IRC log of openacs on 2002-05-23

Timestamps are in UTC.

00:06:15 [til]

argh ... openfts and debian are not friends

00:06:22 [markd2]

:-(

00:10:17 [til]

ah, just found nextsteps.html on openacs.org. hmm, unary error remains

00:47:25 [jim2]

openfts and debian not friends? license?

00:47:46 [til]

no, the configure script needs a lot of hand editing

00:48:17 [til]

maybe this is only with postgresql 7.2 as debian package

00:48:37 [jim2]

which is the woody version?

00:48:43 [til]

have you installed openfts?

00:48:57 [til]

woody ... don't know

00:49:00 [jim2]

I don't remember...

00:49:15 [jim2]

I might have, but I compile everything

00:50:46 [til]

i'm using unstable. yeah, compiling is better propably

00:51:12 [jim2]

(one of debian's relative strong points is their toolchain works due to the package deps among toolchain pieces: it's hard to install conflicting pieces of the toolchain)

00:54:15 [jim2]

seen benadida

00:54:15 [paje]

benadida was last seen on #openacs 86 days, 22 hours, 31 minutes and 31 seconds ago, saying: talk to you later [Sun Feb 24 19:23:42 2002]

00:54:30 [jim2]

-that- long??

00:54:50 [jim2]

can't be... I thought I saw him here about a week and a half ago

00:54:51 [til]

seen ben

00:54:52 [paje]

I haven't seen 'ben', til

00:55:03 [markd2]

paje might have been gone last time he appeared

00:55:03 [paje]

markd2: what?

00:55:11 [markd2]

seen thursdayphylax?

00:55:12 [paje]

thursdayphylax was last seen on #openacs 19 days, 6 hours, 4 minutes and 38 seconds ago, saying: aren't I behind the times [Fri May 3 12:51:31 2002]

00:55:13 [jim2]

true

00:55:19 [markd2]

seen wednesdayphylax?

00:55:20 [paje]

wednesdayphylax was last seen on #openacs 21 days, 42 minutes and 0 seconds ago, saying: hi paje [Wed May 1 18:14:17 2002]

00:55:28 [markd2]

seen sundayphylax?

00:55:28 [paje]

I haven't seen 'sundayphylax', markd2

00:55:49 [jim2]

seen footballphylax

00:55:49 [rbm]

somebody twack me to package openfts/nsopenssl/other-aolserver-modules for Debian

00:55:49 [paje]

I haven't seen 'footballphylax', jim2

00:56:17 [jim2]

rbm: you got in? :)

00:56:28 [rbm]

til: yes, pdflib is proprietary (for commercial purposes)

00:56:40 [rbm]

I mean, it costs money for commercial purposes

00:57:01 [rbm]

jim2: Not yet. I have the T&S test in my indox but haven't had time to reply to it yet.

00:57:24 [rbm]

jim2: we should start working on OpenACS packages

00:57:31 [jim2]

ok

00:57:56 [jim2]

* jim2 is looking at dotlrn...

00:58:10 [jim2]

had to find an oracle I still had up

00:58:25 [til]

* til twacks rbm to package openfts/nsopenssl/other-aolserver-modules

00:58:46 [rbm]

thanks til. I was needing that.

00:59:12 [arjunsanyal]

arjunsanyal (~arjun@pool-151-202-175-216.ny5030.east.verizon.net) has joined #openacs

01:00:09 [rbm]

woot! arjun is here!

01:00:23 [arjunsanyal]

hi

01:00:23 [paje]

hi, arjunsanyal

01:01:04 [rbm]

hey arjun. what's up?

01:01:43 [rbm]

jim2: Is it okay to discuss what I'm thinking to answer to my T&S test?

01:02:18 [jim2]

yes... as long as it's ok for me to decide how I will respond :)

01:02:29 [rbm]

:)

01:02:39 [arjunsanyal]

hey rbm, not much just relaxing

01:02:48 [jim2]

any chance you could send me a copy?

01:03:27 [rbm]

jim2: of the e-mail? Sure.

01:03:36 [jim2]

jim@laney.edu

01:04:01 [jim2]

I need coffee and a coupla other things

01:06:08 [arjunsanyal]

rbm: did you graduate?

01:06:18 [jim2]

argh. my dotlrn is having problems... I got AOLserver/3.3.1+ad13, oracle-8.1.6 and openacs/dotlrn from CVS on may 16th in the late evening...

01:06:33 [arjunsanyal]

jim2: I'm your man

01:06:46 [arjunsanyal]

what's the error?

01:06:46 [paje]

rumour has it the error is generated by the browser

01:06:47 [jim2]

exellent :)

01:07:37 [jim2]

I'm looking at a Subject...

01:07:45 [jim2]

I say New Class...

01:07:47 [rbm]

arjunsanyal: I did.

01:07:59 [arjunsanyal]

rbm: congrats!

01:08:06 [rbm]

arjunsanyal: Thanks :)

01:08:59 [jim2]

I'm doing "subjects" as name of the course, and "classes" are sections of classes... so: subject=java programmming, class=section 1

01:09:02 [jim2]

no problem

01:09:14 [jim2]

"New Class" again

01:09:20 [jim2]

I want section 2

01:09:56 [jim2]

ora8.c:3568:ora_tcl_command: error in `OCIStmtExecute ()': ORA-00001: unique constraint (DOTLRN.CR_ITEMS_UNIQUE_NAME) violated

01:09:56 [jim2]

ORA-06512: at "DOTLRN.CONTENT_ITEM", line 161

01:09:56 [jim2]

ORA-06512: at "DOTLRN.CONTENT_FOLDER", line 34

01:09:56 [jim2]

ORA-06512: at "DOTLRN.FILE_STORAGE", line 85

01:09:56 [jim2]

ORA-06512: at line 3

01:10:31 [arjunsanyal]

jim2: seen it. hold

01:10:31 [paje]

I haven't seen 'it.', arjunsanyal

01:10:57 [rbm]

jim2: sent.

01:11:01 [jim2]

thanks :)

01:11:06 [jim2]

going for coffee...

01:11:18 [jim2]

and there's more problems, stick around :)

01:11:33 [arjunsanyal]

jim2: ok

01:11:42 [rbm]

arjunsanyal: I don't have a real job, so I don't know if I'm really happy that I graduated.

01:12:15 [rbm]

(I didn't look very hard though, since I was thinking of going on to graduate school)

01:12:55 [rbm]

paje: my dotlrn?

01:12:55 [paje]

rbm: bugger all, i dunno

01:13:00 [rbm]

paje: dotlrn?

01:13:00 [paje]

well, dotlrn is running on davb's machine again, and not crashing if you want to look around.

01:13:00 [arjunsanyal]

rbm: after i graduated i took it easy for a long time.... until december. before i got a job

01:14:31 [arjunsanyal]

rbm: sucks

01:14:56 [arjunsanyal]

rbm: where are you from?

01:14:59 [rbm]

Brazil

01:17:43 [til]

wohoo, search finally works!

