October 2009 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] [Seaside] Komanche ?

by Adrian Lienhard

On Oct 12, 2009, at 23:10 , Ramiro Diaz Trepat wrote: > Small clarification. In my case the different VMs did not use the > same port > numbers, they had different ports, indeed one of them did not even > have > Seaside installed. It does not matter what ports the second image opens during startup... the second VM proces would still listen on the same ports like the first one. Adrian > > f > > > On Mon, Oct 12, 2009 at 9:26 PM, John M McIntosh < > johnmci(a)smalltalkconsulting.com> wrote: > >> Ok. >> (a) the first problem is how the 2nd VM is launched. We actually >> use Apple >> Script to re-launch the binary with the supplied image when >> we receive the open document apple event. The behavior of this is >> different between 10.3, 10.4, 10.5 and I assume the 10.6 is the same >> as 10.5 because we had reported an issue with the 10.4 behavior >> they fixed >> in 10.5 >> >> Indications are is that the apple script code executes a execv (but >> we are >> not sure) >> this then retains the open file descriptors so it retains the >> knowledge >> about open sockets. >> >> >> (b) I'm not sure what the behavor would be if you launch VM "A" >> then listen >> on socket 8080. Then launch VM "B" then again listen on socket 8080 >> will both virtual machines "A" and "B" attempt to service socket >> 8080, or >> will VM "B" get told it cannot listen on 8080 because it does not >> own it? >> >> >> Pending is work on a pure objective-c based squeak VM based on the >> the >> iPhone work, which would alter the behaviour. >> But if someone can answer (b) then maybe we can see about an >> interim fix to >> deal with (a) >> >> >> On 2009-10-12, at 1:56 AM, Adrian Lienhard wrote: >> >> Hi Ramiro, >>> >>> Yes, this is an issue of the Mac VM only. >>> >>> It was reported in the sense that I sent a mail with instructions to >>> reproduce the problem to the Squeak-vm mailing list and John >>> acknowledged the problem. I don't know whether John intends to fix >>> this, though. >>> >>> Adrian >>> >>> >> -- >> = >> = >> = >> = >> = >> = >> ===================================================================== >> John M. McIntosh <johnmci(a)smalltalkconsulting.com> Twitter: >> squeaker68882 >> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com >> = >> = >> = >> = >> = >> = >> ===================================================================== >> >> >> >> >> >> _______________________________________________ >> seaside mailing list >> seaside(a)lists.squeakfoundation.org >> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside >> > _______________________________________________ > seaside mailing list > seaside(a)lists.squeakfoundation.org > http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

Oct. 12, 2009

Re: [Pharo-project] Preparing the sprint

by Michael Roberts

Adrian, sorry i did not read your mail carefully ;-) no we don't have an issue on the tracker yet. whilst it is an external package I think it is important enough we maintain compatibility. Is it reasonable for us to track on our tracker? I would say so, but in the past sometimes i think we have closed issues for not being pharo directly. cheers Mike

Oct. 12, 2009

Re: [Pharo-project] Preparing the sprint

by Michael Roberts

Stef, Adrian no it's not related to anything published on SqueakSource. i'll put something in task forces later in the week with some comments. It's not a server, just the core code that runs the tests I put on the wiki. Given the wiki can be updated with svn that's the route i propose. We can also update the issue tracker by marking files in a special way. i haven't tried that yet but could be useful in the future. I think we can auto-create issues, or at least auto-close them. the code is literally only a few lines here & there. I don't think it needs to be that clever. i would prefer to see all energy going into fixing bugs! You'll need to integrate it with something that loads the updates, but I would see that taking more shape once the package system and process of pharo is more mature. what lukas describes sounds interesting. I need OSProcess to spawn svn, unless we have a lighter weight way of doing that? cheers Mike

Oct. 12, 2009

Re: [Pharo-project] [Seaside] Komanche ?

by Ramiro Diaz Trepat

Small clarification. In my case the different VMs did not use the same port numbers, they had different ports, indeed one of them did not even have Seaside installed. f On Mon, Oct 12, 2009 at 9:26 PM, John M McIntosh < johnmci(a)smalltalkconsulting.com> wrote: > Ok. > (a) the first problem is how the 2nd VM is launched. We actually use Apple > Script to re-launch the binary with the supplied image when > we receive the open document apple event. The behavior of this is > different between 10.3, 10.4, 10.5 and I assume the 10.6 is the same > as 10.5 because we had reported an issue with the 10.4 behavior they fixed > in 10.5 > > Indications are is that the apple script code executes a execv (but we are > not sure) > this then retains the open file descriptors so it retains the knowledge > about open sockets. > > > (b) I'm not sure what the behavor would be if you launch VM "A" then listen > on socket 8080. Then launch VM "B" then again listen on socket 8080 > will both virtual machines "A" and "B" attempt to service socket 8080, or > will VM "B" get told it cannot listen on 8080 because it does not own it? > > > Pending is work on a pure objective-c based squeak VM based on the the > iPhone work, which would alter the behaviour. > But if someone can answer (b) then maybe we can see about an interim fix to > deal with (a) > > > On 2009-10-12, at 1:56 AM, Adrian Lienhard wrote: > > Hi Ramiro, >> >> Yes, this is an issue of the Mac VM only. >> >> It was reported in the sense that I sent a mail with instructions to >> reproduce the problem to the Squeak-vm mailing list and John >> acknowledged the problem. I don't know whether John intends to fix >> this, though. >> >> Adrian >> >> > -- > =========================================================================== > John M. McIntosh <johnmci(a)smalltalkconsulting.com> Twitter: > squeaker68882 > Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com > =========================================================================== > > > > > > _______________________________________________ > seaside mailing list > seaside(a)lists.squeakfoundation.org > http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside >

Oct. 12, 2009

Re: [Pharo-project] Preparing the sprint

by Lukas Renggli

> If you can post the code and maybe some comments on how to use it, > that would be great. Maybe we can do some testing and investigation > into the problem during the sprint. > > Do we have an issue for the OSProcess problem or has it been discussed > before? > > We'll discuss where to host this server. We have several options there. We are using Mason from <http://source.wiresong.ca/mc> to build Seaside images. We don't have a server setup yet, but it also works locally. Mason is a very flexible system, it has infrastructure to build images, build SAR archives, build documentation, load Monticello 1 and 2 packages, load Metacello configurations, run tests, run code critics and execute arbitrary scripts, etc. The OB and MC2 distributions are built using Mason for example: <http://www.wiresong.ca/static/releases/>. Lukas -- Lukas Renggli http://www.lukas-renggli.ch

Oct. 12, 2009

Re: [Pharo-project] Preparing the sprint

by Dale Henrichs

RT @tinybuddha: "Do not give your attention to what others do or fail to do; give it to what you do or fail to do." ~Dhammapada ----- "keith" <keith_hodges(a)yahoo.co.uk> wrote: | On 12 Oct 2009, at 17:45, StÃ©phane Ducasse wrote: | | > Mike is it related to the testServer publihsed on squeaksource? | > Because ESUG will also sponsor a summer talk project on the topics. | > So if we could join forces it would be great. | > | > For the server this is not clear yet. Internally we have our server | > that we could use but | > we have to check the security. | > | > I'm waiting that marcus arrives to check that with him and damien. | > | > Stef | | Why why why why why is compete more attractive than contributing. | | Let me explain.... | | Step 1. We would like a test server. | Step 2. Is anyone working on this already? | Step 3. Would anyone like to collaborate to specify and work on a test | server? | | Bob is already written it just needs configuring, TestReporter is | already written, | and has been available for 3 years or more. | | The same thing is happening again and again and again. | | Why did anyone even start working on Metacello when Sake/Packages was | | already there a year before hand. | | Why are the steps above not the default way of working for the | community? It is beyond incredulity. | | Keith | | | | _______________________________________________ | Pharo-project mailing list | Pharo-project(a)lists.gforge.inria.fr | http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Oct. 12, 2009

Re: [Pharo-project] os process crash

by Michael Roberts

this is in the latest pharo core, mac os x 10.5.8, Squeak 4.2.2beta1U. i'll dig around in a spare moment. cheers Mike On Sat, Oct 10, 2009 at 10:42 PM, Schwab,Wilhelm K <bschwab(a)anest.ufl.edu> wrote: > Just for completeness, which OS are you using? Â I had an ugly time with the 09.3 web image crashing out from under ODBC; that was on Linux, though if I get that working, I can also test it on Windows. > > I was distracted with Polymorph, deploying an image for a "release" build of a home-grown utility of mine that now runs on Linux (another small crack in a Redmond sidewalk), and initial porting of more of my software to Pharo. Â That's a long way around to the point, which is that I am hanging my head in shame for not following up on the ODBC crash :) > > A beer at a future meeting says that our two respective problems are related. > > Bill > > > > -----Original Message----- > From: pharo-project-bounces(a)lists.gforge.inria.fr [mailto:pharo-project-bounces@lists.gforge.inria.fr] On Behalf Of Michael Roberts > Sent: Saturday, October 10, 2009 4:12 PM > To: Pharo-project > Subject: [Pharo-project] os process crash > > folks, i'm trying to load OS Process into the latest pharo-core on my mac. it crashes the image when i try and run the tests. Does anyone have sucess on a recent pharo-core? > > thanks, > Mike > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

