| Content of these pages are owned and copyrighted by the poster. | |
| Hosted by: |
|
Note: to get the raw text, simply edit this page and copy out the text between "pre" tags. Authors, when creating log pages, open with a "pre" and don't forget to close with a "pre" or you'll mess up the footer!
### Log session started at Sat Feb 24 22:44:53 2001 ###
[22:45:32] <jdunrue> I'm thinking John put that in there as a place-holder, but I'm not sure if it's a real function.
[22:45:37] <jdunrue> I don't have it at any rate.
[22:46:18] <_Quinn> Hm. That's cute, ProcessServer.c has broken syntax. Maybe I missed something in CVS.
[22:47:16] <_Quinn> !argh
[22:47:26] <jdunrue> Yeah, he needs a #include for pthread.h.
[22:47:38] <_Quinn> I have to rewrite the makefile -- my default gcc isn't 2.95.2
[22:47:54] <_Quinn> er, the .makeme files
[22:47:56] <_Quinn> bbiab
[22:48:17] <jdunrue> What have you been working on lately?
[22:49:03] <_Quinn> sfWiki, actually :)
[22:49:19] <_Quinn> I'm supposed to be working on porting the console drivers from JJOS/decaf, but I haven't had time yet.
[22:49:39] <jdunrue> That's cool. That wiki thing is new to me.
[22:49:45] <_Quinn> I'm in a race with the JOS webmaster to see who finishes their part of the site first :)
[22:49:48] <jdunrue> Ha!
[22:50:30] <_Quinn> (fixing gcc didn't help.)
[22:50:38] <_Quinn> which file is missing pthread.h?
[22:50:50] <jdunrue> ProcessService.c, I believe. Let me check my e-mail.
[22:50:59] <jdunrue> ProcessServer.c
[22:51:02] <_Quinn> It looks like I'm winning, but we'll see. It's a good way to make us get the work done, though.
[22:51:05] <_Quinn> Ah. BBIAB.
[22:51:37] <_Quinn> #include <pthread.h>?
[22:51:48] <jdunrue> yup
[22:52:13] <_Quinn> OK, that worked. But he also forgot to include jvmdi/ in the .makeme file. hold on.
[22:52:59] <_Quinn> No, it looks they aren't in CVS ?!
[22:53:12] <jdunrue> I didn't see them in my last update, either.
[22:53:26] <jdunrue> I got the kernel version compiled, but there are unresolved symbols at run-time.
[22:53:27] <_Quinn> I have jvmdi.c, but not misc, reflections, or monitors -- is your build trying to pick them up?
[22:53:51] <jdunrue> hang on
[22:54:40] <jdunrue> I have to recompile user mode to see that again.
[22:54:45] <_Quinn> Well I committed the fix for ProcessServer.c
[22:54:47] <_Quinn> OK.
[22:55:00] <_Quinn> I'm checking my kernel build, now, too
[22:55:52] <_Quinn> Oh, BTW, are you logging this?
[22:56:17] <jdunrue> Yes. But turn on logging on your end anyway. I'm a IRC newbie and I might
[22:56:33] <jdunrue> have made a mistake. This is the first time I've used KVirc.
[22:56:47] <jdunrue> I too am missing misc, reflection, and monitors.
[22:57:47] <_Quinn> Hm. I'm not an IRC newbie, but I don't how to turn on ircii's logging. I'll futz with it.
[22:58:08] <jdunrue> KVirc has a menu item that I hit to do it, but my log file is not growing.
[22:58:30] <_Quinn> Hm...
[22:58:49] <_Quinn> Well, I think my fix for processerver broke the kernel build. :)
[22:58:59] <jdunrue> oops :-)
[22:59:09] <jdunrue> that's a pun, by the way
[22:59:38] <_Quinn> huh?
[22:59:43] <jdunrue> never mind
[22:59:48] <_Quinn> Sorry, missed it :)
[22:59:57] <jdunrue> Do you have any opinion on my issue with the JDI spec?
[23:00:01] <jdunrue> licensing, that is?
[23:00:26] <_Quinn> Oh, I thought you meant the com.sun.* namespace
[23:00:35] <jdunrue> well, yeah, that's really what I mean
[23:00:44] <jdunrue> I'm spooked by it.
[23:00:53] <jdunrue> But I don't see what alternative there could be.
[23:00:58] <_Quinn> It sure /looks/ fishy :)
[23:01:07] <jdunrue> People must be allowed to implement that stuff. That's the namespace
[23:01:19] <jdunrue> that third-party debuggers would be using.
[23:01:35] <jdunrue> Because that's the spec -- kind of a circular argument.
[23:01:35] <_Quinn> yeah. com.sun.* isn't just implementation stuff, either?