01:19:43 [markd2]

yay

01:19:58 [rbm]

I haven't done anything with openfts yet.

01:20:07 [rbm]

I need to if I want to package it.

01:20:18 [rbm]

* rbm wanders off

01:23:51 [til]

wow, thats weird: whenever i search for words with german umlauts in it, opera wants to save the result page instead of just displaying it

01:27:18 [til]

anybody know how i can make openfts index old content? e.g. news items that have been on the site before openfts was installed?

01:30:10 [jim2]

ok...

01:32:52 [jim2]

here's another...

01:33:09 [jim2]

I'm at /dotlrn/admin/users

01:33:37 [jim2]

I say create

01:34:03 [jim2]

I get the form, and fill it out

01:36:07 [jim2]

I get the email form, and decide not to send, but I want to create another user.

01:36:41 [jim2]

so I go back (reasonable assumtion, I can use my back button? :)

01:37:00 [jim2]

and I find the form, still filled out with the user I just added

01:37:17 [jim2]

so, I say -cool-, I'll just edit that

01:38:17 [jim2]

I do so, and submit... error:

01:38:31 [jim2]

can't use non-numeric string as operand of "!"

01:38:31 [jim2]

while executing

01:38:31 [jim2]

"if {![cc_email_from_party $target_user_id]} {

01:38:31 [jim2]

# create the ACS user

01:38:31 [jim2]

set password [ad_generate_random_string]

01:38:33 [jim2]

set t..."

01:38:52 [jim2]

I know exactly where that is, and can get you the cvs info for that file...

01:39:31 [arjunsanyal]

hmm, haven't seen that one. yeah the cvs info would be great

01:40:35 [jim2]

it's in /packages/dotlrn/www dir, file /user-add.tcl/1.17/Thu May 16 05:17:09 2002//,

01:40:47 [jim2]

about line 121 or 122

01:40:58 [jim2]

HOWEVER

01:42:00 [jim2]

if I would have went back one more time, then clicked on Create New User again, I would have gotten a blank form, filled it out, added the user no problem.

01:42:06 [jim2]

so...

01:42:20 [jim2]

the only consequence of this bug is BBB

01:42:26 [jim2]

Back Button Breakage

01:43:01 [arjunsanyal]

jim2: for the first error, check the vesion of the /pacakges/file-storage/file-storage-init.tcl file. let me try to replicate the user-add bug

01:43:19 [jim2]

ok

01:43:36 [jim2]

(note about versions...

01:44:44 [jim2]

I'm running aolserver 3.3.1 + ad13, oracle-8.1.6 and got openacs and dotlrn from cvs at about the same time, evening of may 16th, pst

01:44:47 [jim2]

)

01:45:17 [til]

jim: in my opinion the assumption that the back button should work is not 100% reasonable in this case.

01:45:33 [jim2]

then new interface is needed :)

01:46:05 [arjunsanyal]

jim2: i agree, it's a bug

01:46:41 [jim2]

if you want to go by the assumptions of Apple Human Interface Guidelines, what you have to do is design the interface in such a way that the error is prevented

01:47:18 [jim2]

til: having said all that, I have a vague tendency to agree, or at least understand

01:47:33 [til]

maybe i just have too many bad experiences, so that i automatically avoid the back button when forms are involved

01:49:40 [jim2]

til: I see your point... thing is, not everyone will do that... me, I like using back :) and there will be -clueless- school staff and faculty who have power and who won't find the situation acceptable, if they can't find a workaround like I did

01:50:30 [til]

so in this case maybe adding a "skip email sending" link that is simply the return-url would do the trick?

01:50:30 [til]

isn't that an acs-subsite issue - not dotlrn?

01:50:47 [jim2]

at the very least, you need buttons on the mail-new-user page that say "don't send it" and "add another user"

01:51:17 [jim2]

and then, some people will -still- use back :)

01:51:26 [jim2]

what can ya do? :)

01:52:19 [arjunsanyal]

til: no i think the back button should work, we want dotlrn users to never see a "request error" page, i'll let yon know about this (it's his department)

01:52:22 [jim2]

oops, we just came up with the same sol'n, and I did so without reading that you did so :)

01:53:31 [arjunsanyal]

jim2: about the file-storage error?

01:53:51 [jim2]

hmmm... maybe this is irrelevent... but the first person I added was John Smith... and the second was his bro, John Lennon...

01:54:01 [jim2]

arjunsanyal: one sec

01:54:40 [jim2]

(I did change email address and ID for Lennon)

01:56:21 [jim2]

the /pacakges/file-storage/file-storage-init.tcl file I can't find yet...

01:56:48 [jim2]

I corrected spelling on packages (well, by never CDing out of that dir :)

01:57:04 [jim2]

but file-storage-init.tcl might be in tcl/

01:57:18 [arjunsanyal]

jim2: yup in /tcl sorry :)

01:57:24 [jim2]

bingo.... /file-storage-init.tcl/1.2/Sun May 12 20:57:02 2002//

01:58:28 [arjunsanyal]

jim2: ahh ok, you need to revert that to v1.1 - you can use this command"

01:58:28 [arjunsanyal]

cvs update -r 1.1 -p file-storage-init.tcl > file-storage-init.tcl", restart, and _i think_ it should fix it

01:59:10 [jim2]

mv file-storage-init.tcl file-storage-init.tcl-version1.2

02:00:04 [jim2]

umm :)

02:00:11 [jim2]

that version has nothing in it :)

02:00:24 [jim2]

except a call to ad_library :)

02:00:59 [arjunsanyal]

jim2: yup that's what we want. of course, I could have said just comment the stuff out... :)

02:01:08 [jim2]

true ;)

02:01:13 [jim2]

ok, hammering.

02:01:34 [arjunsanyal]

I will make dotlrn restarts faster one day

02:01:54 [arjunsanyal]

i do a lot of them

02:02:10 [jim2]

they already -are- faster than one day!

02:02:29 [arjunsanyal]

:)

02:02:52 [jim2]

ok, which bug is that supposed to fix?

02:04:05 [jim2]

add a class under a subject?

02:04:09 [jim2]

add a user?

02:04:25 [jim2]

* jim2 tries both

02:05:55 [arjunsanyal]

sorry, doing 4 things at once. class under a subject

02:06:02 [jim2]

that works now

02:06:06 [arjunsanyal]

cool!

02:06:50 [arjunsanyal]

i sent email to yon about the user-add bug, he'll probably have it fixed tomorrow

02:06:56 [jim2]

adding a community works

02:07:14 [jim2]

(I think I had trouble before)

02:09:13 [jim2]

ok, next:

02:09:31 [arjunsanyal]

great, FYI you can email dotlrn@openforce.net with bug reports - all the dotlrn developers are recipients

02:11:24 [jim2]

OK :) but it's nice to have a quick 2-way with someone in the know before actually filing...

02:12:29 [jim2]

but the way it is now, it's sorta presentable...

02:12:38 [jim2]

I have some suggestions...

