From ghudson at MIT.EDU Thu Dec 1 13:30:40 2005 From: ghudson at MIT.EDU (Greg Hudson) Date: Thu, 1 Dec 2005 13:30:40 -0500 Subject: [Wocky] Re: Server-to-server In-Reply-To: <1129927029.15179.236.camel@egyptian-gods.mit.edu> Message-ID: <200512011830.jB1IUe3L012922@egyptian-gods.mit.edu> On October 21, you wrote: > It looks like server-to-server communication is disabled on > jabber.mit.edu. Is this planned at some point? (or am I > misinterpreting?) Sorry it's taken so long to get back to you about this. My testing indicates: * s2s works between mit.edu and jis.mit.edu * s2s works between mit.edu and 12jabber.com * s2s does *not* work between mit.edu and jabber.org * s2s does *not* work between mit.edu and gmail.com It is well-known that gmail.com does not allow server-to-server Jabber at the current time; see: http://www.google.com/talk/developer.html#service_4 It's disappointing, but we don't have any control over that decision. When I initially tested against jabber.org and found that it didn't work, I assumed the problem must be on our end. I'm rather surprised to find out that the problem seems to be on their end, since we are able to communicate with two other open-registration servers. From saltine at MIT.EDU Thu Dec 1 14:07:45 2005 From: saltine at MIT.EDU (Joseph Calzaretta) Date: Thu, 01 Dec 2005 14:07:45 -0500 Subject: [Wocky] Re: Server-to-server In-Reply-To: <200512011830.jB1IUe3L012922@egyptian-gods.mit.edu> References: <1129927029.15179.236.camel@egyptian-gods.mit.edu> <200512011830.jB1IUe3L012922@egyptian-gods.mit.edu> Message-ID: <6.2.1.2.2.20051201140008.03c72070@hesiod> At 01:30 PM 12/1/2005, Greg Hudson wrote: >* s2s works between mit.edu and jis.mit.edu Is this intermittent? A few weeks ago when I was setting up the Gaim client, I couldn't see anyone @jis.mit.edu via my jcalz at mit.edu account... nor anyone @mit.edu from my jcalz at jis.mit.edu account. Neither could Sam Hartman. Today, the cross-server stuff seems to be working. And presumably it worked in the past as well. So are the servers being reconfigured every so often? Or is there some kind of intermittent something-else going on? Thanks! --Joe From tfitz at MIT.EDU Thu Dec 1 15:45:44 2005 From: tfitz at MIT.EDU (Tom Fitzgerald) Date: Thu, 01 Dec 2005 15:45:44 -0500 Subject: [Wocky] Re: Server-to-server In-Reply-To: <200512011830.jB1IUe3L012922@egyptian-gods.mit.edu> References: <200512011830.jB1IUe3L012922@egyptian-gods.mit.edu> Message-ID: <1133469944.17713.162.camel@sligo.mit.edu> Cool. I was testing with jabber.org too, unfortunately. 12jabber.com will suit me fine. Interestingly, s2s does work between jabber.org and 12jabber.com. > > It looks like server-to-server communication is disabled on > > jabber.mit.edu. Is this planned at some point? (or am I > > misinterpreting?) > > Sorry it's taken so long to get back to you about this. > > My testing indicates: > > * s2s works between mit.edu and jis.mit.edu > * s2s works between mit.edu and 12jabber.com > * s2s does *not* work between mit.edu and jabber.org > * s2s does *not* work between mit.edu and gmail.com > > It is well-known that gmail.com does not allow server-to-server Jabber > at the current time; see: > > http://www.google.com/talk/developer.html#service_4 > > It's disappointing, but we don't have any control over that decision. > > When I initially tested against jabber.org and found that it didn't > work, I assumed the problem must be on our end. I'm rather surprised > to find out that the problem seems to be on their end, since we are > able to communicate with two other open-registration servers. From atticus at MIT.EDU Thu Dec 1 19:01:17 2005 From: atticus at MIT.EDU (Atticus O Gifford) Date: Thu, 1 Dec 2005 19:01:17 -0500 Subject: [Wocky] MIT Adium 0.86 (Beta 1) installer available for testing Message-ID: <20051201190117.km6hayz31x6sgckc@webmail.mit.edu> I've put together the first version of the Mac Adium installer. Next week, this installer, along with the Gaim installer, will have the UI cleaned up match our other MIT installers, perform validation on the username etc. This should allow you to install the application and create an account during the install. You can create more accounts by running the Prefs Installer in /Applications/MIT Adium 0.86/ The Gaim installer is also still available at: I'll plan on coming to the meeting Tuesday if you're having one to discuss possible next steps. Thanks, Atticus From raeburn at MIT.EDU Fri Dec 2 10:58:29 2005 From: raeburn at MIT.EDU (Ken Raeburn) Date: Fri, 2 Dec 2005 10:58:29 -0500 Subject: [Wocky] MIT Adium 0.86 (Beta 1) installer available for testing In-Reply-To: <20051201190117.km6hayz31x6sgckc@webmail.mit.edu> References: <20051201190117.km6hayz31x6sgckc@webmail.mit.edu> Message-ID: <3311EAFB-8F2C-428E-BE2D-9AD855956E00@mit.edu> > If I understand it right, your .htaccess.mit file in that directory limits access to people in two groups, swr-core and wocky. Problem is, the wocky group in Moira and AFS doesn't contain any users, it just contains the one string wocky at mailman.mit.edu. This may explain why I wasn't able to download it. (Or, it may be related to the problems I've been having just recently with email access from my laptop too, if it's something certificate-related.) Fortunately I was able to copy it out of AFS. So, now I've got /Applications/MIT containing MIT Netscape 7.02, and / Applications/MIT Adium 0.86. The Netscape install is kind of old, so maybe things have changed, but does MIT (SWRT) have a consistent policy on what level to install MIT Mac applications at? The installer asked for my Kerberos id, which I gave it. When I started Adium, I then had two raeburn at mit.edu accounts with basically the same configuration data. (I've been using Adium for a while, and already had configured an mit.edu account.) The installer should've checked for that, and either told me nothing needed to be done, or checked that the data was correct and updated (offered to update?) if it was not. So I deleted one of them, the one listed first, which I assume was the old one. Signing on to the account, I'm asked for my password, so I assume this version doesn't have the Gaim patches to use Kerberos authentication. Is there anything interesting to this version besides having MIT's installer and prefs installer on top of the standard Adium distribution? I took a look at the prefs installer. (Why is it buried in yet another subdirectory, one which contains only the one program with the same name as the directory?) It says it will create an Adium account for use at MIT; fine. It also says ``Items will be installed on the disk "Mac OS X"'' -- what items? The account data should be stored wherever the account data is stored under my home directory, regardless of what disk that happens to be on. (If my home directory is not on the boot partition, will this report even be correct?) After I click "skip", the message says I can run "the Adium 0.86 Prefs Installer in your Applications -> MIT Adium 0.86 folder". Except, that's not the name of the program ("MIT Adium 0.86 Prefs Installer"), and it's buried in a sub-folder. Is there a right-arrow glyph that might be used instead of the dash greater-than construct? (The Character Palette shows Unicode code points for Rightwards Arrow ?, Rightwards Double Arrow ?, Rightwards White Arrow ?, and Black Rightwards Arrow ?, for example, in the Arrows and Dingbats tables, but I don't know if these characters are always available; some other code points listed in the Character Palette aren't displayed for me.) The downside for me, semi-paranoid person that I am, is