2008/05/13
IRC [01:31] *** hgcphoeni1 joined the chat.
IRC [01:47] *** hgcphoenix parted the chat.
IRC [02:56] *** cacru1 parted the chat.
IRC [02:56] *** cacru1 joined the chat.
IRC [02:57] *** cacru1 parted the chat.
IRC [02:57] *** cacrus joined the chat.
IRC [10:58] *** holycow joined the chat.
IRC [11:15] *** hgcphoeni1 parted the chat.
IRC [12:24] *** KNRe1 joined the chat.
IRC [14:06] <Dossy> Hi, Rick.
IRC [14:58] *** partymola joined the chat.
IRC [15:46] *** partymola parted the chat.
IRC [15:46] *** partymol2 joined the chat.
IRC [15:46] *** partymola joined the chat.
IRC [15:46] *** partymol2 parted the chat.
IRC [15:52] <KNRe1> hi, Dossy. Just got back from lunch
IRC [15:52] <KNRe1> Thanks for your generosity about letting me put the modules up.
IRC [15:53] <partymola> Dossy is the best :-D
IRC [15:53] <KNRe1> Hi, partymola; what's your handle on the email list? I'm rcobb@knownow.com (well, for another few days)
IRC [15:54] <partymola> juanjose@simpleoption.com
IRC [15:54] <partymola> hello rcobb, nice to meet you )
IRC [15:54] <partymola> :)
IRC [15:54] <KNRe1> Nice to meet you. You'll probably start seeing me post from my other id soon (rick_cobb@ieee.org).
IRC [15:54] <partymola> oh, an IEEE zealot lol
IRC [15:55] <KNRe1> Nah, just a member. I like their email forwarder -- I can change providers whenever I want.
IRC [15:55] <partymola> i was kidding :-)
IRC [15:56] <KNRe1> I hope I've had my IEEE argument back when we were arguing about whether there should be a 1003.3 working group (that'd be around '85, IIRC).
IRC [15:56] <KNRe1> Sorry: my _last_ IEEE argument. I prefer to argue with the IETF :-)
IRC [15:57] <KNRe1> I'm about to drop off, though; just wanted to see if there was an active conversation. I'll be back in an hour or so.
IRC [15:58] *** EvilOtto joined the chat.
IRC [16:00] <EvilOtto> bah, irc has no /review command ...
IRC [16:03] <Dossy> review?
IRC [16:03] <Dossy> yeah. your best bet is the chat logs.
IRC [16:04] <Dossy> Regarding my generosity - I don't know where people get the impression that I'm not encouraging contributions...
IRC [16:04] <EvilOtto> lily, another cmc I'm fond of has a built-in /review command that lets you see what has happened while you were gone.
IRC [16:04] <Dossy> KNRe1: yeah, getting cobb.com wasn't an option, huh :)
IRC [16:04] <Dossy> EvilOtto: yeah ... for large IRC networks, storing that data on the server would be impractical ...
IRC [16:05] <Dossy> I use dircproxy, which stays connected even when my client isn't online - and that is configured to log as well
IRC [16:07] <EvilOtto> lily is a workgroup-sized cmc. Or village sized, as one of the creators put it. It's fantastic for up to a few hundred users.
IRC [16:10] <EvilOtto> I figured out what was causing my first connection to hang using cvs head - the latest pools code does that if minpools == 0.
IRC [16:40] <Dossy> strange.
IRC [16:40] <Dossy> yeah, that code needs to be seriously exercised and all the edge case bugs ironed out.
IRC [16:47] <KNRe1> Never thought you weren't encouraging contributions, Dossy. It's going to be fun to get that stuff up there.
IRC [16:49] <EvilOtto> A written description of the intended pools behavior would be nice too. doc/ appears to be for manpages, wasn't there a place for operation descriptions (like the urlspace trie) at one time?
IRC [17:00] <CIA-51> 03dvrsn * 10aolserver/configure: configure fix to allow building with tcl8.5
IRC [17:13] <Dossy> EvilOtto: Eek, don't edit configure directly. :P
IRC [17:13] <Dossy> It gets autogenerated from configure.in :)
IRC [17:13] <Dossy> and autoconf.
IRC [17:14] <EvilOtto> I was afraid of that... Im not an autoconf guru and aclocal seems to want some other setup to run cleanly
IRC [17:14] <Dossy> Really? Strange.
IRC [17:14] <Dossy> Ignore aclocal "errors". :)
IRC [17:15] <Dossy> You should just be able to run "autoconf" in the top-level AOLserver source dir.
IRC [17:16] <EvilOtto> it complains with "aclocal.m4:39: /usr/bin/m4: Cannot open ../tcl/unix/tcl.m4: No such file or directory"
IRC [17:17] <EvilOtto> which is why I checked in the configure change directly...
IRC [17:19] <EvilOtto> ick. Ok, I symlinked tc.8.5.2 to tcl to get that path to work and ran autoconf, and it changes quite alot in configure...
IRC [17:19] <Dossy> :)
IRC [17:19] <Dossy> yup.
IRC [17:20] <Dossy> What version of autoconf are you using?
IRC [17:20] <Dossy> Go ahead and check it into HEAD. Someone else will complain if it breaks something. Ideally, we shouldn't be committing "configure" - folks should build it locally using autoconf.
IRC [17:20] <Dossy> The argument of "what if they don't have autoconf installed" is pretty lame, today.
IRC [17:21] <EvilOtto> looks like autoconf 2.57
IRC [17:21] <Dossy> Yeah, I have autoconf 2.13 and 2.61.