02:12:54 [jim2]

and at least one more problem, if you'd like to hear them :)

02:14:15 [arjunsanyal]

go for it

02:14:21 [jim2]

OK...

02:16:28 [jim2]

(see... I built something I called the Roster Project, potentially capable of becoming aware of any number of students and the classes they're enrolled in from a large CSV file, including them in a text db and from there generating accounts on an arbitrary number of arbitrary kinds of servers...)

02:16:42 [jim2]

(so I've been here, done some of that :)

02:17:19 [arjunsanyal]

interesting..

02:17:31 [jim2]

first, I like the fact that you allow a bulk transfer of students... but it's not complete. what is needed is a way to get them into their classes.

02:17:35 [jim2]

but

02:17:45 [jim2]

if there's no defined classes, then what?

02:17:54 [jim2]

so you type them in?

02:18:23 [jim2]

better: also have a bulk transfer of classes

02:20:10 [jim2]

I had this running in a four-campus district, where I could identify any section of any class ever taught... by using a key with term, campus and classcode (a four-digit number unique to term and campus)

02:20:23 [jim2]

so, my keys looked like this...

02:20:40 [jim2]

F2001-Laney-1234

02:21:06 [jim2]

(well, a little different, but you can see what's up)

02:21:35 [jim2]

students were identified by their SSN

02:21:53 [jim2]

so their identity was identical on all four campuses

02:21:55 [jim2]

so:

02:23:32 [jim2]

could have deployed this thing such that their login name and password was the same everywhere on the four campuses, and login/passwd pairs would get added to any server equpt they were assigned access by way of their enrolling in a class

02:24:28 [jim2]

further, a specific teacher could ask for a file that could add accounts to that teacher's personal server, if that server was one of those supported by the system

02:24:57 [jim2]

(supporting a new server was fairly simple: you write about 7 perl procs)

02:25:49 [jim2]

now... this college district had a mainframe that had records for all students on all four campuses

02:26:18 [jim2]

and the programmer for that system created a program that could give the list of students for a class

02:26:37 [jim2]

in a fixed-field CSV format

02:26:50 [jim2]

also

02:27:07 [jim2]

she made another download available on a daily basis...

02:27:13 [davb]

davb (dave@alb-24-58-162-82.nycap.rr.com) has joined #openacs

02:27:24 [jim2]

which was the list of courses with their schedules and other info

02:27:47 [jim2]

with me so far?

02:27:52 [jim2]

hi davb

02:27:53 [arjunsanyal]

yup.

02:28:04 [davb]

hi everyone

02:28:19 [markd2]

hi

02:28:41 [jim2]

I was gonna create a calendar app for this district that would let a student choose his courses based on schedule

02:29:45 [jim2]

one idea that was advanced, was "I'm Joe Student, and I have tuesday and thursday available between the hours of 3 and 5 pm. show me all courses on all campuses I can take, in a calendar view.

02:29:49 [jim2]

"

02:31:52 [jim2]

but my idea was, "I'm Joe Potential Student, and I want to try to form a schedule. Help me by giving me a one-week view of my schedule, initially blank, that I can also put my own appointments on. Let me add potential classes to it".

02:33:00 [jim2]

In my Roster Project, I wanted to handle four types of people:

02:33:05 [jim2]

students,

02:33:12 [jim2]

(first priority :)

02:33:18 [jim2]

faculty,

02:33:23 [jim2]

staff,

02:34:02 [jim2]

and system accts (this was once unix-centric, and root, bin, apps, mail, etc were going to be among those)

02:34:35 [arjunsanyal]

ok

02:35:52 [jim2]

so I wanted to be able to create all new, and have my database of staff and faculty added, go to the mainframe for the schedule of classes, add that, then get the students from the mainframe and add them.

02:36:01 [jim2]

following that,

02:37:06 [jim2]

I'd run a query sorta like "select logname,passwd from students group by server-they-belong-to"

02:38:18 [jim2]

and the output of that (NOT A) select was automatically in the form needed by a particular server (most times, a batch file or shell script) to add accounts to that server

02:38:30 [jim2]

next...

02:38:59 [jim2]

I stored timestamps in appropriate places... so I'd know if a student was already added to a server

02:40:02 [jim2]

the upshot of which, was I could run the whole thing again, with students already in the central database I kept, to get new adds and drops, for the purpose of removing them from servers

02:40:09 [jim2]

or adding new people

02:41:14 [arjunsanyal]

this is for access to the local computer or for the calendar app?

02:41:22 [jim2]

both

02:41:47 [jim2]

you'd also try to get new classes or note ones that were apparently cancelled

02:41:50 [the_docwolf]

the_docwolf (~wolf@adsl-34-227-234.bct.bellsouth.net) has joined #openacs

02:41:59 [the_docwolf]

evening, funky cats

02:42:13 [markd2]

dude

02:42:27 [arjunsanyal]

howdy docwolf

02:42:33 [the_docwolf]

hola

02:42:44 [the_docwolf]

greetings to you, paje.

02:42:44 [paje]

the_docwolf: huh?

02:42:50 [the_docwolf]

paje, attack

02:42:51 [paje]

the_docwolf: sorry...

02:42:51 [jim2]

access to local servers is based on "whether student is enrolled in currently active course(s)"

02:43:26 [jim2]

and calendar schedule helper is based on "currently active courses"

02:44:28 [jim2]

further, calendar schedule helper could also be course enroller, i.e., student could then say "OK! I like that schedule! Sign me up for all these classes!"

02:45:33 [denshi]

paje, attack is | <reply> paje knaws furiously on docwolf's leg

02:45:33 [paje]

OK, denshi.

02:45:36 [denshi]

paje, attack

02:45:36 [paje]

paje knaws furiously on docwolf's leg

02:45:50 [jim2]

and, assuming everyone were doing their jobs, they'd have their own access to everything they needed before they walked onto campus

02:45:51 [markd2]

knaws?

02:46:18 [markd2]

paje, oop is ack

02:46:18 [paje]

...but oop is ack or blarg|fnord...

02:46:33 [markd2]

oop is ~/fnord/arg/

02:46:35 [jim2]

knawing is harder than gnawing, given k is harder than g :)

02:46:39 [markd2]

literal oop?

02:46:39 [paje]

markd2: oop =is= ack or blarg|fnord

02:46:52 [jim2]

oop

02:46:54 [jim2]

oop?

02:46:54 [paje]

oop is ack or blarg

02:46:57 [jim2]

oop?

02:46:57 [paje]

oop is fnord

02:47:13 [markd2]

oops is ~=/fnord/arg/

02:47:19 [markd2]

oop is ~=/fnord/arg/

02:47:21 [markd2]

literal oop

02:47:21 [paje]

markd2: oop =is= ack or blarg|fnord

02:47:27 [markd2]

* markd2 sighs

02:47:48 [jim2]

oop =~ s/fnord/arg/

02:47:52 [jim2]

oop?

02:47:53 [paje]

oop is probably ack or blarg

02:47:55 [jim2]

oop?