that I don't appear to have the option of simply copying in one program or folder without running an installer program, as I get with the standard Adium package. Sometimes it's nice to know that (1) nothing's been installed outside that folder, so it's easy to throw it all away at once, and (2) it's not configuring any path-dependent stuff during installation, so renaming it or moving it to another folder is okay. The former may not be a problem if the installer program deigns to leave me a log file to look at, which this one appears to, though it's hard to tell that before running it if you don't already know, and impossible to decide whether to install a program based on whether it installs stuff in shared system folders, replaces system libraries, all that fun stuff you find on Windows. I know I can move Adium.app around without breaking things, but I'm less sure about the MIT Prefs Installer. Ken From atticus at MIT.EDU Fri Dec 2 16:33:19 2005 From: atticus at MIT.EDU (Atticus O Gifford) Date: Fri, 2 Dec 2005 16:33:19 -0500 Subject: [Wocky] MIT Adium 0.86 (Beta 1) installer available for testing In-Reply-To: <3311EAFB-8F2C-428E-BE2D-9AD855956E00@mit.edu> References: <20051201190117.km6hayz31x6sgckc@webmail.mit.edu> <3311EAFB-8F2C-428E-BE2D-9AD855956E00@mit.edu> Message-ID: <20051202163319.054lqobdba1w044o@webmail.mit.edu> Thanks for the detailed results. I've interspersed a few comments below. Quoting Ken Raeburn : >> > > If I understand it right, your .htaccess.mit file in that directory > limits access to people in two groups, swr-core and wocky. Problem > is, the wocky group in Moira and AFS doesn't contain any users, it > just contains the one string wocky at mailman.mit.edu. This may explain > why I wasn't able to download it. (Or, it may be related to the > problems I've been having just recently with email access from my > laptop too, if it's something certificate-related.) Fortunately I > was able to copy it out of AFS. Whoops, I didn't realize wocky wasn't a standard list. For the moment, I've made the installer available to is&t at mit.edu and I'll leave the AFS open as a backup. I'll figure out something a little more appropriate later. > So, now I've got /Applications/MIT containing MIT Netscape 7.02, and > / Applications/MIT Adium 0.86. The Netscape install is kind of old, > so maybe things have changed, but does MIT (SWRT) have a consistent > policy on what level to install MIT Mac applications at? Yes, all our installers install to "MIT " under /Applications. Netscape on Mac was a little wacky (to say the least) and is quite old, so its odd location isn't too surprising. > The installer asked for my Kerberos id, which I gave it. When I > started Adium, I then had two raeburn at mit.edu accounts with basically > the same configuration data. (I've been using Adium for a while, > and already had configured an mit.edu account.) The installer > should've checked for that, and either told me nothing needed to be > done, or checked that the data was correct and updated (offered to > update?) if it was not. You're definitely right about that. That's slated for the B2 installer next week. The main focus of this installer was getting the code together to add an account correctly and making sure it worked. So far so good on that front. Now I need to add validation, doll up the UI (dialogs) and add the rest of the code. I suspect the installer will work something like: 1) Ask for Kerberos username 2) Check for an account of that name 3) If not found, add 4) If found, offer to add anyway, overwrite (update?) the existing or just leave it alone. Hopefully that should all be in the next version.. if you get a chance to look at that one let me know if it works better for you. > So I deleted one of them, the one listed first, which I assume was > the old one. > > Signing on to the account, I'm asked for my password, so I assume > this version doesn't have the Gaim patches to use Kerberos > authentication. Is there anything interesting to this version > besides having MIT's installer and prefs installer on top of the > standard Adium distribution? The installer right now is just the main Adium distribution with configuration added. Now that there are framework installers for Windows and Mac, we can probably discuss additions at the Tuesday meeting (if any). > I took a look at the prefs installer. (Why is it buried in yet > another subdirectory, one which contains only the one program with > the same name as the directory?) That's a hold-over from how we've always done the installers. We used to include a text file in the folder called "What's this?" or something similar. Since most other installers don't have a prefs installer or config wizard or whatever, it seemed easier to abstract it a bit and associate it with a text document. I'm not sure if that's the best way to do things anymore, and we're going to be revamping our mac installer standards over the next 2 months (hopefully). Any ideas or input you have on that front is most welcome. I'm happy to just move the file up a level and not include a description if that seems less confusing. > It says it will create an Adium account for use at MIT; fine. It > also says ``Items will be installed on the disk "Mac OS X"'' -- what > items? The account data should be stored wherever the account data > is stored under my home directory, regardless of what disk that > happens to be on. (If my home directory is not on the boot > partition, will this report even be correct?) I thought I had turned that off, but I'll check again. It may be that since we're using the VISE installer as a wrapper, its text always assumes it will be installing files of some sort. If nothing else, I can just edit the resource string for that text to reflect something more accurate. > After I click "skip", the message says I can run "the Adium 0.86 > Prefs Installer in your Applications -> MIT Adium 0.86 folder". > Except, that's not the name of the program ("MIT Adium 0.86 Prefs > Installer"), and it's buried in a sub-folder. Is there a right-arrow > glyph that might be used instead of the dash greater-than construct? > (The Character Palette shows Unicode code points for Rightwards > Arrow ?, Rightwards Double Arrow ?, Rightwards White Arrow ?, > and Black Rightwards Arrow ?, for example, in the Arrows and > Dingbats tables, but I don't know if these characters are always > available; some other code points listed in the Character Palette > aren't displayed for me.) That's a good point in general. I try not to use that sort of construct, but I should keep that in mind in case I can't get around it. In this case, I think the poor writing (and poor arrow :) of the dialog are a result of slapping together a Beta installer around the configuration script. I definitely planned on rewriting all of that. I'll see what I can polish up for the Beta 2 installer, and I'll use a proper arrow if the need remains. > The downside for me, semi-paranoid person that I am, is that I don't > appear to have the option of simply copying in one program or folder > without running an installer program, as I get with the standard > Adium package. Sometimes it's nice to know that (1) nothing's been > installed outside that folder, so it's easy to throw it all away at > once, and (2) it's not configuring any path-dependent stuff during > installation, so renaming it or moving it to another folder is okay. > The former may not be a problem if the installer program deigns to > leave me a log file to look at, which this one appears to, though > it's hard to tell that before running it if you don't already