00:00:03 <ehird> calamous: limit the number of files on the system?
00:00:10 <ehird> eg it resets if you write more than 100 files
00:00:10 <pikhq> I tend to spend time in #xkcd, on the offchance that there's interesting things there.\
00:00:45 <pikhq> Though the smaller channels on Foonetic are far better; signal/noise ratio and all that.
00:01:20 <ehird> bsmntbombdood: i'm sure Deewiant will be here sometime :P
00:02:23 <calamous> !sh for i in {1..99999}; do ln -s /home/egobot/egobot.hg/foo /tmp/input.$i; done
00:02:36 <ehird> calamous: it took an hour or so
00:02:43 <AnMaster> <ehird> Deewiant: bsmntbombdood wants to know how to use two monitors
00:02:49 <ehird> !sh ls /tmp/input.* | wc -l
00:02:53 <ehird> !sh ls /tmp/input.* | wc -l
00:02:54 <ehird> !sh ls /tmp/input.* | wc -l
00:02:54 <EgoBot> /bin/bash: line 1: :ehird!n=ehird@208.78.103.223: command not found
00:03:06 <calamous> OK, I made it delete them at the end, but that's clearly not good enough.
00:03:19 <ehird> calamous: just limit the number of writes to 100 files
00:03:27 <ehird> bsmntbombdood: tell AnMaster how he failed :-P
00:03:35 <calamous> ehird, You can only limit the number of total open files.
00:03:44 <ehird> calamous: To 100, then.
00:04:04 <ehird> er i misread AnMaster
00:04:18 <calamous> ehird, The number of open files is limited to a very low number.
00:04:27 <calamous> ehird, The problem is that it doesn't keep them open after it writes them.
00:04:28 <AnMaster> which is better if they are different height
00:04:36 <AnMaster> having different height is confusing
00:06:46 <ehird> AnMaster: do you think http://www.newegg.com/Product/Product.aspx?Item=N82E16814187037 can handle dueling monitors
00:07:02 <ehird> the specs seem to imply no
00:08:20 <ehird> seems to cost less, so I assume the 9400 GT is better
00:08:30 <ehird> for values of better equal to still not a good card but bsmntbombdood doesn't want a good card :P
00:09:46 <AnMaster> ehird, I used two TFTs, one vga only and one dvi +vga
00:09:53 <AnMaster> so one went on dvi the other on vga
00:44:39 <bsmntbombdood> my current monitor is vga so that would probably work
00:56:34 <ehird> "Intel(R) Core i7 PC's from £7"
01:15:13 <ehird> bsmntbombdood: Transmission
01:16:38 <ehird> Which is minimal and awesome and <3
01:20:46 -!- oerjan has quit ("Good night").
01:21:56 * bsmntbombdood wonders if he should actually buy ehird's computer
01:22:12 <ehird> you want an upgrade, I got it for the price you wanted, prophet? :-P
01:25:37 <ehird> bsmntbombdood: for the same budget, the only real other options there are is getting an amd processor instead, but that'd just be sacrificing speed and you'd have some of the budget just lying around doing nothing
01:25:56 <ehird> so if you'd upgrade to that budget anyway, it's pretty much what you'd come out with
01:26:08 <ehird> bsmntbombdood: over what?
01:26:27 <ehird> bsmntbombdood: amd's top processors are competitive with core 2 -- the LAST generation of intel processors
01:26:34 <ehird> i7 has the yummy integrated memory
01:26:38 <ehird> so a lot less latency
01:26:44 <ehird> and it has some nice architechtural improvements
01:26:52 <ehird> more speed clock-for-clock under a lot of situations
01:27:25 <ehird> i7 >>> amd phenom black edition 2 electric boogaloo > core 2
01:29:11 <pikhq> I was under the impression that AMD's top processors were competitive with some of the lesser i7 chips, making them great if you're on a budget.
01:29:33 <ehird> pikhq: i haven't seen that; but the i7 was a 2.93ghz
01:29:41 <ehird> so not exactly a lesser i7 chip
01:30:03 <pikhq> AMD's existence for most of the company's history *has* depended on being cheaper than Intel chip, so *shrug*.
01:30:03 <ehird> there's only 3.2ghz and 3.33ghz (the latter i haven't seen sold anywhere) above that
01:30:24 <ehird> pikhq: but that integrated memory controller + ddr3 = definite yum
01:30:37 <ehird> bsmntbombdood: err
01:30:42 <ehird> the i7 comes with it i'm pretty sure
01:31:10 <ehird> i've seen people say "stock cooler" referring to i7s
01:31:10 <pikhq> ehird: Integrated memory controller? I've had one for 3 years, and it was old then. :p
01:31:17 <ehird> pikhq: er, that's not what i meant
01:31:25 <ehird> pikhq: "On-die memory controller: the memory is directly connected to the processor. It is called the uncore part and runs at a different clock (uncore clock) of execution cores. "
01:31:38 <ehird> 20-40% less memory latency has been reported
01:31:46 <ehird> bsmntbombdood: "Cooling DeviceHeatsink and Fan included "
01:31:53 <ehird> from the spec page of the i7
01:32:03 <pikhq> And AMD's done that since they introduced AMD64. Your point? :p
01:32:11 <ehird> pikhq: i'm not sure that's the same thing
01:33:05 <pikhq> Wikipedia says it is.
01:33:13 <ehird> hmm looks like the same thing; still what i've read puts the i7's memory access faster
01:33:40 <ehird> I expect AMD will come out with an i7 competitor _sometime_
01:33:46 <ehird> but atm the i7 is the best choice imo
01:33:55 <pikhq> Can't say I blame Intel for doing it; it *is* a rather easy way to get better memory access speeds.
01:35:04 <ehird> bsmntbombdood: if you do go with it will you go with the x25-m? i'd recommend it if you do have the $300 extra budget (hmm, I distinctly recall asking this before...)
01:36:01 <ehird> bsmntbombdood: yeah, that sounds wise
01:37:02 <ehird> bsmntbombdood: raid-1ing that will leave you with only 500gb of space
01:37:31 <pikhq> Hmm. No wonder the i7's somewhat higher performance; it's the newer architecture.
01:37:43 <ehird> pikhq: Durr, Nehalem's the whole point of an i7 :P
01:37:54 <ehird> bsmntbombdood: newly calculated total inc. ssd: $1689.91
01:38:52 <bsmntbombdood> ehird: what makes you think that mobo is a good choice?
01:39:06 <ehird> bsmntbombdood: first, I went with the stock Intel mobo for i7
01:39:10 <ehird> then i realised it had only 3 ram slots
01:39:14 <ehird> so I searched newegg
01:39:25 <ehird> and that was a high-rated one that's the same as the intel one but w/ 6 ram slots
01:40:27 <ehird> bsmntbombdood: they don't seem to have really changed the intel mobo apart from adding the more slots and sth
01:40:42 <ehird> bsmntbombdood: how many do you need?
01:40:56 <ehird> this one looks bette
01:41:08 <ehird> Cons: As noted in previous reviews, the BIOS needs an update out of box before you can install triple channel RAM. I loaded one stick of RAM and installed windows
01:42:02 * ehird searches newegg for x58
01:42:08 <ehird> bsmntbombdood: but how many pci slost are you gonna use?
01:42:25 <bsmntbombdood> i would like to play with some coproccesors in the future
01:42:40 <ehird> PCI Express 2.0 x16: 3 x PCIe 2.0 x16 (at x16/x16/x4 mode)
01:43:00 <ehird> bsmntbombdood: http://www.newegg.com/Product/Product.aspx?Item=N82E16813131359
01:43:25 * ehird amends his spec document to include these
01:44:10 <ehird> bsmntbombdood: the mobo actually saves money
01:45:18 <ehird> bsmntbombdood: for when you return: http://pastie.org/467175.txt?key=ijtxvejybwknzzyr4yea
01:49:24 <ehird> My conclusion: You can build a top-top of the range PC (sans video card) for around $1.6k.
01:53:13 <pikhq> And the video card isn't really worth it unless you're playing Crysis on 3 different monitors.
01:53:40 <ehird> pikhq: Weeeeeeeeeeeeel, that GeForce 9400 GT seems rather the shit as far as cards you're actually gonna use goes.
01:53:49 <ehird> It's just a stand-in for the missing on-board video on the mobo for this machine.
01:54:33 <ehird> " it ran between 65 and 72 deg C"
01:54:39 <ehird> i wonder if that's under large load
01:54:41 <ehird> because that's a hot card
01:55:08 <pikhq> Eh, as far as graphics cards go, I'm not all that picky.
01:55:22 <ehird> the reviews seem ok
01:55:49 <ehird> bsmntbombdood: the gfx card has an audio card built in
01:55:54 <ehird> that apparently is better than onboard video
01:56:56 <pikhq> My current standards: able to handle Source engine games and able to render 1080p h264 in realtime. And the latter becomes "Can handle 1080p" if my CPU's fast enough to decode 1080p h264 in realtime...
01:57:21 <ehird> My standards: Be the best fanless card you can get.
01:57:39 <ehird> (atm that's the radeon hd 4850)
01:57:42 <pikhq> Completely different standards, but likely to get similar results.
01:57:56 <ehird> pikhq: no way, the radeon 4850 can do much more than that
01:58:06 <ehird> it's the third best single (non-X2) card ATI offer
01:58:23 <ehird> i wish newegg had a uk version
01:58:33 <pikhq> I think it's the best card I could actually see myself bothering to purchase, though.
01:58:46 <pikhq> Much beyond that seems... Silly.
01:59:31 <ehird> pikhq: Heck, even with the 4870X2 -- the best card ATI sells -- Crysis at 19xx-xxxx (I forget what the exact res is) with all settings on high or very high only runs at 20 or so fps
01:59:40 <ehird> Games are wicked demanding
02:00:19 <pikhq> And I play games about a generation behind what's current.
02:00:41 <pikhq> So, it'll be another 5 years or so before I get around to playing Crysis. :p
02:01:08 <ehird> Crysis kind of looks shit :P
02:01:31 <pikhq> Ah, but it's very expensive shit.
02:01:37 <pikhq> And you're thinking 1920x1080.
02:03:32 <ehird> maybe i should build my own system, 'cept i hate thermal paste
02:04:23 <ehird> pikhq: maybe, maybe
02:04:39 <ehird> I'd prefer to pay more to get a silent system from people who know what they're doing in that area, prolly
02:04:46 <pikhq> I build my systems as a matter of habit...
02:05:11 <pikhq> Of course, I upgrade my systems piecemeal, so not exactly what you're dealing with.
02:05:27 * pikhq has a couple of 10 to 15 year old parts
02:16:54 -!- pikhq has quit (Read error: 104 (Connection reset by peer)).
02:17:58 -!- pikhq has joined.
02:30:58 -!- cherez has joined.
02:31:02 -!- cherez has left (?).
02:34:06 -!- Gracenotes has joined.
02:46:43 -!- puzzlet has joined.
02:47:26 -!- puzzlet_ has quit (Remote closed the connection).
02:49:04 <calamous> !c printf("addr of main %p\n", main);
02:49:08 <calamous> !c printf("addr of main %p\n", main);
02:50:04 <pikhq> !c printf("addr of entry point %p\n", __start);
02:50:22 <pikhq> Fik. Different entry point. XD
02:50:38 <pikhq> !c printf("addr of entry point %p\n", _start);
02:50:48 <pikhq> !c printf("addr of entry point %p\n", start);
02:50:54 <pikhq> Sure I'll get it one of these days.
02:52:38 <pikhq> Huh. main *is* the entry point.
03:11:08 -!- puzzlet_ has joined.
03:24:10 -!- puzzlet has quit (Read error: 110 (Connection timed out)).
03:24:11 <bsmntbombdood> ehird: i forgot, i need an optical-disc drive also
03:27:25 <GregorR> OK, now I have to think of a way to fix the shared /tmp/input.* problem
03:35:49 <EgoBot> /bin/ls: /tmp: Function not implemented
03:41:57 -!- Gracenotes has quit (Remote closed the connection).
03:41:58 -!- Gracenotes_ has joined.
03:42:00 -!- Gracenotes_ has changed nick to Gracenotes.
03:57:40 <EgoBot> /bin/ls: /tmp: Function not implemented
03:58:13 <pikhq> What the fnuck is it doing, I wonder...
03:59:47 <GregorR> ls stats or something, Idonno
04:05:04 <EgoBot> /bin/mkdir: created directory `/tmp/home'
04:05:11 <EgoBot> /bin/sh: line 6: cd: -v: invalid option
04:05:25 <GregorR> !psh export HOME=/tmp/home
04:07:58 -!- bsmntbombdood has quit (Connection reset by peer).