02:47:55 [paje]

i think oop is ack or blarg

02:47:56 [jim2]

oop?

02:47:57 [paje]

hmmm... oop is fnord

02:47:57 [denshi]

paje, attack is also | <reply> /me defeats markd2's period epic costume style with his dragon bot style

02:47:58 [paje]

okay, denshi.

02:48:01 [denshi]

paje, attack

02:48:01 [paje]

paje knaws furiously on docwolf's leg

02:48:02 [denshi]

paje, attack

02:48:02 [paje]

/me defeats markd2's period epic costume style with his dragon bot style

02:48:03 [markd2]

literal oop

02:48:03 [paje]

markd2: oop =is= ack or blarg|fnord

02:48:13 [denshi]

paje, literal attack

02:48:13 [paje]

denshi: attack =is= | <reply> paje knaws furiously on docwolf's leg|<reply> /me defeats markd2's period epic costume style with his dragon bot style

02:48:16 [jim2]

denshi: <action> I think

02:48:22 [markd2]

yep

02:48:24 [jim2]

paje, oop =~ s/fnord/arg/

02:48:24 [paje]

OK, jim2

02:48:26 [denshi]

paje, forget attack

02:48:26 [paje]

denshi: I forgot attack

02:48:27 [jim2]

oop?

02:48:28 [paje]

oop is ack or blarg

02:48:28 [jim2]

oop?

02:48:29 [paje]

hmmm... oop is arg

02:48:33 [jim2]

there! :)

02:48:33 [denshi]

paje, attack is also | <action> /me defeats markd2's period epic costume style with his dragon bot style

02:48:33 [paje]

OK, denshi.

02:48:37 [denshi]

paje, attack

02:48:37 [paje]

* paje /me defeats markd2's period epic costume style with his dragon bot style

02:48:41 [denshi]

paje, forget attack

02:48:41 [paje]

denshi: I forgot attack

02:49:27 [denshi]

paje, attack is | <action> commandeers the Kentucky Navy and sinks Pittsburgh.

02:49:28 [paje]

OK, denshi.

02:49:28 [jim2]

s#<action> /me defeats#<action> defeats#

02:49:32 [denshi]

paje, attack

02:49:32 [paje]

denshi: sorry...

02:49:35 [jim2]

there ya go :)

02:49:43 [denshi]

paje, literal attack

02:49:44 [paje]

denshi: attack =is= | <action> commandeers the Kentucky Navy and sinks Pittsburgh.

02:49:52 [denshi]

paje, attack is also | <action> commandeers the Kentucky Navy and sinks Pittsburgh.

02:49:52 [paje]

okay, denshi.

02:49:54 [denshi]

paje, attack

02:49:54 [paje]

denshi: huh?

02:49:55 [denshi]

paje, attack

02:49:56 [paje]

* paje commandeers the Kentucky Navy and sinks Pittsburgh.

02:50:00 [denshi]

huh

02:50:05 [denshi]

paje forget attack

02:50:05 [paje]

denshi: I forgot attack

02:50:12 [arjunsanyal]

what's up docwolf?

02:50:38 [denshi]

paje, attack is also | <reply> You can't fight in here! This is the War Room!

02:50:38 [paje]

OK, denshi.

02:51:12 [jim2]

sheesh :) /me thinks the war room should be the ONLY place fighting is allowed... :)

02:51:16 [denshi]

paje, attack

02:51:17 [paje]

You can't fight in here! This is the War Room!

02:51:31 [denshi]

jim2 is some kind of prevert.

02:51:36 [denshi]

paje, jim2?

02:51:36 [paje]

i heard jim2 was some kind of prevert.

02:51:47 [jim2]

I'm before a vert?

02:52:03 [denshi]

jim2 is also after our precious bodily fluids

02:52:03 [paje]

okay, denshi.

02:52:30 [jim2]

arjunsanyal: OK... so I've spun my twisty tale of campus stuff... what thinks you? :)

02:54:10 [arjunsanyal]

jim2: reading

02:54:53 [jim2]

(the final "problem" is still unsaid yet)

02:57:29 [arjunsanyal]

jim2: ok, see if i got this right, users use the calendar app to populate central user/class db, which is queried by "remote" scripts for access control..

02:59:00 [jim2]

arjunsanyal: well, could happen that way in a district lacking any computerized record keeping...

02:59:39 [jim2]

but in this particular situation, "enrolling" happens on a separate system, most of which I had no access to

03:01:29 [jim2]

so: enroll == transaction on mainframe, get CSV records happens as a consequence of a query to that mainframe... hence, the oracle (hopefully SOON or postgres) would keep records which are duplicates

03:03:43 [jim2]

I would imagine that many campuses are like this...

03:05:28 [jim2]

I can also envision situations where the district data center will entirely deny you access to records necessary to implement pre-access-setup on a campus or even district-wide basis

03:05:51 [jim2]

(which for you is the same as having no computerized records at all)

03:06:21 [arjunsanyal]

so the main goal of this system is pre-access-setup?

03:06:44 [jim2]

well, yeah... it makes the campus bigwigs look good

03:07:36 [jim2]

a student sits down, has his login and password (handed to him over the net somehow, maybe) and he can sit down and start working immediately)

03:07:46 [jim2]

that makes the student look good too

03:08:01 [jim2]

and the student is paying everyone's salary

03:08:42 [jim2]

that's probably the bottom line, monitarily...

03:09:18 [jim2]

(in many -normal- situations, anyway :) aduni considered abnormal)

03:10:20 [jim2]

also, in the strongest and best situations, student success -> teacher success

03:10:38 [markd2]

markd2 has left #openacs

03:11:14 [jim2]

NOT teacher has good rep -> student only successful if follows teacher jumping thru all hoops

03:12:05 [jim2]

in many cases, teachers and admin staff are at odds

03:12:15 [arjunsanyal]

how so?

03:12:19 [jim2]

admins think teachers uppity

03:12:30 [jim2]

teachers think admins are in their WAY.

03:13:13 [jim2]

sometimes teachers ARE uppity

03:13:22 [jim2]

sometimes admins ARE in teachers way

03:13:47 [jim2]

but each think that of the other most of the time, not just sometimes

03:14:36 [arjunsanyal]

so what was the ultimate problem that you mentioned?

03:14:46 [jim2]

it's not ultimate,

03:14:56 [jim2]

it's just the last problem I had to mention :)

03:15:07 [jim2]

I'm in my main calendar view

03:15:22 [jim2]

(which is a portal/portlet)

03:15:30 [jim2]

I can make it disappear

03:15:36 [jim2]

I cannot make it come back

03:15:57 [arjunsanyal]

that's a problem :)

03:16:26 [jim2]

(when I look at the page, what would be the full calendar appears as extra vertical white space, maybe a text line or two)

03:17:00 [jim2]

I tested that: removed the calendar view, there was less vertical space used where it would have appeared

03:17:15 [jim2]

put it back, the white space came back

03:17:32 [jim2]

(but no calendar)

03:17:34 [jim2]

also...

