Towards GNU Classpath 0.07 and 1.0
Hi all, Wanted to share my thoughts on the next snapshot 0.07 and what should be done before we can really start thinking about GNU Classpath 1.0. These are just suggestions so please yell and scream (politely) if you have other ideas about this. 0.07. The best thing for the next snapshot is to just set a date when we want to release it. I propose to just set this at the last week of November (three months from now) since we are already making real progress. Look at the bug/patch database for a number of improvements to object streams, java.text, rmi and javax.naming that are just waiting to get checked in. Much thanks go to the Kaffe team that has been importing Classpath libraries like mad the last couple of weeks and simultaneously debugged them and provided patches! And apologies to anyone whose patches didn't make it for the 0.06 release. If that works out then we can just keep on doing snapshot releases every three months till... 1.0. This is the release everybody is really waiting for. It should give people enough functionality to be useful for creating real projects (from full featured JDKs that people create on top of GNU Classpath) and it should be polished enough that we feel it is worth to maintain it (while we start working towards 2.0). We cannot set a date for it, but we can list the requirements that should be met to make it happen: - At least the same functionality as other 1.1 JDK core classes provide. (We have most of the packages needed to provide, but debugging to get them all working correctly will take some time. The main missing part has always been the AWT (peers) but the gcj hackers are really working hard on them. The wonka visual test tool that Thomas recently imported into Mauve will certainly help with this effort. Since we clearly have already so much more then plain 1.1 we might set the goal a bit higher to include everything that people expect from 1.1 plus the extensions described in the java class libraries second edition, volume 1 supplemental. Plus of course everything from 1.3 and 1.4 that we can clearly show is ready. But getting a minimum 1.1 set of packages all at the same minimum level should be the first goal.) - Clean and stable VM interface that makes implementing a VM and/or compiler based on GNU Classpath easy. (We have a start with the various VM[ClassName] classes inside the vm/reference directory. We should extend these with the proposals for e.g. VMThread and VMSecurityManager that were already proposed on the list and they should be better documented in the vmintegration guide. Ideally even gcj and kaffe would use this VM interface, but the main target are the free VMs that realy exclusively on GNU Classpath as core library implementation.) - Up to date documentation that at a minimum provides a reference like overview of what we have and what each class/method does. And clearly describes what is work in progress/non-functional stubs. (gjdoc really helps here. And our api doc isn't that bad. We really need to write some package.html overview pages for the packages that we feel agree really useful though. We also might need to hack a bit on gjdoc to make do everything we want - like including pictures.) - Testsuite that can be used to see if the libraries work as expected and that we can use to spot regressions while we work on future versions. (We use Mauve for that and it is already pretty good to show regressions on parts of our library, but it needs a bit of cleaning up to make it more generally useful. Also some packages don't have any tests at all.) We aren't that far from 1.0. But polishing everything up and the general rule about "the last 10% takes the other 90% of time" do apply. Hopefully we can come up with a useful task list for 1.0 (we already have a task list on savannah, but I don't if people really use that). But above all, keep having fun, keep on hacking! Cheers, Mark P.S. Speaking about the future of software. Please take the time to read the following important information about patents in Europe: http://swpat.ffii.org/ If you live in Europe and can make time, please join the protest in Brussels next Wednesday (Aug 27 2003): http://swpat.ffii.org/news/03/demo0819/index.en.html If you cannot make it, please join the online demonstration: http://swpat.ffii.org/group/demo/index.en.html And/Or try to help in other ways: http://swpat.ffii.org/group/todo/index.en.html Hope to see some of you in Brussels on Wednesday! Help with lodging and traveling in groups is coordinated at: http://wiki.ael.be/index.php/BigDemo27aug And the #bxl-ffii channel on irc.freenode.org ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath
make distcheck and libtool 1.5a now work
Hi, I got make distcheck and libtool 1.5 (and keeping 1.4.3) to work. make distcheck helps us to see whether a make dist really creates a good release tarball. It points out missing files or make rules which don't cleanup after themselves. In the future we might be able to add automatic Mauve checking to it, but that is a lot more tricky. It does mean that make clean now really removes all generated files (also the .class and glibj.zip file). But since we these files have not been necessary for a long time to bootstrap anyway this shouldn't be a problem. I had to update the libtoolize files (config.guess, config.sub and ltmain.sh) to get libtool 1.5 (and 1.4.3) to work correctly and had to update a few configure.in definitions. If this breaks something for you please let me know. (You will have to do the automagic dance described in the HACKING file after a CVS update.) Cheers, Mark ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath
Re: java.sql.Time#setTime(long)
Hi, On Sun, 2003-08-24 at 20:08, Arnaud Vandyck wrote: > This methode does not exist in classpath but does in J2SE1.4 > > http://java.sun.com/j2se/1.4.2/docs/api/java/sql/Time.html#setTime(long) > > ... and is not deprecated. And it seems to be defined in earlier versions like 1.3. It looks like there is a constructor that does almost exactly what this method does. Could you take a look at that and try to create a patch that implements this method? > Also, Is it normal that the deprecated methode of this class throw an > IllegalArgumentException? Yes, because that is what the documentation says it should do. Cheers, Mark ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath
java.sql.Time#setTime(long)
This methode does not exist in classpath but does in J2SE1.4 http://java.sun.com/j2se/1.4.2/docs/api/java/sql/Time.html#setTime(long) ... and is not deprecated. Also, Is it normal that the deprecated methode of this class throw an IllegalArgumentException? -- Arnaud Vandyck, STE fi, ULg Formateur Cellule Programmation. ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath
Re: [japitools message] how to resolve 'SerialVersionUID'
Arnaud Vandyck wrote: Hi all, if you look at japitools[1], there are some minor problems like this one: org.xml.sax Minor: 3 classes. Minor * class org.xml.sax.SAXException: SerialVersionUID=583241635256073760 in jdk14, but SerialVersionUID=-8251803752662179270 in classpat * class org.xml.sax.SAXNotRecognizedException: SerialVersionUID=5440506620509557213 in jdk14, but SerialVersionUID=1398555601346822626 in classpath * class org.xml.sax.SAXNotSupportedException: SerialVersionUID=-1422818934641823846 in jdk14, but SerialVersionUID=-7373565723188945450 in classpath Can someone tell me how to solve this problems if they can be solved? Add a private final long field SerialVersionUID to the class. Take a look at class documentation to see what it looks like in serialized form. http://java.sun.com/j2se/1.4.2/docs/api/serialized-form.html#org.xml.sax.SAXException See also http://www-124.ibm.com/pipermail/jikes/2001-November/002356.html cheers, dalibor topic ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath
[japitools message] how to resolve 'SerialVersionUID'
Hi all, if you look at japitools[1], there are some minor problems like this one: org.xml.sax Minor: 3 classes. Minor * class org.xml.sax.SAXException: SerialVersionUID=583241635256073760 in jdk14, but SerialVersionUID=-8251803752662179270 in classpath * class org.xml.sax.SAXNotRecognizedException: SerialVersionUID=5440506620509557213 in jdk14, but SerialVersionUID=1398555601346822626 in classpath * class org.xml.sax.SAXNotSupportedException: SerialVersionUID=-1422818934641823846 in jdk14, but SerialVersionUID=-7373565723188945450 in classpath Can someone tell me how to solve this problems if they can be solved? Best Regards, -- Arnaud Vandyck, STE fi, ULg Formateur Cellule Programmation. Footnotes: [1] http://japi.sab39.org/htmlout/h-jdk14-classpath.html ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath
Re: GNU Classpath 0.06 created
Hi, On Sat, 2003-08-23 at 18:07, Per Bothner wrote: > You don't put releases on alpha.gnu.org - you put them on ftp.gnu.org. > You can put pre-releases on alpha.gnu.org, but if you already have > a release there seems little point. My thinking was that the 0.0x releases are actually the pre-releases while we work towards the first real GNU Classpath 1.0 release. People can combine these pre-releases with their own VMs like gcj, kissme, jikesRVM, Jaos, etc already do. But as long as we don't have at least a minimum of functionality to offer (we are hoping that 1.0 will have the equivalent functionality as core libraries that come with some proprietary 1.1 JDKs) and a stable/maintainable VM interface we cannot call it a real release. We do hope that people try these pre-releases out to create compilers and/or VM environments so they can give us feedback what should be improved for the real 1.0 release. (And I also don't have access to ftp.gnu.org) Cheers, Mark ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath
Re: date for release 0.06
Hi, On Sat, 2003-08-23 at 15:57, John Leuner wrote: > When should we expect 0.06 to be released? The tarball can already be found at http://www.klomp.org/mark/classpath/ The CVS tree has been taged and the version number in configure.in has already been bumped to 0.06+cvs. It has been tested and works well with both Kissme and JikesRVM out of the box according to the reports that people send me. But as long as we cannot put it on the gnu ftp server I don't want to announce it widely. Hopefully that can be arranged soon, but communication goes extremely slow since the mail.gnu.org server seems to have a big backlog of messages (probably from the latest Windows worm/virus thingy). Cheers, Mark ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath
Re: GNU Classpath 0.06 created
Mark Wielaard wrote: Although I don't have access yet to alpha.gnu.org to publish the release I have already created the 0.06 release tarball. I also tagged the CVS tree as 'classpath-0_06-release'. You don't put releases on alpha.gnu.org - you put them on ftp.gnu.org. You can put pre-releases on alpha.gnu.org, but if you already have a release there seems little point. -- --Per Bothner [EMAIL PROTECTED] http://per.bothner.com/ ___ Classpath mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/classpath