04:08:33 -!- bsmntbombdood has joined.
04:09:12 <GregorR> !psh wget http://google.com/
04:09:12 <EgoBot> --2009-05-04 03:09:12-- http://google.com/
04:09:13 <EgoBot> Resolving google.com... failed: Name or service not known.
04:10:33 -!- bsmntbombdood has quit (Read error: 104 (Connection reset by peer)).
04:11:06 -!- bsmntbombdood has joined.
04:12:58 -!- bsmntbombdood has quit (Read error: 104 (Connection reset by peer)).
04:21:44 -!- bsmntbombdood has joined.
04:46:47 <kerlo> So, I guess I'm trying to find out where Firefox keeps its profiles under Windows.
04:56:34 <lifthrasiir> somewhere in %userprofile%\Local Settings\Application Data\Mozilla\, iirc.
05:02:12 <kerlo> Users/Fozzie/Local Settings is empty.
05:02:22 <kerlo> No, those forward slashes are not necessary.
05:09:35 <kerlo> I'm not using Windows right now; I want to share a profile between Windows and Linux.
05:33:01 -!- pikhq has changed nick to pikhq_.
05:33:11 -!- pikhq_ has changed nick to pikhq.
05:50:37 -!- bsmntbombdood_ has joined.
05:58:12 -!- bsmntbombdood has quit (Read error: 110 (Connection timed out)).
05:58:56 -!- bsmntbombdood_ has changed nick to bsmntbombdood.
06:01:29 <kerlo> Wow. YouTube under Linux here sounds really fun.
06:04:21 <kerlo> I meant that it sounds very slow and choppy.
06:04:33 <kerlo> I just recorded an hour-long audio file. It took a couple seconds.
06:05:02 <puzzlet_> how about other video websites
06:05:31 <puzzlet_> could be a problem with flash plugin
06:06:46 <kerlo> A problem that also exists with Sound Recorder?
06:07:26 <kerlo> Here, I'll click "record" again.
06:08:09 <kerlo> It says it's been recording for four hours.
06:08:27 <puzzlet_> no, tell me it's not the alsa problem again
06:08:56 <kerlo> This problem appears to be entirely distinct from the problem I had before.
06:09:11 <kerlo> Now it says it's been recording for ten hours.
06:09:18 <kerlo> Maybe some other people have had this same problem.
06:10:40 <kerlo> This is why I don't generally use Linux.
07:06:08 <GregorR> In my experience, the people who have the ever-dwindling set of hardware that Linux doesn't support are the people who think it has terrible hardware support :P
07:06:23 <GregorR> (Which is to say, that sounds like a problem with the sound driver)
07:34:29 <bsmntbombdood> you mean...the people who have problems with linux are the ones who have problems with linux?
07:39:26 -!- GreaseMonkey has joined.
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:07:46 -!- lereah_ has joined.
08:12:19 -!- Deewiant has quit (Read error: 60 (Operation timed out)).
08:21:09 -!- fizzie has quit (Read error: 148 (No route to host)).
08:38:35 -!- Deewiant has joined.
08:53:11 -!- nooga has joined.
08:57:17 -!- WangZeDong has joined.
09:05:02 -!- puzzlet has joined.
09:14:33 -!- Slereah has quit (Read error: 110 (Connection timed out)).
09:17:38 -!- puzzlet_ has quit (Read error: 110 (Connection timed out)).
09:27:07 -!- nooga has quit (Read error: 104 (Connection reset by peer)).
09:57:49 -!- puzzlet_ has joined.
10:10:46 -!- puzzlet has quit (Read error: 110 (Connection timed out)).
10:25:43 -!- fizzie has joined.
10:31:14 -!- GreaseMonkey has quit ("YES -> thor-ainor.it <- THIS IS *DELICIOUS*!").
10:38:03 -!- tombom has joined.
10:42:16 <AnMaster> <GregorR> OK, now I have to think of a way to fix the shared /tmp/input.* problem
10:42:28 <AnMaster> that avoids race conditions too
11:01:21 -!- calamous has quit ("Leaving").
11:54:00 <kerlo> AnMaster: Gnome Sound Recorder.
11:54:15 <AnMaster> no clue about that. Since I don't use Gnome
11:54:26 <AnMaster> I would blame it on gnome probably
11:54:49 <AnMaster> try audacity, it is a good sound recording and editing program.
11:56:21 * kerlo reboots under the alibi of figuring out where the Firefox profile resides
12:16:53 <kerlo> Windows still won't make any connections to the Internet, but I made a symbolic link from a Firefox profile on Linux to one on Windows, and it worked quite magically.
12:20:46 <Deewiant> I just set the profile in Firefox to point to /mnt/...
12:33:58 <AnMaster> why dual boot. The only reason I can think of is massive 3D games.
12:34:36 <AnMaster> for most other windows only stuff you can usually use either wine or virtualbox/vmware-server/whatever
12:41:26 -!- impomatic has joined.
12:42:26 <impomatic> The #corewars channel will be moving to freenode if anyone is interested.
12:42:38 <impomatic> By the way, is anything happening with BF Joust?
13:44:29 -!- puzzlet has joined.
13:45:16 -!- puzzlet_ has quit (Remote closed the connection).
14:26:39 -!- tombom has quit (wolfe.freenode.net irc.freenode.net).
14:26:39 -!- EgoBot has quit (wolfe.freenode.net irc.freenode.net).
14:26:39 -!- Leonidas has quit (wolfe.freenode.net irc.freenode.net).
14:26:39 -!- comex has quit (wolfe.freenode.net irc.freenode.net).
14:26:39 -!- GregorR has quit (wolfe.freenode.net irc.freenode.net).
14:26:43 -!- tombom has joined.
14:26:43 -!- comex has joined.
14:26:43 -!- EgoBot has joined.
14:26:43 -!- Leonidas has joined.
14:26:43 -!- GregorR has joined.
15:03:33 -!- FireFly has joined.
15:06:38 -!- tombom has quit ("Peace and Protection 4.22.2").
15:20:34 -!- Gracenotes has quit (Remote closed the connection).
15:22:02 -!- Gracenotes has joined.
15:27:48 -!- bsmntbombdood has quit (Read error: 104 (Connection reset by peer)).
15:48:04 -!- bsmntbombdood_ has joined.
15:57:43 -!- ais523 has joined.
16:09:54 <ehird> bsmntbombdood_: so optical drive eh
16:09:57 <ehird> I'm sure i can do that :P
16:11:28 -!- lereah_ has quit (Remote closed the connection).
16:11:32 <ehird> bsmntbombdood_: shall I throw in a floppy drive? some firmware upgrades still come on those.
16:11:48 <pikhq> Floppies are cheap...
16:13:26 -!- Dewi has joined.
16:15:44 <ehird> bsmntbombdood_: http://pastie.org/467702.txt?key=7xpp0vdnf1ukje2vbjdq
16:16:00 <ehird> $1709.39, same as before but w/ dvd/cd-reader/writer and floppy drive
16:18:44 <ehird> bsmntbombdood_: btw, graphics cards are useful for vector computation etc as well as just graphics...
16:19:10 -!- impomatic has quit (Read error: 110 (Connection timed out)).
16:23:35 <ais523> meh, what happened to it?
16:23:40 <ehird> ais523: i did !accumulate junk
16:23:51 <ehird> then AnMaster did !accumulate ehird being annoying again, or something
16:23:56 <ehird> and we went at it for a few hours
16:24:26 <ehird> result: AnMaster's ignoring me, keeps saying I'm immature, ignores the fact that it was 50% him, and also that if we were arguing over junk it's obviously a topic we're discussing :D
16:24:28 <AnMaster> a lot less than that. You killed it every time.
16:24:38 <ehird> that ignore lasted a long time there AnMaster
16:24:41 <AnMaster> ehird, GregorR made the bot ignore you too
16:24:46 <ehird> a few hours perhaps
16:24:54 <ehird> AnMaster: yes, to break the chain
16:30:39 <ehird> "Ubuntu To Ship With Super Talent SSDs" ← I totally thought this meant that ubuntu cds would come with an ssd, not the otherway around
16:36:36 <ais523> because there wasn't a decent use for them?
16:36:36 <GregorR> Because there was some issue with them leaving cruft behind I haven't resolved yet.
16:37:03 <GregorR> There's no decent use for most of what EgoBot supports, but I don't want a chicken with its head cut off taking 100% CPU on Codu :P
16:37:40 <GregorR> Anyway, they'll probably come back when I can fix some bugs.
16:37:55 <pikhq> Don't kill Codu with Egobot. ;)
16:38:10 <GregorR> Or alternatively, if I can figure out a way to use chkpt to checkpoint between lines, then I can actually have daemons only run while in use.
16:40:54 -!- iano has joined.
17:08:41 -!- jix has joined.
17:08:49 <lifthrasiir> ehird: if you are interested in esotope-bfc, it now supports exact loop counting for nontrivial deltas.
17:09:02 <ehird> example code/output?
17:09:11 <lifthrasiir> http://pastie.org/private/4z2i9vo3y8kviwet42fma
17:10:16 <ehird> One of the most publicized changes in Windows Me, was that it no longer included real mode MS-DOS.[3] With real-mode support removed, Windows Me can boot up a couple of seconds faster, without loss of Windows functionality. Autoexec.bat and Config.sys are no longer executed during startup by IO.SYS, and the system cannot boot to a MS-DOS command prompt or exit to DOS when Windows has booted. Because of this, applications that needed real mode DOS to run,
17:10:18 <ehird> such as older disk utilities, did not run under Windows Me.
17:10:22 <ehird> if ((mptr[1]%2) != 0) {
17:10:24 <ehird> lifthrasiir: that's hot.
17:10:26 <ehird> is that really general?
17:10:43 <ehird> ais523: take a look at that, it's awesome
17:10:52 <ehird> lifthrasiir: congrats, your BF interp is the est
17:11:15 <pikhq> ehird: Uh, lies. Windows ME shipped with real mode MS-DOS. *However*,, it was not enabled by default.
17:11:17 <lifthrasiir> not yet, i realized that generalization while looking at bfdb ;)
17:11:31 <lifthrasiir> it translates every possible case of while(mptr[k]!=x) { ... mptr[k]+=delta; }.
17:11:48 <ehird> lifthrasiir: bfdb?
17:11:48 <pikhq> Wikipedia lies, then.
17:11:55 <ehird> pikhq: I used ME :-P
17:12:06 <pikhq> I turned on real-mode DOS in ME.
17:12:36 <ehird> lifthrasiir: never heard of it
17:12:44 <lifthrasiir> it needs a fix to compile in recent gcc, but it seems worth looking
17:12:48 <ais523> I'd like to see how esotope-bfc.py handles gcc-bf output, actually
17:13:20 <ais523> presumably it isn't yet good enough to deduce that I'm simulating a hardware stack and a stack of frame pointers?
17:13:40 <lifthrasiir> it cannot handle array operations yet, and i think that is quite challenging
17:14:15 <ehird> lifthrasiir: try compiling with clang, not gcc
17:14:17 <lifthrasiir> anyway, if you are interested see http://hg.mearie.org/esotope/bfc/file/tip/esotope-bfc.py for recent version.
17:14:43 <ehird> that's not c++, lifthrasiir
17:14:44 <ehird> I mean your output
17:14:51 <ehird> but yes, clang's c/c++/obj-c support is mostly okay
17:16:54 <ais523> in other news, it turns out that MS Office's new native ODF support is abysmal
17:17:03 <ais523> for instance, it strips out formulas when loading spreadsheets
17:18:07 <ehird> openoffice is so bloated
17:18:19 <ehird> 17:17 ehird: Does clang/llvm do the optimization of changing some array positions into variables? like, {int x[2];x[0]=1;x[1]=2;return x[0]+x[1]} becoming {int x0,x1;x0=1;x1=2;return x0+x1}
17:18:21 <ehird> 17:18 sabre: ehird: yes
17:18:23 <ehird> 17:18 ehird: sabre: yay!
17:18:26 <ehird> lifthrasiir: yep, use clang.
