Here's the log from the Tuscany weekly chat from yesterday.
The main topic discussed was how to get from the sandbox fork to a single code stream we can all develop into the M2/1.0 release. The two approaches on the table are summarized in these two emails: http://mail-archives.apache.org/mod_mbox/ws-tuscany-dev/200606.mbox/[EMAIL PROTECTED] http://mail-archives.apache.org/mod_mbox/ws-tuscany-dev/200607.mbox/[EMAIL PROTECTED] There was a lot of debate. A suggestion was to call a vote to decide what to do. Two thought that would be a useful approach, two others didn't. No conclusions were reached. ...ant <ant> hi everyone. 5pm on my clock <isilval> it behaves as a forward reference * haleh has joined #tuscany <jmarino> k so in the Java case the proxy calls back onto the reverse chain? <isilval> right <ant> what topics do we have for today? the conversational stuff you two are talking about now, anything else? * kevin has joined #tuscany <ant> I'd like to see if we can get any decission on a shared code base <jboynes> how about how do we use IRC :) <jmarino> sorry, yea, are there other things people would like to chat about too? <isilval> there is a callback invocation handler * monxton has joined #tuscany <jmarino> isilval k so that would call out over a transport then? <isilval> hmmm, yes <jmarino> k then the invocation gets picked up back on the callback VM through a transport listener? <isilval> tho it's not as clean as going from a composite's reference to a composite's service <jsdelfino> hi <jmarino> hi Sebastien <isilval> yeah, I'm using a ws binding as my mental model <jmarino> right. So it sounds like we may be saying similar things, and I like your approach of the bidirectional wire <isilval> it would be great if we could abstract the transport listener ... <jmarino> yea I think we can <isilval> the way it is abstracted by a service <jmarino> that would just be a binding which is a system service <jmarino> on the way back to the callback instance, it would be just a regular dispatch <jmarino> you could theoretically callback over a different binding <jmarino> transport <isilval> right * kgoodson has quit IRC (Read error: 104 (Connection reset by peer)) <jmarino> k so we don't monopolize other discussions, do you want to post maybe on the wiki or an email, the steps as you see it and then we can iterate on how that fits with the various subsystems? <isilval> that sounds great ... <jmarino> k cool <isilval> do I need auth to post to the wiki? <jmarino> no just and account <jmarino> you can register <isilval> k, will do <jmarino> the editing takes a little bit to get used to... <ant> can we talk about what to do about getting to a shared codebase now? what to do with the elephant? <jmarino> just look at some of the pages for examples <jmarino> lol elephant <isilval> k, thx <jmarino> sure <jsdelfino> if u go to the page you want to edit and you're not logged on, there's a toolbar on the top right if I remember correctly with a button to register and log in <ant> The project has stalled over this <ant> There's only been two approaches suggested - use the sandbox, or start a fresh stream merging in the best bits of M1/Sandbox/new develeopment <ant> This has been going on for weeks now, we've all looked at the code and debated what to do <jmarino> I think Sebastien's was to start over, right? Not a merge <ant> I think the time has come, we need to choose <ant> I'd like to call a vote on this now, it will be done by next Monday, and then we can start moving forward again <jmarino> not on IRC though :) <ant> I know it may be devisive, but the current situation is becoming unhealthy <jsdelfino> yes I'm not really proposing a merge <ant> no, it would be on the dev list <ant> What do you all here think, is it time to vote? <jmarino> I don't think we know enough about the alternatives yet <ant> (please all air you opinion now, don't just sit back being quiet) <jmarino> IMO we're not ready for a vote <jmarino> also I don't think we should have one..we should reach consensus <haleh> What do we need to know to get a voting <jmarino> specifics <jmarino> we also should be able to come to consensus <ant> i understand that, i'm interested in everyone else who's here views <jsdelfino> +1 for starting fresh, step back, discuss the design again, and rebuild the runtime through baby steps driven by scenarios from me * bhdaniel has joined #tuscany <jmarino> jsdelfino,.... <jboynes> I think we made progress on using scenarios <jboynes> there was some good discussion on that over the weekend <jli1> +1 for moving sandbox forward and starting adding more into sandbox <jmarino> could you respond to why you think your criticisms in the email can incrementally be done through core2? <jmarino> cannot sorre :) <ant> guy's th