know, > and impossible to decide whether to install a program based on > whether it installs stuff in shared system folders, replaces system > libraries, all that fun stuff you find on Windows. I know I can move > Adium.app around without breaking things, but I'm less sure about > the MIT Prefs Installer. That's also a good point. When we write installers on the Mac platform, the primary reason is usually to add customization. The problem is, we haven't thought up any system to handle customization that is substantially more clever than the sytem we're using (a prefs installer) without being more invasive or path-dependent. The only time we install items to non-Application folder locales is when the vendor installer does the same thing. The new TSM installer is a good example: it installs binaries in unix locations as well as under /Library/Application Support. Adding a log file is trivial, but as you say it doesn't solve the initial uneasiness. Part of the problem is balancing experienced users' trepidations with novice users' skittishness. Too much information and too many decisions tends to make novice users nervous, whereas the reverse seems to be true as experience increases. We're going to try to move all our installers to .pkg installers in the near future for those users who are using Apple Remote Desktop to manage their users. I suspect that will only obscure the installers' workings even further, however. If you have any ideas or more issues you'd like addressed about installers in general (or this installer obviously) please send them along. It's hard to get input on these things, and the more info we have moving forward with a redesign the better. Thanks for all the info, Atticus P.S. The prefs installer only adds accounts to ~/Library/Application Support/Adium 2.0/... btw (in case you were still wondering). From hallisey at MIT.EDU Mon Dec 5 15:49:45 2005 From: hallisey at MIT.EDU (Joanne Hallisey) Date: Mon, 5 Dec 2005 15:49:45 -0500 Subject: [Wocky] Meeting Tuesday, Dec. 6 at 3:00 Message-ID: <3898751F-5A07-4064-8013-9A0E7D3D7AA9@mit.edu> Hello, There will be a Jabber meeting Tuesday, Dec. 6 at 6:00 in W92-225. Agenda 1. Sponsor check in: Probable expansion of the pilot 2. Progress on survey questions 3. Progress on installers; any issues reported 4. Other Thanks, Joanne -------------------------------------------- Joanne Hallisey IS&T W92-153 3-1894 hallisey at mit.edu From jchrepta at MIT.EDU Mon Dec 5 23:01:54 2005 From: jchrepta at MIT.EDU (Darien J Chrepta) Date: Mon, 5 Dec 2005 23:01:54 -0500 Subject: Fwd: [Wocky] Invitation to participate in the Jabber messaging pilot Message-ID: <20051205230154.9krwqe37kuwwsg84@webmail.mit.edu> Hello IS&T I'm interested in participating. I'll check out the Jabber page. Best wishes, Jay Chrepta ----- Forwarded message from tregan at MIT.EDU ----- Date: Mon, 5 Dec 2005 20:50:47 -0500 From: Theresa M Regan Reply-To: Theresa M Regan Subject: Fwd: [Wocky] Invitation to participate in the Jabber messaging pilot To: Theresa M Regan Greetings, IS&T is currently piloting an instant messaging service, Jabber. If you are interested in participating, please read the following note which highlights the service and support options during this pilot. Questions, comments and/or suggestions may be posted to Cheers, Theresa p.s. please note... one's login account is: kerberos principal at mit.edu, i.e., tregan at mit.edu Begin forwarded message: > From: "Joanne M. Hallisey" > Date: October 18, 2005 10:59:10 AM EDT > To: IT-Partners at mit.edu, IT-Lead at mit.edu, helpstaff at mit.edu, > helpstudents at mit.edu, athena-rcc at mit.edu, athena-rcc-students at mit.edu > Cc: wocky at mit.edu > Cc: is&t at mit.edu > Subject: [Wocky] Invitation to participate in the Jabber messaging > pilot > > Hello, > > IS&T invites you to participate in a pilot of a Jabber Instant > Messaging service. Jabber is an instant messaging system that > allows MIT users to communicate with one another and Jabber users > elsewhere on the Internet. Jabber also supports creation and > joining of chat rooms for group communication. Additionally, many > Jabber clients also offer support for other instant messaging > systems such as AIM (AOL) and YIM (Yahoo). > > The intention of the pilot is to learn more about requirements for > providing this as an enterprise messaging service and understand > some of the support and service issues. The pilot will focus on > setting up the server environment and delivering installers for > GAIM and AdiumX (software clients for Linux, Macintosh and Windows). > > The Jabber pilot team will provide limited best effort support, and > pilot participants are strongly encouraged to give feedback. As the > pilot progresses, the team will send updates and post them to the > web pages listed below. The pilot will run through December 2005. > IS&T will start a more formal release effort if the pilot is > successful. > > The following draft web pages provide download information, > configuration instructions and information about known issues. > > Jabber Service Page > > > Jabber Stock Answer Branch > > > Please note: > As always, protect your password. Do not allow the client program > to store your password. > > For questions or more information please contact the team via the > Wocky mailing list: > Wocky at mit.edu > > > Thank you. > For the Jabber Team, > -- > Joanne Hallisey > _______________________________________________ > Reminders: > - Make off-campus connections through the Virtual Private Network. > For information on the VPN see network/vpn.html> > > - Take advantage of the MIT Windows Automatic Update Service. For > information on WAUS see updates/> > > - Take advantage of a recent enhancement to MIT's Spam Screening > service which allows for personally setting the number of days > "spam" is aged prior to auto deletion. For more complete > information and to set your personal options see web.mit.edu/ist/services/email/nospam/> > > -- > Joanne Hallisey > Sr. Project Manager > MIT - Information Services and Technology > 617-253-1894 > _______________________________________________ > Wocky mailing list > Wocky at mit.edu > http://mailman.mit.edu/mailman/listinfo/wocky ----- End forwarded message ----- -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.mit.edu/pipermail/wocky/attachments/20051205/32e9e7f7/attachment.htm From hallisey at MIT.EDU Wed Dec 7 09:28:13 2005 From: hallisey at MIT.EDU (Joanne Hallisey) Date: Wed, 7 Dec 2005 09:28:13 -0500 Subject: [Wocky] Meeting Notes: Jabber Message-ID: Theresa extended the invitation to additional piloteers. One person responded. We have asked Mark for usage/server statistics; waiting for his response. Survey - reviewed questions - Joanne will send them to wocky at the end of the week. This will give the newest users some time to use Jabber. Server to server: did some tests. It works fine on some public jabber servers. Jabber.org has had some problems and has not been working gmail.com (googletalk) does not do server to server right now. 12jabber.com; jis.mit.edu work. Heather will set up some information/FAQs for server to server. Atticus: Saw Ken Raeburn's responses to the AdiumX and is making changes. Did not see any responses to Windows Gaim installer. Hope to have updates out in 2 or 3 days for next round of testing. Windows installer should work in the domain. Gaim will appear in the Start Menu like other MIT installers. MIT configurations in Tools submenu consistent with other MIT installers. Run usability testing. Joanne will set up some testing. Heather Anne will do some documentation. Greg has