurent.eschenauer
> at gmail.com on Gtalk.

Added the gmail one.

Cheers,
--
Tuomas

[Laurent Eschenauer]

Discovery seemed to work nicely and just "slaped" you (endpoint

http://eschenauer.be/salmon/user/laurent). Did you get it?

Got your ping ! Thank you. So, it works, but strangely not from identi.ca, i'll have to dig deeper.

I failed at verifying your signature, but that is another story. I captured the packet and will try to debug this later. One of the reason may be that you are not removing padding when base64url encoding the pieces of M (this has changed lately).

> Can you
> already confirm me that you are receiving slaps from identi.ca ?

Yes I am. (http://www.lobstermonster.org/profile/koski)

Hmm.. will look into their source code to see if I miss something.

Thanks Tuomas !

Laurent

[Tuomas Koski]

Hi,

On 16 March 2011 11:28, Laurent Eschenauer <laurent.e...@gmail.com> wrote:
>> Discovery seemed to work nicely and just "slaped" you (endpoint
>> http://eschenauer.be/salmon/user/laurent). Did you get it?
>
> Got your ping ! Thank you. So, it works, but strangely not from identi.ca,
> i'll have to dig deeper.
> I failed at verifying your signature, but that is another story. I captured
> the packet and will try to debug this later. One of the reason may be that
> you are not removing padding when base64url encoding the pieces of M (this
> has changed lately).

yeah, this might be possible. I have not touched it so much since the
SWAT0 times. My plan is to update all of these in the near future now
when I have a better idea how the architecture should be done.

On 16 March 2011 11:28, Laurent Eschenauer <laurent.e...@gmail.com> wrote:
>> > Can you
>> > already confirm me that you are receiving slaps from identi.ca ?
>>
>> Yes I am. (http://www.lobstermonster.org/profile/koski)
>
> Hmm.. will look into their source code to see if I miss something.

It's quite messy (sorry!) so you might loose time there (the minime
source i mean).

Maybe it would be the best to ask directly from the statusnet-dev mailing list.

Cheers,
--
Tuomas

[Evan Prodromou]

Laurent,

We did an upgrade from StatusNet 0.9.6 to 0.9.7beta a few weeks ago. I'm testing unsolicited slaps (no subscription) across StatusNet sites and seeing the same problem.

I'll see if we can get that fixed ASAP.

-Evan

[Brion Vibber]

On Wed, Mar 16, 2011 at 6:47 AM, Evan Prodromou <ev...@status.net> wrote:

Laurent,

We did an upgrade from StatusNet 0.9.6 to 0.9.7beta a few weeks ago. I'm testing unsolicited slaps (no subscription) across StatusNet sites and seeing the same problem.

On further investigation, StatusNet 0.9.7 seems to be working ok among itself; at least we can find no problems doing several ad-hoc tests both within our hosted cluster and between that and a StatusNet 1.0.x test site.

Trying to poke Laurent's site now I just get the HTTP 503 response so can't fully test; let us know when it's back online and we'll be happy to help with testing & debugging.

If discovery all goes right, be aware that StatusNet 0.9.7 will attempt two slaps, first using the current magic sig formula, then again using an earlier salt spec if the first one fails to validate. If you're not even seeing the first slap, then something's probably going awry in discovery...

-- brion vibber (brion @ status.net)

[James Walker]

One thing (potentially) holding things up is - if you've moved the
location of your salmon endpoint at all StatusNet caches that value
very aggressively. I've found running a local StatusNet instance is
super helpful for debugging stuff like this.

There are a few other places where discovery will fail if certain bits
of info don't match - which are tough to debug remotely ...

--
James Walker :: http://walkah.net/

[eschnou]

Thanks Evan, Brion and James for looking into this.

 

> On further investigation, StatusNet 0.9.7 seems to be working ok among
> itself; at least we can find no problems doing several ad-hoc tests both
> within our hosted cluster and between that and a StatusNet 1.0.x test site.