evote will be on the dev list <jli1> :--) <ant> what i'm looking for here now, is if enough people agree its time to call a vote <rfeng> I assume vote can only be cast by committers only, right? <jboynes> I think we just about reached consensus there <ant> no, everyone should vote <ant> but only committers votes are binding <jmarino> wait...what is binding? <jmarino> ah k <jmarino> jsdelfino, can you respond to my question? <rfeng> If we're going to vote, can we give jsdelfino a chance to present his ideas like what we did for the sandbox? <jsdelfino> jim, everything is always possible, restarting from scratch is possible, evolving M1 is possible, evolving the sandbox is possible, with infinite time and resources, everything is always possible, the question is really more what is the *best* path for the project and community moving forward <jmarino> I don't think it is that much refactoring - Map -->List? <ant> guys, there's 16 people on this chat, could any of you other than the usual 4 of us voice an opinion? <jmarino> yes agreed that Sebastien should present...now I'll be quite :) <kevin> my sense is that we will end up in a very similar place whether we merge or start fresh ... <jsdelfino> and I'm claiming that, after having done an M1, this is the right time to rethink the design and restart with a fresh stream reusing some pieces and integrating design ideas from all of us, that's what I'm proposing, it's very simple and natural IMO <kevin> the main question as Sebastien said is which is least effort <ant> no thats not it IMO, its also about getting community involvement <jsdelfino> the main question is to allow others than the usual 4 to participate..., starting fresh gives us a better path to that in my view <jsdelfino> ant, yes :) I agree <jmarino> well others have been participating (quite again) <kevin> right, "best" is better than "least effort". i agree <jsdelfino> I think we should give this effort a chance <jboynes> I think we should give incremental improvement a chance <jmarino> why can't we refactor core2? technical specifics, please? <rfeng> I suggest we should first evaluate jsedelfino's technical proposal and then decide the mechanics (merge or start fresh) <ant> rfeng, i think sebastien summed it up in this email: <ant> http://mail-archives.apache.org/mod_mbox/ws-tuscany-dev/200606.mbox/[EMAIL PROTECTED] <jmarino> that's not specifics - it's all motherhood and apple pie stuff <ant> Jeremy here in this: <ant> http://mail-archives.apache.org/mod_mbox/ws-tuscany-dev/200607.mbox/[EMAIL PROTECTED] <jmarino> Raymond has a good point <jmarino> Sebastien, would you present for us? <ant> i think thats been done on the mailing list already <isilval> jsdelfino, I am not sure why the sandbox is not a good enough point to go forward/evolve from; from a practical perspective, I was starting to get used to M1, then the sandbox; starting over a third time is not clear to me is the most productive option, assuming no compelling technical reasons to start over <jboynes> architecturally, sandbox is very similar to M1 <jmarino> no ant it hasn't been done <jmarino> I'm looking for architecture ideas <rfeng> right, same from me <jboynes> that should be the easiest path forward for all <ant> so... <jmarino> Sebastien is asking us to make an important decision. We need to give him the opportunity to present in detail what he is proposing <ant> should i be calling a vote or not? <jmarino> technical detail <jmarino> what basis are we going to make a decision on? <jsdelfino> jim, I think you're missing something, the point here is not for one person to present an achieved technical proposal against another technical proposal represented by a sandbox codebase, the point is to discuss as a community the technical direction and architectural ideas, and use these discussions to rebuild the runtime <jsdelfino> we reached a point with M1 <jmarino> exactly so what are yours? <jmarino> you are making a claim.. <jboynes> you are pre-supposing he community wants to rebuild the runtime <jmarino> that the core2 is not good starting point. Why? <jboynes> we can discuss technical issues and architecture but they should be in the context of what's there <jboynes> only when we all agree we're stuck should we embark on a rewrite <jli1> As I have been working on the celtix binding, i would like to comment on the spi stuff. the new SPI, wiring stuff are more clean and better designed than M1. <jboynes> thanks <jli1> So no matter we start fresh or from sandbox, i feel the current spi code base in sandbox is a good starting point <jboynes> I would like to see more scenarios presetned <jsdelfino> I am not presupposing anything, I looked at M1 and sandbox, they have many of the same problems IMO, and I'm saying that now we