03:17:47 [jim2]

moved (missing) calendar view to another page

03:17:57 [jim2]

added extra white space to that page

03:18:00 [jim2]

removed...

03:18:06 [jim2]

white space went away

03:18:14 [jim2]

moved it back to calendar page...

03:18:28 [jim2]

white space came back to calendar page

03:18:56 [jim2]

so there's something I'm removing/replacing/moving...

03:19:07 [jim2]

it just doesn't look like a calendar :)

03:20:19 [jim2]

now, if you want an ultimate problem, here it is: there's no reference docs! I go to /doc, no dotlrn (or new-portal for that matter) package shows up on the right side

03:21:01 [arjunsanyal]

how did you make it disappear? through the "customize" page or by hitting the shade button?

03:21:26 [jim2]

(hence /packages/{new-portal,dotlrn*}/admin/doc doesn't exist, right?)

03:21:48 [arjunsanyal]

yes, there's nothing in there yet

03:21:50 [jim2]

I think in one case I shaded it, and in another I hit the x button

03:22:50 [jim2]

admittedly, this is my first day of looking at all of this... maybe I'm missing something obvious... but I can't find it yet?

03:26:45 [arjunsanyal]

nope, I just checked and you have found an bug i didn't know about. did you try to do the same thing with other portlets?

03:27:33 [jim2]

actually, just now did it

03:27:52 [jim2]

I added a community called comFOOnity

03:28:13 [jim2]

I turned out to be a member of comFOOnity...

03:28:18 [jim2]

so I went there

03:28:31 [jim2]

looked at its full calendar view

03:28:55 [jim2]

shaded it

03:29:21 [jim2]

it then redirected me to the main comFOOnity page

03:29:57 [jim2]

went back to comFOOnity's full calendar view

03:30:02 [jim2]

whitespace

03:30:50 [jim2]

(perhaps where it meant to display a window with the unshade button on it)

03:31:23 [arjunsanyal]

yup saw the same thing

03:31:59 [arjunsanyal]

right, it should show the "blank" portlet with the unshade button

03:32:34 [jim2]

interesting: here's the page source between the header and footer

03:32:36 [jim2]

<!-- Sloan Header End -->

03:32:36 [jim2]

<p></p>

03:32:36 [jim2]

<!-- A simple 1-column thingy. -->

03:32:36 [jim2]

<table border=0 width="100%">

03:32:36 [jim2]

<tr>

03:32:37 [jim2]

<td valign=top width="50%">

03:32:41 [jim2]

<br>

03:32:43 [jim2]

</td>

03:32:45 [jim2]

</tr>

03:32:47 [jim2]

</table>

03:32:49 [jim2]

<!-- Sloan Footer Begin -->

03:33:09 [arjunsanyal]

BTW, thanks, this is all very useful testing for us

03:33:58 [arjunsanyal]

yup, that's a 1-column layout template with nothing in it. and those are my comments :)

03:34:24 [jim2]

np :) and thank you, your presence and attention motivates me to test and report, given said testing and reporting will actually be used :)

03:34:55 [jim2]

* jim2 is burned out on not being listened to when using automatic bug mechanisms :)

03:35:55 [jim2]

are you working with Berklee?

03:36:04 [arjunsanyal]

yeah, they can be a drag, but nonetheless, we will have one (probably lars's) soon

03:36:33 [jim2]

I have no problem with that, as long as I get feedback

03:37:00 [jim2]

(and timely feedback that shows that my input was at least considered :)

03:37:50 [arjunsanyal]

about Berklee. me: no, OF: i don't know

03:38:47 [jim2]

reason I ask... is I'm a bassist, and I've done a LOT of tutoring in the areas of theory and harmony; I wrote a drill study that can be factored

03:39:15 [jim2]

in fact, I did said factoring, but not into separate packages

03:40:31 [jim2]

(doing so would be fairly trivial)

03:41:59 [jim2]

then you'd have single-string questions, and questions that have an ordered list of things as answer (where you'd click on list items to add them to the list)

03:42:25 [denshi]

denshi has quit (Remote closed the connection)

03:44:26 [arjunsanyal]

who are some of your fav bassists?

03:44:29 [jim2]

I have an example of that in what might be called production use: a key signature drill that has about 5 or 6 different question types

03:44:38 [jim2]

Best: Gary Willis

03:44:47 [arjunsanyal]

what band?

03:44:54 [jim2]

Jeff Berlin is OK too :)

03:45:04 [jim2]

Tribal Tech and his own project

03:45:43 [denshi]

denshi (~chatzilla@cs6625176-26.austin.rr.com) has joined #openacs

03:45:50 [arjunsanyal]

the most amazing bass performance i've seen is that guy from Level 42 - mark something?

03:45:56 [jim2]

a LOT of people are suggesting he's beating Jaco at his own game :)

03:46:25 [jim2]

The guy in Dave Weckl band is good

03:46:33 [jim2]

Tommy Kennedy

03:47:18 [jim2]

he has this funny trick; he plays this amazing line while moving his hand up and down the bass neck really fast

03:49:08 [jim2]

A local guy here is real good, his name is Mark VanWageningen

03:50:00 [jim2]

Darryl Jones

03:52:47 [jim2]

I just searched; the guy you're talking about is Mark King

03:52:54 [jim2]

never heard him play

03:53:59 [jim2]

looks like he has his own project

03:55:21 [jim2]

jim2 has changed the topic to: OpenACS: Free Web Toolkit <denshi> my god, it's full of geeksKing Group

03:55:29 [jim2]

oops

03:55:33 [arjunsanyal]

yup that's him - really amazing. Check out level 42, if you get the chance

03:55:44 [jim2]

jim2 has changed the topic to: OpenACS: Free Web Toolkit <denshi> my god, it's full of geeks

03:56:09 [jim2]

Mark King Group

03:57:34 [arjunsanyal]

don't know about his solo stuff, but i can't imagine his talent has declined any

03:57:44 [jim2]

what kind of stuff do they play?

03:59:09 [arjunsanyal]

the sound was fairly mainstream 80's pop, just with amazing musical talent - oh and the drummer is also legendary

04:00:42 [jim2]

they have a hell of a lot of gigs lined up

04:01:01 [arjunsanyal]

the mark king group?

04:01:11 [jim2]

umm

04:01:13 [jim2]

not sure

04:03:40 [jim2]

looks like at least some of the gigs are level 42 gigs

04:05:02 [davb]

wow, jim2, you have been very busy

04:05:25 [jim2]

I have?? :)

04:05:40 [jim2]

sheesh :) and I just finished my coffee!

04:05:43 [davb]

ok, you have been typing alot of stuff into the channel

04:05:49 [jim2]

heh :)

04:06:10 [jim2]

yes, that's a bit unusual/sporadic lately :)

04:06:51 [jim2]

davb: located a few dotlrn issues they're gonna take a look at

04:07:03 [jim2]

maybe one in new-portal

04:07:13 [davb]

ah. very cool.

