[00:12:55] ** _Maniac has left IRC ("Leaving")
[02:31:00] ** gpciceri has joined us
[03:59:42] ** gpciceri has left IRC (Read error: 110 (Connection timed out))
[06:35:44] ** jack-e has joined us
[06:38:06] <jack-e> morning
[08:42:03] ** gbay has joined us
[09:02:21] ** gbay has left IRC ("It's not like I'm leaving or anything...")
[09:55:38] ** hazmat has joined us
[10:29:09] ** hazmat has left IRC (Read error: 110 (Connection timed out))
[10:43:35] ** sremington has joined us
[11:10:55] ** vlado has joined us
[11:20:16] ** vlado has left IRC ("Leaving")
[12:31:26] ** hazmat has joined us
[13:08:47] ** hazmat has left IRC (Read error: 110 (Connection timed out))
[13:55:24] ** _jpl_ has joined us
[13:55:30] <_jpl_> Hi all
[13:57:46] ** gpciceri has joined us
[13:57:59] ** gpciceri has left IRC (Read error: 104 (Connection reset by peer))
[15:18:08] ** hazmat has joined us
[15:39:13] ** redLED has joined us
[15:39:45] <Maniac> hello redLED
[15:39:50] <redLED> hi :)
[15:40:01] <Maniac> welcome to #PEAK
[15:41:36] <redLED> thanks :)
[15:44:47] * redLED poits, is there no precompiled install package for win32 py2.3 available, or did i manage to carelessly miss it?
[15:47:22] ** gbay has joined us
[15:49:11] ** hazmat has left IRC (Read error: 60 (Operation timed out))
[15:49:39] <redLED> since i'm using activestate activepython (2.3) under windows, and if there's no 2.3 precompiled, that renders me in need of MS visual studio :(
[15:49:55] <sremington> redLED: there doesn't appear to be a 2.3 exe
[15:58:32] <Maniac> eh?
[15:58:48] <Maniac> oh, only a win32-py2.2
[15:58:53] <Maniac> hmm
[15:59:16] <Maniac> redLED, apparently you don't need MS visual studio
[15:59:35] <Maniac> http://randomthoughts.vandorp.ca/syncato/WK/blog/758?t=item
[16:00:02] <Maniac> if your' willing to use cygwin or mingw
[16:00:44] <redLED> that's quite allright, i do have visual studio 6 (with which activepython is being built stock) somewhere around here, and im quite sure it'll preform better than cygwin and mingw...
[16:00:56] <redLED> so i'll go do python setup.py bdist_wininst, upload it somewhere, and we'll pick up from there :)
[16:02:17] <redLED> PyProtocols seems to be lacking a windowish 2.3 compile as well, so might just kill two birds in one run while i'm at it.
[16:08:16] <redLED> i'll be righttt back.
[16:08:19] ** redLED has left IRC ("Leaving")
[16:17:55] ** gbay has left IRC ("It's not like I'm leaving or anything...")
[16:18:22] ** redLED has joined us
[16:18:37] <redLED> There we go.
[16:18:49] <redLED> available at http://futureinquestion.net/resources/dist/ in case any lost soul besides myself ever comes around asking
[16:35:55] <_jpl_> Indeed, others are going to appreciate a Win32 binary package for PEAK.
[16:36:15] <_jpl_> I've been needing one myself and haven't had the time so put one together, so let me be the first to offer thanks. :)
[16:49:01] <Maniac> redLED, you are a gentleman
[16:49:44] <Maniac> should we poke PJE and get python 2.3 versions up there, especially w.r.t pyprotocols ?
[16:52:53] <_jpl_> pje only supports 2.2 for the time being.
[16:53:24] <etrepum> cause he's CRAZY ;)
[16:53:53] <etrepum> some of the 2.2-bugfixes he has in PEAK are horrible
[16:55:22] <redLED> i would'nt be so harsh :) it makes sense from an adaptation perspective, 2.2 is what comes with most linux distros out of the box
[16:55:25] <redLED> even fedora core 1 is 2.2.3
[16:55:58] <_jpl_> What are some of the 2.2-bugfixes?
