hepserver-devel Mailing List for Hep Message Server

SourceForge Headquarters 1320 Columbia Street Suite 310 San Diego, CA 92101 +1 (858) 422-6466

Call out Gouranga be happy!!!
Gouranga Gouranga Gouranga ....
That which brings the highest happiness!!

I've got a new release of Hep available at 
http://www.fettig.net/projects/hep/.  This is pretty much all-new code, 
and the architecture is a LOT better.  It may not be a lot to look at 
yet, but it's as functional as the 0.3 series was.  So give it a try.

I'm also releasing a new library called Yarn 
(http://www.fettig.net/projects/yarn/), a seperate package containing 
all the messaging-system-abstraction bits that used to be in Hep.  Yarn 
lets you work with RSS and Atom feeds, weblogs, Maildir directories, 
etc., using a common Folder/Message API.  It supports authentication, 
sharing connections between users, and tracking metadata between 
sessions.  It's fully unicode-aware.  It uses plugins to make it easy to 
add support for new protocols and file formats.

Yarn is a little short on docs at the moment, but if you're feeling 
curious there's some code in the examples/ directory that will get you 
started.  I'll have some tutorials  up on my website soon, though.

Abe

Hi,

I've just started using Hep within the last few days and it's very
awesome stuff; viewing blog entries from my email client, mutt, is
just great. And I look forward even more to being able to send out
replies from within mutt to the Hep server and have them show up as
comments on the blogs (though I have no idea how feasible this is
with current RSS/XML-RPC/etc. blog APIs, I'm still a newbie when
it comes to such things).

So, to tangent to the point of my email, my boss mentioned today
that he'd really like to send email into our wiki, e.g. auto-append
stuff to a page like 'ToDo' or what not.

At first I thought writing a simple script to poll an email address
and post the results directly into the page wouldn't be hard, Hep
came to mind as a more elegant (fun) way to go about doing it.

E.g. have a folder like 'wikis/foo' that email could be sent into,
which triggers sending in to the folder's desintation, a HTTP
post into a wiki.

(I read the general source -> folder -> destination overview on the
webpage, but I haven't looked closely yet to really understand how
they relate, fire events amongst themselves, etc.)

Is such a thing fairly doable with the current state of Hep? Is it
at the point I could realistically start playing around with a wiki
destination, etc., or are you still in the midst of refactoring and
would rather have the dust settle for awhile before letting other
people meander through the code?

Thanks,
Stephen

Hi folks,

This list has been pretty quiet of late, so I wanted to pop in with a
quick update on Hep.

There are a couple of changes brewing that you should know about. 
First, I've decided to start taking advantage of more of Twisted -
plugins, the woven web development framework, authorization, etc. 
Basically, if it's in Twisted, and relevant to Hep, I'm going to use it,
rather than re-inventing the wheel myself.  I have enough experience
with Twisted now to feel comfortable that the developers know what
they're doing (and generally are smarter than me :-)).  Hopefully Hep
will end up being a good example of How To Do Cool Things With Twisted. 
For more details, see the Twisted mailing list, starting with this
message:

http://www.twistedmatrix.com/pipermail/twisted-python/2003-June/004452.html

The other change coming has to do with connections and folders.  In the
current Hep CVS, when you create a connection you automatically get a
folder under "Connections" that contains a complete archive of all past
and present messages from that URL.  I've come to realize that there's
some flaws with that design, the formost of which is a lack of
flexibility.  Some people may want to continue to use the POP3
store-and-download model for pulling messages, or have new messages be
automatically sent to an email address.

So what I'm doing is seperating connections from folders completely. 
Connections let you access the current messages at a URL.  Folders let
you store messages on the Hep server.  Living in between folders and
connections will be Filters.  Filters will be message handling plugins
that decide what happens when a new message appears on a connection. 
You'll be able to choose which filters are active on which connections.

Examples:

* Archive Mailing Lists: When a message from a mailing list is recieved,
put it in a folder named "Mailing Lists/<ListName>", creating the folder
if necessary
* Send new messages to this URL (useful for aggregating RSS feeds into a
single weblog, or sending messages to an e-mail address
* Spam Filtering
* Archive News: Put messages in a folder with the same name as the
message source (this would give the same results as the current Hep
CVS).