17:19:39 <ehird> @openoffice.org users: try abiword & gnumeric :P
17:19:54 <ais523> gnumeric's ODF support is also awful, I think
17:20:02 <ais523> but then, it doesn't claim to do it correctly
17:20:07 <ais523> it looks like they've only just started implementing it
17:20:16 <ehird> I mostly meant abiword
17:20:22 <ehird> but then I thought people might complain about spreadsheets
17:20:26 <ehird> so i added gnumeric
17:20:52 <ehird> ais523: openoffice is pretty bad though; it doesn't pick up fonts in ~/.fonts for instance IME
17:20:58 <ehird> which is really irritating
17:21:06 <ehird> bloody portability :)
17:21:16 <ais523> openoffice manages to not follow the conventions on any system
17:21:21 <ais523> the difference is, on Windows, I'm glad it doesn't
17:21:41 <ehird> I always thought that openoffice was Java because it was so badly integrated... but Java is more integrated and less ugly!
17:21:48 <ehird> I don't even know how they did it
17:21:56 <ais523> openoffice is I suspect mostly written in Java
17:22:08 <ais523> or by people who write in the same style, anywa
17:22:12 <ehird> well, it's a Sun project
17:22:19 <ehird> but it doesn't use Swing
17:22:23 <ehird> and I don't think it uses SWT
17:22:30 <ehird> or if it does use either, it hacks them weirdly
17:22:34 <ehird> it's what eclipse uses
17:22:45 <ais523> oh, I like AWT anyway, though
17:22:54 <ais523> it's easily the best of the windowing systems Java has come up with so far
17:23:08 <ais523> AWT = Java is a programming language
17:23:28 <ais523> personally I think treating Java as a programming language makes more sense
17:23:57 <ehird> ais523: you can make swing use native widgets
17:24:01 <ehird> not native text rendering, though
17:24:45 <ais523> native-widget swing would be an improvement, certainly
17:24:46 <fizzie> I did a bit of SWT once; it wasn't too shabby. It's pretty much just a wrapper for native GUI toolkits over JNI; except of course using the SWT abstractions.
17:24:59 <ehird> that's java's c ffi
17:25:06 <ehird> well, not much of an ffi, you have to tailor your code to it
17:25:08 -!- jix has quit (Remote closed the connection).
17:25:11 <ehird> c extension interface
17:25:24 <ehird> well, not specifically C
17:25:29 <ehird> but it'd be a pain to use anything else
17:25:37 -!- jix has joined.
17:25:39 <ehird> it has some c++ stuff i think though
17:26:01 <ais523> someone link Java to Befunge-98 via C and INTERCAL
17:26:03 <ehird> to hell with portable GUI libs, anyway
17:26:06 <fizzie> Yes, in C++ code you can pretend you're calling Java classes, or something.
17:26:08 <ehird> you can't automate UI guidelines
17:26:24 <ehird> you need to have a separate UI for each platform if you want to have a good UI
17:26:53 <ais523> the Gnome/KDE mirrorflip is one of the most annoying things about that
17:27:15 <ehird> ais523: gnome/kde's minor compared to anything_on_linux vs OS X
17:27:22 <ehird> you basically have to completely forget your other UI to go either way
17:27:31 <ehird> i get using a portable toolkit if you just want something that works
17:27:43 <ehird> but just use Tk or something; that's really easy to use, it's pig ugly most of the time though
17:27:48 <ais523> really? the most annoying thing for me when I had to use OS X on someone else's laptop for a bit was muddling the apple and control keys
17:27:50 <ehird> but that doesn't matter if you just want something that works
17:27:53 <ais523> and the lack of pgup/pgdn
17:27:57 <pikhq> It integrates well with everything but Linux.
17:28:05 <ehird> tk is horrid on os x
17:28:07 <ehird> it's fine on windows
17:28:12 <ais523> so it was really keyboard layout confusion, rather than anything to do with the UI itself
17:28:19 <ehird> ais523: for the user, there may not be much difference at first sight
17:28:25 <pikhq> Hrm, last time I used Tk on OS X, it looked like native widgets...
17:28:26 <ehird> but the philosophy is massively different
17:28:30 <ehird> pikhq: yes, to a degree
17:28:38 <ehird> but half-heartedly doing something on OS X is a bit laughable
17:28:52 <ehird> they pretty much only get buttons and checkboxes right
17:29:04 <ehird> and integration goes far further than how the widgets look anyway
17:29:20 <pikhq> Hrm. Either I used some odd OS X-only fork, or it's gone downhill over time.
17:29:37 <pikhq> Because I remember it being, well, native.
17:29:38 <ehird> pikhq: Or you're not an obsessive-compulsive OS X user (redundant) that notices things :-)
17:29:48 <ehird> For instance, it uses a white background and the window has no padding by default.
17:29:56 <ehird> And text boxes are totally non-native
17:30:01 <pikhq> I'm picky about things looking right. ;)
17:30:03 <ehird> I suspect it of merely rendering things that look like native widgets
17:30:07 <ais523> for me, it's not widget look that matters, but behaviour
17:30:08 <ehird> not actually using them
17:30:14 <ehird> ais523: yep, tk gets both wrong
17:30:22 <ais523> e.g. on Gnome I can right-click on a scrollbar tab to scroll to the other end
17:30:32 <ais523> that doesn't work anywhere else I've tried, but it's really useful
17:30:38 <pikhq> I *distinctly* remember a pinstripe background (it was old OS X).
17:30:42 <ais523> (as in, if I right-click on scroll-up, it scrolls ot the top)
17:32:44 <pikhq> It was some "Tcl/Tk Aqua" with batteries included Tcl distribution. Probably some fork to make it act right, then.
17:37:27 <ehird> Nah, it's in mainline, must just have fallen behind
17:37:31 <ehird> Or at least, it comes with OS X
17:41:19 <ais523> you know how in C, if you allocate a variable, it isn't initialized and you get whatever happens to be in memory at the time?
17:41:27 <ais523> you can do that in Perl6 too, but you have to request it explicitly
17:41:51 <ais523> is it the default, though?
17:42:01 <GregorR> No, in D you get zero'd stuff.
17:42:10 <Deewiant> By default integers are zero, floats are nan, chars are all-bits-1
17:42:54 <pikhq> 0xFE? That's an odd choice.
17:42:57 <ais523> I do like the idea of lazy exceptions, though
17:43:13 <pikhq> ais523: That's all bits.
17:43:29 <ais523> not "all bits but one one"
17:43:35 <pikhq> I read that as all bits minus one.
17:43:45 <Deewiant> I don't understand how "all bits" means a numeric value :-P
17:43:48 <pikhq> Okay, 0xFF is slightly less odd.
17:44:01 <pikhq> (I'd just go with 0x00, myself)
17:44:08 <fizzie> So 255 is "less odd" than 254? Odd definition for odd.
17:44:13 <Deewiant> Possibly 0xffffffff for UTF-32
17:44:29 <pikhq> fizzie: 254 just seems arbitrary.
17:44:42 <ais523> the idea behind lazy exceptions is that if something goes wrong when calling a function, it returns a return value that throws an exception if you try to use it
17:45:06 <ais523> but you can look at it to see if it's a genuine return value or a lazy exception
17:45:27 -!- jix has quit (Success).
17:45:39 <ehird> I wonder what Perl 6 would look like if you chopped off the sigils and the semicolons.
17:45:55 <ais523> Perl 6 has both sigils and twigils, though
17:46:03 <ehird> Well, fuck them too, whatever they are
17:46:09 <ais523> with the result that it would be hideously ambiguous if you chopped the sigils off
17:46:15 <ais523> a twigil is sort of a sigil for a sigil
17:46:28 <pikhq> Vaguely close, though.
17:46:43 <ehird> Perl 6 is pretty shit
17:46:50 <fizzie> Deewiant: So how does a non-initialized variable declaration look like?
17:47:01 <ais523> as in, $var is a scalar variable, $.var is a scalar property of a class
17:47:09 <Deewiant> fizzie: <type> variable = void;
17:47:19 <ais523> ehird: Perl6 is just even further into the realms of perlness than perl5, I rather like it
17:47:55 <ehird> Perl 6: http://svn.perl.org/perl6/pugs/trunk/examples/life.pl
17:47:57 <ehird> Ploferl 6: http://pastie.org/467798.txt?key=i3fnmc5yygtz8yly0k1zq
17:49:23 <pikhq> I don't think it's all that Ploflike.
17:49:38 <pikhq> Though I'm sure it could be implemented in PSL.
17:49:46 <ehird> pikhq: It's not too far off
17:50:06 <ais523> that is insanely similar, of course
17:50:15 <ais523> except, I suspect various other bits of Perl6 wouldn't translate
17:50:30 <ehird> 17:50 pikhq: Vaguely similar.
17:50:30 <ehird> 17:50 ais523: that is insanely similar, of course
17:50:36 <ais523> that seems to have been using just the Perl5y bits
17:50:49 <ehird> that's because most of the perl 6 bits are useless
17:50:57 <ais523> what's Plof's OO like?
17:51:01 <ais523> what are its regexps like?
17:51:08 <ehird> Plof's OO is, uh, Javascript.
17:51:15 <ehird> Its regexps are PCRE.
17:51:20 <ehird> But you can only have them in the bnf declaration thing.
17:51:20 <ais523> so really rather different, then
17:51:26 <pikhq> Plof's OO is basically the Javascript method done better.
17:51:34 <ehird> Er, pikhq, it barely differs.
17:51:37 <ehird> Apart from removing constructors.
17:52:26 <pikhq> Also, there's PCRE bindings for Plof.
17:52:57 <ais523> does Plof have typed variables?
17:53:18 <pikhq> Erm. That's not done yet.
17:53:32 <pikhq> ais523, the type system consists of the object system.
17:53:46 <pikhq> Variables are untyped.
17:53:49 <ais523> my guess is that Plof will model a subset of Perl6
17:54:00 <ais523> actually, fwiw, /every language in existence/ will model a subset of Perl6
17:55:33 <pikhq> plofbnf { top = /flimble/ => plof { print("Flimble!\n") } }
17:56:02 <ehird> lifthrasiir: how is clang working?
17:56:09 -!- Hiato has joined.
17:56:18 <lifthrasiir> i was working on other optimization passes.
17:56:23 <ehird> ais523: no, perl 6 misses a lot of things
17:56:36 <ehird> lifthrasiir: i'm upset because i can't think of a way to improve on what you're doing :-)
17:57:16 <ais523> <perl6 spec in svn> my Cat|Dog Fish $mitsy = new Fish but { Bool.pick ?? .does Cat !! .does Dog };
17:57:32 <ehird> MORE SYNTAX! MOOOOOOOOAR SYNTAX!
17:57:37 <ehird> SYNTAX SOLVE EVERYTHING!
17:58:04 * pikhq can add that to Plof
17:58:17 <lifthrasiir> ehird: it only performs a basic optimization. it even didn't remove codes after infinite loop... :S
17:58:29 <lifthrasiir> i just implemented that and pushed to hg repo.
17:58:37 <AnMaster> <ais523> actually, fwiw, /every language in existence/ will model a subset of Perl6 <-- so true...
17:58:37 <ehird> lifthrasiir: it's the most anything does atm
17:58:57 <ehird> turning things into expressions, polynomials, for loops
17:59:14 <AnMaster> <ais523> <perl6 spec in svn> my Cat|Dog Fish $mitsy = new Fish but { Bool.pick ?? .does Cat !! .does Dog }; <-- what does it do...
17:59:45 <ehird> AnMaster: it's a variable that's both a fish and a (cat xor dog)
18:00:03 <ehird> you know. like in real life
18:00:17 <ehird> lifthrasiir: "I implement this interface" I think
18:00:21 <ais523> to be precise, it generates a Fish, then makes a copy of it with either catness or dogness applies, at random
18:00:37 <ais523> ehird: .does is an abbreviation for $_.does
18:00:46 <ais523> *catness or dogness apllied
18:00:48 <ehird> ais523: that's just awful
18:00:59 <ais523> ehird: we were discussing "true but false" earlier
18:01:05 <AnMaster> ais523, sorry, but this is *worse* than intercal
18:01:11 <ehird> AnMaster: no, it's not
18:01:13 <ais523> which in Perl6, is a scalar value identical to true, apart from being false
18:01:16 <lifthrasiir> i thoroughly read perl 6 specs a year ago, but cannot think of it. heck.
18:01:17 <ehird> if you think that you don't know intercal :)
18:01:20 <AnMaster> ehird, this specific aspect of it at least.