have a good opportunity to do better by restarting with less heritage from M1 or sandbox and getting more people involved... since we're done with M1 and all free to work on this now <jmarino> Sebastien, can't we just start from there and I think you bring up some really good points that we can improve core2 with? <jboynes> that would help us decide whether we are in a position to proceed <jboynes> if you're not presupposing anything, why can't we continue from where we are? <jmarino> we build stuff off the scenarios on the wiki (we can add more) <isilval> can we assume that the architectural ideas in tuscany.architecture.v4.ppt would not be impacted by a rewrite, or is that one of the reasons for one? <jmarino> I believe Sebastien is proposing we start from scratch so nothing in the preso applies, ex nihilo <jsdelfino> yes, and the scenario work is part of this exercise <jmarino> so can't we just start from core2? <jsdelfino> scenarios will help us put what we do in perspective <jmarino> what's the issue? <jboynes> if it fulfills the scenarios <jboynes> or can be made do so <jboynes> through incremental improvement <ant> i think starting from somewhere other ethan the sandbox would make it eaiser for others to engage. Thats just my opinion, i think we just need to vote and be done with it <jmarino> why? people have started engaging? <jmarino> also having something there to reference is easier IMO <ant> so then they'll all vote for the sandbox. by early next week we know <haleh> Why did we re-write core 2? Is it because M1 could not be tailored? Do we have the same issue with the improvements? <jmarino> I contend we don't have the same issue with the improvements <jmarino> all of the things Sebastien brought up, assuming they are correct, can be incrementally added to core <jboynes> perhaps you should read that thread ant linked to <jsdelfino> I'm re-rereading jeremy's proposal, the interesting thing is that it says that core2 restarted a fresh stream :) <jsdelfino> so I could ask the same question, why did core2 restart a fresh stream? <jmarino> Seabstien, again, what is the big deal with just starting with core2? <jboynes> wasn't quite fresh - it pulled in a lot from M1 <jli1> hi, one question. do we have a specific date for M2? or we will have infinite time to do whatever we want <jboynes> I don't have a date although I liked Jim's 6 week proposal <jsdelfino> the date is open IMO, I think we release M2 when we're happy with it <jboynes> as I said, architecturally it is very similar to M1 so it was easy to pull stuff over <jboynes> if we change the architecture as jsdelfino is proposing I don't think we will get as much reuse <jboynes> and if we don't change the architecture, then why start over? <haleh> Jeremy, following up on what you just said. If the architecture is so different, how can Core be tailored to fit it? <kevin> so is the main question whether the current architecture has become unwieldy? <jboynes> that's why I don't think we will get reuse <jmarino> I'm confused. Jeremy was responding to whom? <jboynes> whereas with the sandbox the architecture was the same so the freshness was just refactoring to clean up fragility <ant> I don't think reuse is despreately important. As you've said the two of you wrote the bulk of the sandbox in under two weeks <jmarino> no not under two weeks <jmarino> where did you get that? <ant> with all of us helping we should be able to get something going in a month or two you'd think <jmarino> look at svn log <jmarino> if core2 is usable why start over? <haleh> Is the time the issue or the technical merit of the proposal? <jsdelfino> but the thing is that this was written just by 2 people, not the whole group was involved, and there were not many discussions on its design, since the rest of us was busy with M1 <ant> From a discussion with jeremy at apachecon (unless i misunderstood) <jmarino> it's not a time issue <jmarino> you misunderstood <jmarino> yes there were discussions on design...anyone could have jumped in <ant> i agree its not a time issue, starting a fresh will take a bit longer in the short term but that doesn't matter in the long run <jmarino> you also told me you didn't want to work on core anymore, if you remember <haleh> OK. then let's stay focus on the technical merit and also community involvement. <jboynes> and of course we didn't really change the design <jsdelfino> so I think restarting a new stream now, with the experience of M1, the experience of the sandbox, scenarios to help make sense of the direction, and more of us involved will get us in a better position <jsdelfino> to build a longer term community <jmarino> why? <jmarino> doesn't engaging core2 do the same? <jmarino> while things could be easier (docs, etc.) there is something