I've also been slowly but surely improving a lot of the core Hep classes
to make the whole thing nicer and more flexible.  Some of my changes
have gone into CVS, but many have not as they break things and/or
require files to be moved.  I've been holding off for the sake of people
running Hep from CVS.  So a warning:  If you've been running Hep out of
CVS, you should either stop updating for a while, or be prepared to have
things break.

I've just finished of a very busy month at my day job, so I should have
more time going forward.  Expect to see some cool things soon!

Abe

On Fri, 2003-06-06 at 02:09, Wari Wahab wrote:
> Abe Fettig wrote:
> 
> >Sorry for taking so long to reply...
> >
> No problem
> 
> >There isn't any way to configure this on a per-connection basis, but the
> >default is 45 minutes IIRC.  I haven't ever had a problem with
> >Slashdot... 
> >
> Hmmm, that's wierd, but where is the configuration for it, I think the 
> old hep server uses 15 mins for grabbing feeds, and now I migrated the 
> datafiles with the new one. Slashdot give me warnings for about 5 times, 
> and later denies serving me for 72 hours.

See heplib/agents/scanner.py:

	FREQUENCY = 60*45

> >are you pulling more than 1 RSS feed?
> >  
> >
> I've got about 50++ feeds

Sorry, I meant to say "Are you pulling more than 1 feed from Slashdot?" 
I remember reading about a similar problem somewhere a while ago, where
a person was subscribed to multiple feeds at Slashdot, and Slashdot
(incorrectly) thought they were pulling the same feed several times.

Abe

On Fri, 2003-06-06 at 02:33, Wari Wahab wrote:
> Was just looking at the source today to see what's happening with Hep, 
> then I saw this wierd code in hep.py itself.
> 
> from heplib.agents import scanner
>     for agent in [scanner]:
>         pass
>         agent.start(data)
> 
> I'm not sure why this is done this way instead of just doing 
> scanner.start(data). Any future plans for this? Sorry if you find this 
> to be a nitpick :)

Yeah, in the future there will be more agents than scanner, that's why I
did it that way.  also the 'pass' line is in there to make it easy to
comment out the agent.start line, if I want to test Hep without any
agents running.

Also, I've been thinking about getting rid of the distinction between
agents and servers.  I'll probably do so soon.

Abe

Was just looking at the source today to see what's happening with Hep, 
then I saw this wierd code in hep.py itself.

from heplib.agents import scanner
    for agent in [scanner]:
        pass
        agent.start(data)

I'm not sure why this is done this way instead of just doing 
scanner.start(data). Any future plans for this? Sorry if you find this 
to be a nitpick :)

-- 
Regards: Wari Wahab
Senior R&D Engineer
Celestix Networks
http://www.celestix.com/

Abe Fettig wrote:

>Sorry for taking so long to reply...
>
No problem

>There isn't any way to configure this on a per-connection basis, but the
>default is 45 minutes IIRC.  I haven't ever had a problem with
>Slashdot... 
>
Hmmm, that's wierd, but where is the configuration for it, I think the 
old hep server uses 15 mins for grabbing feeds, and now I migrated the 
datafiles with the new one. Slashdot give me warnings for about 5 times, 
and later denies serving me for 72 hours.

>are you pulling more than 1 RSS feed?
>  
>
I've got about 50++ feeds

-- 
Regards: Wari Wahab
Senior R&D Engineer
Celestix Networks
http://www.celestix.com/
vim: syntax=mail tw=72

Sorry for taking so long to reply...

There isn't any way to configure this on a per-connection basis, but the
default is 45 minutes IIRC.  I haven't ever had a problem with
Slashdot... are you pulling more than 1 RSS feed?