IRC [17:21] <Dossy> I think the old configure script was generated with autoconf 2.1x
IRC [17:22] <EvilOtto> I guess hits is the difference between cvs head and a release ... the release should definitely have configure. but then we get back to why we can't jsut run ./configure...
IRC [17:23] <EvilOtto> I also don't have anything other than linux handy to check that change on
IRC [17:24] <EvilOtto> the old one was generated by autoconf 2.59, accordinf to it
IRC [17:24] <Dossy> oh? ok - then generating with autoconf >= 2.50 is ok then :)
IRC [17:25] <EvilOtto> heh
IRC [17:25] <CIA-51> 03dvrsn * 10aolserver/ (configure configure.in): changed configure.in, regenerated configure
IRC [17:25] <Dossy> thanks, jeff :)
IRC [17:25] <EvilOtto> :)
IRC [17:26] <Dossy> I'm going to cvs up and test over here.
IRC [17:26] <Dossy> Are you going to update ChangeLog and commit it too?
IRC [17:27] <EvilOtto> sure. The naviserver list had been gently encouraging committers to do that too.
IRC [17:28] <Dossy> yeah. using cvs2cl is okay, as long as you don't use it to regenerate the whole changelog - just your most recent commit.
IRC [17:28] <Dossy> But, yeah - it's pretty important/useful to maintain a ChangeLog
IRC [17:29] <CIA-51> 03dvrsn * 10nssqlite3/nssqlite3.c:
IRC [17:29] <CIA-51> fixed a bug where statement cancellation before any query is run
IRC [17:29] <CIA-51> would crash the server when a Ns_DbHandle was not zeroed out when
IRC [17:29] <CIA-51> allocated (which it isn't guaranteed to be, since it is malloced
IRC [17:29] <CIA-51> not calloced).
IRC [17:29] <Dossy> Woo! :)
IRC [17:30] <EvilOtto> I see that everyone else has been updatign the changelog too
IRC [17:31] <EvilOtto> last entry is just over a year ago
IRC [17:31] <Dossy> Cool, CVS HEAD still builds here.
IRC [17:31] <Dossy> lol :)
IRC [17:31] <Dossy> Lets not go there. :P
IRC [17:34] <CIA-51> 03dvrsn * 10nssqlite3/ChangeLog: updated changelogs
IRC [17:34] <CIA-51> 03dvrsn * 10aolserver/ChangeLog: updated changelogs
IRC [17:42] <Dossy> Think we're brave enough to merge HEAD down to aolserver_v45_bp and tag 4.5.1?
IRC [17:44] <partymola> Dossy, let me compile it on BSD :)
IRC [17:47] <partymola> it compiles :)
IRC [17:55] <Dossy> partymola: woo :)
IRC [17:55] <Dossy> yeah, I should set up an AOLserver buildbot farm
IRC [17:57] <partymola> Dossy, it compiled against tcl8.5 perfectly
IRC [17:57] <partymola> and usually i had to compile some files manually setting -nostartfiles, etc.
IRC [17:57] <partymola> but this time it worked perfectly at the first try
IRC [17:57] <partymola> so it's a very big woooo :D
IRC [18:04] <Dossy> Sweet.
IRC [19:06] <Dossy> EvilOtto: Otherwise, have you had any issues or trouble with nssqlite3?
IRC [19:08] <EvilOtto> yes, one other: you cannot properly run a select that returns no rows. The driver interprets that result as a dml operation, so iterating over the (0) rows, or 0or1row just throws an error.
IRC [19:13] <EvilOtto> when looking into this I did learn that a db driver cannot have both a DbFn_Select and a generic DbFn_Exec (or a DML abd Exec) because the generic Exec function is checked for first. This seems backward to me (e.g., a driver might want to ensure that select is only select at the parsing stage), but at least one other person (forget who) thought that having separate select/dml/exec is...
IRC [19:13] <EvilOtto> ...redundant (including the higher-level interfaces like ns_db 0or1row and ns_db select) and everything should be just exec
IRC [20:09] <KNRe1> Hey, Dossy, if you're still there... my sourceforge ID is "cobbr2". I definitely need to be added as a developer/committer. Thanks.
IRC [20:11] <Dossy> OK.
IRC [20:12] <Dossy> yeah
IRC [20:12] <Dossy> I don't use anything but ns_db exec, personally.
IRC [20:12] <Dossy> Poor Rick - was there already someone "cobbr" at SourceForge? :P
IRC [20:13] <Dossy> OK, Rick - I've added "cobbr2" to the SF project w/ CVS access.
IRC [20:14] <KNRe1> Thanks. No, actually, it's my oldest handle. In high school, in '73, we were allowed to create 10 files on the DECtape for our PDP-8...
IRC [20:14] <KNRe1> first four letters of your last name, first letter of your first, and one digit.
IRC [20:15] <KNRe1> I'm a Jr., so... cobbr2 was the handle. And that was where my little adaptive system written in BASIC for predicting football games lived :-)
IRC [20:16] <Dossy> Ah, OK - cool!
IRC [20:16] <partymola> KNRe1, you're talktive, eh!
IRC [20:16] <partymola> well, good night everyone
IRC [20:16] <Dossy> Good night, partymola.
IRC [20:39] *** partymola parted the chat.
IRC [21:11] *** hgcphoenix joined the chat.
IRC [21:29] *** KNRe1 parted the chat.
IRC [22:16] <Dossy> Ugh. Debian OpenSSL security bug announcement is sad. :-(