18:01:39 <AnMaster> <ais523> which in Perl6, is a scalar value identical to true, apart from being false <-- err.. what
18:01:41 <ehird> you can probably implement comefrom in perl
18:01:45 <ehird> AnMaster: boolean value
18:01:50 <ehird> object True is true as a boolean
18:01:53 <ehird> object False is false as a boolean
18:01:55 <ehird> object 3 is true as a boolean
18:02:04 <ehird> (true but false) is an object identical to True, but it's false as a boolean
18:02:14 <ehird> yeah, it's just an edge-case
18:02:20 <iano> I added return stack access to my False variant
18:02:32 <ais523> AnMaster: makes a modified copy
18:02:36 <ehird> AnMaster: (3 but false) is an object identical to 3 but false as a boolean
18:02:46 <iano> for better stack access and custom control structures
18:02:47 <ais523> that's different in some way, to do with properties or roles or traits or whatever
18:02:48 <ehird> it makes perfect sense...
18:03:00 <ehird> iano: it may be TC
18:03:13 <ais523> it depends on what sort of access
18:03:14 <iano> ehird: exactly
18:03:23 <AnMaster> ehird, So 3 is an object representing 3, but booleans are not objects?
18:03:33 <ehird> AnMaster: you misunderstand me
18:03:40 <pikhq> In Plof: (var tmp = new(True);tmp.ifTrue = Bool.ifTrue;tmp.ifFalse = False.ifFalse;)
18:03:41 <ehird> what booleanity an object has is an aspect of it
18:03:52 <ehird> True is an object which has nothing other than booleanity-true
18:04:07 <ehird> So (true but false) is True but with booleanity-false
18:04:19 <pikhq> Takes a bit more work than that; no opOr, opAnd, or opNot.
18:04:22 <AnMaster> ehird, what about primitive datatypes. Do they exist, or is everything an object.
18:04:23 <ehird> this is rather useless, ofc
18:04:29 <ais523> atm, I'm trying to work out what (true but false) stringifies as
18:04:37 <ehird> AnMaster: i think everything's an object, apart from the array/scalar/hash distinction; ask ai
18:04:49 <ais523> it depends on whether true has its stringification as "true" bundled with it, or whether it's calculated automatically from its truth
18:05:07 <ais523> ehird: actually, in Perl6 everything can be treated as an object or as a nonobject
18:05:07 <ehird> LOL, Jeff Atwood uses a password in the dictionary
18:05:11 <ehird> His idiocy is unbounded
18:05:29 <ais523> to the extent that it's rather difficult to tell if anything is boxed or not at any given moment
18:06:13 <ehird> Jeff Atwood is a retarded microsoft drone who has a shitty blog where he is a moron extraordinaire.
18:06:18 <ais523> isn't Jeff Atwood that Microsoft fan who spends all his time trying to be famous?
18:06:19 <ehird> He owns stack overflow with Joel Spolsky.
18:06:41 <ehird> Joel Spolsky's an idiot but he's said some clever things in his time... Jeff Atwood, uh, not so much.
18:06:49 <pikhq> In Plof: (var tmp = new(True);tmp.ifTrue = Bool.ifTrue;tmp.ifFalse = False.ifFalse;tmp.opAnd = False.opAnd;tmp opOr = False.opOr;tmp opNot = False.opNot;)
18:06:53 <Deewiant> He said it wasn't "really" a dictionary password
18:07:05 <ehird> Deewiant: Yeah it was one of those non-dictionary dictionary passwords
18:07:06 -!- tombom has joined.
18:07:11 <ehird> I like how he calls people dummies in the same sentence
18:07:18 <ehird> Real smart there today Atwood
18:07:25 <pikhq> Or, to make that an operator:
18:07:37 <ais523> beh, Perl6 isn't in the Ubuntu repos
18:07:53 <ehird> ais523: err, rakudo IS the canonical perl6
18:07:59 <ehird> ais523: anyway, is Parrot?
18:08:13 <ais523> but I mean, Perl6 isn't finished yet
18:08:16 <ais523> but Rakudo exists today
18:08:16 <ehird> get parrot, then http://rakudo.org/how-to-get-rakudo
18:08:24 <ehird> Perl6 is not an implementation
18:08:30 <ehird> Rakudo is the "perl.org" implementation of perl 6
18:08:36 <ehird> ie the sort-of-canonical one
18:08:43 <ehird> ais523: anyway, get parrot then http://rakudo.org/how-to-get-rakudo
18:08:45 <ais523> there can't be an implementation of Perl6 yet, because it isn't finished; Rakudo presumably implements the work-in-progress spec
18:09:03 <ehird> ais523: don't get parrot
18:09:04 <ehird> it does it for you
18:09:16 <ehird> git clone git://github.com/rakudo/rakudo.git && cd rakudo && perl Configure.pl --gen-parrot && make
18:11:08 <ehird> ais523: I would like to point out that Perl 6's other name is PEYHM.
18:11:32 <ais523> no results for that either
18:11:48 <ehird> ais523: I'm not being too serious: http://www.pugscode.org/images/pugs.small.png
18:11:57 <ehird> It's more like Pugs' other name for PErl 6, though.
18:12:24 <ais523> aww, Perl6 no longer allows ' as a namespace separator
18:12:28 <pikhq> opBut = (x, y as Bool) {var tmp = new(x) : Bool;tmp.ifTrue = y.ifTrue;tmp.ifFalse = y.ifFalse;tmp.opAnd = y.opAnd;tmp.opOr = y.opOr;tmp opNot = y.opNot;}
18:12:29 <ais523> admittedly, that's been deprecated for ages
18:12:32 <pikhq> plofbnf { plof_unopr => plof_but; plof_but = plof_but "but"w plof_but_next => plof {opBut($0, $1)}; plof_but = plof_but_next => plof { $0 };}
18:12:41 <pikhq> And yes, I am completely mad.
18:12:56 <ehird> ais523: done that git/cd/perl/make dance to get rakudo?
18:12:58 <ehird> it handles everything for you
18:13:02 <ais523> ehird: yes, it's running atm
18:13:08 <ais523> I checked the commands and the website first, thuogh
18:13:48 <ehird> --gen-parrot is code for "launch nuclear weapons", obviously
18:15:27 <ehird> "its performance is excellent (in terms of both compilation speed and - literally - execution speed)"
18:16:08 * ehird looks at Rakudo building and decides to look away because it's probably awful and bloated and meta
18:16:10 <ais523> I think we're referencing the same website
18:16:25 * pikhq is pleased at having implemented the but operator in Plof)
18:16:36 <ais523> and the build's mostly full of cc and parrot invocations
18:16:45 <ais523> pikhq: but it doesn't do all of what Perl6's but does
18:17:07 <ehird> 18:16 ais523: I think we're referencing the same website ← ofc
18:17:17 <pikhq> That takes more than two lines of silly stuff hashed out in an IRC buffer.
18:17:20 <ais523> can you make it apply roles to classes?
18:17:41 <ais523> that's probably its major use
18:17:47 <pikhq> No, that takes a more sophisticated opBut.
18:17:48 <ehird> Leeeeeeeeet's gooooooooooo!
18:17:58 <ais523> although I look forward to testing things like "4 but Callable
18:18:00 <pikhq> The syntax definition there is just fine, though.
18:18:13 <ehird> Could not find non-existent sub false
18:18:19 <ehird> there is no "false"
18:18:31 <ehird> The but operator can only be used with a role or enum value on the right hand side
18:18:33 <ais523> > say (true but false)
18:18:34 <ais523> Could not find non-existent sub false
18:18:36 <ais523> > say (True but False)
18:18:37 <ais523> The but operator can only be used with a role or enum value on the right hand side
18:18:39 <ais523> > say (True but false)
18:18:40 <ais523> Null PMC access in isa()
18:18:49 <ehird> > say (4 but Callable)
18:18:54 <ehird> (4 but Callable)()?
18:19:01 <ehird> ais523: invoke() not implemented in class 'Integer'
18:19:11 <ehird> this thing is slow
18:19:18 <ehird> i can feel the delay from typing to error
18:19:44 <ais523> rakudo feels rather buggy to me
18:19:51 <ais523> compared to the spec, at least
18:20:32 <ehird> will try the rakudo tests and the perl6 tests from pugs
18:21:36 <ehird> All tests successful.
18:21:36 <ehird> Files=29, Tests=236, 56 wallclock secs ( 0.16 usr 0.12 sys + 47.38 cusr 3.89 csys = 51.55 CPU)
18:21:51 <ehird> /Users/ehird/Downloads/rakudo/parrot/parrot perl6.pbc --target=pir --output=Test.pir Test.pm
18:22:41 <ehird> "This is one of our favourites. The value of NULL on a DeathStation 9000 varies from one compile to the next. It might be all-bits-zero and it might not."
18:22:52 <ehird> "So this code: memset(&weapon, 0, sizeof weapon); wasn't quite as robust as it might have been"
18:22:56 <ehird> 0 as a pointer is always NULL
18:23:00 <ehird> that's correct code
18:23:03 <ehird> THE DEATHSTATION IS WRONG!
18:23:10 <ais523> the second argument of memset isn't a pointer
18:23:19 <ais523> 0 as a char != 0 as a pointer, necessarily
18:24:17 <ehird> ais523: all these spectests are succeeding atm
18:24:24 <ehird> bsmntbombdood_: hi
18:24:34 <ais523> ehird: I doubt they're in sync with the actual spec, though
18:24:45 <ehird> ais523: they're reasonably in sync.
18:24:50 <ehird> the spec doesn't change much
18:24:51 <ehird> mostly it's hot air
18:26:39 <ais523> heh, Perl6 doesn't just have unary plus, it also has unary concatenate
18:28:38 <lifthrasiir> http://pastie.org/467828 esotope-bfc is, i think, getting much slower over the revisions... :S
18:28:58 <ehird> lifthrasiir: how does it do lostkng?
18:29:05 <ais523> Statement not terminated properly at line 1, near "\x{c2}\x{ab}b\x{c2}\x{bb}).per"
18:29:27 <ais523> the error message strongly implies I passed correct UTF-8 to it
18:29:36 <lifthrasiir> ehird: given link has a table of time taken to compile lostkng.
18:29:40 <ehird> ais523: tell it you're using utf-8, then
18:29:45 <ehird> lifthrasiir: but what's the output
18:29:47 <ais523> maybe it's trying to interpret my utf8 as latin-1, for some reason
18:30:05 <ehird> lifthrasiir: perhaps!
18:30:07 <ehird> eh, just like to esotope
18:30:35 <ais523> $ perl6 --encoding=utf8
18:30:38 <ais523> Statement not terminated properly at line 1, near "\x{c2}\x{ab}b\x{c2}\x{bb}).per"
18:30:39 <ehird> link to esotope anyway lifthrasiir :-P
18:30:55 <ehird> ais523: i highly doubt it doesn't support that syntax; it's something else
18:31:05 <lifthrasiir> ehird: http://hg.mearie.org/esotope/bfc/raw-file/tip/esotope-bfc.py
18:31:46 <ais523> $ LC_ALL=en_US.UTF-8 perl6 --encoding=utf8
18:31:49 <ais523> Statement not terminated properly at line 1, near "\x{c2}\x{ab}b\x{c2}\x{bb}).per"
18:31:50 <lifthrasiir> when it once became stable i'll release it to avoid pasting that link everyday. :p
18:32:01 <ais523> if Rakudo has some way to indicate an encoding, it isn't an obvious one
18:32:28 <ehird> lifthrasiir: ImportError: cannot import name namedtuple
18:32:36 <ais523> ehird: "Running "make spectest" will import relevant portions of the official Perl 6 test suite from the Pugs repository (http://svn.pugscode.org/pugs/t/spec/) and run all of the tests that are currently known to pass."
18:32:46 <ais523> so it only tests the bit they've implemented against the spec
18:32:49 <ais523> no wonder it was passing
18:32:59 <ehird> shrug, rakudo is quite complete
18:33:04 <ehird> you seem to think it's a lot worse than it i
18:33:11 <lifthrasiir> ehird: sorry, namedtuple is not in use in current revision. remove "from collections import namedtuple" line from the code.
18:33:24 <ais523> AnMaster: Perl6 work-in-progress bytecode compiler
18:33:39 <ehird> how come you're the only person who gets the privilege to ask questions that you could easily find out, nobody else?
18:34:20 -!- iano has quit.
18:34:34 <ehird> % python esotope-bfc.py LostKng.b > LostKng.b.c
18:34:38 <ehird> lifthrasiir: this is taking >11s...
18:35:21 <ais523> > my Int $x = (fail "testing");
18:35:24 <ais523> Scope not found for PAST::Var '$x' in
18:35:25 <ais523> I'm finding bugs everywhere
18:35:28 <ais523> and yes, that is the entire error message
18:35:36 <ehird> ais523: Are you sure you know the spec more than they do?