[16:56:04] <etrepum> yes, but you would have to pay me quite a bit to care about 2.2
[16:56:18] <etrepum> I certainly wouldn't do it for fun
[16:57:51] <etrepum> I don't remember exactly where they are, but I remember seeing some awful stuff
[16:58:12] <etrepum> I haven't used 2.2 in at least 6 months, so I sort of just gloss over this stuff
[16:58:28] <Maniac> etrepum, i dont' think it's a crazy thing, i think it's a work factor beyond his control at this time thing :)
[16:58:46] <etrepum> it is a crazy thing, and it sucks that he would have to do it for work :)
[16:59:16] <redLED> thats what work is all about :) people giving you $'s to do all the nasty stuff you would never consider doing in your spare time =D
[16:59:29] <etrepum> it doesn't have to be that way
[17:00:10] <etrepum> or rather, not everyone is willing to accept that ;)
[17:01:31] <redLED> acceptance is just a matter of taking a good look at the pile of bills your mailbox manages to collect over a given time period
[17:01:46] ** gpciceri has joined us
[17:02:00] ** gpciceri has left IRC (Client Quit)
[17:02:01] <etrepum> well, I quit my job, and now I'm taking on freelance projects, but only the ones that interest me
[17:02:48] <etrepum> no problems yet
[17:04:14] <redLED> i guess that means you have some interests which coincidentially collide with actual business needs :) something more people struggle hard to find :p
[17:04:19] <_jpl_> Is there enough Python work like that in NYC?
[17:04:55] <etrepum> it's not low level "do this in python work", it's more of a "we need this, do it" sort of work, that I choose Python for
[17:05:29] <Maniac> etrepum is independantly wealthy so he gets to do whatever he wants
[17:05:34] <Maniac> etrepum, what's your pycon topic?
[17:06:20] <redLED> i do business development consultancy alltogether :p i'm only into python so i can show up infront of a team of engineers and explain them what i want them to do not having to walk out of there with everyone pointing fingers at me and laughing :)
[17:07:19] <etrepum> eh, I'm full on middle class, but I get paid well because I do good work :)
[17:08:15] <etrepum> it's easy to do whatever you want so long as you're young and single, the problem is finding something you want to do
[17:08:40] <Maniac> well, you have to be reasonably smart too
[17:08:54] <etrepum> ok, fine, I assume everyone here is
[17:09:01] <Maniac> ha, fair enough
[17:09:06] <redLED> i'd go for wildlife photography :p but that market is saturated =/
[17:09:20] <redLED> so is unfortunately the market for paid betatesting positions for MMORPGs :(
[17:09:31] <etrepum> lol
[17:09:57] ** hazmat has joined us
[17:10:03] <Maniac> ah macpython bob
[17:10:07] <etrepum> I change my mind every couple years
[17:10:10] <etrepum> yeah
[17:10:48] <etrepum> I don't know if I could do something specific for more than 2 or 3 years at a time
[17:11:19] <etrepum> (and enjoy it, at least)
[17:11:21] <Maniac> gah! cygwin won't let me delete it 
[17:11:40] <etrepum> cygwin is pure evil
[17:12:17] <_jpl_> a necessary evil in my environment
[17:12:31] <Maniac> etrepum, ssh
[17:12:32] <etrepum> I got away with unxutils and mingw
[17:12:36] <Maniac> ssh
[17:12:46] <etrepum> ssh + *nix works better, of course
[17:13:55] <Maniac> i meant shh and i typed it twice 
[17:14:00] <etrepum> haha
[17:14:03] <Maniac> time for more coffee
[17:14:07] <etrepum> no kidding
[17:14:10] <etrepum> jolt makes gum now, too
[17:14:24] <Maniac> so how do you delete files when they don't let you in this stupid operating system known as windows
[17:14:30] <etrepum> reboot
[17:14:39] <Maniac> i did
[17:14:50] <redLED> i pay $5.30 per bottle of Bawls on my side of the world :(
[17:14:50] <etrepum> odd
[17:15:06] <redLED> Maniac: do you have any *win* services which depend on cygwin running at boot time?
[17:15:07] <etrepum> that's terrible
[17:15:30] <Maniac> redLED, good point
[17:15:35] <Maniac> probably sshd :)
[17:15:36] <Maniac> grr
[17:15:41] <etrepum> haha, sshd on windows
[17:15:47] <etrepum> you poor guy
[17:16:20] <_jpl_> likewise here
[17:16:44] <_jpl_> But it beats the hell out of managing 1000+ Windows machines without sshd
[17:17:02] <etrepum> true
[17:17:21] <etrepum> don't they have some sort of evil DCOM scriptability nonsense that you can use to administrate machines with?