On Mon, 2003-05-05 at 22:16, Wari Wahab wrote:
> Is there a way to configure Hep's default rss grabbing timing to be more 
> than 30 mins (actually if I'm not wrong, the default is still at every 
> 15 minutes right?)
> 
> ______________________________________________________________________
> From: Slashdot
> Subject: Your Headline Reader Has Been Banned
> Date: 04 May 2003 07:28:08 +0000
> 
> Your RSS reader is abusing the Slashdot server. You are requesting pages more often than our terms of service allow. Please see the FAQ link for more information, and if you email us, include your IPID MD5: 349c2dc95fedbd189a6a51ed03a2625d.
> 
> http://slashdot.org/faq/accounts.shtml#ac1050
> 

On Thu, 2003-05-15 at 15:01, Curtis Seyfried wrote:
> Abe,
> Well it works. I did what you recommended and got "can't open file" and / 
> or "syntax error".  I eventually found that there was something wrong with 
> my Python 2.2 installation, so I reinstalled it and Twisted and then HEP 
> worked fine.

Cool, I'm glad it wasn't something on my end :-)

> I'm reading through the material you've written on HEP and the app looks 
> very interesting and promising. It'll take me a while of playing with it to 
> see how I can use it.
> 
> Can I create custom message sources ?  like pdf's from a specific Website 
> ?  Can I use HEP to index, or at least make sense of  some large 
> collections of HTML files - news articles,  documents, manuals, etc. - that 
> reside on my HD ?

A lot of the stuff I've been writing lately applies to Hep 0.4, which
hasn't been released yet.  I've done a lot of work on this new version,
and it basically works, but there are still a few features to add, and
bugs and issues with performance that I'm trying to iron out before I do
a release.  If you're interested in playing with the current code, it's
available through CVS - see the April archive on my website for details
on how to check out and run the code.

Creating custom message sources is certainly possible.  Are you thinking
of pulling the text from PDFs, or accessing the raw files?  At the
moment Hep is set up to handle text and html messages only.  I've
thought about adding support for images and other binary formats like
PDF someday, but that would definitely be a post-1.0 feature - in other
words, at least 6 months away.

As for using Hep to access documents on your hard drive - the problem
there is security implications.  From a technical standpoint, the only
thing preventing you from accessing on-disk files today is that I've
specifically disabled file:// Urls.  This is because some people run Hep
as a multi-user server, and they might not want everybody with a Hep
account to be able to arbitrarily access files on their Hep server's
disk.

I've been mulling the idea of creating an "Admin" account in Hep, that
could do things like set up connections to local files, and possibly
share them with other users.  That would probably give you what you want
without opening a big security hole.

Abe

Abe,
Well it works. I did what you recommended and got "can't open file" and / 
or "syntax error".  I eventually found that there was something wrong with 
my Python 2.2 installation, so I reinstalled it and Twisted and then HEP 
worked fine.

I'm reading through the material you've written on HEP and the app looks 
very interesting and promising. It'll take me a while of playing with it to 
see how I can use it.

Can I create custom message sources ?  like pdf's from a specific Website 
?  Can I use HEP to index, or at least make sense of  some large 
collections of HTML files - news articles,  documents, manuals, etc. - that 
reside on my HD ?