18:35:40 <ehird> I bet your code is invalid
18:35:41 <AnMaster> <ehird> how come you're the only person who gets the privilege to ask questions that you could easily find out, nobody else <-- What are you talking about.
18:35:52 <ehird> lifthrasiir: doesn't seem to do much more than bf2c.hs on lostkng
18:35:58 <ais523> ehird: I have the spec open atm...
18:36:06 <ehird> ais523: report the bugs then
18:36:15 <ehird> lifthrasiir: http://esoteric.sange.fi/brainfuck/impl/compilers/bf2c.hs
18:36:17 <ais523> it's possible that the REPL runs each line in a separate scope
18:36:23 <ehird> it says that on the page
18:36:29 <ehird> the get rakudo page
18:36:34 <ehird> each line is a separate compilation unit
18:36:37 <ehird> so subs persist, but not vars
18:36:44 <ehird> lifthrasiir: until recently the best compiler
18:36:53 <ehird> lifthrasiir: does move-shifting, variable elimination, polynomialization...
18:37:00 <ehird> BF optimization is a field in itself...
18:37:28 <ais523> > {my Int $x = (fail "testing"); say $x+1;}
18:37:38 <ais523> presumably, attempting to output an unthrown exception doesn't throw it
18:38:14 <ais523> well, yes, but they're meant to throw themselves whenever you try to do something with them that's obviously intended not to be done with an exception
18:38:59 <ehird> lifthrasiir: does your compiler work on [>]?
18:39:33 <ehird> lifthrasiir: what, it fails on valid code?
18:39:35 <lifthrasiir> i'm thinking about how to optimize such code.
18:39:42 <ehird> i just mean does it work :-P
18:39:46 <ehird> lifthrasiir: also, [>] is pretty easy
18:40:11 <lifthrasiir> of course it does work on every valid code, just not get optimized
18:40:14 <ehird> lifthrasiir: [>] is mptr = memchr(mptr, 0, (size ofmemptr))
18:40:24 <lifthrasiir> and you're right, but how about [>>] or variants?
18:40:24 <ehird> which is mptr-mem, I think
18:40:47 <lifthrasiir> i think combined array analysis is needed for complete optimization.
18:40:56 <ehird> and that sounds hard :-)
18:41:37 <lifthrasiir> for example the memory cells 7+2*k is actually one array, and optimized to be stored as like etc.
18:42:05 <AnMaster> <ehird> lifthrasiir: [>] is mptr = memchr(mptr, 0, (size ofmemptr))
18:42:11 <AnMaster> yes, I said that yesterday iirc
18:42:35 <ehird> lifthrasiir: yeah, that'd make it turn into "real" c
18:42:40 <ehird> if you analyzed data structures
18:42:44 <ehird> to get variables, arrays, etc out of it
18:42:48 <tombom> what'#s the best compil;er nowtrhen
18:42:55 <ehird> tombom: esotope-bfc
18:43:15 <ehird> tombom: & compiling the generated c with clang/llvm
18:43:32 <lifthrasiir> i'm not sure, but it does certain degree of optimization
18:43:40 <ehird> lifthrasiir: you should optimize the 1,X,1,X,1,X,0 array format
18:43:57 <ehird> AnMaster: console(v)
18:44:11 <GregorR> Ah yes, computer science has entirely destroyed the previous meaning of the word console :P
18:44:56 <GregorR> (OK, I guess consoles in the noun sense predate computers by a bit)
18:45:55 <lifthrasiir> ehird: btw, pastie says "Your paste cannot be larger than 100 kb. Sorry." so is this a new goal? :p
18:45:56 <Deewiant> Where 'a bit' = a few centuries, yese
18:45:57 <ais523> hah, IBM's offering companies $8000 for each Sparc processor they replace with a Power processor
18:45:59 <AnMaster> GregorR, what is the name for those triangular things you put under shelves to mount them on the wall
18:46:16 <AnMaster> but I don't remember if it is console on English as well
18:46:20 <ehird> i forgot people still use non-x86s in non-embedded environment
18:46:32 <GregorR> I'm not sure what you're referring to.
18:46:47 <ais523> AnMaster: they're normally called shelf brackets in English
18:46:55 <GregorR> It might be that Swedish furniture technology is advanced to the point that you have words for things we don't even have.
18:46:59 <ais523> not that I often need to call them anything
18:47:32 <AnMaster> ais523, nor do I. I don't think I have any shelves mounted that way in the house...
18:47:48 <ais523> I do, but have never felt a need to name the things holding the shelf up
18:49:30 <AnMaster> ais523, well, if you are going to buy some it might be useful to know
18:50:13 <ais523> ah, first google result for "shelf brackets": http://www.wickes.co.uk/Shelf-Brackets/Metal-Brackets/icat/tsmetalbrack
18:50:16 <ehird> GregorR: yeah ikea is like the LHC
18:51:32 <AnMaster> ais523, right. Was thinking of larger ones in wood though. Usually triangular. Sometimes ornamented(sp?)
18:51:45 <ais523> they're all shelf brackets either way, though
18:51:51 <ais523> it seems wickes only make the metal ones
18:53:13 <ehird> AnMaster: remember when I showed you a cooler which you thought was mad because it was big and heavy and would fall?
18:53:15 <ehird> AnMaster: cpu cooler
18:53:22 <ehird> AnMaster: here's the cooler I'm currently planning on using: http://www.silentpcreview.com/files/images/prolima-megahalems/12.jpg
18:53:35 <ehird> women and children, shield your eyes
18:53:58 <AnMaster> ehird, it looks like it would put quite a large load on the mobo
18:54:12 <AnMaster> especially if the mobo is mounted vertically
18:54:14 <GregorR> "<ehird> women and children, shield your eyes" // must just be a giant penis you attach to your CPU
18:54:28 <ehird> AnMaster: the attaching system is pretty solid
18:54:58 <ehird> AnMaster: "A backplate is placed beneath the CPU socket and bolts with threads on both sides are secured to it." "Two aluminum side bars are placed over the bolts and nuts are used to tighten them. There are a second set of holes on these bars for use with the LGA1366 backplate." "The heatsink is then placed on top of the CPU and a crossbar is fitted above the mounting plate. Large spring-loaded bolts are then screwed into the side bars. This is the only
18:55:00 <ehird> step that requires any tools — the rest is done by hand."
18:55:05 <AnMaster> ... Ok ... I never understood what this "That's what she said" was about.
18:55:17 <AnMaster> is it just random xkcd reference, or?
18:55:18 <ais523> > my $foo = "Bar"; my $Bar=4; say $::("MY::$foo");
18:55:20 <ais523> say requires an argument at line 1, near " $::(\"MY::"
18:55:23 <ehird> AnMaster: it's a sexual innuendo
18:55:27 <ais523> ehird: I just copied that example from the spec
18:55:33 <GregorR> AnMaster: It's meant to suggest that something is sexual innuendo when it really isn't.
18:55:45 <ehird> AnMaster: "That's so small." "That's what SHE said [in bed, referring to your penis]." ← example
18:56:00 <GregorR> AnMaster: Since I said the cooler must look like a giant penis, I decided that "<ehird> AnMaster: the attaching system is pretty solid" must be innuendo.
18:56:08 <ais523> I wonder if anyone's ever tried that riposte to a woman?
18:56:11 <AnMaster> ehird, backplate hm.. Hope it isn't a conductive one.
18:56:20 <ehird> AnMaster: those megahalems are 820g with the crossbar, bolts and fan clips, apparently
18:56:26 <ehird> 790g for just the heatsink
18:56:41 <ais523> that's in terms of weight, presumably
18:57:22 <AnMaster> ais523, isn't it upper case G for acceleration iirc
18:57:25 <ais523> I thought so, I wasn't quite sure there wasn't a second meaning though
18:58:11 <ais523> it might be reasonable propulsion caused by a nuclear explosion
18:58:17 <ehird> hmm, apparently with one fan on the megahalems running at only 800 RPM, with full load of a top of the range core 2 quad (NOT i7), it runs at 62c
18:58:31 <ehird> so probably 65-70c for an i7 at full load
18:58:33 <AnMaster> ais523, anything involving a (former) human being
18:58:40 <ais523> is megahalem a trade name, or does it have a technical meaning?
18:58:44 <ehird> question is if you can drop the fan
18:59:01 <ehird> AnMaster: "Megahalems" is the brand of those giant heatsinks for the Core i7s, which use the Nehalem architechture
18:59:24 <ais523> AnMaster: Did you mean: Did you mean: ais523:
19:00:08 <ehird> the megahalems are a lot less hot in silentpcreview's review
19:00:16 <AnMaster> ais523, no. Shouldn't there be a dot at the end
19:00:28 <ais523> AnMaster: ehird said "AnMaster:"
19:00:37 <ais523> which is what you were trying to correct in the first place
19:00:38 <AnMaster> ais523, I perform word splitting on :
19:00:39 <ehird> so I'm assuming that dropping the fan would be easy
19:00:49 <ais523> AnMaster: I give context in corrections
19:00:56 <ehird> you'd get 30-40C or so on average and higher on load
19:01:06 <ais523> just like diff -u >> diff for patching programs
19:01:26 <AnMaster> ais523, ok. I should have done /<AnMaster> Did/s/: /: $/
19:02:01 <ais523> replace end-of-string with start-of-string?
19:02:05 <ais523> does that even make sense?
19:02:25 <AnMaster> <AnMaster> ais523, ok. I should have done /<AnMaster> Did/s/: /: $/
19:02:27 <ehird> ugh, I can't believe I'm actually considering applying my own gigantic heatsink
19:02:28 <AnMaster> <AnMaster> ais523, ok. I should have done /<AnMaster> Did/s/: /: ^/
19:02:36 <AnMaster> <AnMaster> Did you mean: ais523
19:02:40 <ais523> I don't get why you want a ^ in the replacement anyway
19:02:40 <AnMaster> <AnMaster> Did you mean: ^ais523
19:02:50 <ais523> yes, but surely ehird didn't mean "^ais523"
19:03:06 <ais523> oh, you're trying to anchor the context to the start of the line/
19:03:12 <ais523> lines start with : not ^ on IRC
19:03:12 <ehird> "In keeping with the spirit of the [H] we are once again doing hardware heat measurment. This means drilling a very small path into an expensive CPU to place our thermocouple in. This is by far the best way to test coolers and the only way here at the [H]." Holy shit, they're insane
19:03:19 <ehird> http://enthusiast.hardocp.com/image.html?image=MTIzMjU1MjM1MlEyZ1NFQzg5ckJfMV81X2wuanBn
19:03:21 <ais523> so s/^/:/ in your most recent metacorrection
19:03:26 <ehird> They cut open a cpu and put a wire into it
19:03:34 <AnMaster> ais523, not the way the client displays it
19:03:43 <AnMaster> <AnMaster> Did you mean: :ais523
19:03:45 <ais523> AnMaster: for client display, you want >
19:03:57 <ais523> unless your client displays IRC messages as regexps?
19:05:46 <ais523> well, I was deliberately trying to make it absurd
19:09:39 <ehird> "The fan completely blocks any RAM from being installed in the first slot and even the second slot blocks the retention clip from properly holding the fan. "
19:09:46 <ehird> Hope my mobo's big enough to avoid THAT.
19:12:25 -!- BeholdMyGlory has joined.
19:14:15 -!- AnMaster_ has joined.
19:16:14 <ehird> lifthrasiir: test with mandelbrot.b
19:16:20 <ehird> lostkng is big but not intensive
19:16:38 <lifthrasiir> is that http://www.menuetos.net/mandel.txt ?
19:16:54 <ehird> lifthrasiir: http://swapped.cc/bf/
19:17:00 <ehird> http://swapped.cc/bf/files/mandelbrot.b
19:17:21 <ehird> also, ew, menuetos :-P
19:24:01 -!- AnMaster has quit (Connection timed out).
19:26:52 -!- AnMaster_ has changed nick to AnMaster.
19:27:20 <ais523> Scope not found for PAST::Var '$x' in
19:28:26 <Deewiant> AnMaster: I forwarded you an e-mail earlier today, FWIW.