there to work with <jboynes> and people can engage on specific things (as they have been) rather than starting with blue-sky discussions <ant> so...shall i call a vote to find out <jsdelfino> I'm ok with a vote <haleh> do we have all the technical issues answered to take a vote? <jsdelfino> you guys already know how I'll vote :) <jmarino> no we don't <ant> its not just a technical issue!!!! <jmarino> IMO Sebastien has not pointed to the issues which prevent core2 from being used <jmarino> yes it's not just technical but that is a huge factor <ant> no, its just _a_ factor <jboynes> right and forcing a vote is being divisive rather than consensus building <jmarino> yes exactly <jsdelfino> it's one of the factors <jmarino> and I think we need to be careful as that may destroy the project <jmarino> IMO it demonstrates we can't reach consensus <ant> if we don't its just the sandbox by default, as as soon as anyone takes the smallest look you say they've engaged in it <ant> its not an elepant in the sandbox its the Borg Collective <jmarino> no not necessarily <jmarino> come on...stop that <jmarino> I don't think we need to be nasty..that is not community building <jsdelfino> again, my main proposal here is to take a new look at the design of our runtime, the spec is changing, we're just starting to look at deployment and I'm not sure anybody here has a clear picture on the deployment story, we've seen complexity and runtime brittleness in M1 <jsdelfino> so I'd like our whole group <jmarino> so, what's wrong with using core2 to do that? <jsdelfino> (and not just one or two or three of us) <jsdelfino> to be able to attack the next wave of design / implementation of the runtime <jboynes> I've been asking for deployment ideas for weeks <jsdelfino> and go hack core2 is really not the right way to do IMO <jboynes> that would be a good area to start technical discussion <jsdelfino> it's already a big codebase <jboynes> its independent of which codebase we start with <jsdelfino> no the codebase you start with is an important factor * Venkat has joined #tuscany <jboynes> why? wouldn't we start with end-user scenarios? <jsdelfino> we saw that at the beginning of the project, when we started with 2 existing codebases <ant> hi Venkat <jsdelfino> and changing them was a nightmare <Venkat> Hello Ant <jsdelfino> changing core2 now (or changing M1) will be very difficult <jboynes> merging them was a nightmare, we're beyond that now <jmarino> why is it difficult? <jmarino> Sebastien, excuse my frustration, but you are making assertions without examples <jsdelfino> no need to be frustrated, I started to give some examples in an email I sent yesterday, and got pretty offensive responses accusing me of dumping a huge codebase <jsdelfino> it was just a bunch of interfaces <jboynes> (time check, some of us need to drop off in 10 minutes for an SCA spec call) <jmarino> was my response offensive? <ant> i have to leave soon. I've still not heard many voices saying yes or no on calling a vote? <jmarino> none of those reasons entail droping core2 <jsdelfino> to illustrate some design ideas and trigger a discussion on the recursive model <jmarino> things like Map-->List? <rfeng> did anyone have a chance looking into Sebastien's model code? <jboynes> I had a look yes <ant> me too <Venkat> me too ... partly <jsdelfino> I think it's very important that we all understand the new recursive model that we'll have to support <ant> i liked it <jmarino> ant, why do you like it better than core2? <jsdelfino> the code/interfaces I dropped is a support for discussion <jsdelfino> and like I said it's just a strawman... we have an opportunity here <jsdelfino> get ALL of us involved in this design <jsdelfino> more eyes will give us a much better result <jboynes> jsdelfino: why not engage in a discussion on the sandbox code? <jsdelfino> and more fun for everybody in the group <ant> jboynes, we've been doing that for many weeks now, on th emailing list, apachecon, phone calls etc <jboynes> all the points in your email could have been made in that context <jsdelfino> my email outlined a number of issues in both M1 and/or the sandbox, I'm not sure I understand what you mean... <jboynes> why does List vs. Map warrant a rewrite? <jboynes> why does not supporting include warrant a rewrite? <ant> it was a much longer email than just that <jboynes> why does the location of Remotable warrant a rewrite <jsdelfino> you're picking on individual things here, look at the the sum of them... <jboynes> these are all things that can be incremental improvements <jmarino> yes ant, why id any of the things mentioned warrant a re-write? <jmarino> even in total? <ant> no one is saying anything new here. I really don't see a way forward without a simple vote. Is there any