From mcconnell at dpml.net Wed Sep 2 01:57:15 2009 From: mcconnell at dpml.net (Stephen J. McConnell) Date: Wed, 02 Sep 2009 18:27:15 +0930 Subject: meeting regarding jpkg output In-Reply-To: <231998536.608381251477213144.JavaMail.root@zmail01.collab.prod.int.phx2.redhat.com> References: <231998536.608381251477213144.JavaMail.root@zmail01.collab.prod.int.phx2.redhat.com> Message-ID: <1251881835.9918.13.camel@gloria> On Fri, 2009-08-28 at 12:33 -0400, jon.vanalten at redhat.com wrote: > Hello, > > My name is Jon VanAlten and I am sitting here with Man Lung Wong. We > are interns at Red Hat and hope to contribute to the efforts of > project Jigsaw wrt streamlining with the RPM process. > > This is not intended as a full meeting synopsis. > > Today we had a meeting as a follow-up to: > http://mail.openjdk.java.net/pipermail/jigsaw-dev/2009-July.txt > > The main purpose of this meeting was to come up with a potential > output format for the jpkg tool, when run with appropriate flags to > emit module info rather than create a complete .deb/.rpm package. > > It seems that the required module info consists primarily of key-value > pairs, ie it is simple in nature. The main bits that we are concerned > with are: > > Dependency (possibly multiple) > Export (possibly multiple) > Name > Version > > Also mentioned was the possibility of including further information in > the future, such as a copyright notice or other items. > > So, after considering more robust formats such as XML, it was agreed > that a simple shell script exporting environment variables (which can > then be sourced by other tools later in the process) would be best. > Of course such a script would be encoded in a charset supporting > international characters. > > After the meeting we sat down to consider this solution, and realized > that we do have something just a little bit more complicated than > key-value pairs. This is because the dependencies and exports of a > jigsaw-based package may not be (in fact in most cases will not be) > singular. So we had the idea that a standard separator be used for > these variables that may need to be assigned multiple values. This > separator should be a character that is not permitted to appear in the > name of a dependency/export, and not conflict with any other special > characters. Dalibor, do you think this is a good idea, and if so do > you have any suggestions, are there already any restricted characters > for module names, or other possible future module info? Jon: Assuming that Dependency, Export, Name, and Version are equivalent to the notions of 'ModuleRequires', 'ModuleProvides', 'QualifiedIdentifier' (the module name), and 'ModuleVersion' as declared in the Project Jigsaw: Language changes for Modules document [1], I would suggest that the special character you are looking for is ','. If I understand things correctly, the module name and version are treated as opaque identifiers (@,,... Cheers, Steve. > > cheers, > > Jon VanAlten > Man Lung Wong > > [1] http://openjdk.java.net/projects/jigsaw/doc/language.html -- Stephen J. McConnell mailto:mcconnell at dpml.net http://www.dpml.net From mwong at redhat.com Fri Sep 4 07:55:18 2009 From: mwong at redhat.com (Man Wong) Date: Fri, 4 Sep 2009 10:55:18 -0400 (EDT) Subject: What is the next step? Message-ID: <144932836.720081252076118867.JavaMail.root@zmail04.collab.prod.int.phx2.redhat.com> Hi Dalibor, My name is Man, which Jon mentioned in the last email regarding what we discussed in the call last Friday. We are just wondering what is the next step for jigsaw and is there any particular part that Jon and I should get started on? Thanks. Man Lung Wong From Dalibor.Topic at Sun.COM Thu Sep 10 16:39:09 2009 From: Dalibor.Topic at Sun.COM (Dalibor Topic) Date: Fri, 11 Sep 2009 01:39:09 +0200 Subject: What is the next step? In-Reply-To: <144932836.720081252076118867.JavaMail.root@zmail04.collab.prod.int.phx2.redhat.com> References: <144932836.720081252076118867.JavaMail.root@zmail04.collab.prod.int.phx2.redhat.com> Message-ID: <4AA98E1D.1000309@sun.com> Man Wong wrote: > Hi Dalibor, > > My name is Man, which Jon mentioned in the last email regarding what we discussed in the call last Friday. > We are just wondering what is the next step for jigsaw and is there any particular part that Jon and I should get started on? Thanks. Hi, I'm currently working on getting the jpkg tool to output the metadata, and on getting it to build on Fedora 11 - the latter part will take the better part of the next couple of days. Once that all works, then you should be able to start playing with it in the Fedora RPM toolchain. cheers, dalibor topic -- ******************************************************************* Dalibor Topic Tel: (+49 40) 23 646 738 Java F/OSS Ambassador AIM: robiladonaim Sun Microsystems GmbH Mobile: (+49 177) 2664 192 Nagelsweg 55 http://openjdk.java.net D-20097 Hamburg mailto:Dalibor.Topic at sun.com Sitz der Gesellschaft: Sonnenallee 1, D-85551 Kirchheim-Heimstetten Amtsgericht M?nchen: HRB 161028 Gesch?ftsf?hrer: Thomas Schr?der, Wolfgang Engels, Wolf Frenkel Vorsitzender des Aufsichtsrates: Martin H?ring From mr at sun.com Tue Sep 15 11:07:52 2009 From: mr at sun.com (Mark Reinhold) Date: Tue, 15 Sep 2009 11:07:52 -0700 Subject: Post-build modularization scripts and class lists Message-ID: <20090915180752.AA48ADE2@eggemoggin.niobe.net> I've copied these to http://cr.openjdk.java.net/~mr/jigsaw/jdk.make.modules/ for now; they'll be pushed into the Jigsaw forest in due course, in the jdk/make/modules directory. The class lists are the very crude, preliminary cut of rt.jar which I did for demo purposes at JavaOne back in June. In contrast to Mandy's analysis, these lists are bottom-up. The boot (aka base) module, in particular, is the static transitive closure over API and implementation dependences of the java.{lang{,.ref},io,net,util} packages and the sun.launcher package, plus some additional classes needed to satisfy dynamic dependences (i.e., via Class.forName). - Mark From mr at sun.com Wed Sep 30 11:21:06 2009 From: mr at sun.com (Mark Reinhold) Date: Wed, 30 Sep 2009 11:21:06 -0700 Subject: Jigsaw tools repository Message-ID: <20090930182106.93955864@eggemoggin.niobe.net> I've created a side repository for miscellaneous tools, such as the dependence analyzer which Alan and Mandy have been hacking on: http://hg.openjdk.java.net/jigsaw/tools In the long term we'll most likely merge these into the JDK itself, but for now it will be convenient to be able to work on them standalone. - Mark From mcconnell at dpml.net Wed Sep 30 11:37:52 2009 From: mcconnell at dpml.net (Stephen J. McConnell) Date: Thu, 01 Oct 2009 04:07:52 +0930 Subject: Jigsaw tools repository In-Reply-To: <20090930182106.93955864@eggemoggin.niobe.net> References: <20090930182106.93955864@eggemoggin.niobe.net> Message-ID: <1254335872.8360.6.camel@gloria> Mark: Is there some info somewhere concerning the dependence analyzer tool activities? It is a subject I'm rather interested in with respect to build automation tooling and runtime service management. Cheers, Steve. On Wed, 2009-09-30 at 11:21 -0700, Mark Reinhold wrote: > I've created a side repository for miscellaneous tools, such as the > dependence analyzer which Alan and Mandy have been hacking on: > > http://hg.openjdk.java.net/jigsaw/tools > > In the long term we'll most likely merge these into the JDK itself, but > for now it will be convenient to be able to work on them standalone. > > - Mark > From mr at sun.com Wed Sep 30 11:59:32 2009 From: mr at sun.com (Mark Reinhold) Date: Wed, 30 Sep 2009 11:59:32 -0700 Subject: Jigsaw tools repository In-Reply-To: mcconnell@dpml.net; Thu, 01 Oct 2009 04:07:52 +0930; <1254335872.8360.6.camel@gloria> Message-ID: <20090930185932.57DA6864@eggemoggin.niobe.net> > Date: Thu, 01 Oct 2009 04:07:52 +0930 > From: "Stephen J. McConnell" > Is there some info somewhere concerning the dependence analyzer tool > activities? It is a subject I'm rather interested in with respect to > build automation tooling and runtime service management. Stay tuned -- Alan and Mandy have only just begun working on this in a serious way. - Mark