19:30:02 <AnMaster> " Your interpretation would be considered undefined at best. The specs say that the child should execute before the parent next executes but does not specify immediately before, only that it must execute an instruction before the parent executes its next instruction. In fact it does not even state whether the child ip needs to execute in the same tick as the t command that created it, or execute it
19:30:02 <AnMaster> for the first time in the next tick. Both CCBI and Rc/Funge-98 execute it the next tick, but again, which tick it executes in is undefined and only really has importance when using TRDS. "
19:30:34 <AnMaster> if we are going to go undef like that almost half of the GOOD in mycology will turn into UNDEF
19:31:14 <ais523> I'm pretty certain that process creation taking more than one tick would be a DS9K implementation of Befunge-98
19:31:44 <ais523> oh, what is it about then?
19:32:08 <ehird> AnMaster: who wrote that?
19:32:17 <ais523> almost cetainly mike riley
19:32:22 <AnMaster> ehird, do you agree with it or not
19:32:23 <ais523> although there isn't any direct evidence for that
19:32:25 <ais523> just an implication from contetx
19:32:33 <ehird> he writes better not on IRC then
19:32:40 <Deewiant> Well, the original message was that if you have, say IPs [ 1 2 3 ], then if 2 forks to create 4 you should get [ 1 2 4 3 ] and not [ 1 2 3 4 ]
19:32:47 <ehird> maybe his irc client inserts the ,,,, ... and ehhehehehehs
19:33:04 <ehird> Deewiant: yep, that's right
19:33:18 <ehird> yes but it's the right option :-)
19:33:20 * AnMaster tries to figure out DS9K interpretations for the fingerprint acryonyms
19:33:47 <ehird> FPSP: Fire-Powered Shattering Pulser
19:33:58 <Deewiant> And Mike's "it does not even state" is about whether, when the child IP is created, it executes for the first time in that same tick or the next one
19:34:01 <ehird> but with more fire and explosion
19:34:52 <ehird> MVRS: Multiverse tools, including D: destroys the current universe.
19:35:55 <ehird> REXP: Rectal Excavation eXtra Powertools
19:36:04 <ehird> Used in hospitals to control a rectal excavation machine.
19:36:28 <ehird> "SOCK": Securely Operated Checkless K-Bolts
19:36:38 <ehird> SUBR: Submarine control system.
19:36:54 <ehird> AnMaster: a weapon of some sort; it sounds like a long, thin bolt that's highly explosive or somethin
19:37:02 <ais523> AnMaster: I think they're explosive bolts used to hold parts of rockets together
19:37:03 <ehird> it doesn't actually exist, but that's what it sounds like
19:37:06 <ehird> ais523: what, really?
19:37:10 <ehird> i just invented them!
19:37:10 <ais523> which are blown up to separate the stages
19:37:16 <ais523> ehird: they may be called something else
19:37:23 <ehird> http://www.google.com/search?client=safari&rls=en-us&q=K-bolt&ie=UTF-8&oe=UTF-8
19:37:26 <ehird> yeah, I just invented them ;-)
19:37:37 <ehird> TRDS: controls the *real* TARDIS.
19:37:54 <ais523> ehird: wow, that was really weird
19:37:55 <ehird> UNIX: Castration utilities.
19:37:59 <ais523> I followed your link, and only got one advert
19:38:04 <ais523> then I removed the client=safari, and got lots
19:38:13 <AnMaster> ehird, err. That isn't a backronym
19:38:19 <ehird> if i s/safari/firefox/ i get more
19:38:24 <ehird> i get 2 with safari
19:38:26 <ais523> with client=epiphany, I get 2
19:38:38 <ehird> AnMaster: How is that fail?
19:38:40 <ais523> and client=lynx gives 6
19:38:45 <ehird> I know trds means tardis.
19:38:45 <ais523> I tried it before you suggested that
19:39:35 <ais523> and ie7 shows none at all
19:39:44 <ais523> wait, or 1 on the refresh
19:39:49 <ais523> I suspect it's just giving random numbers of ads
19:40:03 <ais523> nope, ie7 never gives more than one ad
19:40:28 <ais523> and ie8 is not giving more than one now either, even after about 8 refreshes
19:41:08 -!- oerjan has joined.
19:41:28 <AnMaster> ais523 or is it statistical significant?
19:43:49 <ais523> it probably depends on other things like the IP
19:44:02 <ais523> I mean, Google have a user-agent, why would they care about the useragent specified in the address bar?
19:44:37 -!- tombom_ has joined.
19:44:40 <AnMaster> why is the user-agent even in the address bar
19:47:14 <ehird> AnMaster: done with JS I think
19:47:17 <ehird> to avoid server side load
19:47:58 <AnMaster> ehird, err. How does it have to parse less that way
19:48:10 <ehird> AnMaster: it doesn't
19:48:14 <ehird> AnMaster: but they're already parsing the query string
19:48:21 <ehird> they're probably just throwing away the user-agent header
19:48:25 <ehird> and besides, it's easier to read
19:48:34 <ehird> Internet Explorer Mozilla AWESOMEFEST (aaa;;:43847 NETSCAPE)
19:48:45 <ehird> AnMaster: that's part of it
19:49:07 -!- tombom has quit (Read error: 60 (Operation timed out)).
19:49:08 -!- tombom_ has changed nick to tombom.
19:50:02 <ehird> AnMaster: i've seen worse; a particular idiot thought that including the highest Acid standards test they pass would be best... even though the acid tests are 100% edge cases, and are DELIBERATELY INVALID to test browsers' handling of invalid pages
19:50:10 <ehird> as in, just having the UA be "Acid 2" or "Acid 3"
19:50:27 <ehird> the acid tests aren't even official w3c!
19:50:40 <AnMaster> ehird, err acid 1 doesn't have any invalid bits iirc
19:51:01 <ehird> AnMaster: barely anything doesn't pass acid 2 these days
19:51:04 <ehird> but my point stands
19:51:22 <Deewiant> My Firefox (3.5b4) doesn't pass Acid2
19:51:33 <Deewiant> I haven't tried in safe mode, though, so it could be an extension
19:51:47 <AnMaster> I guess it might be a beta bug
19:51:56 <ehird> acid tests don't matter anywhere
19:51:59 <ais523> Deewiant: Acid2, or Acid3?
19:52:02 <ehird> a load of talk and nothing useful or worthwhile
19:52:05 <Deewiant> The nose is a pixel or so too big and the chin is one block too tall
19:52:19 <Deewiant> Even the latest alphas don't pass Acid3.
19:53:08 <Deewiant> With extensions on, my browser doesn't pass Acid1.
19:53:23 <AnMaster> Deewiant, you have odd extensions then
19:53:44 <Deewiant> I refuse to allow pages to set the font size of monospaced fonts.
19:54:19 <Deewiant> Because they are invariably either way too small or way too big, because they assume a different font than I specify.
19:55:44 -!- FireFly has quit ("Later").
19:59:27 -!- FireFly has joined.
20:00:02 <pikhq> I refuse to allow pages to set the font.
20:00:22 <pikhq> You've got a choice between sans serif, serif, and monospace.
20:00:58 <pikhq> AnMaster: Someone using that gets shot.
20:01:20 <ehird> 20:00 pikhq: I refuse to allow pages to set the font. ← My designer sense is tingling, and it wants to throttle you.
20:01:35 <ehird> Please assume a relaxed position so that it may commence.
20:02:05 <Deewiant> I used to have that on but it broke too many pages
20:02:14 <ais523> IMO, designers should design pages so that with any reasonable sane settings for the user, it'll display as the user wants
20:02:59 <Deewiant> IMO designers shouldn't touch the 'font-foo' CSS settings at all.
20:02:59 <ehird> OTOH, a font with different metrics fucks up the whole vertical/horizontal rhythm entirly.
20:03:21 <ehird> Which is not a problem unless you actually, y'know, design.
20:03:29 <ehird> When it all breaks down.
20:03:35 <ehird> Deewiant: Agreed wholeheartedly.
20:03:44 <ehird> Irrelevant in this case, however.
20:03:45 <AnMaster> ehird, you can't assume the user will have any of the fonts
20:03:52 <ehird> AnMaster: That's why you specify fallbacks.
20:04:03 <AnMaster> you can only assume there is some sans-serif, some serif, some monospace
20:04:18 <ehird> you can't assume anything
20:04:26 <pikhq> ehird: I've done this because far, *far* too many 'designers' misuse it.
20:04:41 <AnMaster> ehird, well true. But lynx doesn't support CSS anyway :P
20:04:44 <ehird> pikhq: They ruin it for everyone else :P
20:04:48 <AnMaster> so all you can assume is "a font"
20:04:59 <pikhq> AnMaster: elinks supports CSS. ;)
20:05:52 <pikhq> ehird: What I'd like is a way to refuse allowing sites to set the font, with a whitelist for sites that aren't really retarded with font selection.
20:06:25 <AnMaster> pikhq, make an extension.. called nofont
20:07:21 <pikhq> Very damned tempting.
20:07:41 <pikhq> Opposite hands. ;)
20:09:15 <Deewiant> Since you'd have to whitelist the site anyway to check whether it's retarded or not :-P
20:09:58 -!- olsner has joined.
20:11:53 <fizzie> Tempt-ping; it sends seductive ICMP messages.
20:13:16 <AnMaster> Deewiant, why not have a setting for "default allow/disallow"
20:17:27 <Deewiant> Because that complicates things
20:17:33 <Deewiant> You have to keep track of both a blacklist and a whitelist
20:17:42 <AnMaster> Deewiant, yes that allows exceptions to0o
20:17:44 <Deewiant> Unless you just want to wipe it or invert its meaning whenever the setting is changed :-P
20:21:52 -!- Hiato has quit ("Leaving.").
20:23:42 -!- kar8nga has joined.
20:26:58 <AnMaster> Deewiant, just add the complement to the list when the option is changed ;P
20:40:01 <oerjan> !underload (S):(u)(:*)(:*)::**^^(re :-P)**~^
20:40:01 <EgoBot> Suuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuure :-P
20:40:20 <ais523> heh, a link someone pasted on Slashdot to try to explain the Daily Mail to Americans: http://www.google.co.uk/search?q=site%3Adailymail.co.uk+%22ministers+are+considering%22
20:40:49 <ais523> oerjan: I'm sure that program could be simplified
20:41:00 <ais523> actually, that must have been deliberately obfuscated
20:41:05 <ais523> or you would have used more than one S
20:41:47 <oerjan> once i realized (S)S could be rewritten, the rest was irresistible
20:42:35 <oerjan> oh of course (:*)(:*) = (:*):
20:42:47 <ais523> and :::** will boggle people somewhat
20:44:03 <pikhq> !underload (S):(u)(:*):::**^^(re...)**~^
20:44:04 <EgoBot> Suuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuure...
20:44:42 <oerjan> hm in fact that's a nice compression of 256
20:45:04 <ais523> (:*):::**^ is the compression
20:45:14 <ais523> the second ^ is using the 256 to duplicate the u
20:45:20 <pikhq> !underload (:*):::**^
20:45:44 <pikhq> Church numerals...
20:45:52 <pikhq> Make my head hurt.
20:46:04 <ais523> I like Church numerals
20:46:25 <AnMaster> I think we need yet another odd number system
20:46:25 -!- tombom has quit ("Peace and Protection 4.22.2").
20:46:34 <AnMaster> I liked that idea I had some days ago
20:46:50 <AnMaster> that the interpreter selected the minimal possible base to interpret the number in
20:46:54 <pikhq> ais523: They're brilliant, they just screw with my head.
20:46:58 <AnMaster> so 30 was in base 4, 45 was in base 6
20:47:04 <oerjan> !underload (X):*:*:*(~S:^):^
20:47:04 <EgoBot> XXXXXXXXError: Stack underflow in ~
20:47:15 <fungot> XXXXXXXX ...out of stack!
20:47:35 <ais523> they seem to be disagreeing about what in particular underflowed
20:47:41 <ais523> EgoBot and thutubot, that is
20:47:48 <ais523> AnMaster: "underflow", not "overflow"
20:48:00 <oerjan> AnMaster: i just wanted to check if it would print anything before the error
20:48:28 <oerjan> so you can use that for printing the whole stack like in the others
20:48:37 <AnMaster> ais523, where should it really have underflown
20:49:11 <ais523> if a ~ succeeds, the subsequent S can't possibly underflow
20:49:16 <ais523> because there must be at least two stack elements
20:49:41 <ais523> or to be more precise, a bug in its error-handling
20:50:59 -!- kar8nga has quit (Remote closed the connection).
20:51:13 <ais523> oh, it didn't say ...' out of stack
20:52:42 <fizzie> fungot doesn't say what underflows.