One thing (potentially) holding things up  is - if you've moved the

location of your salmon endpoint at all StatusNet caches that value
very aggressively. I've found running a local StatusNet instance is
super helpful for debugging stuff like this.

Indeed, I keep on moving things around :-) I'll setup my own instance of status.net, it will definitively make it easier to debug. 

Cheers,

Laurent

[James Walker]

Cool - keep up the great work! I'm excited to see node-ostatus coming along :)

[Brion Vibber]

Yeah that can do it -- I'll think about how to fix our end to rerun discovery on failure, should help in such situations.

- brion

[eschnou]

- I post from the identi.ca web UI a message with mention of @lau...@eschenauer.be

- I see in my logs a webfinger request followed by a request for the user atom stream

But I never see any incoming salmon request, and don't see other errors.

Problem solved. Having my own statusnet instance and stepping through the code did help. I solved my problem by adding the following link in my activities atom stream:

<link href="http://eschenauer.be/salmon/user/laurent" rel="http://salmon-protocol.org/ns/salmon-replies"/>

I thought that this rel value was deprecated :-) Any plans to switch/support rel='salmon' ?

Looking in more details in the status.net discovery flow, I have another issue: When you do a webfinger discovery, if a feed is discovered, then statusnet switch to a feed based discovery and ignore the content of the XRD. This explains why my salmon links exposed in my XRD never got discovered.

http://gitorious.org/statusnet/mainline/blobs/master/plugins/OStatus/classes/Ostatus_profile.php#line1651

I hope to have some time over the weekend to look into signatures. 

Thanks for your help,

Cheers,

Laurent

[Tuomas Koski]

Hi Laurent,

On 18 March 2011 09:26, eschnou <laurent.e...@gmail.com> wrote:
>> - I post from the identi.ca web UI a message with mention of
>> @lau...@eschenauer.be
>> - I see in my logs a webfinger request followed by a request for the user
>> atom stream
>> But I never see any incoming salmon request, and don't see other errors.
>
> Problem solved.

Good news!

On 18 March 2011 09:26, eschnou <laurent.e...@gmail.com> wrote:
>Having my own statusnet instance and stepping through the
> code did help. I solved my problem by adding the following link in my
> activities atom stream:
> <link href="http://eschenauer.be/salmon/user/laurent"
> rel="http://salmon-protocol.org/ns/salmon-replies"/>
>
> I thought that this rel value was deprecated :-) Any plans to switch/support
> rel='salmon' ?
> Looking in more details in the status.net discovery flow, I have another
> issue: When you do a webfinger discovery, if a feed is discovered, then
> statusnet switch to a feed based discovery and ignore the content of the
> XRD. This explains why my salmon links exposed in my XRD never got
> discovered.
> http://gitorious.org/statusnet/mainline/blobs/master/plugins/OStatus/classes/Ostatus_profile.php#line1651

Dose this "service" discovering your data correctly:
http://www.madebymonsieur.com/ostatus_discovery/ ?

On 18 March 2011 09:26, eschnou <laurent.e...@gmail.com> wrote:
> I hope to have some time over the weekend to look into signatures.

I start to work on this the coming weeks too, so all the feedback is
much appreciated.

Cheers,
--
Tuomas

[Matthias Pfefferle]

Indeed, Status.NET uses the Atom-Feed if you add one in your host-meta-
file. I ran in a very similar problem while working on the OStatus
WordPress Plugin and I solved it by adding all necessary meta-infos to
the Atom-Feed (salmon-links, hub-links, Activity-Stream Extension,
PoCo, ...). You can have a look here: http://notizblog.org/author/matthias-pfefferle/feed/atom/

To implement the old spec it is important to add an
<activity:subject>...</activity:subject> to your feed, the new spec
uses the enriched <author> tag:

<author>
<name>Max Mustermann</name>
<uri>http://example.com</uri>
<activity:object-type>http://activitystrea.ms/schema/1.0/person</
activity:object-type>
<link rel='alternate' type='text/html' href='http://example.com' />
<link rel='avatar' type='image/jpeg' media:width='300'
media:height='300' href='...' />
<link rel='avatar' type='image/jpeg' media:width='96'
media:height='96' href='...'/>
<link rel='avatar' type='image/jpeg' media:width='48'
media:height='48' href='...'/>
<link rel='avatar' type='image/jpeg' media:width='24'
media:height='24' href='...'/>
<poco:preferredUsername>Max-Mustermann</poco:preferredUsername>
<poco:displayName>Max Mustermann</poco:displayName>
<poco:note><![CDATA[...]]></poco:note>
</author>

I hope this could help!

