AfC | Is there a canonical page describing the Policy being applied for Java packages (in the Debian Policy sense of the word)? | 07:43 |
---|---|---|
persia | AfC, http://www.debian.org/doc/packaging-manuals/java-policy/ might be what you seek. | 07:45 |
AfC | persia: sweet. I needed the Debian answer too, so thanks very much. Ubuntu following that (at least in theory, etc)? | 07:46 |
persia | Yes. The primary significant difference between Ubuntu Java packages and Debian Java packages is tweaks to the build process to ensure builds on the Ubuntu buildds. There are a couple packages that god more attention, but not very many, and the changes should be in line with that policy. | 07:48 |
AfC | persia: great. | 07:48 |
AfC | persia: (Mostly I was looking because I was writing up an example of someone secifying .jar and .so installed locations, and figured I'd use Debian et al as the example. I knew about you lot whacking .sos into /usr/lib/jni, but wasn't sure where you were putting .jars these days) | 07:49 |
persia | Yeah, that's probably a good guide. The general idea is to have a common classpath, and make sure everything ends up in the classpath. | 07:50 |
persia | The hardest part is coordinating with upstreams to ensure everyone is using the same set of common libraries. | 07:50 |
AfC | Uh, I'm not sure specifying -classpath '/usr/share/java/*' is a good idea. That's an awful lot of chugging, even if the VM in question does mmap. | 07:55 |
persia | You mean because there could be *lots* of libraries there? | 07:58 |
AfC | Well, yeah | 08:09 |
persia | Then the JRE just has to get better about it. We use /usr/lib for all the libraries for C, and it works. :) | 08:10 |
persia | Theoretically, it should only be loading classes when they will be required anyway, so it shouldn't need to pull the unreferenced ones. | 08:11 |
AfC | persia: I'm not here to argue, and you were quite helpful above, but I'm afraid you really need to check your assumptions on this one. To locate a class it has to completely parse a .jar file - and that involves mmapping a compressed file which is no great pain but then it has to uncompress it (eats memory) to then finding out what's there, and only THEN taking the load and link actions. | 08:15 |
AfC | Anyway, I'm off back to the normal universe. Thanks for your help. | 08:15 |
persia | AfC, Right. That's not ideal. | 08:15 |
persia | Have a good day. | 08:15 |
javaprince22 | Java Guru's in d house please? Got a small problem that needs solvubf | 11:24 |
persia | javaprince22, For Java coding, you probably want ##java | 11:24 |
javaprince22 | persia, hi. Am new to the whole irc thing . so how exactly do i get to ##java? Pls don't laugh | 11:31 |
persia | /join ##java | 11:32 |
javaprince22 | Thx. Persia. Are u a Java programmer too? | 11:33 |
persia | I'm not very strong in Java, although I've written some. | 11:33 |
javaprince22 | But u are a programmer? Wat language? I tried d /join ##java says authentication required | 11:34 |
persia | Right. Authenticate to freenode. See http://freenode.net/faq.shtml#contents-userregistration | 11:34 |
javaprince22 | Long literature to read. | 11:41 |
javaprince22 | Persia, where else can i chat with Java guru's? | 11:49 |
persia | That's really the best place. | 11:50 |
Kyran | does anyone here know why the latest version of sun-java6-plugin requires firefox? | 13:58 |
Kyran | no one? | 14:06 |
Kyran | anyway, I feel this is a packaging error, but I have no idea where and how to report it | 14:07 |
persia | Kyran, Apparently not. | 14:07 |
Kyran | :) | 14:07 |
persia | It would be a bug against the source package, but I can't find the binary package sun-java6-plugin, so I don't know which source. | 14:08 |
=== LucidFox_ is now known as LucidFox | ||
persia | Most of us are focusing on OpenJDK though : does the same happen with the OpenJDK plugin? | 14:08 |
Kyran | I'll check | 14:09 |
Kyran | there's no openjdk-6-plugin package in my list | 14:10 |
Kyran | maybe the plugin is in another package for the openjdk | 14:11 |
persia | Which release are you running? | 14:11 |
Kyran | kubuntu 8.04 | 14:12 |
persia | Oh, right, then it doesn't have the new OpenJDK. Yeah, file a bug. | 14:12 |
persia | It's fairly unlikely to be fixed for 8.04, but if it's still present in 8.10, it can be referenced as a significant issue that has been around for a while. | 14:13 |
Kyran | care to help me a bit in how to file a bug? | 14:15 |
Kyran | I'm quite new to those things | 14:15 |
Kyran | been lurking around for a while though | 14:15 |
persia | The easiest way is to run ubuntu-bug at the command line. | 14:17 |
Kyran | I'll give that a shot | 14:19 |
Kyran | so it's a bug in sun-java6-plugin? | 14:19 |
persia | If that's the package that depends on firefox, and you think it shouldn't, yes, except that I can't find such a package. | 14:19 |
Kyran | it's in multiverse | 14:21 |
Kyran | (being closed source) | 14:21 |
persia | Ah, and it's i386-only, which is why I couldn't find it (I'm running amd64). Yeah, that'd be the package to file the bug against. | 14:22 |
Kyran | ok thanks :) | 14:23 |
Kyran | launchpad shows there's a fix released for the bug in march | 14:26 |
persia | Odd. That should have been in 8.04. You might want to ask the folk on #ubuntu-bugs about the best procedure for reporting that a fix doesn't seem to have worked. | 14:27 |
Kyran | I've read the report and it says something like this | 14:28 |
Kyran | it's fixed upstream | 14:28 |
Kyran | but distributions aren't allowed to package that version | 14:28 |
Kyran | (yet) | 14:28 |
persia | Oh, that's frustrating. One of the reasons I only like some parts of multiverse :) | 14:29 |
Kyran | I havent' had much problems with multiverse yet | 14:30 |
Kyran | up until now things just worked (tm) | 14:30 |
persia | Lots of it works, some doesn't. For a large chunk of things that don't work, we can fix it, as long as we follow certain guidelines. For the rest, it's just annoying. | 14:31 |
Kyran | jup | 14:33 |
Kyran | it's good that java went open source | 14:33 |
Kyran | but I make the occasional java app, so I like to test against sun's implementation also | 14:33 |
persia | Indeed. OpenJDK is much easier to patch :) | 14:33 |
persia | OpenJDK is now Sun's implementation, but it will take a while for that to propagate everwhere. | 14:33 |
Kyran | most likely | 14:36 |
Kyran | I've always liked java | 14:36 |
Kyran | it sort of makes sense | 14:37 |
persia | Usually, except when it doesn't :) | 14:38 |
Kyran | :) | 14:40 |
Kyran | so are you an active ubuntu packager, or just a friendly lad helping people out here? | 14:43 |
persia | I should be the former, but end up being more of the latter. | 14:46 |
Kyran | life tends to do that | 14:49 |
andreahack | hello :D | 18:45 |
slytherin | andreahack: hi | 18:46 |
andreahack | hi slytherin :D | 18:46 |
slytherin | andreahack: what brings you in this part of IRC? :-) | 18:47 |
andreahack | well one moment that i don't speak english very well:P | 18:48 |
andreahack | i am here because do not have to do better | 18:50 |
andreahack | :P | 18:50 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!