Oct. 12, 2009

Re: [Pharo-project] Preparing the sprint

by keith

On 12 Oct 2009, at 17:45, StÃ©phane Ducasse wrote: > Mike is it related to the testServer publihsed on squeaksource? > Because ESUG will also sponsor a summer talk project on the topics. > So if we could join forces it would be great. > > For the server this is not clear yet. Internally we have our server > that we could use but > we have to check the security. > > I'm waiting that marcus arrives to check that with him and damien. > > Stef Why why why why why is compete more attractive than contributing. Let me explain.... Step 1. We would like a test server. Step 2. Is anyone working on this already? Step 3. Would anyone like to collaborate to specify and work on a test server? Bob is already written it just needs configuring, TestReporter is already written, and has been available for 3 years or more. The same thing is happening again and again and again. Why did anyone even start working on Metacello when Sake/Packages was already there a year before hand. Why are the steps above not the default way of working for the community? It is beyond incredulity. Keith

Oct. 12, 2009

Re: [Pharo-project] [Seaside] Komanche ?

by John M McIntosh

Ok. (a) the first problem is how the 2nd VM is launched. We actually use Apple Script to re-launch the binary with the supplied image when we receive the open document apple event. The behavior of this is different between 10.3, 10.4, 10.5 and I assume the 10.6 is the same as 10.5 because we had reported an issue with the 10.4 behavior they fixed in 10.5 Indications are is that the apple script code executes a execv (but we are not sure) this then retains the open file descriptors so it retains the knowledge about open sockets. (b) I'm not sure what the behavor would be if you launch VM "A" then listen on socket 8080. Then launch VM "B" then again listen on socket 8080 will both virtual machines "A" and "B" attempt to service socket 8080, or will VM "B" get told it cannot listen on 8080 because it does not own it? Pending is work on a pure objective-c based squeak VM based on the the iPhone work, which would alter the behaviour. But if someone can answer (b) then maybe we can see about an interim fix to deal with (a) On 2009-10-12, at 1:56 AM, Adrian Lienhard wrote: > Hi Ramiro, > > Yes, this is an issue of the Mac VM only. > > It was reported in the sense that I sent a mail with instructions to > reproduce the problem to the Squeak-vm mailing list and John > acknowledged the problem. I don't know whether John intends to fix > this, though. > > Adrian > -- = = = ======================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Twitter: squeaker68882 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================

Oct. 12, 2009

Re: [Pharo-project] Preparing the sprint

by Adrian Lienhard

Hi Mike, If you can post the code and maybe some comments on how to use it, that would be great. Maybe we can do some testing and investigation into the problem during the sprint. Do we have an issue for the OSProcess problem or has it been discussed before? We'll discuss where to host this server. We have several options there. Adrian On Oct 12, 2009, at 18:25 , Michael Roberts wrote: > hi I can't make the sprint but I have the basic core of a test server > that runs the tests and updates the wiki. I just need to be able to > spawn svn shell commands which I'm stuck on due to osprocess crash. I > can post the code in task forces if someone is interested. > > To turn it into a server we need a host and some scripts to kick it > off. Do we have any hardware as a group? > cheers mike > > On Monday, October 12, 2009, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr > > wrote: >> Yes I want to cut some dirty cycles >> Network could be a first start and after the ugly morphic ;-} >> On Oct 12, 2009, at 3:01 PM, Simon Denier wrote: >> >>> >>> Also during the sprint, me and Jannik would like to run an >>> experiment >>> running eDSM (a matrix visualization of dependencies) to refactor >>> some >>> package dependencies (heads up for those who will attend WCRE this >>> week, eDSM is on schedule Wednesday :)). For people interested (we >>> will contact or you can contact us), the experiment will take >>> around 3 >>> hours and you get to do what you want to do in Pharo, but with our >>> tool :) >>> >>> >>> On 10 oct. 2009, at 10:01, StÃ©phane Ducasse wrote: >>> >>>> Hi guys >>>> >>>> I would like to have several kinds of tasks so that people can pick >>>> up >>>> Now let me know if you have other points >>>> >>>> - fixing stuff for 1.0 >>>> - integrating pending for 1.1 >>>> - OB related UI >>>> - fixing more elaborated points for 1.1 >>>> - package structure >>>> - improving the system (smalllint?) >>>> - Popup discard >>>> - TextEditor from Squeak/CUIS >>>> - ToolBuilder changes >>>> - checking squeak fixes >>>> - Preparing release candidate >>>> - fixing _ >>>> - free fonts >>>> - Dreaming about the future >>>> - filesystem? Rio/ >>>> - newcompiler >>>> - true package >>>> - management of noncore packages >>>> - browsers >>>> - release/integration process >>>> >>>> BTW what is costing me most of my time is cleaning the inbox. >>>> Can we script that because moving these files is a pain. >>>> >>>> Syeg >>>> >>>> _______________________________________________ >>>> Pharo-project mailing list >>>> Pharo-project(a)lists.gforge.inria.fr >>>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>> >>> -- >>> Simon >>> >>> >>> >>> >>> _______________________________________________ >>> Pharo-project mailing list >>> Pharo-project(a)lists.gforge.inria.fr >>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Oct. 12, 2009