[17:17:39] <Maniac> hey, it's not my main computer it's running in vmware
[17:17:39] <redLED> you could use one of them mouse-click capture tools to manage them sequentially via terminal service, while praying to $diety durring the proccess that all the icons are still in the same place :D
[17:19:26] <redLED> etrepum: i'm not sure about DCOM's usability for it, but the new win 2k3 has something called 'Windows Scripting Host' which is Microsoft's idea on a bash shell with ability to poke windows APIs
[17:19:41] <etrepum> you could call it "woosh"!
[17:20:01] <etrepum> I'm sure they already have that, but not the shell for it
[17:20:16] <Maniac> wow this is making me mad
[17:20:18] <etrepum> all of the .msc panels should be networkable
[17:20:32] <redLED> well, for now there's Microsoft WMI
[17:20:44] <redLED> it's pretty decent, but too much overhead for the average sysadmin to bother himself with
[17:20:50] <etrepum> probably
[17:20:58] <etrepum> the average sysadmin is an idiot, anyways
[17:21:19] <Maniac> can you reboot to a command prompt in win2k?
[17:21:34] <etrepum> not a very usable one
[17:22:01] <redLED> "shutdown /L /R /Y /C"
[17:22:01] <etrepum> it's mostly read only, there's only like one problem you can repair with it.. because it's "secure"
[17:22:04] <redLED> from command prompt
[17:22:10] <etrepum> at least from what I remember
[17:22:51] <Maniac> so i cant delete these files? i need a copy of knoppix
[17:22:59] <redLED> /Local, /Reboot /Yes (automatically) /Close applications and offer user to save
[17:23:15] <redLED> (its what the letters stand for)
[17:23:19] <etrepum> I think it depends on where the files are
[17:23:47] <etrepum> that shutdown doesn't boot into a command prompt though, does it?
[17:23:58] <redLED> oh, boot *into* a command prompt...
[17:24:01] <etrepum> yes
[17:24:02] <redLED> sorry, misread
[17:24:38] <redLED> that would include modifying the bootldr with some swiches, but should be doable as well
[17:24:45] <redLED> not something i'd remember better than google.com though.
[17:25:04] <etrepum> awwz yeah windows
[17:25:17] ** sremington has left IRC (Read error: 60 (Operation timed out))
[17:25:32] <Maniac> well, it's been booting in 'safe mode' for several minutes....
[17:25:37] <Maniac> doen't seem so safe right now
[17:26:03] <Maniac> how do you delete a directory tree in windows ?
[17:26:05] <Maniac> deltree 
[17:26:08] <Maniac> ?
[17:26:15] <etrepum> try holding cmd-s when it books, that should take you into single user mode and let you login as root ;)
[17:26:18] <etrepum> s/books/boots/
[17:27:28] <redLED> http://support.microsoft.com/default.aspx?scid=kb;en-us;q239780
[17:27:29] <redLED> there
[17:28:08] <redLED> (incase you need to do it remotely or automatically for some reason beyond my understanding heh)
[17:28:51] <etrepum> These switches are included if you press F8 and choose one of the modes.
[17:28:57] <redLED> idd
[17:28:58] <etrepum> that sounds easier
[17:29:03] <_jpl_> etrepum: regarding evil DCOM scriptability, sure, but who the hell wants to do that crap if you can automate with bash scripts and the like?
[17:29:19] <Maniac> how the hell do i delete these 'protected files?'
[17:29:20] <etrepum> it sure sounds like you just want to use a different OS
[17:29:33] <_jpl_> And actually Windows Scripting Host came with Win2k
[17:29:45] <etrepum> but not a bash-like console
[17:29:57] <Maniac> it's not letting me delete .lnk files wtf
[17:30:05] <_jpl_> Of course *we* (the Ops group) do, but the apps we support are written for Windows
[17:30:47] <_jpl_> Thankfully the powers-that-be have recently alloted time in the schedule to port one app to Linux, which will allow us to ditch Win2k on ~800 machines.