20:52:42 <fungot> fizzie: " fnord?" she inquired. so, as i thought good, ' as many as five nights fnord warmth, you know.'
20:52:53 <fizzie> It's just a generic "out of stack" message.
20:52:57 <ais523> I suspect it has some sort of underflow token that's being swapped onto the stack
20:53:03 <ais523> and you get an underflow error when you try to use it
20:53:13 <ais523> I didn't do that deliberately, but thutubot's complicated enough that I might have done it by mistake
20:55:25 <AnMaster> <ais523> I suspect it has some sort of underflow token that's being swapped onto the stack
20:55:25 <AnMaster> <ais523> and you get an underflow error when you try to use it
20:55:38 <oerjan> seems only ~ gets fooled, and only for the second element
20:55:42 <ais523> yes, but that's not the same as knowing how it works
20:56:04 <AnMaster> ais523, so how did you detect underflow then
20:56:08 <oerjan> so probably ~ just checks for one argument
20:56:50 <ais523> AnMaster: if the program underflows, none of the regexps match, so it falls out of the main loop unexpectedly
20:56:55 <oerjan> ah * has the same bug :D
20:57:10 <AnMaster> what does * do then. I don't remember
20:58:32 <oerjan> those are the only two-argument commands
20:59:44 <oerjan> interesting that * doesn't mess up more though, since it actually combines the two args rather than just swapping them
21:01:13 <ais523> my current guess is that * and ~ both do the same thing as ! if there's only one element on the stack
21:01:23 <ais523> which would be weird, but not completely implausible
21:01:57 <AnMaster> ais523, what does ! do now again...
21:03:12 <oerjan> +ul (!):( drops)~^( accidentally)SS
21:03:53 <AnMaster> looks like it dropped certainly :P
21:04:11 <oerjan> yeah the accident was in the printing part
21:04:15 <AnMaster> ais523, is there any underload in underload implementation
21:04:54 <oerjan> +ul ((Underload in Underload)S)^
21:04:55 <AnMaster> ^ul (!):( drops)~^( accidentally)SS
21:05:12 <AnMaster> oerjan, well that is cheating... I meant a non-eval implementation
21:05:16 <oerjan> SS should have been *S
21:05:23 <oerjan> AnMaster: underload has no input
21:05:41 <AnMaster> but surely you could embed the input in some way
21:05:44 <AnMaster> ^ul (!):( drops)~^( accidentally)*S
21:05:48 <AnMaster> +ul (!):( drops)~^( accidentally)*S
21:05:58 <AnMaster> !underload (!):( drops)~^( accidentally)*S
21:06:08 <ais523> AnMaster: there's an Underload interp in BF, and a BF-minus-input to Underload compiler
21:06:13 <ais523> you could trivially combine the two
21:06:32 <oerjan> you could code it in binary with : and ^, like i did with the rule 110 automaton
21:06:53 <AnMaster> but you couldn't do it as a base64 encoded string.
21:07:05 <AnMaster> Since strings in underload are more like atoms
21:07:28 <oerjan> yep, no way of decoding except by running
21:07:52 <ais523> Underlambda takes that further, incidentally, there's no way to do anything with information in Tier 1 except by running it
21:08:08 <ais523> if you want to get at the details of the code, you load a library that puts metadata into things
21:08:11 <AnMaster> that is irritating. Makes it require more thought to be sure if it is TC or not.
21:08:12 <ais523> it's a lot cleaner that way
21:08:29 <oerjan> um wait was it : and ^ i used or was that the other option i gave up
21:08:31 <ais523> AnMaster: it's pretty easy to compile SKI into Underload
21:09:00 <ais523> Underlambda just uses the function from stacks to stacks as its basic data type
21:09:06 <ais523> so it's very purely concatenative
21:09:12 <ais523> whereas Underload is slightly dirtier due to S
21:09:27 <ais523> (in Underlambda, S serialises a function, but the form of the serialisation is implementation-defined)
21:09:31 <oerjan> yeah it was : and ^, the other was : and a i think
21:09:44 <AnMaster> ais523, I could compile befunge into underload easily if there is no input...
21:10:09 <ais523> I would have thought it was impossible due to ?
21:10:14 <AnMaster> OUTPUT="$(./cfunge $1)"; echo "($OUTPUT)S"
21:10:16 <ais523> although I suppose you could stick a PRNG in there somehow
21:10:19 <oerjan> AnMaster: those are as far as i can tell the only pairs of commands whose sequences can be cleanly decoded
21:10:28 <ais523> also, that fails on infinite loops
21:10:36 <AnMaster> ais523, true. But it was a joke
21:10:43 <ais523> it was an /incorrect/ joke, though
21:10:55 <ais523> besides, what if there are parens in Befunge's output?
21:10:59 <ais523> you didn't even escape properly
21:11:28 <AnMaster> it was a "initial import" rather
21:11:59 <Deewiant> ehird: http://www.tgdaily.com/html_tmp/content-view-42283-135.html
21:12:39 <ehird> Deewiant: http://www.ramsan.com/products/ramsan-440.htm
21:15:17 <oerjan> !underload (u)(~:S( )S:*~:^):^
21:15:30 <oerjan> still no infloop printing
21:15:38 <oerjan> +ul (u)(~:S( )S:*~:^):^
21:15:39 <thutubot> u uu uuuu uuuuuuuu uuuuuuuuuuuuuuuu uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu ...too much output!
21:16:05 <fungot> u uu uuuu uuuuuuuu uuuuuuuuuuuuuuuu uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu ...too much output!
21:16:11 <thutubot> u uu uuuu uuuuuuuu uuuuuuuuuuuuuuuu uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu ...too much output!
21:16:15 <oerjan> ^ul (^^:^^^:^^^^^:^^^^^^:::^^^^^^^^:::^^^:^^^^::^)()~((())~:a~*):a~*~^!(~((!())(!:^(^)*)(!!:^(!^)*))~*^!!^):^(~((()())(:a~*:(*(!^)(:)S)~*~(!*(^)(^)S)~*):a~**((!^)~^!^)(!(^)~^^))~*^( )S!!a:(*)*~(~*)**^~*(()()(!)()(!)(:a~*:(!^(!^((!^)*)(!(^)*))(!^((^)*)(!(^)*)))~*~(!^(!^((!^)*)(!(^)*))(!^((^)*)(!(!^)*)))~*):^)~*^!!!!!!~:^):^
21:16:17 <fungot> ^^:^^^:^^^^^:^^^^^^:::^^^^^^^^:::^^^:^^^^::^ :^^^:^^^:::^^^::::^::^^::::::^::^^:^^^::^:^^ ^^:^^^:^::^^:^:::^^:^^^:::::^^:^^^^^:^:^^^^^ :^^^:^^^:^^^^^::^^^^^:^::::^^^^^:::^^^^^:::: ^^:^^^:^^^:::^:^^:::^^^:::^^:::^::^^:::^:::: ^^^^:^^^:^::^^^^^::^^:^::^^^::^^:^^^::^^:::^ :::^^^:^^^:^^:::^:^^^^^:^^:^:^^^^^:^:^^^::^^ ::^^:^^^: ...too much output!
21:16:42 <AnMaster> I mean, how can you track that
21:17:05 <oerjan> AnMaster: that's the rule 110 automaton
21:17:20 <AnMaster> multiline would make it easier
21:17:34 <oerjan> yes but the irc bots don't allow that
21:18:06 <oerjan> (it would require giving a newline in the input, for one thing)
21:18:27 <AnMaster> oerjan, EgoBot can load from urls
21:18:35 <AnMaster> fungot can't yet. it was planned however
21:18:35 <fungot> AnMaster: " a far fnord one than mine, said arthur. " why, they're only a pack of cards!"
21:18:42 <oerjan> it shows fine in irssi though, due to line wrapping
21:18:56 <ais523> fungot: Arthur in Wonderland!
21:18:56 <fungot> ais523: " but he needn't run over me!" or to any loud cry, such as " fnord me!" i said to myself " that's very curious!" she cried. ' yet i must sell my sunday fnord the fnord almost unnatural fnord which arthur met the woman who had won his heart, and outgrabe in despair, took to pointing out places for himself, and feebly asked " is that great yellow fnord fairyland?"
21:19:03 <fungot> Available: agora alice* darwin discworld europarl ff7 fisher ic irc jargon lovecraft nethack pa speeches ss wp
21:19:11 <ais523> AnMaster: I don't know, ask the bot
21:19:15 <fizzie> "Any loud cry, such as 'fnord me!'."
21:19:23 <oerjan> AnMaster: well for EgoBot only the first line would be on channel, too
21:19:40 -!- MigoMipo has joined.
21:19:41 <AnMaster> oerjan, you can use it outside the printed bits
21:20:14 <oerjan> i considered making a "quine" version but there is an enormous amount of extra noise
21:20:50 <oerjan> the rule 110 automaton
21:21:15 <oerjan> but something that would do the automaton between two bots
21:21:38 <AnMaster> oerjan, can you fit that in one line
21:21:40 <oerjan> actually i think they don't allow enough output for it
21:22:03 <fizzie> You can do length-unlimited fungot programs if you want to do a bit of scripting which uses ^str add in a privmsg. I don't think those have a length limit. (You can also use that to make fungot run out of memory, so don't do it.)
21:22:04 <fungot> fizzie: how it happened. ' that fnord full of them. ' i'm sure mine only works one way,' and then its eyes looked fnord" and an " o."
21:22:43 <AnMaster> oerjan, err (!^((!^)*)(!(^)*))(!^((^)*)(!(^)*))) <-- seems repeated there. Can't you use : somehow to cut that down?
21:23:13 <AnMaster> surely it can be made shorter somehow
21:24:05 <oerjan> AnMaster: that's the encoding of the automaton rules
21:24:19 <AnMaster> oerjan, still, can't you make it shorter
21:27:31 <oerjan> maybe a bit of tweaking
21:27:55 <oerjan> ((^)*)(!(^)*) = ((^)*):(!)~*
21:28:36 <oerjan> a bit complicated for saving one char :D
21:30:07 <ais523> in Underlambda, higher tiers have a few charsaving abbreviations
21:30:14 <ais523> in particular, ` for ~^ and & for ~*
21:31:44 <oerjan> AnMaster: thing is, the representation of a bit in that tree depends not just on what the bit is, but also on whether it is left or right branch
21:32:40 <AnMaster> ais523, hm... would be funnier it it looked like combined
21:32:53 <ais523> yes, but that's less useful
21:33:00 <ais523> oh, ' is another typesaving character
21:33:11 <ais523> likewise for '^ == (^) and for any other character
21:33:19 <ais523> also, I suspect '(a) == ((a))
21:33:24 <ais523> because there's no other obvious meaning
21:33:47 <oerjan> AnMaster: or rather, every right branch contains an extra ! to get rid of the left branch still on the stack
21:33:58 <ais523> AnMaster: but that would make the input /longer/
21:34:11 <AnMaster> oerjan, one char saved is one place up in anagolf gained!
21:34:42 <AnMaster> ais523, only if you count bytes rather than chars
21:34:59 <ais523> oerjan: if the rules were long enough, presumably it would be shorter to write a recursive !-adder
21:35:11 <ais523> AnMaster: yes, but golfing servers do count bytes
21:36:37 <oerjan> ais523: oh and also there is a recursion on the stack, because you look at the last two bits and one bit, each of which have two elements on the stack for accepting next bit
21:37:20 <fizzie> You can add a combining ^ into the mathematical ~ operator, ∼. That should give a reasonably-rendered ~^-combination. Would look something like ∼̂. Maybe the ascii-~ would work too. Though combining characters tend to render uglily.
21:38:42 <coppro> fizzie: is that actually a combining circumflex over a tilde?
21:38:54 <fizzie> It should be, but I might've gotten them in the wrong order.
21:38:57 <ais523> it's a combining circumflex over something that looks like a tilde but isn't
21:39:02 <ais523> why not just use a genuine tilde?
21:39:15 <fizzie> Because the TILDE OPERATOR is obviously greater than some crummy ASCII tilde.
21:40:01 <fizzie> Besides, the mathematical operators block has, for those situations where a regular ~ is too weak, also the ∾ -- "inverted lazy S = most positive".
21:40:59 <fizzie> Heh, there's also a Unicode codepoint with the official name "NOT TILDE", but sadly it's just a tilde-with-a-slash (≁) and not something that's really not tilde, like, say, a duck.
21:41:14 <coppro> First one is an ascii tilde, second is a TILDE OPERATOR
21:41:18 <ais523> the one on the left looks better to me, that's the ASCII tilde based on how it looks
21:41:29 <fizzie> They render pretty similarly here.