At 5/15/2003, you wrote:
>Hi Curtis,
>
>Thanks for your interest in Hep, and for sending a nice detailed bug
>report. It sounds like you've done everything right so far.  To debug
>the problem you're having, the best thing to do is run Hep from the
>command line:
>
>cd e:\program files_MC\Hep
>python hep-add-user.py
>
>This will let you see the error that Hep is generating.  Send me the
>error, and I'll try to figure out what's going on.
>
>Thanks!
>
>Abe
>
>On Thu, 2003-05-15 at 06:59, Curtis Seyfried wrote:
> > Hi, Abe,
> > I was reading your weblog posts about HEP and it sounds very
> > interesting. I downloaded the zip file and unzipped it into E:\program
> > files_MC\Hep
> >
> > OH, I am running a win2K pro standalone workstation box SP3 with all
> > updates, patches and other stupid stuff you must do to keep a working
> > MS Win machine. I have a Dell Dimension XPS T700r  PIII-700, 512 meg.
> > RAM. DSL for Internet connection. MS Office 2000 pro, Open Office 1.2
> > beta. Symantec Norton AV 2003 w/ NU 2003, Zone Alarm Pro v.3.7.143
> > firewall along with Symantec Norton Internet Security and Tiny
> > Firewall as backup firewalls.
> >
> > Python 2.2 is installed on D:\program files4\Python22, and Twisted is
> > installed into the same directory.  I use Eudora pro 5.2.1 for e-mail
> > and have IE6, MoZZilla 1.3 and 1.4 Opera 7 and  a few other browsers.
> > I have Apache HTTP Server installed, configured and runable, though I
> > am not using at the moment, also Jakarta Tomcat 4.1.8 installed
> > configured and runable only when I run the apps it is used for.  Also
> > a custom config of Tomcat called Schroedinger, it has a Java Swing GUI
> > which allows you to manage Tomcat and view all error messages.
> > *******
> > I configured the hep.ini file to point to a data directory in HEP's
> > directory, and it is created. I configured the 1st line of both hep.py
> > and hep-add-user.py to read     #!D:/program files7/Python22/python
> >
> > I then tried to run hep-add-user.py, a python CM window opens, some
> > messages fly by and it closes. No prompts for username and password. I
> > tried anyway to run hep.py same results.
> >
> > Do you know how I can get python to log the error messages that fly by
> > when I run hep-add-user.py ?
> >
> > Is there anything you can recommend that I might be doing wrong ? I'm
> > not familiar with Python for programming, I use it to run ZOPE and
> > Easy Publisher.
> > *******
> > You mentioned ZOE and Spaces. I have both. I use ZOE to index and
> > search all my e-mail, it is great, and still a work in progress.  I'm
> > also using PopFile to filter and classify my incoming e-mail.
> >
> > Spaces is now Clevercactus, a cat, and has progressed well but not to
> > the point where I trust it enough to use as a sole e-mail client. But,
> > it is coming along well and will make a very good OUTLOOK replacement
> > with additional features. Being Java it won't be so susceptible to
> > virus attacks waged against OutLook. I have Outlook installed but
> > stopped using it because of too many problems.
> >
> >
> >
> > ------------------
> > All Outgoing mail, downloaded files and e-mail attachments are
> > certified
> > Virus Free.  Checked by Symantec Norton Anti-virus 2003 using the
> > latest virus definition
> > list.
> > ------------------------
> > All Incoming mail, downloaded files and e-mail attachments are
> > certified
> > Virus Free.  Checked by Symantec Norton Anti-virus 2003 using the
> > latest virus definition
> > list.
> > Curtis Seyfried, B.Sc., MA., Paralegal <
> > mailto:cur...@ve...>
> > Bronx, New York, NY.
> > Anti-Spam sites to goto:
> > http://thisurlenablesemailtogetthroughoverzealousspamfilters.org or
> > http://wecanstopspam.org
>--
>Abe

Yeah, in general, let's say that I like technical articles but not so much the personal stuff.  I could have hep flag what it deems "non-technical" and throw that in a different folder, or I could say that I want to categorize Java stuff separately.  Or what could be really cool is to come up with my own categories and have hep assign incoming posts to categories based on what it thought the category should be.  To do this using bayesian filtering, I think that you'd have to build a chain, or maybe a tree, of filters.  For instance, you could create a heirarchy like this: 

everything  
   Technical 
      Java 
         J2EE 
      .NET 
      Functional Programming 
   Non - Technical 
      Sports 
         Football 
         Baseball 
      Politics 

The filters would take a best guess at where a message belonged and put it in the correct folder.  It's basically the reverse of how bloggers put categories on messages to organize their content - could I do the same thing on incoming messages?  

Like I say, this is all very much in the speculative stage.  It could definitely be done on the client side, with something like a mail reader plugin.  I'm thinking that hep might have a better view of the incoming data though and provide a nice integration point.  Plus, with IMAP support, I could expose all those categories as folders and provide the same service at a central server.  What I need to do is get off my butt and do some prototyping (darn day job gets in the way!), see if I can get any encouraging results, and then worry about how to integrate it... 