[17:31:07] <etrepum> that's awesome
[17:31:37] <redLED> i see that the negative sentiment toward the world's favorite evil empire and it's operating systems presist all across freenode :)
[17:31:37] <_jpl_> hell yeah
[17:32:02] <_jpl_> part of the delay was no doubt because of the fact that MSN was like 70% of our income until recently
[17:32:21] <etrepum> isn't msn going away or something?
[17:32:28] <_jpl_> no
[17:32:35] <_jpl_> I don't know
[17:32:42] <_jpl_> who cares...
[17:33:32] <redLED> i'll have to retire, getting midnightish here
[17:33:37] <redLED> thanks for the nice convo :)
[17:33:44] * _jpl_ waves
[17:34:27] ** redLED has left IRC ("Leaving")
[17:35:18] <_jpl_> ugh, I now have to finish the conversion tool I started which moves data from Access to Postgres; the database was cobble together by someone who is most definitely not a database developer, or even a programmer for that matter.
[17:35:25] <etrepum> oh god
[17:36:12] <_jpl_> So he used Access's lookup feature on almost every damn field in this one table.  Try doing outer joins with 15 or so lookup tables...
[17:36:25] <etrepum> I think I'm going to go watch TV or something, and forget you ever mentioned that
[17:37:18] <_jpl_> There's not that much data, so I can read it all into dicts and assign values to fields record-by-record as I convert the main table.
[17:37:55] <_jpl_> Yeah, Jerry Springer would be preferrable to finishing this task...
[17:38:15] <etrepum> you should be able to get at it easily via ADODB or something
[17:38:21] <etrepum> but still.. gross.
[17:38:31] <_jpl_> Of course the day or two I spent reverse engineering the database I was ready to end my existence.
[17:38:44] <etrepum> I've felt that pain before
[17:38:59] <etrepum> I've had to work on databases built by accenture monkeys
[17:39:06] <etrepum> not access, but just as bad
[17:39:19] <Maniac> crock of shit they're all set readonly
[17:39:46] <_jpl_> Yeah, I already wrote the code which reads all the data into dicts, now I just have to turn it into something sensible.  There's a nice DBAPI compliant ADO driver for Python I've been using.
[17:40:26] <etrepum> I think I just talked to it directly with COM when I needed it, because I was using so little of it that I didn't see a need to pay the egenix guy for it
[17:41:12] <etrepum> er wait, that was ODBC not ADO
[17:41:27] <etrepum> well, I used ADO over COM directly, instead of mxODBC, anyways
[17:41:41] <etrepum> (directly as in win32all)
[17:42:06] <_jpl_> That sounds painful.
[17:42:42] <etrepum> nah, not at all really
[17:43:00] <etrepum> DBAPI isn't exactly the best thing since sliced bread anyways
[17:43:10] <_jpl_> True
[17:43:36] <etrepum> using ADO directly gave me a sane way to call stored procedures and crap, too
[17:43:59] <etrepum> so really, using something DBAPI based would probably have been more painful
[17:44:10] <_jpl_> None of those, thankfully.
[17:44:26] <etrepum> a database without stored procedures probably doesn't need to live in an RDBMS :)
[17:44:40] <_jpl_> Hmm, I wonder if I can install the Win32 PEAK dist redLED just put together and use the DMs I've already created for the target database...
[17:44:43] <Maniac> etrepum, that's a strange view
[17:45:09] <etrepum> what's the point of using an RDBMS if you've just got tables and indexes?
[17:45:19] <Maniac> ACID
[17:45:26] <Maniac> transactions etc.
[17:45:36] <etrepum> you can get ACID and transactions elsewhere without too much pain
[17:45:52] <Maniac> i dont' find RDBMS painfull
[17:45:59] <Maniac> i'm actually a fan
[17:46:05] <_jpl_> hey, I just found a use for dict.popitem()  :)
[17:46:30] <etrepum> probably not one that's any better than itering over something
[17:47:04] <etrepum> out of process RDBMS is just a pain if you're not really using it
[17:47:12] <_jpl_> etrepum: An RDBMS usually makes sense for most data in a corporate environment where people expect to look at it with Access and other user-fiendly tools...