[23:01:58] <jdunrue> Not as far as I can tell. They make it sound like *that* is the spec.
[23:02:09] <jdunrue> So I'm going to move ahead and assume we will be OK.
[23:02:21] <jdunrue> Let 'em sue me :-)
[23:02:47] <_Quinn> Yeah :)
[23:03:03] <_Quinn> OK, I just talked the kernel into building.
[23:03:34] <jdunrue> I get a bunch of c-library type of functions unresolved when I do the insmod thing.
[23:04:03] <_Quinn> I haven't tried insmoding it yet. It hasn't blown up recently, but I wanted to check the other two builds first.
[23:04:17] <jdunrue> BTW, where can I go to get RJK?
[23:04:43] <jdunrue> I've got a separate machine here at home that I can test stuff on and not mess up my main machine.
[23:05:58] <_Quinn> The missing files only show up in the user build, so I bet John is testing them out there.
[23:06:21] <_Quinn> good question. I think SourceForge. I'll go check -- I got emailed the source they're using
[23:07:04] <jdunrue> I found Robert's user name on SourceForge, but he isn't listed as having a project with the initials RJK
[23:07:15] <jdunrue> last time I checked.
[23:07:28] <_Quinn> *shrug* Let me check if I still have the archive lying around.
[23:07:42] <_Quinn> I do -- did you want it emailed?
[23:07:48] <jdunrue> yes, please!
[23:07:55] <_Quinn> I have two machines too, but not two keyboards and monitors :(
[23:08:18] <jdunrue> If I could only get my network straightened out, then I'd really be in good shape.
[23:08:41] <_Quinn> It's en-route.
[23:08:46] <jdunrue> ok, thanks
[23:08:49] <_Quinn> Yeah, same here :)
[23:09:13] <jdunrue> I am trying to avoid buying another NIC, but I can't get aliasing to work with ipchains
[23:09:34] <jdunrue> I can get the alias for eth0 alright, but ipchains doesn't recognize it
[23:09:35] <_Quinn> OK, I can't get RJK to build... investigating
[23:09:53] <jdunrue> a topic for LinuxHelp anyway :-)
[23:10:02] <_Quinn> I know almost nothing about networking, unfortunately.
[23:10:58] <jdunrue> I got the rjk archive, thanks again.
[23:11:10] <jdunrue> Any idea how far along this is?
[23:11:36] <_Quinn> RJK? Further than the JJOS kernel, and it has threads, but I don't know much more than that.
[23:13:51] <jdunrue> On another topic... has anyone heard from the guy that was starting the Opal project, which I
[23:14:05] <jdunrue> think is/was supposed to be a GUI for JOS.
[23:14:17] <jdunrue> Nothing from him on the list, lately.
[23:14:27] <_Quinn> /I/ sure haven't.
[23:15:05] <jdunrue> I'm personally glad to have stumbled upon something I can work on (JOSDI) that I'm pretty sure won't
[23:15:10] <jdunrue> get my employer up-tight.
[23:15:44] <_Quinn> who do you work for?
[23:15:47] <jdunrue> Sybase
[23:17:07] <_Quinn> Ah.
[23:17:12] <_Quinn> OK, here's the RJK problem.
[23:17:13] <_Quinn> struct native_thread_lock {
[23:17:13] <_Quinn> tOBREF object;
[23:17:13] <_Quinn> struct native_thread_lock* next;
[23:17:13] <_Quinn> kuint thread_id;
[23:17:13] <_Quinn> int iHasCondition;
[23:17:14] <_Quinn> };
[23:17:30] <_Quinn> About which the compiler is furiously complaining, just before the kuint.
[23:17:31] <_Quinn> Ideas?
[23:17:42] <jdunrue> what's the nature of the complaint?
[23:17:49] <jdunrue> syntax errors?
[23:17:57] <_Quinn> Oh, yeah, hold on a second :)
[23:18:45] <_Quinn> vm/threads_native.h:27: parse error before `kuint'
[23:18:45] <_Quinn> vm/threads_native.h:27: warning: no semicolon at end of struct or union
[23:18:45] <_Quinn> vm/threads_native.h:33: parse error before `}'
[23:19:02] <_Quinn> (I pasted in from the preprocessed version.)
[23:20:25] <jdunrue> not sure, I'm doing some digging...
[23:20:50] <_Quinn> maybe it doesn't like the 'struct' on the previous line?
[23:21:38] <jdunrue> I /thought/ that was legal.