-----Original Message----- 

> Abe Fettig wrote: 

So, for example, you could make a "Really Interesting News" folder, and 
put messages that you thought were interesting in there, and then the 
Bayesian filter would compare each incoming message to the contents of 
that folder and determine if the message was Really Interesting or not. 
Is that the sort of thing you'd want to do? 

On Mon, 2003-05-12 at 00:20, Gordon Weakliem wrote:
> Abe Fettig wrote:
> 
> >>2) I've been thinking about experimenting with integrating Bayesian 
> >>classification into Hep.  I've been studying the source for
> Spambayes, 
> >>but I'm wondering if/how it could fit into Hep's architecture.  I
> guess 
> >>the easiest answer is to put Spambayes' POP3 filter in front of the
> POP3 
> >>protocol - do you think that there's a way to get it all into the
> same 
> >>package?  I think that to get Bayesian filtering to work, you'd have
> to
> >>    
> >>
> >
> >My feeling is that it probably makes sense to put any kind of
> filtering
> >into the Hep core, rather than tying it to POP3 or another protocol. 
> >Could you give me a better idea of what you could use Bayesian
> filtering
> >for?  I know basically how it's used for spam detection, but to the
> best
> >of my knowledge nobody is using Hep to read e-mail at the moment...
> >  
> >
> The point is more managing information - Graham's paper was
> specifically 
> targeting spam but since then there's been various people talking
> about 
> extending the idea to just general classification.  My idea was 
> specifically that you could create categories and then have a program 
> tag incoming messages based on what category it thought they should be
> in.  I have no idea if it would work, but I've been kicking the idea 
> around and thought that it would be worth experimenting with.

So, for example, you could make a "Really Interesting News" folder, and
put messages that you thought were interesting in there, and then the
Bayesian filter would compare each incoming message to the contents of
that folder and determine if the message was Really Interesting or not.
Is that the sort of thing you'd want to do?

Abe Fettig wrote:

>>2) I've been thinking about experimenting with integrating Bayesian 
>>classification into Hep.  I've been studying the source for Spambayes, 
>>but I'm wondering if/how it could fit into Hep's architecture.  I guess 
>>the easiest answer is to put Spambayes' POP3 filter in front of the POP3 
>>protocol - do you think that there's a way to get it all into the same 
>>package?  I think that to get Bayesian filtering to work, you'd have to
>>    
>>
>
>My feeling is that it probably makes sense to put any kind of filtering
>into the Hep core, rather than tying it to POP3 or another protocol. 
>Could you give me a better idea of what you could use Bayesian filtering
>for?  I know basically how it's used for spam detection, but to the best
>of my knowledge nobody is using Hep to read e-mail at the moment...
>  
>
The point is more managing information - Graham's paper was specifically 
targeting spam but since then there's been various people talking about 
extending the idea to just general classification.  My idea was 
specifically that you could create categories and then have a program 
tag incoming messages based on what category it thought they should be 
in.  I have no idea if it would work, but I've been kicking the idea 
around and thought that it would be worth experimenting with.

On Fri, 2003-05-09 at 16:18, Gordon Weakliem wrote:
> A couple questions:
> 
> 1) What's the status of IMAP support in Hep?  The comment in 
> heplib.servers.imap says "includes a full-blown twisted imap server, 
> which should get migrated upstream eventually." - is this waiting on 
> Twisted?

That comment dates back a few months to when I thought I'd have to
implement IMAP support myself.  Since then IMAP has been added to
Twisted (it's in CVS, but not in a release yet).  So now I can use
Twisted's IMAP protocol module.  The only thing keeping IMAP support out
of Hep at this point is my schedule - it's probably 10 or so hours of
work to get it up and running.

> 2) I've been thinking about experimenting with integrating Bayesian 
> classification into Hep.  I've been studying the source for Spambayes, 
> but I'm wondering if/how it could fit into Hep's architecture.  I guess 
> the easiest answer is to put Spambayes' POP3 filter in front of the POP3 
> protocol - do you think that there's a way to get it all into the same 
> package?  I think that to get Bayesian filtering to work, you'd have to

My feeling is that it probably makes sense to put any kind of filtering
into the Hep core, rather than tying it to POP3 or another protocol. 
Could you give me a better idea of what you could use Bayesian filtering
for?  I know basically how it's used for spam detection, but to the best
of my knowledge nobody is using Hep to read e-mail at the moment...