[17:47:13] <etrepum> I like PostgreSQL, I think MySQL is garbage
[17:47:15] <Maniac> my blog has a memory leak
[17:47:26] <Maniac> etrepum, i use sqlite for embedded stuff 
[17:47:45] <etrepum> sqlite seemed kind of cute, I haven't used it much
[17:47:51] ** hazmat has left IRC (Connection timed out)
[17:47:56] <_jpl_> its definitely 'cute'
[17:48:03] <Maniac> then you don't have to run a RDBMS 'process'
[17:48:05] <_jpl_> an interesting piece of software
[17:48:11] <etrepum> I've used software that depends on it, but I haven't written anything against it
[17:48:50] <Maniac> but if you're already using RDBMS for other things then it's pretty minor to tack on other databases
[17:49:11] <etrepum> unless you want to store a tree or linked list :)
[17:49:31] <etrepum> it's all possible in SQL, of course, but it's probably braindead and slow to traverse
[17:50:45] <etrepum> SQLObject looks cool.. I wrote something like it last time I used lots and lots of SQL from Python
[17:51:26] <_jpl_> It doesn't look all the cool to me
[17:51:49] <etrepum> I like ORM
[17:52:10] <etrepum> as long as you can get down to SQL when you want to
[17:52:15] <_jpl_> With peak.model and peak.storage you get most of what SQLObject does.
[17:52:24] <etrepum> except you have to write the SQL statements yourself
[17:52:40] <_jpl_> I've written a TableDM which acts as a baseclass for other DMs for that.
[17:53:04] <_jpl_> All you have to do is specify the list of fields, the keys, and the table, much like SQLObject.
[17:53:20] <etrepum> I haven't seen TableDM
[17:53:25] <_jpl_> I've been meaning to clean it up a bit and check it in to peakplace CVS.
[17:54:00] <_jpl_> I posted an early version of it to the mailing list back in November I think...
[17:54:30] <etrepum> I wasn't reading the list back then
[17:54:44] <etrepum> I don't think
[17:55:12] <_jpl_> Ah, no, it was October.  It's in the archive.
[17:55:37] * jack-e has also a SQLEntityDM in his cvs (but old binding.api) based on the ideas/code of PyDO .. it's not in much in use though ..
[17:55:42] <jack-e> hello btw.
[17:55:59] <etrepum> and apparently pje plans to obsolete both soon
[17:56:05] <jack-e> yeah :)
[17:56:15] <_jpl_> Hey jack-e
[17:56:22] <jack-e> i did not spend too much work in it ..
[17:56:32] <jack-e> hey _jpl_ 
[17:56:37] <_jpl_> Well, there's no way of knowing when that'll happen.
[17:56:46] <_jpl_> "when he needs it"
[17:56:55] <etrepum> he was talking about it the other day
[17:57:12] <_jpl_> It sounds like it'll be really nice when he does get to it.
[17:57:13] <jack-e> right .. but datamanagers definitly need refactoring to coop with peak.events properly
[17:57:21] * _jpl_ nods
[17:57:38] <etrepum> the model stuff seems to be a bit specific too
[17:57:42] <etrepum> I mean, no mappings, wtf?
[17:58:26] <_jpl_> yeah... apparently it mostly does what he needs it to right now, otherwise he'd have added a lot more to peak.model by now.
[17:58:54] <etrepum> well from the way he phrased it, it sounds like he is using it directly from some kind of UML serialization or something
[17:59:04] <_jpl_> We have field mappers in TableDM
[17:59:09] <etrepum> and the UML spec/software just doesn't support much
[18:00:06] <jack-e> the idea is to use XMI-Exports (XML Representation  of MOF MetaObjectFacilty) to create peak.model source-code
[18:00:23] <_jpl_> I'd still like to see that in action...
[18:00:34] <jack-e> this code is then configured with aspects that implement the behaviour
[18:00:56] <_jpl_> Is that how pje is using it?
[18:01:02] <jack-e> all the parts from peak.metamodels.UML* are generated code from the UML-1.x XMI-Files
[18:01:03] <etrepum> well fine, but what good is XMI if it doesn't support mappings? :)
[18:01:18] <jack-e> that would be one way of using it .. it's not yet 100% done
[18:01:32] <jack-e> use structs
[18:01:43] <etrepum> that's awful
[18:01:50] <_jpl_> What specifically do you mean by mappings?