04:08:01 [jim2]

also, he fixed a few things in one shot, couldn't add courses or something like that, commented out some code, bingo... all works now

04:09:12 [jim2]

arjunsanyal: now about this tree structure of courses you have... different campuses might want to do other things with them, and some might want to add concepts... consider this:

04:10:05 [jim2]

in this district, they have campuses, courses, sections of courses, and sessions of sections

04:10:59 [jim2]

some campuses might want to keep track of each session separately, to make the calendar work correctly:

04:11:54 [jim2]

- you might have classes part of a particular term that only go half way

04:12:15 [jim2]

- or perhaps some course has only -one- meeding

04:12:34 [jim2]

+ testing and evaluation

04:12:46 [arjunsanyal]

one meeting? sign me up, let's get this over in one shot!

04:13:37 [jim2]

+ orientation meeting of a course that otherwise takes place as some form of home study with credit

04:14:15 [jim2]

+ small subject, only needs one meeting

04:15:30 [jim2]

- there are school holidays, so no sessions take place then (or maybe a few do; depends on situation)

04:16:09 [jim2]

so it might be nice for an -accurate- school calendar to have records of separate sessions

04:16:17 [jim2]

also...

04:16:39 [jim2]

there are different types of sessions... lecture and lab come to mind right away

04:17:15 [jim2]

the mainframe has its idea of what kind of sessions happen when and where

04:18:41 [jim2]

"where" brings up another topic...

04:18:58 [jim2]

you might incorporate a room reservation system

04:19:53 [the_docwolf]

the_docwolf has left #openacs

04:19:55 [jim2]

which also implies the need to give the server a list of rooms that are on the campuses

04:20:19 [jim2]

(another possible bulk-upload task)

04:21:42 [jim2]

and finally, a list of servers, together with their server type, and a list of courses that would have students added to those servers

04:22:30 [jim2]

(the way I did it, is if the course list was empty for a particular server, then add students from all courses to that server)

04:24:57 [arjunsanyal]

jim2: I don't want to be a wet blanket, but dotlrn is more focused on web communities than the actual task of managing a large educational institution, so while the ideas you suggest are great, they might fall out of the scope of dotlrn

04:25:55 [jim2]

ok, then I didn't understand until now what dotlrn was for :)

04:26:43 [jim2]

so if the community is the focus, is a "community" a "subsite"?

04:26:59 [davb]

so if a professor wanted a community for a class, he could set it up?

04:27:15 [davb]

as opposed to managing the whole campus from dotlrn?

04:28:12 [jim2]

right, and a professor should then be allowed to mount apps on communities he creates

04:28:16 [arjunsanyal]

not a problem, in general, i would say that if a problem makes you think "hey, openacs (or a module), would be great for that", dotlrn would probably be good for that thing too in the context of education

04:29:09 [davb]

so dotlrn is more a tool for a community or class to use?

04:29:42 [davb]

jim: i am getting intermedia errors on an openacs oracle install, but it seemed to work fine anyway. did you notice any?

04:30:23 [jim2]

take a look at http://12-233-225-152.client.attbi.com/key-signatures

04:31:12 [jim2]

davb: I'm not yet familiar with intermedia,,, but I didn't notice any problems

04:31:39 [davb]

ok. it was on the package loading screen. i'll ignore it for now unless it causes a problem :)

04:31:50 [jim2]

which package?

04:32:02 [arjunsanyal]

davb: yes, have you looked at http://helice.mit.edu:8080

04:32:46 [davb]

content repository

04:32:46 [paje]

content repository is especially for content, keeping it in the content-item tables.

04:32:49 [davb]

heh

04:32:53 [davb]

not yet, i'll look.

04:33:31 [jim2]

"Thank you. Your registration request has been sent to the System Administrator."

04:33:40 [davb]

arjunsanyal: neat. I setup dotlrn back when it was first released. It has alot of interesting potential, and I love new-portal.

04:34:07 [davb]

looks like me and rbm will have to hack aolserver ourselves if we want it to behave nicely with tcl.

04:34:46 [arjunsanyal]

jim2: you are now a full external dotlrn user

04:34:56 [jim2]

arjunsanyal: thanks :)

04:35:51 [jim2]

oop :) it did the same thing as mine :)

04:36:10 [jim2]

(the disappearing calendar trick :)

04:36:31 [davb]

ok, see you later, it's late over here.

04:36:47 [arjunsanyal]

davb: thanks, all the bugs in new-portal are my fault. see you later

04:36:56 [davb]

:)

04:36:59 [davb]

davb has quit ("Client Exiting")

04:38:03 [arjunsanyal]

jim2: yup works there too. I'll be looking into it tomorrow. I must be going too. goodnight all

04:39:56 [arjunsanyal]

arjunsanyal has quit ("Download Gaim [http://gaim.sourceforge.net/]")

04:48:43 [jim2]

dang, found something else

04:49:14 [jim2]

ok, I'm idle for awhile

07:09:00 [jim2]

jim2 has quit ("[x]chat")

09:16:14 [hazmat]

hazmat has quit (Connection timed out)

09:33:01 [denshi]

denshi has quit (Read error: 110 (Connection timed out))

11:10:50 [denshi]

denshi (~chatzilla@cs6625176-26.austin.rr.com) has joined #openacs

13:26:38 [davb]

davb (~chatzilla@rrcs-nys-24-97-22-203.biz.rr.com) has joined #openacs

13:26:48 [davb]

hi there

13:29:31 [denshi]

morning davb

13:38:04 [docwolf]

docwolf (~wolf@adsl-34-227-234.bct.bellsouth.net) has joined #openacs

13:38:13 [denshi]

paje, attack

13:38:13 [paje]

You can't fight in here! This is the War Room!

13:39:42 [denshi]

docwolf, seen rolf lately?

13:40:38 [docwolf]

hi

13:40:46 [docwolf]

rolf is asleep, as best i can determine

13:41:26 [denshi]

then gotham lies undefended....

13:41:53 [denshi]

okay. he was talking up some formbuilder macros he said he wrote

13:43:06 [docwolf]

hmm

13:43:09 [docwolf]

that i don't know about

13:43:18 [docwolf]

when he wakes up in 3-4 hours, i'll ask him

13:56:33 [jim2]

jim2 (~jim@12.240.101.86) has joined #openacs

14:11:53 [cro]

cro (~cro@defiant.nca.asu.edu) has joined #openacs

14:25:20 [denshi_]

denshi_ (~chatzilla@cs6625176-26.austin.rr.com) has joined #openacs

15:23:44 [beezo]

hey guys

15:23:50 [docwolf]

beezo?

15:24:06 [denshi_]

paje, attack beezo

15:24:06 [paje]

denshi_: huh?

15:25:39 [beezo]

paje, tell denshi_ his shorts are riding a bit too high for the rest of us

15:25:40 [paje]

OK, beezo.

15:26:10 [beezo]

denshi_: have you tried this gentoo distro?

15:26:19 [denshi_]

beezo, I'm giving an apache2 lecture at the local LUG tonight

15:26:29 [beezo]

wow!

