2008-08-20
| 09:38 | cemerick | it appears that all of the invoke overloads emitted by genclass (when the genclass specification includes a reference to IFn or something that implements/extends it) attempt to invoke super.invoke with two more arguments than they should -- this leads to the 19- and 20-arg invoke implementations including references to 21- and 22-arg IFn invoke signatures, which don't exist. |
| 12:34 | cemerick | does anyone else see a significant delay in google groups' email gateway (both sending and receiving)? Posted messages get to me about two days after they're visible on the site, and messages I send in by email can take anywhere from 5 minutes to many hours to appear online. |
| 12:35 | kotarak | I noticed that, but that was a one time issue for me. A mail of mine took 12hours to show up. It was even overtaken by another one. |
| 12:35 | cemerick | it's persistent for me. Totally irritating, especially since email is the only way to attach files to a post. |
| 12:36 | kotarak | Yeah. That's also annoying. |
| 12:40 | Chouser | 2 days!? I hadn't noticed that. |
| 12:41 | cemerick | Yup. I'm reading the alter vs. commute thread online because it hasn't shown up in my email yet. |
| 12:41 | cemerick | or, the latest stuff hasn't. |
| 12:44 | Chouser | I've gotten all of that via email. I don't think Rich's last note was more than a few miutes late. |
| 13:30 | albino | setting up mailman is definitely preferred |
| 13:52 | cemerick | google groups is far preferable to mailman w.r.t. everything aside from its mail handling |
| 13:54 | albino | why is that? |
| 14:13 | cemerick | mailman is painful in every aspect -- administration, UI, archival access, etc., etc. |
| 14:15 | albino | heh |
| 14:15 | albino | I thought the opposite |
| 14:16 | cemerick | Interesting. I thought everyone hated mailman archives. They have a very gopher-esque quality to them. ;-) |
| 14:25 | albino | my personal hate goes to the web arvhives that don't have a thread view |
| 14:25 | albino | that always kills me |
| 15:31 | solkis | Question: What is the best way to add third-party jars to the clojure installation? I've previously setup command-line and emacs support for clojure via a script but now I want to start to play with it a bit. Do I need to modify the -cp flag of my script to point to each jar I may want to use or is there a way that can dump jars into a directory and they will be available? |
| 15:38 | Chouser | I think that's actually a Java question. Whatever the normal way is to add .jars to your classpath, do that. |
| 15:38 | Chouser | No, I don't know what that is. :-) It may very from one OS to another, or even between distros. I'm not sure. |
| 15:39 | kotarak | In Mac (and Unix I think) java respects the CLASSPATH env var. |
| 15:39 | cemerick | Yeah, Clojure is a Java "app" just like any other. At some point, you need to inject the other jars into your classpath (either -cp or CLASSPATH env variable, etc) |
| 15:40 | solkis | Ok, thanks. |
| 16:21 | albino | there's a cli agrument you can pass to use a directory full of jars, isn't there? |
| 16:22 | albino | and -cp if you want to pass each jar on the command line |
| 16:30 | solkis | not sure... I think -cp or -classpath will search in directories for class files but will not automatically add any jars or zips it finds to the classpath. |
| 16:35 | albino | http://today.java.net/pub/a/today/2005/04/26/extending.html <- -Djava.ext.dirs= according to that |
| 16:40 | solkis | ok, yes... I have seen that used |
| 16:46 | solkis | another easy question: is there a way to use bit-or functionality when I have multiple arguments? I'm checking out clojure while using a little SWT and when trying to create a shell you need to bitwise or some various options... but this won't work because bit-or only takes two arguments: (bit-or (. SWT CLOSE) (. SWT TITLE) (. SWT ON_TOP ) |
| 16:46 | solkis | I could nest another bit-or but I didn't know if there was an easier way |
| 16:48 | albino | no clue really, but can't you write a function that takes a variable amount of arguments and returns the bit-or of all the args? |
| 16:48 | kotarak | maybe: (reduce bit-or [(. SWT ....)])? |
| 16:51 | solkis | I'll investigate these approaches... I'm new to clojure (and lisp) :) |
| 16:53 | kotarak | I don't know how far to push this, but...: (reduce bit-or (map #(. SWT %) '(CLOSE TITLE ON_TOP))) |
| 16:53 | kotarak | Probably bad style. |
| 16:54 | solkis | (reduce bit-or [(. SWT CLOSE) (. SWT TITLE) (. SWT ON_TOP )]) definitely does it! |
| 16:55 | kotarak | I don't like the multiple (. SWT ...), but it's probably pushing things... |
| 16:56 | solkis | I like your 2nd example too, especially in terms of me learning more here, but it didn't run... I'm looking to see why |
| 16:56 | kotarak | Oops. Ok. Know why. |
| 16:57 | kotarak | It's wrong. |
| 16:57 | kotarak | As I said: pushing things.# |
| 16:58 | kotarak | I shouldn't chat at that time of the day... |
| 21:48 | Chouser | if those are class statics, you cant say [SWT/CLOSE SWT/TITLE SWT/ON_TOP] instead |