been looking at the confirmation step when you add a buddy. Looked at presence/privacy for other IM solutions. Right now it is hard to make it possible for people who are at MIT to see your presence without a confirmation. Will likely continue the pilot to get more experience and to understand more about the Jabber community. Next meeting: Dec. 19 at 2:00 in . -------------------------------------------- Joanne Hallisey IS&T W92-153 3-1894 hallisey at mit.edu From rboes at plant.mit.edu Wed Dec 7 14:11:20 2005 From: rboes at plant.mit.edu (Robert Boes) Date: Wed, 7 Dec 2005 14:11:20 -0500 Subject: [Wocky] Jabber login Message-ID: Hi, I was passed an invitation to test jabber by someone who's in IT Partners. (I am not.) I downloaded the software, registered myself as a Jabber account, but couldn't login. I was using my Kerberos password. Is the test of the product limited to a certain group of people, or am I doing something wrong? Thanks for any help. Bob Robert G. Boes Senior Systems Planner MIT Campus Planning and Design 77 Massachusetts Ave. Room NE49-2100 Cambridge 02139 (617) 452-2992 rboes at mit.edu -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.mit.edu/pipermail/wocky/attachments/20051207/05fbfb16/attachment.htm From hallisey at MIT.EDU Wed Dec 7 14:43:32 2005 From: hallisey at MIT.EDU (Joanne Hallisey) Date: Wed, 7 Dec 2005 14:43:32 -0500 Subject: [Wocky] Jabber login In-Reply-To: References: Message-ID: Hello Robert, There is some information here. http://web.mit.edu/ist/services/ messaging/jabber.html -------------------------------------------- Joanne Hallisey IS&T W92-153 3-1894 hallisey at mit.edu On Dec 7, 2005, at 2:11 PM, Robert Boes wrote: > Hi, > I was passed an invitation to test jabber by someone who's in IT > Partners. (I am not.) I downloaded the software, registered myself > as a Jabber account, but couldn't login. I was using my Kerberos > password. Is the test of the product limited to a certain group of > people, or am I doing something wrong? > Thanks for any help. > Bob > > Robert G. Boes > Senior Systems Planner > MIT Campus Planning and Design > 77 Massachusetts Ave. Room NE49-2100 > Cambridge 02139 > (617) 452-2992 > rboes at mit.edu > > _______________________________________________ > Wocky mailing list > Wocky at mit.edu > http://mailman.mit.edu/mailman/listinfo/wocky From ghudson at MIT.EDU Mon Dec 12 16:06:02 2005 From: ghudson at MIT.EDU (Greg Hudson) Date: Mon, 12 Dec 2005 16:06:02 -0500 Subject: [Wocky] Status update on gaim GSSAPI patches Message-ID: <200512122106.jBCL62sr018709@egyptian-gods.mit.edu> I decided to set up a source repository for gaim to manage the various changes we are prototyping. It's in /afs/dev.mit.edu/project/jabber/repos/gaim. This repository doesn't have a trunk; it has a branches/upstream for the stock Gaim 1.5.0 source and a branches/gssapi-only for Simon's more recent patch with our additions. I expect to create a branches/cyrus-sasl branch for working on Simon's older patch if we continue to believe that's the direction to go in. However, I'm currently prototyping with the GSSAPI-only patch since that's temporarily easier to work with and might be appropriate for a local deployment. Qing, I took your change to get the connect server instead of hardcoding jabber.mit.edu. Thanks. On requesting the password only when it is needed: currently, the machinery for requesting the user's password on demand is in a sub-block of gaim_connection_connect(), which also creates the connection. I will need to move that out into a separate function in src/connection.c. In the Jabber code, I've determined that all authentication goes through auther jabber_auth_start or jabber_auth_start_old (the latter for old-style IQ auth, which doesn't support GSSAPI). I can request the password unconditionally in jabber_auth_start_old, and conditionally depending on whether the server and client support GSSAPI in jabber_auth_start. Those functions will have to be broken up into two pieces since requesting the password is a stop-and-wait-for-a-callback operation. From ghudson at MIT.EDU Wed Dec 14 01:57:53 2005 From: ghudson at MIT.EDU (Greg Hudson) Date: Wed, 14 Dec 2005 01:57:53 -0500 Subject: [Wocky] Status update on gaim GSSAPI support Message-ID: <200512140657.jBE6vrow008218@egyptian-gods.mit.edu> I've written and checked in prototype code to make gaim request a password when the connection isn't using GSSAPI. I think we now have something we could deploy locally as a Gaim 1.5.0 enhancement, if we want to go that route. I'll start working on the older Cyrus SASL patches and see if I can beat them into shape, and I'll starting talking with Simon and the Gaim developers about how we might integrate this stuff upstream. From jcolon at MIT.EDU Wed Dec 14 10:30:55 2005 From: jcolon at MIT.EDU (=?iso-8859-1?Q?Jessica_Col=F3n?=) Date: Wed, 14 Dec 2005 10:30:55 -0500 Subject: [Wocky] Trillian Message-ID: <007801c600c3$5fe23200$4900aa12@econ.ms.mit.edu> Hello, My suggestion: Allowing access and/or post configuration instructions for Jabber through Trillian. http://www.ceruleanstudios.com/learn/ =] __________________________________________________ Jessica Col?n Massachusetts Institute of Technology Department of Economics 50 Memorial Drive Building E52, Room 373 Cambridge MA 02142 Email jcolon at mit.edu Phone 617.253.3807 Fax 617.253.1330 -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.mit.edu/pipermail/wocky/attachments/20051214/6c139319/attachment.htm From morzinski at MIT.EDU Wed Dec 14 13:10:56 2005 From: morzinski at MIT.EDU (Jacob Morzinski) Date: Wed, 14 Dec 2005 13:10:56 -0500 Subject: [Wocky] Trillian In-Reply-To: Message from =?iso-8859-1?Q?Jessica_Col=F3n?= of "Wed, 14 Dec 2005 10:30:55 EST." <007801c600c3$5fe23200$4900aa12@econ.ms.mit.edu> References: <007801c600c3$5fe23200$4900aa12@econ.ms.mit.edu> Message-ID: <200512141810.jBEIAvfn026864@kamp-krusty.mit.edu> I took a brief look at Trillian, but could not proceed. The "basic" version of Trillian does not support their Jabber plugin; only the "pro" version supports Jabber. Without a Jabber-enabled version of Trillian to test on, I can't suggest how to set up Jabber. All I can suggest is that you take the Gaim configuration instructions, and adapt them as best as you can. Regards, Jacob Morzinski jmorzins at mit.edu On Wed, 14 Dec 2005, Jessica Col?n wrote: > My suggestion: Allowing access and/or post configuration > instructions for Jabber through Trillian. > http://www.ceruleanstudios.com/learn/ From ghudson at MIT.EDU Wed Dec 14 13:56:52 2005 From: ghudson at MIT.EDU (Greg Hudson) Date: Wed, 14 Dec 2005 13:56:52 -0500 Subject: [Wocky] Trillian In-Reply-To: <200512141810.jBEIAvfn026864@kamp-krusty.mit.edu> References: <007801c600c3$5fe23200$4900aa12@econ.ms.mit.edu> <200512141810.jBEIAvfn026864@kamp-krusty.mit.edu> Message-ID: <1134586612.12162.258.camel@egyptian-gods.mit.edu> On Wed, 2005-12-14 at 13:10 -0500, Jacob Morzinski wrote: > Without a Jabber-enabled version of Trillian to test on, I can't > suggest how to set up Jabber. All I can suggest is that you take > the Gaim configuration instructions, and adapt them as best as > you can. You can also try adapting the instructions at http://www.google.com/support/talk/bin/answer.py?answer=24077 in order to triangulate from another direction. Of course, you'd use username at mit.edu instead of username at gmail.com, and you'd use jabber.mit.edu in place of talk.google.com. From pbh at MIT.EDU Wed Dec 14 17:56:08 2005 From: pbh at MIT.EDU (Paul B. Hill) Date: Wed, 14 Dec 2005 17:56:08 -0500 Subject: [Wocky] SASL vs. GSSAPI Message-ID: <200512142256.jBEMubQI008183@outgoing.mit.edu> Hi, The Windows Oracle Calendar software (TechTime) has a configuration option for a couple of authentication methods. These are gssapi:kerberos5, SASLKerberosv4, CST Basic Authentication, and Standard Authentication. (Actually the choice are dependant on what other libraries appear on your system.) When gssapi:kerberos5 is selected the client will use the CMU SASL libraries if they are present. However, if the CMU SASL libraries are not present, then the client will look for gssapi.dll and use that directly. If we want to provide authentication patches to Gaim, then a similar strategy might be the best way to get the Gaim developer community to accept the changes. The clients should see if the SASL libraries are present. If they are, then any of the SASL mechanisms installed could be used. If the SASL libraries are not able to be loaded, then see if the GSSAPI libraries can be loaded directly. Paul From ghudson at MIT.EDU Thu Dec 15 02:49:34 2005 From: ghudson at MIT.EDU (Greg Hudson) Date: Thu, 15 Dec 2005 02:49:34 -0500 Subject: [Wocky] SASL vs. GSSAPI In-Reply-To: <200512142256.jBEMubQI008183@outgoing.mit.edu> References: <200512142256.jBEMubQI008183@outgoing.mit.edu> Message-ID: <1134632974.12162.283.camel@egyptian-gods.mit.edu> One of the issues with dynamic loading is that the Gaim code needs to know the signatures of each function it wants to load. Simon's strategy was to include a copy of the krb5 gssapi.h, hacked up so that instead of declaring functions, it declared types. I don't think that approach would be acceptable to the Gaim people, and I'm curious what TechTime does. I suspect that whatever we send upstream will need to use a more conventional strategy of linking to dependency libraries. Other protocol plugins have their own library dependencies, I believe, and I don't think they use dynamic loading. I can look into that more. From ghudson at MIT.EDU Thu Dec 15 18:38:58 2005 From: ghudson at MIT.EDU (Greg Hudson) Date: Thu, 15 Dec 2005 18:38:58 -0500 Subject: [Wocky] More on Gaim dependency architecture Message-ID: <200512152338.jBFNcwZY030869@egyptian-gods.mit.edu> I looked into how Gaim handles dependencies on non-ubiquitous external libraries today. Here are my conclusions: * Gaim dynamically loads its own plugins, including protocol plugins. There is no issue with needing to know function signatures for this kind of loading, since Gaim gets to define the plugin ABI itself. Gaim is generally tolerant of failure to load any given plugin, and just doesn't offer the plugin's functionality. * Protocol plugins with dependencies are nothing new. The Zephyr, silc, and Bonjour plugins all have dependencies on external libraries. The Unix binary packaging strategy is to build the protocol plugin so that it links against the external library, and then put the resulting plugin in a separate package so that the core gaim package doesn't have to depend on libsilc or whatnot. * However, all of those protocol plugin dependencies are mandatory. You cannot have a Gaim Zephyr plugin without libzephyr. For our work, the GSSAPI or Cyrus SASL dependency would be optional. Unix binary packagers would want to be able to distribute the Jabber plugin without making it depend on Kerberos libraries. * Gaim has solved this problem in the past with SSL. What they did was create a separate plugin purely for SSL support (two of them, actually, for different external SSL libraries) and a layer in the core code which takes care of loading the plugin and offering or not offering SSL support as available. I think we would need to do the same thing for SASL. The existing built-in DIGEST-MD5 code could be moved into this new layer, and we could have plugins to wrap either libgssapi_krb5 or libsasl2 or both. * It will be substantially easier to write this new layer if we don't have to worry about supporting security layers, and just rely on SSL to take care of protecting the data stream. But I'm willing to do it either way. I don't think the Gaim developers would care. From awillis at MIT.EDU Fri Dec 16 14:00:51 2005 From: awillis at MIT.EDU (Albert Willis) Date: Fri, 16 Dec 2005 14:00:51 -0500 Subject: [Wocky] Adium 0.87 released Message-ID: <423B1588-1868-443C-B56B-B6551A2B4620@MIT.EDU> You can see all of the changes at http://trac.adiumx.com/wiki/ AdiumVersionHistory. -- Al From atticus at MIT.EDU Fri Dec 16 16:32:53 2005 From: atticus at MIT.EDU (Atticus O Gifford) Date: Fri, 16 Dec 2005 16:32:53 -0500 Subject: [Wocky] MIT Adium 0.87 (Beta 2) available for testing Message-ID: <20051216163253.dh5c8gpyq4u8088g@webmail.mit.edu> I've updated the Adium installer to incorporate a number of the changes Ken recommended (among other tweaks) and to perform basic input validation. The installer also includes the new 0.87 binary which seems to work fine so far as I've tested it. If a version of the installer with 0.86 is desired, just let me know and I'll re-upload it: Our plan is to revise our Mac installer requirements in general over the next few weeks. I think that going forward this installer will probably receive the new polish as well if it's finished in time, but we can discuss that before anything is released. Have a good weekend, Atticus From pbh at MIT.EDU Fri Dec 16 16:44:05 2005 From: pbh at MIT.EDU (Paul B. Hill) Date: Fri, 16 Dec 2005 16:44:05 -0500 Subject: [Wocky] google talk future directions... Message-ID: <200512162144.jBGLiT4M022862@outgoing.mit.edu> >From : 4. What other communication services will you federate with? We plan to support open server-to-server federation. We do believe, however, that it is important to have the safeguards in place to ensure that we maintain a safe and reliable service that protects user privacy and blocks spam and other abuses. We are using the federation opportunity with EarthLink, Sipphone and other partners to develop a set of best practices by which all members of the federated network can work together to ensure that we protect our users while maximizing the reach of the network. We are also eager to hear from other people in the industry about how best to build a federation model that is open, scalable, and ensures best-in-class user experiences. If you have thoughts on federation or suggestions for how we can better enable open communications, please share them with us at the Google Talk Interoperability Google Group. ---- Some of the traffic on the Google Talk Interoperability Google Group seems to indicate that the change will happen in the very near future. Supporting evidence is that Google started publishing the SRV records on December 9th, but at the moment the ports aren't open. --- Somewhat orthogonal... The people from Google Talk have been cooperating with people from the Jabber Foundation to develop specifications for VoIP signaling and media encoding. -----Original Message----- From: sip-bounces at ietf.org [mailto:sip-bounces at ietf.org] On Behalf Of fwmiller at cornfed.com Sent: Friday, December 16, 2005 4:11 PM To: sip at ietf.org Subject: [Sip] Jabber VoIP specs For those that may not have seen this, the Jabber Foundation has published their VoIP signaling and media encoding specs. These specs have been reconciled with Google Talk if I read the introductory comments correctly. I'd be interested in any discussion of the design of these protocols by the SIP gurus here... http://www.jabber.org/jeps/jep-0166.html http://www.jabber.org/jeps/jep-0167.html FM _______________________________________________ Sip mailing list https://www1.ietf.org/mailman/listinfo/sip This list is for NEW development of the