> 3) One more thing, it looks like Hep isn't quoting the From: line when I 
> do a POP3 retrieve, the first example I took from a trace retrieving 
> mail from my ISP, the 2nd is a trace from Hep.  In Moz mail, this causes 
> the from line to get displayed incorrectly:
> 
> From: "PC Connection" <pcc...@pc...>
> From: Danno Ferrin: ... Speling Errors
> 
> I was going to submit a patch instead of a bug report, but I'm still not 
> up to speed on the code :-(

Thanks for being interested in the code.  I've been working on improving
the messaging.Message class, and improved address handling is going to
be part of this - I'll be checking it in this week.

Abe

A couple questions:

1) What's the status of IMAP support in Hep?  The comment in 
heplib.servers.imap says "includes a full-blown twisted imap server, 
which should get migrated upstream eventually." - is this waiting on 
Twisted?

2) I've been thinking about experimenting with integrating Bayesian 
classification into Hep.  I've been studying the source for Spambayes, 
but I'm wondering if/how it could fit into Hep's architecture.  I guess 
the easiest answer is to put Spambayes' POP3 filter in front of the POP3 
protocol - do you think that there's a way to get it all into the same 
package?  I think that to get Bayesian filtering to work, you'd have to

3) One more thing, it looks like Hep isn't quoting the From: line when I 
do a POP3 retrieve, the first example I took from a trace retrieving 
mail from my ISP, the 2nd is a trace from Hep.  In Moz mail, this causes 
the from line to get displayed incorrectly:

From: "PC Connection" <pcc...@pc...>
From: Danno Ferrin: ... Speling Errors

I was going to submit a patch instead of a bug report, but I'm still not 
up to speed on the code :-(

Is there a way to configure Hep's default rss grabbing timing to be more 
than 30 mins (actually if I'm not wrong, the default is still at every 
15 minutes right?)

On Sun, 2003-05-04 at 21:50, Wari Wahab wrote:

> I've got one more suggestion to ask: Follow directions of Errors 
> 301 and 302. I think the old hep does it, at least that's what I 
> thought. But yeah, a 302 should just follow the directions and hit the 
> previous URL on the next request (IE follows the new URL, stupid IE), 
> and for a 301, try to change the configuration file to make the new link 
> permanent.
> 
> It's annoying to get this message:
> 
> 301 Moved Permanently
> 
> Moved Permanently
> The document has moved here.

On my TODO list, although the 301 permanent redirect will be a bit of
work to implement.

Abe

Abe Fettig wrote:

>Wow, nice bug report.  It's like a usability study, without the effort
>of me actually watching someone use Hep and taking notes.  Thanks!
>
Coming from someone who've done support, I guess this is much better 
than saying, "Hey, I didn't do anything and now this program doesn't 
work, fix it!" :)

Anyway I've got one more suggestion to ask: Follow directions of Errors 
301 and 302. I think the old hep does it, at least that's what I 
thought. But yeah, a 302 should just follow the directions and hit the 
previous URL on the next request (IE follows the new URL, stupid IE), 
and for a 301, try to change the configuration file to make the new link 
permanent.

It's annoying to get this message:

301 Moved Permanently

Moved Permanently
The document has moved here.

Hi Wari,

Wow, nice bug report.  It's like a usability study, without the effort
of me actually watching someone use Hep and taking notes.  Thanks!

Now to address the issues you bring up:

> I'm not sure what connection means, so I added http://fettig.net as a 
> connection (hoping it would auto discover the rss) 

Not very intutive, is it?  I'll try to make this more descriptive, and
add a link to a relevent help page that describes connections.

Basically, you're supposed to enter a url, of one of the following
types:

http://server/file.xml
pop3://user:pass@server
blogger://user:pass@server/path/to/RPC/blogID - note that blogID should
be url-escaped, without slashes, so for pyblosxom blogs '/' becomes
'%2F'.