[23:21:49] <_Quinn> So did I... :)
[23:22:10] <_Quinn> well, pulling the struct out gives me
[23:22:11] <_Quinn> vm/threads_native.h:18: parse error before `native_thread_lock'
[23:22:12] <_Quinn> vm/threads_native.h:18: warning: no semicolon at end of struct or union
[23:22:12] <_Quinn> vm/threads_native.h:27: warning: type defaults to `int' in declaration of `thread_id'
[23:22:12] <_Quinn> vm/threads_native.h:27: warning: data definition has no type or storage class
[23:22:12] <_Quinn> vm/threads_native.h:33: parse error before `}'
[23:22:34] <jdunrue> gcc 2.95.2 thinks it is OK, too.
[23:22:51] <jdunrue> I wonder if there is a missing declaration for tOBREF?
[23:23:06] <jdunrue> I meant, gcc thinks the example I just cooked up is OK.
[23:23:31] <jdunrue> {
[23:23:31] <jdunrue> struct xxx* y;
[23:23:31] <jdunrue> };
[23:23:37] <jdunrue> oops
[23:23:39] <_Quinn> Really? I thought I was using gcc 2.95.2; maybe I blew my rewrite.
[23:23:41] <_Quinn> bbiab
[23:23:51] <jdunrue> struct xxx
[23:23:51] <jdunrue> {
[23:23:51] <jdunrue> struct xxx* y;
[23:23:51] <jdunrue> };
[23:23:57] <jdunrue> seems to compile OK for me
[23:25:51] <_Quinn> OK, I'm trying something else right now.
[23:26:14] <jdunrue> ok, I'm coding, so take your time.
[23:26:24] <_Quinn> No, I'm getting the same error, even with gcc-2.95.2 used explicitly.
[23:26:35] _Quinn tests the user build.
[23:27:27] <_Quinn> Heh, though I saw a warning: "Inelegant code used, ignoring."
[23:27:35] <_Quinn> OK, user build likes me.
[23:27:48] <_Quinn> (except for the missing bits)
[23:28:17] <jdunrue> I'm constantly surprised by the kinds of messages gcc spits out.
[23:28:49] <jdunrue> Sometimes, it's like -- hey you idiot, you forgot a semi-colon, here let me put it in there for you.
[23:28:57] <_Quinn> No, I'm pretty sure I was hallucinating :)
[23:29:22] <_Quinn> Wouldn't it be nice if it actually did, though? :)
[23:29:54] <jdunrue> I don't know, I'd be leery of any kind of compiler that "smart" :-)
[23:30:52] <jdunrue> speaking of C compilers, do you suppose it will be a goal for JOS (name your flavor) to
[23:31:04] <jdunrue> be self-hosting at some point. A long time down the road, I know.
[23:32:57] <_Quinn> Actually, JOS will self-hosting once we get a filesystem working, I think: I'm pretty sure Sun's javac is written in java.
[23:33:07] <_Quinn> Or did you mean C also?
[23:33:20] <jdunrue> but the kernels are in C and asm
[23:33:26] <_Quinn> Something about kuint is really just pissing gcc off.
[23:33:46] <jdunrue> the vm is in C, too
[23:34:01] <jdunrue> I think gcc just doesn't like you this evening.
[23:34:09] <_Quinn> If I change it to /*ku*/int, it works fine until a function has kuint as a return value.
[23:34:20] <_Quinn> And the RJK build is working for you? (Or have you not tried it?)
[23:34:34] <_Quinn> I'm tempted to believe in personal enmity too, at this point.
[23:34:59] <jdunrue> haven't tried it. I've been poking at my own code. I haven't looked at whether the RJK
[23:35:22] <jdunrue> code needs to be in a special place, or what other funny things need to be done for it all to build right.
[23:37:15] <jdunrue> a ha
[23:37:28] <jdunrue> I see at least one magic path in .makemeRJK
[23:37:46] <_Quinn> I changed it, but I didn't change it right.
[23:38:08] <_Quinn> And John is also using a more recent version of RJK, it looks like
[23:38:16] <_Quinn> sorry, less recent
[23:38:55] <jdunrue> life on the bleedin' edge, man... gotta love it :-)
[23:39:44] <_Quinn> Yeah. I don't think John has quite wrapped his head around using CVS with/for other developers :)
[23:39:52] <_Quinn> OK, that time it went right together.
[23:40:08] <_Quinn> *sigh*
[23:40:28] <_Quinn> I'm also curious as to why current/k* are there, but whatever...
[23:40:36] <jdunrue> gcc has deigned to accomodate you
[23:40:47] <_Quinn> Yes, it aims to be mysterious.
[23:40:50] <jdunrue> CVS is non-intuitive, I have found.
[23:40:58] <jdunrue> I have to deal with lots of people that have never used it.
[23:41:01] <_Quinn> Like a fine woman, you must court gcc carefully
[23:41:30] <_Quinn> It's non-intuitive, but I'm talking more about things like hardcoding your home directory. That's what variables (or relative paths) are for.
[23:41:39] <jdunrue> Oh yeah, gotcha.