core SIP Protocol Use sip-implementors at cs.columbia.edu for questions on current sip Use sipping at ietf.org for new developments on the application of sip From hallisey at MIT.EDU Mon Dec 19 09:15:34 2005 From: hallisey at MIT.EDU (Joanne Hallisey) Date: Mon, 19 Dec 2005 09:15:34 -0500 Subject: [Wocky] Initial Evaluation of Jabber IM Message-ID: <811ADEE8-2050-43A4-9F71-1A85502582AC@mit.edu> Hello, We would like to get an initial evaluation of your experience with Jabber. Please answer the following questions and send your responses to wocky at mit.edu. If possible, send responses by December 19, 2005. After we have assessed the experience of the pilot, we will determine next steps. If you know people who are using Jabber and are not on the wocky list, please feel free to send this survey to them. Thanks very much for your feedback. Joanne -------------------------------------------- 1. Which client did you use? Gaim for Windows Gaim for Linux AdiumX for Macintosh iChat (Mac OS 10.4) 2. How did you acquire the client software? Download from vendor site Download from MIT site Pre-installed with OS 3. Please rate the following for the client that you used: (can use a 1 - 5 scale) MIT download and installation Creation of "buddy" list Initiation of IM conversation IM conversation/chat history Creation of multi-user Chat Speed Reliability Login 4. Please list other features or functionality that you would like to have in an IM service 5. Please comment on any problems you may have had with the IM pilot service 6. Please add any suggestions/comments -------------------------------------------- Joanne Hallisey Sr. Project Manager Information Services and Technology W92-153 617-253-1894 hallisey at mit.edu From morzinski at MIT.EDU Mon Dec 19 13:09:22 2005 From: morzinski at MIT.EDU (Jacob Morzinski) Date: Mon, 19 Dec 2005 13:09:22 -0500 Subject: [Wocky] Initial Evaluation of Jabber IM In-Reply-To: Message from Joanne Hallisey of "Mon, 19 Dec 2005 09:15:34 EST." <811ADEE8-2050-43A4-9F71-1A85502582AC@mit.edu> References: <811ADEE8-2050-43A4-9F71-1A85502582AC@mit.edu> Message-ID: <200512191809.jBJI9Nw4011180@kamp-krusty.mit.edu> > -------------------------------------------- > 1. Which client did you use? Mostly Gaim for Solaris/Athena, a very little Gaim for Windows > 2. How did you acquire the client software? Gaim for Solaris/Athena is part of the Athena release. > 3. Please rate the following for the client that you used: (can use > a 1 - 5 scale) > MIT download and installation Very good > Creation of "buddy" list Somewhat poor > Initiation of IM conversation Somewhat good > IM conversation/chat history Somewhat poor > Creation of multi-user Chat Somewhat poor. > Speed Neutral > Reliability Somewhat good > Login Somewhat poor > 4. Please list other features or functionality that you would like to > have in an IM service '() > 5. Please comment on any problems you may have had with the IM pilot > service Learning to use @mit.edu "handles" was an adjustment, but once it was learned it was easy enough to remember. Creating a chat room in Gaim easy, but configuring a chat room is hard. This may be something we have to live with. > 6. Please add any suggestions/comments '() From pbh at MIT.EDU Mon Dec 19 14:51:25 2005 From: pbh at MIT.EDU (Paul B. Hill) Date: Mon, 19 Dec 2005 14:51:25 -0500 Subject: [Wocky] Initial Evaluation of Jabber IM In-Reply-To: <811ADEE8-2050-43A4-9F71-1A85502582AC@mit.edu> Message-ID: <200512191951.jBJJpUjO009349@outgoing.mit.edu> -------------------------------------------- 1. Which client did you use? Gaim for Windows Exodus for Windows Imov for PocketPC 2003 2. How did you acquire the client software? Download from vendor site (imov, and Gaim on XP) Download from MIT site (Gaim on Vista beta) 3. Please rate the following for the client that you used: (can use a 1 - 5 scale) MIT download and installation Download was OK. There was a z-buffer problem while running the installer on Vista. A configuration screen that required user input was displayed behind all other windows making it look like the installer had hung. Creation of "buddy" list Surprised by the consistently poor UI in each of the clients. There were no onscreen hints to indicate that the user should enter a qualified username (@mit.edu) in the dialog box. By default clients did not indicate that the user being added was offline and therefore didn't appear in the buddy list by default. Most clients didn't indicate that the buddy being added had to authorize this step before messages could be sent. Initiation of IM conversation Easy after buddy approval step had occurred. IM conversation/chat history Chat history is useful. It would be more useful if there was a way to see history that occurred before the server was last restarted. The imov PocketPC client does not provide the ability to join chat rooms from what I can see. It appears to be limited to messaging between individual users. Creation of multi-user Chat I prefer the Exodus UI to the Gaim UI for this task. Speed Reliability Appears to work OK with users on the same server. It is difficult for a user to determine the source of problems when interacting with resources on different servers. In many cases it works, in some cases it doesn't, but the clients don't provide enough information for the end user to understand why. Login So far I have only used SSL. I would prefer GSSAPI. 4. Please list other features or functionality that you would like to have in an IM service I'd like to understand how to use bots running at other sites. I'd like to be able to create bots at MIT. 5. Please comment on any problems you may have had with the IM pilot service 6. Please add any suggestions/comments -------------------------------------------- Joanne Hallisey Sr. Project Manager Information Services and Technology W92-153 617-253-1894 hallisey at mit.edu _______________________________________________ Wocky mailing list Wocky at mit.edu http://mailman.mit.edu/mailman/listinfo/wocky From ghudson at MIT.EDU Mon Dec 19 19:10:54 2005 From: ghudson at MIT.EDU (Greg Hudson) Date: Mon, 19 Dec 2005 19:10:54 -0500 Subject: [Wocky] More updates on gaim's Jabber authentication Message-ID: <200512200010.jBK0AsaV026190@egyptian-gods.mit.edu> I've modified the gaim code which dynamically loads the GSSAPI module so that it can work for win32 and various different Unix GSSAPI libraries. I also took out the hardcoded /usr/athena/lib dependency. I've put a build and a source tree checkout in /afs/dev.mit.edu/project/jabber/gaim-krb5. Qing, can you test this code under Windows? While corresponding with Simon and feeding back my changes, I learned that Mozilla is apparently shifting to a strategy of dynamically loading the GSSAPI module, using the same approach (a bundled gssapi.h). I was able to verify that with the current Mozilla code. So it's possible that this approach would be reasonable for Gaim as well. The bundled gssapi.h in Mozilla is identical. Simon said that he still uses Cyrus SASL for Linux local deployments (perhaps because his SASL-using patch has security layer support), but the direct GSSAPI approach "was the only way to get Gaim with GSSAPI working on Windows (and of getting libgaim working on Mac OS X, so that Adium would run)." I don't know what the specific issues were. Simon says he has attempted to get the Gaim developers interested in his changes, but has had no luck. From lurking on the gaim-devel list, it looks like they are pretty focused in their 2.0 release, so I will give it a try after that's out. In the meantime, we may want to think about distributing