[18:01:54] <jack-e> peak.util.struct :)
[18:02:03] <etrepum> key->value
[18:02:13] <etrepum> "foo":"bar"
[18:02:16] <_jpl_> Yes, but at what level?
[18:02:39] <etrepum> referencedKeyType = model.String
[18:03:27] <_jpl_> such that...?
[18:03:43] <etrepum>     class parties(Mapping):
[18:03:43] <etrepum>         singularName = 'party'
[18:03:43] <etrepum>         referencedKeyType = model.String
[18:03:43] <etrepum>         referencedType = 'Party'
[18:03:44] <etrepum>         referencedEnd = 'presidentialRace'
[18:04:22] <jack-e> didn't you already implement it ?
[18:05:16] <_jpl_> So that Party.presidentialRace on the referenced end is assured to be a model.String?
[18:05:32] <jack-e> i don't miss the Mapping that much as i miss the DateTime Field ;-) .. there is still a way to go .. we could provide such types/components in a shared library on peakplace ...
[18:05:44] <etrepum> yes, I did already implement it
[18:05:54] <etrepum> I have a Date wrapper
[18:06:04] <jack-e> me too ;-)
[18:06:11] <etrepum> that bridges datetime.date and mx.DateTime
[18:06:24] <etrepum> as well as ISO date strings
[18:06:35] <_jpl_> I have a couple of handy types as well.
[18:06:36] <Maniac> yay just built c extensions for python in windows
[18:06:58] <etrepum> I haven't needed datetime or time for anything in this project, so I didn't write the wrappers for those
[18:07:18] <etrepum> well the problem with mapping is that it's fundamental
[18:07:20] <etrepum> datetime is not
[18:07:34] <_jpl_> I wrote a type called EasyBoolean which normalizes things like 'yes', 'n', 1, etc. to True/False.
[18:07:39] <jack-e> that's true .. maybe that's why UML doesn't support it ..
[18:08:11] <etrepum> for mappings I needed to add getForKey, setForKey,e tc.
[18:08:20] <etrepum> date wouldn't need new accessors
[18:09:00] <_jpl_> Why wouldn't you put that sort of thing in a DM rather than in the model class itself?
[18:09:22] <etrepum> because why the hell should it go in a DM?
[18:09:27] <etrepum> it's DM independent
[18:09:49] <Maniac> cheerio , home time
[18:09:50] <_jpl_> Then I'm still not following what you're doing.
[18:09:54] * Maniac leaves the room
[18:10:01] <jack-e> how do you store a mapping-attribute in your backend ?
[18:10:21] <etrepum> it doesn't matter how I store it
[18:11:28] <etrepum> I'm not using SQL at all at the moment, if that's what you're asking.. 
[18:12:52] <jack-e> you could extend your MappingType to delegate __get/setitem__ calls to the underlying datastructure ..
[18:12:54] <etrepum> in SQL you would probably want to store it differently depending on what the key and value is
[18:13:46] <etrepum> unless the underlying data structure already supports that :)
[18:15:32] <jack-e> i've done quite a few apps with peak.model (some finished, some not). in all cases where i wanted use a mapping i could also use a new DataType (e.g. Contact, ContactAddress, ...)
[18:16:18] <jack-e> it is much more expressive and reusable for with Adaption
[18:16:36] <etrepum> you can adapt the keys just like you adapt the values
[18:17:51] * jack-e is tired and has a long meeting 2morrow ..
[18:18:04] <jack-e> cu 
[18:18:13] jack-e is now known as jack-e|away
[18:18:15] <etrepum> of course you can store any mapping as a list of 2-tuple structs
[18:18:40] <etrepum> and you can also represent an integer as a string
[18:18:43] <etrepum> but that doesn't mean that you should :)
[18:34:54] ** sremington has joined us
[18:56:42] ** _Maniac has joined us
[20:06:04] <_jpl_> augh... I forgot that in this miserable mess of a database the 'designer' even used lookup fields for things that should be values, like disk size, amount of memory(!!!)
[20:53:53] ** hazmat has joined us
[21:35:35] ** hazmat has left IRC (Read error: 60 (Operation timed out))
[22:03:41] ** gbay has joined us
[22:42:35] ** gbay has left IRC ("It's not like I'm leaving or anything...")