[Andrea Messina]

Hi Mathias. What is the URI to follow your blog??

http://notizblog.org: --> http://notizblog.org/?webfinger-uri=http%3A%2F%2Fnotizblog.org%2F
return a BAD XML
matthias-...@notizblog.org -->
http://notizblog.org/?webfinger-uri=acct%3Amatthias-pfefferle%40notizblog.org
TOO.

I think you should setup an XRD

[Matthias Pfefferle]

Hi Andrea,

yo can follow every profile page on my blog (mine is
http://notizblog.org/author/matthias-pfefferle/) or
username@blogdomain (mine is pfef...@notizblog.org) but, as you
mentioned it, i think i should also add the nice-name (what is used to
build the profile url (matthias-...@notizblog.org))

On 18 Mrz., 12:10, Andrea Messina <messinand...@gmail.com> wrote:
> Hi Mathias. What is the URI to follow your blog??
>
> http://notizblog.org:-->http://notizblog.org/?webfinger-uri=http%3A%2F%2Fnotizblog.org%2F
> return a BAD XML

> matthias-pfeffe...@notizblog.org -->http://notizblog.org/?webfinger-uri=acct%3Amatthias-pfefferle%40notiz...

[Andrea Messina]

I have noticed some problem during the subscription process. Especially, regarding with the Salmon follow message.

<?xml version="1.0" encoding="UTF-8" ?>

<entry xmlns="http://www.w3.org/2005/Atom" xmlns:activity="http://activitystrea.ms/spec/1.0/" xmlns:poco="http://portablecontacts.net/spec/1.0"> 

<id>domain:tilab.no-ip.org,follower_id:http://tilab.no-ip.org/teamlifesvn/profile.php?uid=215,topic:http://notizblog.org/author/matthias-pfefferle/feed/atom/,date:2011-03-18T14:08:09+00:00</id> 

<title>Follow</title> 

<published>2011-03-18T14:08:09+00:00</published>

<updated>2011-03-18T14:08:09+00:00</updated> 

<content type="html">ndria is now following Matthias Pfefferle.</content> 

<author> 

 <activity:object-type>http://activitystrea.ms/schema/1.0/person</activity:object-type>

 <uri>http://tilab.no-ip.org/teamlifesvn/profile.php?uid=215</uri> 

 <name>ndria</name> 

 <link rel="alternate" type="text/html" href="http://tilab.no-ip.org/teamlifesvn/profile.php?uid=215"/> 

 <link rel="avatar" href="http://tilab.no-ip.org/teamlifesvn/albums/userpics/100000215/user_100000215_avatar.jpg"/>

 <poco:preferredUsername>ndria</poco:preferredUsername> 

 <poco:displayName>ndria</poco:displayName> 

</author> 

<activity:verb>http://activitystrea.ms/schema/1.0/follow</activity:verb> 

<activity:object> 

 <activity:object-type>http://activitystrea.ms/schema/1.0/person</activity:object-type> 

 <id>http://notizblog.org</id> 

 <title>Matthias Pfefferle</title> 

 <link rel="alternate" type="text/html" href="http://notizblog.org/author/matthias-pfefferle/"/> 

 <poco:preferredUsername>matthias-pfefferle</poco:preferredUsername> 

 <poco:displayName>Matthias Pfefferle</poco:displayName> 

</activity:object>

</entry>

RESPONSE:

HTTP info:

Array

(

    [url] => http://notizblog.org/author/matthias-pfefferle/?salmon=endpoint

    [content_type] => text/html; charset=UTF-8

    [http_code] => 400

    [header_size] => 661

    [request_size] => 213

    [filetime] => -1

    [ssl_verify_result] => 0

    [redirect_count] => 0

    [total_time] => 1.28203

    [namelookup_time] => 0.172203

    [connect_time] => 0.22128

    [pretransfer_time] => 0.22134

    [size_upload] => 2552

    [size_download] => 38

    [speed_download] => 29

    [speed_upload] => 1990

    [download_content_length] => -1

    [upload_content_length] => 2552

    [starttransfer_time] => 0.279228

    [redirect_time] => 0

)

I am compliant with statusnet & friendika.

2011/3/18 Matthias Pfefferle <pfef...@gmail.com>

pfef...@notizblog.org

--
Politecnico di Torino

Andrea Messina

[Matthias Pfefferle]

Oops, it seems to be a bug in the response, but I received both
following messages.

On 18 Mrz., 14:10, Andrea Messina <messinand...@gmail.com> wrote:
> I have noticed some problem during the subscription process. Especially,
> regarding with the Salmon follow message.
>
> <?xml version="1.0" encoding="UTF-8" ?>
> <entry xmlns="http://www.w3.org/2005/Atom" xmlns:activity="http://activitystrea.ms/spec/1.0/" xmlns:poco="http://portablecontacts.net/spec/1.0">

> <id>domain:tilab.no-ip.org,follower_id:http://tilab.no-ip.org/teamlifesvn/profile.php?uid=215,topic:http://n...

> </id>
> <title>Follow</title>
> <published>2011-03-18T14:08:09+00:00</published>
> <updated>2011-03-18T14:08:09+00:00</updated>
> <content type="html">ndria is now following Matthias Pfefferle.</content>
> <author>
>  <activity:object-type>http://activitystrea.ms/schema/1.0/person
> </activity:object-type>
>  <uri>http://tilab.no-ip.org/teamlifesvn/profile.php?uid=215</uri>
>  <name>ndria</name>
>  <link rel="alternate" type="text/html" href="http://tilab.no-ip.org/teamlifesvn/profile.php?uid=215"/>

>  <link rel="avatar" href="http://tilab.no-ip.org/teamlifesvn/albums/userpics/100000215/user_100...

> 2011/3/18 Matthias Pfefferle <pfeffe...@gmail.com>
>
> > pfeffe...@notizblog.org
>
> --
> *Politecnico di Torino*
> *Andrea Messina*

[Matthias Pfefferle]

fixed, now it should return a http 20x code. (follow-messages and @-
replys: 202 Accepted and comments: 201 created)

On 18 Mrz., 14:10, Andrea Messina <messinand...@gmail.com> wrote:
> I have noticed some problem during the subscription process. Especially,
> regarding with the Salmon follow message.
>
> <?xml version="1.0" encoding="UTF-8" ?>
> <entry xmlns="http://www.w3.org/2005/Atom" xmlns:activity="http://activitystrea.ms/spec/1.0/" xmlns:poco="http://portablecontacts.net/spec/1.0">

> <id>domain:tilab.no-ip.org,follower_id:http://tilab.no-ip.org/teamlifesvn/profile.php?uid=215,topic:http://n...

> </id>
> <title>Follow</title>
> <published>2011-03-18T14:08:09+00:00</published>
> <updated>2011-03-18T14:08:09+00:00</updated>
> <content type="html">ndria is now following Matthias Pfefferle.</content>
> <author>
>  <activity:object-type>http://activitystrea.ms/schema/1.0/person
> </activity:object-type>
>  <uri>http://tilab.no-ip.org/teamlifesvn/profile.php?uid=215</uri>
>  <name>ndria</name>
>  <link rel="alternate" type="text/html" href="http://tilab.no-ip.org/teamlifesvn/profile.php?uid=215"/>

>  <link rel="avatar" href="http://tilab.no-ip.org/teamlifesvn/albums/userpics/100000215/user_100...