a locally-modified version of Gaim. Two people were talking on Zephyr about playing with Gaim 2.0.0beta1, and say it has issues authenticating to jabber.mit.edu. I will look into that this week. From hallisey at MIT.EDU Tue Dec 20 12:19:31 2005 From: hallisey at MIT.EDU (Joanne Hallisey) Date: Tue, 20 Dec 2005 12:19:31 -0500 Subject: [Wocky] Next Meeting Message-ID: Hello, I will be away until Dec. 28. While I had hoped we would meet this week, I did not have a chance to send out a reminder for yesterday. My apologies. In the next week, please consider what our next steps are for this pilot. If we decide to continue in Pilot, what is our focus? If we decide to roll out a service, what are some of the risks, and, how do we minimize, avoid etc. You can also consider what the impact of a no-go decision is. We can have a discussion around this next week. I have reserved W92-125 on Wednesday, Dec. 28 at 2:00 PM for our next meeting. Please let me know if this time does not work. Thanks, and have a wonderful holiday. Joanne -------------------------------------------- Joanne Hallisey IS&T W92-153 3-1894 hallisey at mit.edu From ghudson at MIT.EDU Tue Dec 20 14:31:17 2005 From: ghudson at MIT.EDU (Greg Hudson) Date: Tue, 20 Dec 2005 14:31:17 -0500 Subject: [Wocky] Flaw in reciprocal buddy-adding Message-ID: <200512201931.jBKJVHF2015747@egyptian-gods.mit.edu> Until today, I never actually added anyone to my Gaim buddies proactively (except for myself). I always waited until they added me, at which point after I authorized them I was asked if I wanted to reciprocally add them, and I said yes. Gaim does not seem to request authorization when you reciprocally add a buddy. I just tried this with systest (using @mit.edu on all the JIDs, of course): systest adds ghudson as a buddy ghudson receives an authorization dialog ghudson reciprocally adds systest as a buddy systest does *not* receive an authorization dialog systest appears offline to gaim, or "Not authorized" if I l-click If I r-click and "(Re-)request authorization" from systest, systest gets an authorization dialog and all becomes kosher. I assume this is a Gaim bug, but it could conceivably be a jabberd bug if the server is expected to provide reciprocal access. I will put this on my list of things to dig into. From ghudson at MIT.EDU Tue Dec 20 15:24:14 2005 From: ghudson at MIT.EDU (Greg Hudson) Date: Tue, 20 Dec 2005 15:24:14 -0500 Subject: [Wocky] Gaim 2.0 and Jabber authentication Message-ID: <200512202024.jBKKOEtf016294@egyptian-gods.mit.edu> It looks like Gaim integrated Simon's Cyrus SASL support (not the dynamically-loaded GSSAPI support) for Jabber on December 17. Then they released the beta. Then on December 19, they reverted Simon's change to make passwords optional in the Jabber plugin, since that change broke password authentication if you're not storing the password. The code which handles requesting the password for a new connection has changed a bunch between 1.5 and 2.0, so I'll have to rework my patch to allow requesting a password after the connection is opened. I'll also have to rework the Jabber side of that patch against the Cyrus SASL code instead of the GSSAPI code. From ghudson at MIT.EDU Wed Dec 21 01:49:52 2005 From: ghudson at MIT.EDU (Greg Hudson) Date: Wed, 21 Dec 2005 01:49:52 -0500 Subject: [Wocky] [Fwd: Re: [jabberd] Online Game Group is pleased to announce the release of palaver 0.1.] Message-ID: <1135147792.12162.396.camel@egyptian-gods.mit.edu> We should keep in mind that there are alternatives to mu-conference if we want better group chat functionality. I'm assuming, perhaps optimistically, that palaver interfaces with Jabberd 2. This particular one would likely require us to install the Twisted framework, but it might allow us to ditch the JCR in return. (The JCR, which we have a repository for alongside the jabberd 2 repository in /afs/dev/project/jabber, is a way of making jabberd 1.4 components interface with jabberd 2; in our case, we use it to make mu-conference work. I don't particularly like its build system and would be happy to see it go.) -------------- next part -------------- An embedded message was scrubbed... From: Christopher Parker Subject: Re: [jabberd] Online Game Group is pleased to announce the release of palaver 0.1. Date: Tue, 13 Dec 2005 10:07:57 -0600 Size: 4633 Url: http://mailman.mit.edu/pipermail/wocky/attachments/20051221/442f6655/attachment.eml From ghudson at MIT.EDU Wed Dec 21 10:32:16 2005 From: ghudson at MIT.EDU (Greg Hudson) Date: Wed, 21 Dec 2005 10:32:16 -0500 Subject: [Wocky] [Fwd: [Gaim-devel] Jabber, SASL, passwords, binary packages] Message-ID: <1135179136.12162.416.camel@egyptian-gods.mit.edu> I realized this morning that I cannot easily adapt my password-request support to Simon's Cyrus SASL patch, because Cyrus SASL will be requesting the password in a callback which must return the information immediately. Since we can't put up modal dialogs in Gaim given it's architecture, we need another approach. Also, I'm afraid that binary packagers on some platforms (particularly Windows) won't build in Cyrus SASL support. So I sent this message to gaim-devel proposing a gaim_sasl layer which (a) can have a direct dynamically-loaded GSSAPI implementation when the Cyrus SASL plugin isn't available, and (b) will "just know" when a given mechanism is going to require a password. Here's the message I sent. -------------- next part -------------- An embedded message was scrubbed... From: Greg Hudson Subject: [Gaim-devel] Jabber, SASL, passwords, binary packages Date: Wed, 21 Dec 2005 10:25:53 -0500 Size: 6316 Url: http://mailman.mit.edu/pipermail/wocky/attachments/20051221/f0dbac45/attachment.eml From raeburn at MIT.EDU Thu Dec 22 19:07:28 2005 From: raeburn at MIT.EDU (Ken Raeburn) Date: Thu, 22 Dec 2005 19:07:28 -0500 Subject: [Wocky] jabber for web browsers? Message-ID: <8F0792D1-DB13-45BD-86EF-3789903611A0@MIT.EDU> Is anyone looking at web-based solutions for people unable to install software on systems? (E.g., one friend of mine at a company where policy forbids it; borrowing a friend's laptop and not wanting to mess around with the configuration.) There seem to be a couple of approaches in freeware that wouldn't require sending the user's password off to some third-party site. If I'm reading it right, JabberApplet lets you install on your web server Java code that a browser can run to connect back to a Jabber server on the same server host. Then there's JWChat, which appears to be an AJAX-based web interface, which on the back end plugs into your Jabber server. I don't know anything more about them, like whether they'd be practical for us to use securely, but having a friend I'd like to chat with who can't install software on her machine at work, I'm getting interested in the issue... Ken From hallisey at MIT.EDU Wed Dec 28 08:59:19 2005 From: hallisey at MIT.EDU (Joanne Hallisey) Date: Wed, 28 Dec 2005 08:59:19 -0500 Subject: [Wocky] Meeting Wed. at 2:00 Message-ID: Hello, Before I left for vacation, I had scheduled a meeting for today, Wednesday at 2:00 in W92-125. If this time is not good, please let me know what other time this week will work. I would like to review progress and write our recommendation to Theresa. Thanks, Joanne -------------------------------------------- Joanne Hallisey IS&T W92-153 3-1894 hallisey at mit.edu From ghudson at MIT.EDU Wed Dec 28 16:19:07 2005 From: ghudson at MIT.EDU (Greg Hudson) Date: Wed, 28 Dec 2005 16:19:07 -0500 Subject: [Wocky] Flaw in reciprocal buddy-adding In-Reply-To: <200512201931.jBKJVHF2015747@egyptian-gods.mit.edu> Message-ID: <200512282119.jBSLJ7aG013327@equal-rites.mit.edu> > I assume this is a Gaim bug, but it could conceivably be a jabberd > bug if the server is expected to provide reciprocal access. I will > put this on my list of things to dig into. It's a Gaim bug. Here's what happens: 1. Alice adds a buddy Bob 2. Bob receives an authorization request for Alice's subscription 3. Bob approves the request (message sent to server) 4. Bob's Gaim process sees that Alice is not currently a buddy, and displays the reciprocal buddy-adding dialog. (Like all Gaim dialogs, this one is non-modal, so control returns to the main loop.) 5. Bob's Gaim receives a roster push from the server containing a buddy entry for Alice, with a subscription type of 'From'. Bob's Gaim stuffs that into the 'Buddies' group since no group is specified. 