[23:42:21] <jdunrue> I'm going to give this a whirl and see if it will build for me. what did you end up having to do?
[23:42:27] <jdunrue> RJK that is.
[23:43:13] <_Quinn> Oh. I had replaced <john's home dir>/rjk.20010204 with <path-to-my-copy>rjk.20010204, but I had rjk-20010211 :)
[23:43:56] <jdunrue> hang on, let me get this file tweaked for my setup, and we'll see what happens
[23:45:06] _Quinn hangs
[23:46:16] <jdunrue> <compiling>
[23:46:48] <jdunrue> well, whatcha know... kuint strikes again
[23:47:07] <_Quinn> Argh!
[23:47:10] <_Quinn> Same place?
[23:47:28] <jdunrue> line 55 of vm/garbage.c
[23:47:47] <jdunrue> uh, waitaminnut
[23:48:13] <jdunrue> I mean line 26 of vm/threads_native.h
[23:49:20] <jdunrue> so yeah, same place
[23:49:28] <_Quinn> Hm. CVS is showing a change in that file on my machine, but the diff is empty.
[23:49:37] <_Quinn> I'll commit it anyway, and let you try it.
[23:50:47] <_Quinn> Try now
[23:50:54] <jdunrue> ok
[23:51:15] <jdunrue> <compiling>
[23:51:36] <jdunrue> same error
[23:52:04] <_Quinn> OK. And you've the k*.h files in current/?
[23:52:39] <jdunrue> kimplementation.h, kinterface.h am I missing anything?
[23:53:11] <jdunrue> ok, i see the problem
[23:53:22] <_Quinn> nope.
[23:53:25] <_Quinn> you do?
[23:53:32] <_Quinn> You have sharper eyes than I :)
[23:53:56] <jdunrue> hang on a sec.
[23:54:23] <jdunrue> so is this compiling OK for you now?
[23:55:03] <jdunrue> I was wrong.
[23:55:07] <_Quinn> Yes...
[23:55:14] <jdunrue> I thought he didn't have -I. in the list of include paths.
[23:55:42] <_Quinn> grep .makemeRJK and make sure you didn't miss any 'jewel's?
[23:56:17] <jdunrue> oh boy, do I feel silly
[23:57:10] <jdunrue> the case of the bad search-and-replace.
[23:57:17] <jdunrue> let me try that again
[23:58:00] <_Quinn> Well, I did it twice, so don't feel too bad :)
[23:58:19] <jdunrue> :-)
[23:59:04] <jdunrue> needed to compile RJK itself, duh...
[23:59:13] <jdunrue> OK, it's built.
[23:59:25] <jdunrue> what does it boot with?
[00:00:14] <_Quinn> GRUB
[00:00:20] <jdunrue> oh...
[00:00:24] <_Quinn> I should have a boot-disk image lying around here somewhere
[00:00:34] <jdunrue> I guess I should break down and get grub installed on my other machine.
[00:00:52] <jdunrue> yeah, a boot-disk would be cool.
[00:01:07] <jdunrue> this would be an image you can just dd onto a floppy, right?
[00:02:24] <_Quinn> yes.
[00:02:34] <_Quinn> sorry, I'm looking at the -Wall stuff simultaneously
[00:02:46] <_Quinn> I'll email it.
[00:03:01] <jdunrue> ok, thanks.
[00:04:36] <jdunrue> btw, looks like my IRC log is functioning as expected, just missed the first exchange or two.
[00:04:44] <_Quinn> mail's away
[00:05:01] <_Quinn> Well, I suspect they weren't of earth-shaking importance anyway :)
[00:05:17] <jdunrue> Hate to be a party crasher, but I should probably get going in a minute.
[00:06:27] <_Quinn> That's fine. Two in the morning is late enough for me :)
[00:06:35] <_Quinn> What's it for you, 6?
[00:06:53] <jdunrue> midnight, I'm in Colorado. you must be East Coast?
[00:07:03] <_Quinn> (and we got everthing compiled already)
[00:09:09] <jdunrue> Earthlink's mail server is not cooperating with me. I might have to wait until morning to check for your boot disk.
[00:09:21] <_Quinn> OK.
[00:09:37] <jdunrue> So, I will bid you adieu. Thanks for dropping in :-) Nice to meet you, virtually speaking.
[00:09:41] <_Quinn> Yes. You said -4 GMT, so I though Europe for (+4 GMT)
[00:09:47] <_Quinn> llater!
[00:09:51] <jdunrue> bye
[00:10:05] <_Quinn> (which, of course, would make it more like ten over there, but whatever)
### Log session terminated at Sun Feb 25 00:10:10 2001 ###
| Content of these pages are owned and copyrighted by the poster. | |
| Hosted by: |
|