Gustavo/Raph,
After getting Raph's response, I also had my suspicion on DIPort (though I do not think I store any constraint variables there - just some log messages). After eliminating that I reduced my warning messages from three to two. So there may be some other variables. I am not sure what other things I should be looking after. Of course there are some FD variables in the code, but whether they are part of solution that gets back to the master process I am not sure. Can you point me to some suggested reading sections to clear up this point? Another thing, I am still unable to have ssh in my windows XP. I have installed cygwin and ssh - but still having "port 22: connection refused" error. Also so far my localhost is the only machine I am doing parallel search. Appreciate your help. Ashis ________________________________ From: [email protected] [mailto:[email protected]] On Behalf Of Gustavo Gutiérrez Sent: Tuesday, June 02, 2009 10:06 AM To: Mozart users Subject: Re: Parallel Search / release 1.4.0 Can you comment something on the kind of information you intend to have in DIPort ? Remember that you cannot send constraint variables to ports. Can you isolate the behavior to an small program? this will be also helpful. Regards, Gustavo On 02 Jun 2009, at 15:19, Maity, Ashis K wrote: Thanks Gustavo. Changing back to release 1.3.2 did help. Now I can progress a bit further. However, I am getting the following error message and the process is getting killed. *** Warning: marshaling a variable as a resource! *** Warning: marshaling a variable as a resource! *** Warning: marshaling a variable as a resource! *** Internal Error: Please send a bug report to [email protected] *** *** with the following information: *** version: 1.3.2 *** platform: win32-i486 *** date: 20080123 Malicious message received: redirect MsgContainer::unmarshal: bad Oz term FYI, I am starting parallel Search in the following way: functor SchedWrapper export Script import SchedRef at 'SchedulerFunctor.ozf' %works define Script = {SchedRef.makeProblem ProblemDescription DIPort} end try E = {New Search.parallel init(localhost:1#sh)} catch X then {Browser.browse incatch} {Browser.browse X} end {E trace(true)} Solution = {E one(SchedWrapper $)} {E close} Any idea what could be the error? Thanks, Ashis -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Gustavo Gutiérrez Sent: Monday, June 01, 2009 3:01 PM To: Mozart users Subject: Re: Parallel Search / release 1.4.0 On 01 Jun 2009, at 14:20, Maity, Ashis K wrote: > I was trying parallel search using version 1.4.0 and was having > trouble. > Going through the mailing list, it appears that I need to use version > 1.3.2 for Parallel Search. Is this right? What are the features > would I > be losing if I go back to version 1.3.2? 1.4.0 provides a new implementation of the distribution subsystem, so if you are concerned only with constraint programming you are not loosing anything. This fact also explains why the parallel search is broken in 1.4.0 and is because it relies on the implementation of the distribution system. Regards, Gustavo _________________________________________________________________________________ mozart-users mailing list [email protected] http://www.mozart-oz.org/mailman/listinfo/mozart-users _________________________________________________________________________________ mozart-users mailing list [email protected] http://www.mozart-oz.org/mailman/listinfo/mozart-users
_________________________________________________________________________________ mozart-users mailing list [email protected] http://www.mozart-oz.org/mailman/listinfo/mozart-users