6. Bob finishes filling out the reciprocal buddy-adding form, thus asking Gaim to add Alice to his 'Buddies' group. (If Bob chooses a different group, then things actually work.) 7. Bob's Gaim sees that Alice is already a buddy, due to the roster push in step 5, and does nothing. Bob may as well have clicked cancel on the reciprocal buddy-adding dialog. There is clearly an impedance mismatch between Jabber, where you can have a buddy entry but not have requested subscription rights, and Gaim, which was built around the simpler AIM model. Also, it's not clear to me that Gaim is making a wise decision by sticking ungrouped roster entries in the 'Buddies' group. Confusing matters, the conditional in step 7 only appears on the gaim 1.5 branch, along with a bunch of other work which never seems to have made it to the mainline. So Gaim 2.0 probably doesn't manifest this bug, but might manifest other bugs instead. I will test some candidate fixes soon, and will also try this scenario and some related scenarios in the Gaim 2.0 beta. From hallisey at MIT.EDU Thu Dec 29 09:49:07 2005 From: hallisey at MIT.EDU (Joanne Hallisey) Date: Thu, 29 Dec 2005 09:49:07 -0500 Subject: [Wocky] Meeting Notes: Jabber - Dec. 28. 2005 Message-ID: Jabber Meeting Notes December 28, 2006 Recommendation to Theresa for next steps on Jabber. There are few things that would make the adoption of Jabber less successful. The lack of single sign on support for an MIT local service would lead some Athena users to resist. Another issue is that Gaim does not seem to request authorization when you reciprocally add a buddy. Greg will be investigating this, but it is not a show stopper. We have also noticed since the pilot user group is somewhat small, not many people are using Jabber after initial experiences. There needs to be more testing of the group chat functionality. There needs to be a better understanding of the server to server functionality. The project will continue in pilot. By the end of January the team would like to begin work on support and service plans. Some tasks that will prepare for this follow: - Do a focus group during IAP to get moere community feedback - Joanne will talk to Training and IS&T Comm Teams. - Establish a Support Plan - Talk to training and Pubs - Create Quick Start training for users - FAQs, and Self Help documentation - Training for Computing Help Desk - Establish a Service Plan - Server maintenance plan - talk to Mark - Determine change processes - Develop Communication Plan Greg also looked into web solutions> - There is a jwchat for Ajax solution - uses http polling or binding - possible to modify this to use certificates - need to look into this possibility A better solution may be: - jabber applet - because it is written in java could not have Kerberos support - need to use passwork over ssl Greg will set up and test to see if this is a viable solution. We will need to determine how to introduce this. (I would suggest asking one of the interested groups to be a tester) -------------------------------------------- Joanne Hallisey IS&T W92-153 3-1894 hallisey at mit.edu -------------------------------------------- Joanne Hallisey Sr. Project Manager Information Services and Technology W92-153 617-253-1894 hallisey at mit.edu From pbh at MIT.EDU Thu Dec 29 10:08:12 2005 From: pbh at MIT.EDU (Paul B. Hill) Date: Thu, 29 Dec 2005 10:08:12 -0500 Subject: [Wocky] Meeting Notes: Jabber - Dec. 28. 2005 In-Reply-To: Message-ID: <004f01c60c89$b04124f0$b103bc12@pbhtablet> >A better solution may be: >- jabber applet - because it is written in java could not have >Kerberos support - need to use passwork over ssl >Greg will set up and test to see if this is a viable solution. We >will need to determine how to introduce this. (I would suggest asking >one of the interested groups to be a tester) Actually Java applets can use Kerberos and GSS. GSS/Kerberos support was first added in 1.3 although there were a number of issues in the older implementations. SASL support was first added in 1.4. There was a big improvement from JDK 1.4 to 1.5. It is already possible to create a single sign-on Java applet when the ticket cache is file based when using 1.5. With 1.6, a number of improvements are being made. On Windows Vista with 1.6 and KfW installed the ability to have single sign-on while using the in memory ticket cache will be possible. Paul From ghudson at MIT.EDU Thu Dec 29 21:32:45 2005 From: ghudson at MIT.EDU (Greg Hudson) Date: Thu, 29 Dec 2005 21:32:45 -0500 Subject: [Wocky] Server-to-server notes Message-ID: <200512300232.jBU2WjXW029627@egyptian-gods.mit.edu> I did some testing today regarding server-to-server with jabber.org. My findings: * wocky.mit.edu to jabber.org works * jis.mit.edu to jabber.org works * mit.edu to jabber.org fails: mit.edu -> jabber.org fails with a service-unavailable error jabber.org -> mit.edu fails with a remote-server-timeout error My best guess right now is that jabber.org does not support SRV records. I suppose we could run a port forwarder on mit.edu to work around servers like that, but I'm not sure how NIST would feel about the idea. From ghudson at MIT.EDU Fri Dec 30 13:41:13 2005 From: ghudson at MIT.EDU (Greg Hudson) Date: Fri, 30 Dec 2005 13:41:13 -0500 Subject: [Wocky] Flaw in reciprocal buddy-adding In-Reply-To: <200512282119.jBSLJ7aG013327@equal-rites.mit.edu> Message-ID: <200512301841.jBUIfD4o014351@egyptian-gods.mit.edu> > Confusing matters, the conditional in step 7 only appears on the > gaim 1.5 branch, along with a bunch of other work which never seems > to have made it to the mainline. So Gaim 2.0 probably doesn't > manifest this bug, but might manifest other bugs instead. > I will test some candidate fixes soon, and will also try this > scenario and some related scenarios in the Gaim 2.0 beta. As I expected, Gaim 2.0 beta exhibits differently-buggy behavior. With the older code (which is in the newer release), the reciprocal buddy add works, but the local buddy list winds up with two entries for the same buddy, one of which is functional and one of which is not. The redundant entry does not exist on the server, so if you quit and restart Gaim, the redundant entry is cleaned up and you wind up with just one, functional buddy entry. I came up with a fix for the Gaim 1.5 behavior which does the right thing: if you're trying to add a buddy which already exists, tell the server to add it, but don't redundantly add it to the local buddy list. I've checked this into the Gaim repository in /afs/dev.mit.edu/project/jabber, in anticipation of a local deployment with some Jabber fixes and GSSAPI support. There was an old bug entry for the old bug, so I submitted my findings there. See: https://sourceforge.net/tracker/?func=detail&atid=100235&aid=1041829&group_id=235