15:26:29 [denshi_]

as for the gentoo, no, but everyone says I should

15:26:30 [beezo]

killer

15:26:36 [beezo]

it's *really* nice

15:26:43 [beezo]

at least the portage system is really nice

15:26:53 [denshi_]

two wins w/ the lecture: new contacts, and oacs tie-in

15:27:02 [beezo]

what the BSD people have been bragging about

15:27:05 [beezo]

yeah, that's really really good

15:27:06 [denshi_]

here's a question, though

15:27:15 [beezo]

how big is the LUG?

15:27:41 [denshi_]

they've seen a couple hundred people show up for voting before, but only about 40 people show for meetings

15:27:46 [denshi_]

so I'll say 40

15:27:49 [beezo]

40 people is *huge*

15:28:11 [beezo]

for groups like that, if you get 10 active members you're a strong club

15:28:15 [beezo]

40 people is very big

15:28:18 [denshi_]

well, it's texas. nerds need to stick together.

15:28:27 [beezo]

what's your question?

15:28:31 [beezo]

haha

15:28:54 [denshi_]

well, I can get up and talk about context switching costs and how fast threaded servers are, blah, blah, blah

15:29:20 [beezo]

kind of like prairie dogs? keep a look out for predators, or in your case, gun-toting cowboy fans?

15:29:27 [denshi_]

but I'd rather talk up the new things you can do with a threaded server: db pooling, memoized fxns, caching, etc

15:29:49 [denshi_]

the "etc" is the question. what features are in OACS that couldn't be there without aolserver's design?

15:29:56 [beezo]

do you think anyone there will be familiar with threaded servers?

15:30:08 [docwolf]

denshi: how literate is your audience?

15:30:11 [beezo]

like, i saw that websphere has pooled JDBC connections

15:30:18 [docwolf]

i mean, from my perspective, all i'd really want to know

15:30:22 [beezo]

but websphere is apache with some fixin'

15:30:25 [docwolf]

is why i should use 2.0 vs. 1.3

15:30:33 [docwolf]

concretely, what the benefits will be.

15:30:45 [beezo]

docwolf's got a good point

15:30:50 [docwolf]

and what disasters can occur when moving from 1.3 to 2.0

15:30:51 [denshi_]

he does

15:30:59 [docwolf]

(this is not a theoretical problem for us, btw...)

15:31:01 [beezo]

unless you're dealing with CS people, that's all anyone will want to know

15:31:28 [denshi_]

that's a good point, b/c personally I don't really care about other people's servers. i just want to show off the new toys

15:31:28 [beezo]

i was pretty impressed when i saw that websphere has db pooling, btw

15:31:44 [docwolf]

i would take a very business-like approach

15:31:53 [docwolf]

assume everyone has some familiarity with apache 1.3

15:31:58 [beezo]

i heard a lot of people say websphere is the best java server out there, but i figured it was just better marketing

15:32:00 [docwolf]

and explain why 2.0 is nothing like it, and why it's so much better

15:32:03 [rbm]

ooom

15:32:06 [beezo]

maybe they actually did put some nice things in there

15:32:09 [beezo]

hey rbm

15:32:10 [docwolf]

moo to you, rbm

15:32:22 [davb]

hi rbm

15:32:27 [rbm]

hi all

15:32:42 [denshi_]

rbm, GRE scores come back yet?

15:33:13 [rbm]

denshi_: Yes. I had my scores right after I finished the test (it's computer-based only)

15:34:07 [denshi_]

hmm. maybe by the time I take them I'll be able to get my scores, acceptance, and scholarship all immediately after finishing.

15:34:20 [denshi_]

maybe I should build such a machine.

15:34:42 [denshi_]

it'll look like a big ATM, and be capable of dispensing diplomas.

15:34:49 [rbm]

denshi_: :)

15:35:27 [denshi_]

docwolf, my 4 plank demo is: changes, new architecture, things that broke, new toys

15:35:47 [docwolf]

cool!

15:36:57 [beezo]

denshi_: you may want to add an overview of apache1.3

15:37:04 [beezo]

to give context

15:37:18 [beezo]

also, are you going to have slides or an outline of this? i would love to see those

15:37:36 [denshi_]

I'll make some simple slides

15:37:40 [denshi_]

I hate making slides.

15:37:49 [denshi_]

I'd rather just jot on the whiteboard

15:37:58 [denshi_]

but my drawing sucks, too

15:40:48 [denshi_]

http://www.norvig.com/Gettysburg/

15:40:48 [oacs-chump]

A: http://www.norvig.com/Gettysburg/ from denshi_

15:41:01 [denshi_]

A: If Abraham Lincoln had PowerPoint

15:41:03 [oacs-chump]

added comment A1

16:25:10 [beezo]

beezo has quit ("Client Exiting")

16:26:23 [beezo]

beezo (~talli@pool-162-83-236-44.ny5030.east.verizon.net) has joined #openacs

17:35:05 [denshi_]

denshi_ has quit ()

17:35:41 [denshi]

denshi has quit (Read error: 104 (Connection reset by peer))

18:24:15 [davb]

jim or jim2 around?

18:33:27 [markd2]

markd2 (~Snak@h166-102-041-128.ip.alltel.net) has joined #openacs

18:48:42 [hazmat]

hazmat (~ender@adsl-66-123-57-58.dsl.lsan03.pacbell.net) has joined #openacs

18:48:48 [hazmat]

join #codeit

18:52:19 [denshi]

denshi (~chatzilla@cs6625176-26.austin.rr.com) has joined #openacs

18:52:27 [denshi]

oooom

19:01:33 [markd2]

hmm.. looks like not much happening on #codeit

19:01:56 [denshi]

what is #codeit for?

19:02:01 [markd2]

dunno

19:02:08 [markd2]

maybe something zopelike

19:06:16 [talli]

talli has quit (Read error: 110 (Connection timed out))

19:12:45 [davb]

denshi: http://radio.weblogs.com/0100136/ another lisp fan'sweblog, interesting stuff (for met anyway :)

19:12:46 [docwolf]

docwolf has quit (Read error: 104 (Connection reset by peer))

19:24:03 [talli]

talli (~talli@xd84b5c59.ip.ggn.net) has joined #openacs

19:26:52 [davb]

hi talli

19:28:17 [davb]

whats the syntax to checkout one file from cvs?

19:29:39 [rbm]

can you checkout just one file?

19:29:50 [davb]

or update rather...

19:29:51 [rbm]

www.refcards.com has a handy cheat sheet

19:29:56 [davb]

thanks

19:31:30 [markd2]

cvs update file-name

19:32:32 [davb]

ok. didn't seem to work, but I am sure I screwed it up :)

19:33:20 [markd2]

heh

19:33:25 [markd2]

get any kind of error back?

19:34:03 [davb]

probably, but I ignored it :)

19:34:10 [davb]

I re checked out the whole thing anyway.

19:34:36 [davb]

which i guess is the equivilant or reinstalling an OS to fix an application error...