alternative? <jboynes> how about we work together to make what we have better? <jmarino> ant I would be *very* careful with that <jmarino> it will show we can't reach consensus <jmarino> it may very well ruin the project <jsdelfino> I'm going to say it again :) my proposal is to restart a fresh stream with more of us involved in the actual design of M2, and go with consensus on technical decisions <kevin> more time spent to reach consensus seems best <jboynes> so lets work together instead - let's lay out scenarios that matter to all of us and look at whether we can support them <jsdelfino> this is very different from core2 where just a small subset of the whole group was involved <jsdelfino> and now we're struggling to understand what to do with it * lresende has quit IRC (Read error: 104 (Connection reset by peer)) <jboynes> so, how about we work together to make what we have better? * lresende has joined #tuscany <jsdelfino> that's what I'm proposing, we have bits and pieces in M1, sandbox, various ideas that I'm sure people have in their heads, their laptops or their sandboxes <jsdelfino> and I'm proposing that the whole group work together to build M2 from that <jmarino> why not do it with core2 as a starting point? What's the problem? <jboynes> let's define some of your scenarios <jboynes> and fix some of the issue that you raised <jboynes> working together to make things better <jsdelfino> why can't we start a fresh stream and bring small pieces into it, one (baby) step at a time, I think it'll help a lot <Venkat> My HUMBLE opinion is that we park this code base discussion and recap on the architecture and design goals of the core. We may then use the code that best implements this design... maybe from both codebases. If everybody has understood the design goals well I don't see why there must be this argument over implementation approaches. <jsdelfino> I'm +1000 on that! <jsdelfino> this is the time to think through the design goals again, scenarios, design goals and new design ideas <jboynes> thing is, at an architectural level we don't have two codebases <jmarino> so we continue evolving both streams? <jsdelfino> hacking existing codebases is not going to get us therre <jboynes> really, why not? <jboynes> s/hacking/improving/ of course <jmarino> let's go with the scenarios and hack the code and see what works better? <jboynes> let's define the scenarios and see if what we have supports them <jmarino> can we do that and call it a day? <jboynes> we've already added some over the weekend, let's add more so we get a bigger picture <jmarino> btw congrats to Dan on becoming a dad! <Venkat> jboynes, if I were to understand if any of the codebases does the right thing or not... what am I supposed to get back to... the specs which is not public ... what else.. <rfeng> that's great news :-) <dkulp> Thanks. <kevin> congrats! <jsdelfino> hey congrats dan <jboynes> congrats (again) :) <jboynes> Venkat: that's what the scenarios help define <kevin> sleep is overrated <dkulp> :-) <jboynes> they shape what we want the project to do <jmarino> so let's go with the scenarios and hack away. Sebastien, it would be great if you put some in <jmarino> can we close now, it's kind of late here? <ant> I'm still interested in getting it over and done with a vote. but i have to go now so no more time to debate * ant is now known as ant_away <jsdelfino> I have to go too <jboynes> do you agree with the apporach? * Venkat has left #tuscany <jboynes> are you willing to collaborate on it? <jsdelfino> is this a question for me? I'm the one who proposed using scenarios... <jsdelfino> ok I really have to go now <jmarino> yea I thought it was a great idea and got busy this week on it <jsdelfino> I hope you donmind * halehM has joined #tuscany <jsdelfino> don't mind my question, but... do we have any scenarios yet? <jmarino> Sebastien, add yours to the wiki...a bunch of other have as well <jboynes> http://wiki.apache.org/ws/Tuscany/TuscanyJava/Scenarios?highlight=%28tuscany%29%7C%28scenario%29 <halehM> these are bunch of scenario names, no content :) <jsdelfino> :) <jsdelfino> the only one I can see is rick's eagerinit scenario <jmarino> there is also content <jmarino> try conversations * kevin has quit IRC <jmarino> I filled the ones I'm interested in <jboynes> so we have plenty of opportunity :) <jmarino> some others have added too like Meeraj and the guy from Oracle * Venkat has joined #tuscany * haleh has quit IRC (Read error: 110 (Connection timed out)) <jsdelfino> Like I had said in my initial proposal, I'm think about higher level end to end scenarios (like Bigbank but simpler) I'll check that in SVN. <jsdelfino> I really have to go now <jsdelfino> ttyl <jmarino> why not fill a scenario out * jsdelfino has left #tuscany <jmarino> I understand what you want... <jmarino> I'm interested in more lower level tech scenarios * dkulp has quit IRC (Read error: 104 (Connection reset by peer)) <jmarino> we can have both <jboynes> jmarino: he left <jmarino> too bad <jmarino> well maybe we should wrap up? <jmarino> sounds like we work through scenarios and hack code until we understand things better? <jmarino> a vote seems to be the wrong thing to do IMO <jmarino> comments? <jboynes> I think we need to give consensus building a bit more of a chance <jmarino> ant? <jboynes> he left as well <jmarino> anyone else? * jli1 has left #tuscany <Venkat> but please, why not discuss the design from the scenarios.. scenarios are usecases... <Venkat> but the design is what will get you closer to implem. isn't it <jboynes> agreed - we just need scenarios to be documented <jmarino> I've got to get going <Venkat> I am most certain that if we had a design in place then you could convince on the merits of core2 in a much better way .. <Venkat> a way in which the whole lot of us can follow... <jboynes> I agree we need more documentation <Venkat> otherwise to be frank I am totally lost in these discussions and feel bad about not being able to add any value <jmarino> sorry gotta run <jboynes> ttfn <jmarino> Venkat, we need more docs but also feel free to ask questions on the list too <jmarino> k really gotta run <jboynes> Venkat: are there any areas that you have been looking at <jboynes> ? * jmarino has quit IRC <jboynes> anything we can do to help you fell less lost? <Venkat> well, I was sort of getting use to M1 and infact tried the RMI binding and posted that as well... but then * halehM has quit IRC <Venkat> there is now M2 that I must start picking up ... * bhdaniel has left #tuscany <Venkat> So instead of starting with the code .. I thought if I could understand the design or strucutre of M2 then I can start drilling down <jboynes> ant mentioned the rmi binding at ApacheCon but I must confess that I had missed it <jboynes> how about trying to get it working with the sandbox code? <Venkat> no probs. it was rudimentary, but I intended to improve it with community inputs / feedback <jboynes> good stuff, and again I apologize for missing it <jboynes> did you send it to the list or is it in a jira? <Venkat> posted it as JIRA <jboynes> 467? <Venkat> yes <jboynes> is the code attached the latest? <Venkat> yes... <jboynes> ok, I have the code on my machine, let me take a look through <jboynes> will you be around today? <Venkat> I work from India and it is late into the night now :-) <jboynes> ok - wasn't sure if it was late night or early morning <Venkat> ok. So let me know your opinions over the mailing list.. I will take it from there. Can I log off then? or is there anything else? <jboynes> should I look at migrating this to the sandbox impl or is that something you would prefer to do? <Venkat> I would like to do that as it will give me a feel of the new implementation. I suppose the programming model for extensions may not have changed very drastically from this ? <Venkat> I mean from what existed in M2 <jboynes> not drastically but there are some detail differences <jboynes> I'll post some suggestions to the list <Venkat> may be how it plugs into the core along the IoC points may have changed... <Venkat> oh sure.. that will be great.. thanks <jboynes> jliu: it would help if you could recap your experience with the celtix binding as well <jboynes> I should be around this evening my time, until then ... <Venkat> I guess jliu left sometime back <jboynes> probably, it's even later for him :) <Venkat> Oh I see... :) <jboynes> ttyl <Venkat> sure bye :) * Venkat has quit IRC * monxton has left #tuscany * kevin__ has joined #tuscany * pombreda has quit IRC (Connection timed out) * rfeng has quit IRC (Read error: 110 (Connection timed out)) * kevin__ has quit IRC * ThorPrime has joined #tuscany * ThorPrime has left #tuscany * david-w has joined #tuscany * lresende has quit IRC (Read error: 110 (Connection timed out)) * rfeng has joined #tuscany * jmarino has joined #tuscany * jmarino has quit IRC (Client Quit) * pombreda has joined #tuscany * kevin__ has joined #tuscany * kevin__ has quit IRC * isilval has quit IRC ("Trillian (http://www.ceruleanstudios.com";) * jsdelfino has joined #tuscany * kevin has joined #tuscany * jmarino has joined #tuscany * kevin has quit IRC * jsisson has quit IRC (Read error: 110 (Connection timed out)) * kgoodson has joined #Tuscany * rfen1 has joined #tuscany * jliu has quit IRC (Read error: 104 (Connection reset by peer)) * rfeng has quit IRC (Read error: 110 (Connection timed out)) * jmarino has quit IRC * kgoodson has quit IRC (Read error: 110 (Connection timed out)) * Disconnected Session Close: Fri Jul 07 00:58:22 2006