> I'm then greeted with 
> tracebacks, basically informing me that hep did not manage to create the 
> directory called 
> data/users/wari/messages/connections/http%3A/fettig.net/cur/ or 
> something like that. So I looked into the source and changed all the 
> os.mkdir() to os.makedirs() and that managed to keep hep quiet for a 
> while (diff attached). 

There are several problems here.  One, Hep doesn't check to make sure
it's data directory actually exists before it starts to create
sub-directories in it.  Two, the default directory, ./data, isn't
created by default when you check out Hep CVS.  And three, instead of
the much simpler os.makedirs() I'm using my crufty checkDirTree
function.

I'll apply your patch, and go one step further and get rid of all the
checkDirTree() stuff.

> So I go on again, creating a connection at the 
> Configuration page, hit the go button, I'm then greeted with a blank 
> textbox and a submit button, typed in http://fettig.net again, the 
> browser then hangs there. Nevermind, load up the 'messages' page again, 
> this time I found the http://fettig.net/ and it loaded the default 
> debian page that I have :) Interesting.

The blank textbox is where you're supposed to enter the name of your
connection.  I know there's no way to tell that at the moment - sorry!
(Actually the box is supposed to be filled in with a suggested name for
the connection, but at the moment that's broken for RSS).  Also
auto-detection of RSS feeds isn't working yet.

Finally, the redirect after adding the feed stopped working a while ago
- I think it broke with the Twisted upgrade.  I've been meaning to fix
it, but haven't yet.

> Subscribe bookmarklet does not work yet, 

TODO, once I fix the other new-connetion stuff it will be trivial.

> when I try to delete the old 
> fettig.net connection I got the page pointing to 
> http://localhost:5080/config/config/editconnection and I got the message 
> "No such resource" Go I guess I have to delete the connection myself (If 
> I know howto).

There's currently no way to edit and delete existing connections, which
is really annoying.

The good news is that I'm planning to work on the config pages this
weekend.  So all the problems you mentioned should be fixed soon.

> PS: I would still like to have the old style of viewing the blog entries 
> in reverse chronological order.

You can read all new messages by using the "New Messages" view.  I might
also add a view for "Today's Messages", or something like that, to let
you easily read messages by date.

Actually I've been doing some thinking about various ways of presenting
messages in the web interface, and I have some ideas for improving the
message viewer.  But I've got some other things on my list which are
higher priority.

Thanks for the bug report!

Abe

I've gotten hep and the messagging library, install twisted 1.0.4 and 
Lupy and tried out hep on a very fresh copy (No data directory).

After figuring out where to put what, I ran hep-add-user. Typed in the 
username and password, run hep, and hep runs nicely, got to 
localhost:5080 and everything runs fine. This is where the nightmare 
begins :)

I'm not sure what connection means, so I added http://fettig.net as a 
connection (hoping it would auto discover the rss) I'm then greeted with 
tracebacks, basically informing me that hep did not manage to create the 
directory called 
data/users/wari/messages/connections/http%3A/fettig.net/cur/ or 
something like that. So I looked into the source and changed all the 
os.mkdir() to os.makedirs() and that managed to keep hep quiet for a 
while (diff attached). So I go on again, creating a connection at the 
Configuration page, hit the go button, I'm then greeted with a blank 
textbox and a submit button, typed in http://fettig.net again, the 
browser then hangs there. Nevermind, load up the 'messages' page again, 
this time I found the http://fettig.net/ and it loaded the default 
debian page that I have :) Interesting.

Configuration: http://fettig.net/?flav=rss [go], goes back to that 
particular page again where there's one textbox and a submit button. 
Enter http://fettig.net/?flav=rss, the page hangs again, but after 
clicking on the new fettig.net entry in messages, I got your RSS entry, 
thank god for that :).

Subscribe bookmarklet does not work yet, when I try to delete the old 
fettig.net connection I got the page pointing to 
http://localhost:5080/config/config/editconnection and I got the message 
"No such resource" Go I guess I have to delete the connection myself (If 
I know howto).