19:34:45 [talli]

hey davb

19:34:59 [davb]

Matrix trailer is online...

19:35:06 [davb]

http://whatisthematrix.warnerbros.com/rl_cmp/trailer_qt.html

19:35:07 [oacs-chump]

B: http://whatisthematrix.warnerbros.com/rl_cmp/trailer_qt.html from davb

19:37:47 [davb]

B: quicktime only

19:37:47 [oacs-chump]

added comment B1

19:38:47 [cro]

they showed the Matrix trailer as a preview before Attack of the Clones

19:39:51 [davb]

cool.

19:39:56 [davb]

* davb has no speakers

19:39:57 [davb]

argh

19:40:35 [markd2]

time to start saying how cool the trailer dialog is

19:44:35 [talli]

is attack of the clones any good?

19:44:42 [markd2]

yes

19:44:55 [markd2]

not nearly as painful as EpI

19:45:11 [markd2]

some of the dialog is atrocious

19:45:21 [markd2]

but it isn't nearly as horrific as much of the little-kid scenes in EpI

19:45:31 [talli]

any naked people?

19:46:02 [markd2]

oh yeah. 3P0 and R2 get it on

19:46:25 [markd2]

it's amazing the 'devices' that R2 is carrying

19:47:53 [markd2]

markd2 has quit ("wheeeee")

19:50:05 [cro]

The Yoda-fight scene is worth the cost of admission.

19:53:20 [cro]

cro has left #openacs

20:00:04 [davb]

i was unable to download the trailer because it was taking forever!

20:27:22 [davb]

bbl

20:27:24 [davb]

davb has quit ("ChatZilla 0.8.7 [Mozilla rv:1.0rc2/20020510]")

20:28:14 [denshi]

paje, jim2?

20:28:14 [paje]

jim2 is some kind of prevert. or after our precious bodily fluids

20:29:18 [rbm]

ep2 is much better than ep1, but still leaves something to be desired IMHS

20:38:28 [denshi]

hey rbm, go see the Royal Tenenbaums

20:38:54 [denshi]

and what does IMHS stand for?

20:40:47 [rbm]

sorry, s/IMHS/IMHO/ :)

21:17:13 [rbm]

So how do you guys feel about forming a task force to fix AOLserver+Tcl integration issues?

21:17:20 [rbm]

I mean _really_ fix it.

21:17:34 [rbm]

Not being able to gracefully load Tcl modules into AOLserver is a huge lose for all of us, and frankly, I don't think the DCI team cares.

21:29:43 [denshi]

tell me again who the DCI team is

21:29:56 [denshi]

and what's wrong with AOLserver + Tcl integration now?

21:30:23 [rbm]

denshi: Digital City Inc IIRC

21:30:30 [rbm]

denshi: Are you on the AOLserver list?

21:30:52 [denshi]

you know, I tried several times and was defeated by their poor listserv code

21:31:16 [denshi]

I was also thwarted in my attempts to read the archives

21:31:20 [rbm]

denshi: I can forward you an e-mail that talks about it. address?

21:31:28 [denshi]

denshi@pobox.com

21:32:16 [rbm]

denshi: bounced it your way

21:34:14 [rbm]

denshi: here's what I got:

21:34:15 [rbm]

----- The following addresses had permanent fatal errors -----

21:34:15 [rbm]

<toddg@mail.ma.utexas.edu>

21:34:16 [rbm]

21:34:17 [rbm]

----- Transcript of session follows -----

21:34:19 [rbm]

554 5.0.0 Too many hops 11 (10 max): from <roberto@brasileiro.net> via wormwood.pobox.com, to <toddg@mail.ma.utexas.edu>

22:19:59 [davb]

davb (dave@alb-24-58-162-82.nycap.rr.com) has joined #openacs

22:23:05 [davb]

hi there

22:27:58 [beezo]

hey davb

22:30:07 [denshi]

rbm, still here?

22:30:11 [denshi]

paje, rbm?

22:30:12 [paje]

rbm is a CS messiah who will crush all the EJB posers and their kindred and lead us all to a new era of coder enlightenment.

22:30:58 [rbm]

denshi: yes

22:31:13 [rbm]

hehehe. Didn't know paje had that definition of me :)

22:31:45 [denshi]

read the email

22:31:53 [denshi]

do you know why the problem is what it is?

22:32:46 [rbm]

not in details. But I know that AOLserver loads Tcl modules the "wrong" way, which messes up how modules are expecting to find the environment to be loaded.

22:34:05 [denshi]

well, I just read the 3 ways aolserver has instantiated interps in the past, so it sounds like b/c the way they clone interps in 3 is the problem

22:36:41 [rbm]

yep

22:56:43 [denshi]

but I don't understand what was wrong with the aolserver1 approach: init a pool of interpreters

22:57:17 [denshi]

the major diff seems to be that aolserver3 has a 'copy global symbols' operator on interp cloning.

22:57:32 [denshi]

but you still need the nsv_ commands to alter the globals

22:57:34 [rbm]

it's not just the issue of the number of intersp, but how they are started

22:57:48 [denshi]

that's what I said

22:57:56 [denshi]

the enumeration of them is irrelevant

22:58:50 [denshi]

the diff is between their init routines, but nsd1 & nsd3 are much more similar to each other than nsd2.

22:59:18 [denshi]

so if nsd1 is at least theoretically capable of loading tcl extensions, then shouldn't we be able to make nsd3 work?

22:59:39 [rbm]

I never used nsd1 to know if it was capable or not.

22:59:46 [rbm]

I used AS 2 and 3

23:00:29 [rbm]

how would you know how nsd1 loaded interps? the source for it was never released (nor for nsd2 for that matter)

23:18:23 [davb]

hi

23:18:23 [paje]

hi, davb

23:19:33 [davb]

i have also been thinking about this.

23:19:58 [davb]

the interesting part is that adp pages have some global variables set that tcl pages do not.

23:22:06 [denshi]

rbm: read Donaldson's slides on the Digital City arch

23:22:19 [rbm]

denshi: You mean Davidson's? :)

23:22:26 [denshi]

he describes the 3 models of interp allocation nsd has tried

23:22:30 [davb]

ok the one this is tcl_version is set in an adp page.

23:22:34 [denshi]

nsd1 sounds like how I'm doing mod_scheme

23:22:50 [rbm]

I just think that we should really tackle the problem. I don't think DCI will ever do it.

23:23:04 [davb]

sure, but will they accept a patch?

23:23:08 [rbm]

davb: Yes.

23:23:15 [davb]

ok, cool.

23:23:43 [rbm]

Did you read the "thread allocator" thread which I referred to in my post? Kriston asked Zoran to post his patch to sourceforge

23:23:48 [davb]

so just to clarify, this is the whole issue of tcl namespace, package require, etc...?

23:24:02 [rbm]

davb: Yes, but the issue is deeper than you thought.

23:25:36 [davb]

it is in how the tcl interpreters are created.

23:28:38 [rbm]

bbiab