When I started up hep I got:
Scanner: filled queue ['http://fettig.net', 'cvs', 
'http://roughingit.subtlehints.net/pyblosxom', 
'http://fettig.net/?flav=rss']
HTTP: Connected to http://:80/.
HTTP: Connected to http://:80/.
HTTP: Connected to http://:80/.
HTTP: Connected to http://fettig.net:80/?flav=rss.

Ah, that's what happened to the rest :)

So anyway, the patch is all I am able to contribute right now. Which is 
nothing much. I'm not sure how hep works now :) One thing for sure, it 
definitely looks good now.

PS: I would still like to have the old style of viewing the blog entries 
in reverse chronological order.

I've set up a mailing list for CVS commits -
hep...@cv....

If you'd like to be automatically notified of changes to Hep CVS, sign
up at https://lists.sourceforge.net/lists/listinfo/hepserver-commits

Abe

That's it, I was using Twisted 1.0.1, upgraded to 1.0.3 and everything 
works like a charm!
Abe Fettig wrote:

>On Tue, 2003-04-15 at 22:55, Gordon Weakliem wrote:
>  
>
>>Hep is downloading my RSS feeds (verified that with a sniffer), but when 
>>I connect via email, I get no messages.  When I go in via the web 
>>interface and click on a feed (under "connections"), I get either: Scan 
>>failed: [Failure instance: Traceback! 
>>twisted.internet.error.ConnectionLost, Connection lost ] or Scan failed: 
>>[Failure instance: Traceback! twisted.python.failure.DefaultException, 
>>Incomplete response ]
>>    
>>
>
>Hep works fine for me on my Windows test box (running Windows 2K, Python
>2.2.1, and Twisted 1.0.3)  What version of Twisted do you have?  If it's
>earlier than 1.0.3, would you upgrade and see if that fixes the problem?
>(If it does I'll put a note on the Hep webpage that 1.0.3 is required)
>
>Abe
>  
>

On Tue, 2003-04-15 at 22:55, Gordon Weakliem wrote:
> Hep is downloading my RSS feeds (verified that with a sniffer), but when 
> I connect via email, I get no messages.  When I go in via the web 
> interface and click on a feed (under "connections"), I get either: Scan 
> failed: [Failure instance: Traceback! 
> twisted.internet.error.ConnectionLost, Connection lost ] or Scan failed: 
> [Failure instance: Traceback! twisted.python.failure.DefaultException, 
> Incomplete response ]

Hep works fine for me on my Windows test box (running Windows 2K, Python
2.2.1, and Twisted 1.0.3)  What version of Twisted do you have?  If it's
earlier than 1.0.3, would you upgrade and see if that fixes the problem?
(If it does I'll put a note on the Hep webpage that 1.0.3 is required)

Abe

> 
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Hepserver-devel mailing list
> Hep...@li...
> https://lists.sourceforge.net/lists/listinfo/hepserver-devel

Hep is downloading my RSS feeds (verified that with a sniffer), but when 
I connect via email, I get no messages.  When I go in via the web 
interface and click on a feed (under "connections"), I get either: Scan 
failed: [Failure instance: Traceback! 
twisted.internet.error.ConnectionLost, Connection lost ] or Scan failed: 
[Failure instance: Traceback! twisted.python.failure.DefaultException, 
Incomplete response ]

2002	Jan	Feb	Mar	Apr	May	Jun	Jul	Aug	Sep	Oct (4)	Nov (20)	Dec
2003	Jan (34)	Feb (13)	Mar (6)	Apr (18)	May (12)	Jun (6)	Jul (1)	Aug	Sep	Oct	Nov	Dec
2004	Jan	Feb	Mar	Apr	May	Jun	Jul (1)	Aug	Sep	Oct	Nov	Dec
2005	Jan	Feb	Mar	Apr	May (1)	Jun	Jul	Aug	Sep	Oct	Nov	Dec

hepserver-devel Mailing List for Hep Message Server

hepserver-devel — Mailing list for Hep developers