21:41:36 <coppro> the only difference here is that the OPERATOR is antialiased
21:41:42 <coppro> possibly a different font
21:41:51 * ais523 wonders why everyone is guessing that Jeff Atwood's password is "orange"
21:42:46 <oerjan> ais523: orange u glad that isn't your password?
21:43:01 <ais523> oerjan: where does that joke come from, I never understood it
21:43:14 <oerjan> it's a knock knock joke iirc
21:43:30 <ais523> substitution of "orange" for "aren't" isn't particularly obviously funny
21:43:36 <ais523> although I've seen it several times in jokes
21:43:44 <fizzie> Here's the ASCII tilde with a combining tilde both above and below: ~̰̃.
21:43:46 <ais523> for me it's become funny due to the repetition
21:44:09 <oerjan> ais523: http://www.listphile.com/Best_Knock_Knock_Jokes_of_All_Time/Orange_Banana
21:44:42 <oerjan> ais523: also, knock knock jokes are _supposed_ to be puns like that
21:44:54 <ais523> I still don't get the pun
21:45:02 <oerjan> hm the banana part isn't though
21:45:03 <ais523> I know they're supposed to be puns
21:45:09 <ais523> but orange doesn't sound anything like aren't
21:45:19 <ais523> nor does aren't sound like something orange-related
21:45:24 <oerjan> ais523: "aren't you glad i didn't say banana"
21:45:25 <AnMaster> ais523, so, any underlambda implementation yet
21:45:37 <oerjan> ais523: i suppose it depends on dialect
21:46:06 <ais523> I can't even imagine how "orange" could sound like "aren't"
21:46:36 <oerjan> you have to mingle the two words
21:46:55 * ais523 wonders if he's being meta-trolled
21:47:28 <ais523> all the vowels are different, nearly all the consonants are different...
21:47:36 <oerjan> ais523: oh you're british, maybe you don't have an r sound in aren't?
21:47:39 <AnMaster> <fizzie> Here's the ASCII tilde with a combining tilde both above and below: ~̰̃. <-- what about ~ with ^ above
21:47:57 <AnMaster> how do you make those combining ones
21:47:57 <ais523> AnMaster: loads of those have been pasted earlier
21:48:05 <ais523> AnMaster: and via character map
21:48:10 <ais523> you mean, you don't have a GUI program for htat?
21:48:21 <ais523> AnMaster: the combiners are there too
21:48:36 <fizzie> I've usually been using gucharmap, which is a gnomey sort of program.
21:48:44 <fizzie> I'm sure there's a KDE thing too.
21:48:53 <oerjan> ais523: i suspect it works best in a texas accent, or something midwest US like that
21:48:55 <fizzie> It might even have a built-in character-select-o-tron, for all I know.
21:49:13 <fizzie> Just remember to stick combining characters after the thing you want them to combine with.
21:50:15 <AnMaster> will compile it and try it later
21:50:51 <oerjan> any americans around who can tell us whether "aren't" really sounds anything like "orange"?
21:51:28 <AnMaster> different in Oxford English I think
21:51:34 -!- MigoMipo has quit ("I must sleep!").
21:51:56 <pikhq> *Maybe* in some exaggerated parody of Texas English...
21:53:00 <oerjan> so GWB gets it perfect? :)
21:53:56 <ais523> oh, if it needs an American accent, possibly I can imagine how it works
21:54:11 * oerjan should have seen that AnMaster comment coming
21:54:40 <oerjan> AnMaster: famous ex-president with bad speech
22:03:26 <GregorR> ORANGE YOU GLAD I DIDN'T SAY BANANA?
22:03:41 <ais523> GregorR: oerjan linked me to it, I'd never seen it before
22:04:05 <GregorR> Oh I was just responding to <oerjan> any americans around who can tell us whether "aren't" really sounds anything like "orange"?
22:04:31 <AnMaster> GregorR, yay back to square one
22:04:43 <AnMaster> that was the start of the discussion duh
22:05:03 <GregorR> I have a program looking for all potential 26-letter pangrams right now.
22:05:06 <oerjan> GregorR: ais523 being british couldn't imagine how that pun actually works
22:05:33 <GregorR> Fail Brittania (couldn't resist the pun urge :P )
22:05:54 <ais523> GregorR: there was a 26-letter pangram as the answer to a massive puzzle on Agora that lasted five weeks
22:05:58 <GregorR> AnMaster: A sentence containing every letter of the alphabet, e.g. The quick brown fox jumps over the lazy dog.
22:06:02 <oerjan> This sentence contains 43 'a's, 9 'b's, ..., and 16 small jumping rats.
22:06:12 <AnMaster> GregorR, that is a huge search space
22:06:24 <oerjan> oh not self-describing ones
22:06:32 <ais523> "Zing! Vext cwm fly jabs Kurd qoph."
22:06:43 <ais523> oerjan: I don't think there's a 26-letter self-describing pangram
22:06:45 <GregorR> And the trick is not to use lame ones like that :P
22:06:47 <AnMaster> GregorR, you need to check if they are gramtically valid.
22:06:48 <oerjan> those Kurds are always getting it
22:06:58 <GregorR> AnMaster: Yeah, that's gonna be the PITA part X-P
22:06:59 <oerjan> ais523: well not in english. maybe Rotokas.
22:06:59 <ais523> oerjan: actually, it's just a Kurd who drew the qoph
22:07:18 <ais523> someone should write a minimum-length pangram in Chinese
22:07:22 <AnMaster> GregorR, what about: The brown fox quick over jumps the lazy dog.
22:07:39 <GregorR> AnMaster: That's not a _26-letter_ pangram.
22:07:49 <ais523> there's a minimum-length pangram in Japanese kana, IIRC, which makes a lot of sense and even scans correctly, IIRC they use it to order the alphabet
22:08:03 <AnMaster> GregorR, any example of such as 26 one
22:08:05 <GregorR> AnMaster: Which is to say, contains ever letter exactly once.
22:08:15 <GregorR> <ais523> "Zing! Vext cwm fly jabs Kurd qoph."
22:08:16 <pikhq> GregorR: Been watching Code Geass or something?
22:08:24 <GregorR> I've seen two or three, but they're all hyper-lame.
22:08:35 <ais523> AnMaster: yes, although some of them are rather obscure
22:08:45 <AnMaster> ais523, too obscure for aspell even
22:08:50 <ais523> that's one of the few isogram pangrams which actually makes grammatical sense
22:09:00 <GregorR> Nobody outside Wales would ever use "cwm"
22:09:06 <ais523> AnMaster: "zing" is the noise a bullet makes as it whizzes past you
22:09:15 <pikhq> I thought that was Welsh.
22:09:21 <AnMaster> like all words but "fly" there
22:09:28 <pikhq> (I don't know of any other languages that use w as a vowel...)
22:09:30 <GregorR> pikhq: It's a Welsh loan word, yes.
22:10:04 <ais523> it's a sort of valley shaped like a semicircle
22:10:36 <ais523> Vext is an old spelling of "vexed"
22:10:37 <AnMaster> fly jabs I think I know, unless they mean something odd
22:10:40 <ais523> which means confused, or annoyed
22:10:57 <ais523> qoph is a letter in the Hebrew alphabet
22:11:08 <ais523> and a Kurd is a member of the Kurdish ethnolinguistic group
22:11:14 <ais523> Deewiant: I thought so too, but I looked it up
22:11:35 <AnMaster> so. what does the entire sentence mean
22:11:40 <GregorR> Of course "vexed" means confused.
22:12:07 <oerjan> you didn't explain cwm
22:12:09 <GregorR> Confused flying valleys jab those instances of the letter qoph that are (quite inexplicably) Kurdish.
22:12:11 <ais523> AnMaster: Zing! An angry fly that lives in a semicircular valley jabs the Hebrew letter qoph, as drawn by a Kurd
22:12:14 <ais523> oerjan: I did, just earlier
22:12:24 <ais523> GregorR: it's ambiguous, my meaning is slightly more sensible
22:12:26 <GregorR> Oh, I'm sorry, yes, fly as in a noun, bleh
22:13:15 <AnMaster> what about one using space at most once ;P
22:13:52 <AnMaster> how did anyone come up with that one
22:14:10 <AnMaster> how the hell do you pronounce "cwm"
22:15:02 <ais523> AnMaster: the w's a vowel, it sounds like about half an oo
22:15:49 * GregorR wonders how AnMaster pronounces "book" :P
22:15:52 <AnMaster> Deewiant, doesn't help me if I don't know how it is supposed to sound
22:16:07 <AnMaster> GregorR, not like a double single o no
22:16:19 <AnMaster> I just can't figure out how to pronounce half an o
22:16:31 <GregorR> AnMaster: Only the Welsh truly know.
22:16:32 <AnMaster> Deewiant, is that u the Swedish u?
22:17:20 <GregorR> Well, y'know, those Swedes speak in a blurry way.
22:17:42 <Deewiant> I can't think of a Swedish word with /u/ :-/
22:17:52 <AnMaster> works great of ASCII and åäöÅÄÖ but not anything else really
22:17:54 <Deewiant> Anyway, it's like the Swedish o
22:18:11 <AnMaster> Deewiant, so Swedish o is /u/?
22:19:04 <GregorR> IIII'm a monoglot and I'm OH-KAY, I sleep all night and I work all day.
22:19:13 <AnMaster> Deewiant, so ais523 was just misleading me then
22:19:25 <kerlo> A 26-letter self-describing pangram: A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R, S, T, U, V, W, X, Y, Z.
22:19:30 <Deewiant> More likely he's confused about the vowel sounds in his own language :-P
22:19:50 <AnMaster> Deewiant, I still can't figure out cwm, I mean... what is the c there. is it s or k sound
22:20:06 <kerlo> C is only S before E, I and Y.
22:20:09 <kerlo> W is not E, I or Y.
22:20:14 <AnMaster> Deewiant, so like "kom" in Swedish?
22:20:37 <Deewiant> Right, Swedish o is *not* always /u/
22:20:43 <kerlo> And the letter "w" is almost always the long "oo" sound, I think. Like, um...
22:20:51 <AnMaster> Deewiant, yeah it is more like å in there
22:20:56 <oerjan> i vaguely recall welsh c is always k
22:21:54 <AnMaster> ok.... So I have something like 4 totally different descriptions of how cwm is pronounced now
22:22:21 <kerlo> Ideally, everyone would understand me if I called a goose a gwç.
22:22:27 <oerjan> heh, "The letter "k" was in common use until the sixteenth century, but was dropped at the time of the publication of the New Testament in Welsh, as William Salesbury explained: "C for K, because the printers have not so many as the Welsh requireth". This change was not popular at the time."
22:22:36 <AnMaster> Deewiant, and that is like "kom" but with the other o sound? As in "gol"?
22:22:52 <kerlo> AnMaster: in IPA, /u/ is the "oo" of "goose".
22:23:01 <ais523> so in other words, it isn't spelt "kwm" because old printers didn't have enough ks?
22:23:20 <AnMaster> ais523, now you are just being absurd.
22:23:29 <ais523> AnMaster: I was replying to oerjan
22:23:32 * AnMaster wonders why "om" at end of words in Swedish make a short o sound.
22:23:43 <kerlo> Okay, maybe I'm weird.
22:24:05 <oerjan> ais523: so it seems :D
22:24:40 <kerlo> No, "goose" isn't /gʊ:s/, unless either http://en.wikipedia.org/wiki/Help:IPA_for_English is lying or my knowledge of English is very fail.
22:24:56 <fizzie> Wiktionary's IPAfication of "goose" is /guːs/ actually.
22:25:14 <kerlo> The word "foot" is /fʊt/.
22:25:33 <Deewiant> Yeah, that's definitely a different sound
22:25:37 <Deewiant> But I think /u/ is not what I thought it was
22:25:57 <AnMaster> so spelled in English it would be koom?
22:26:15 <coppro> dictionary.com lists it as koom
22:26:30 <kerlo> Yeah, it would probably be "koom".
22:26:36 <AnMaster> I think I hit a hidden TP reference there...
22:27:00 <Deewiant> Okay, I just don't get how Finnish "kuu" is /ku:/ while English "food" is /fu:d/
22:27:03 <AnMaster> I guess "koom valley" in the Discworld is a pun on cwm
22:27:05 <kerlo> Rhyming with "boom", not the first half of "woman".
22:27:26 <kerlo> Too bad I don't know any Finnish.
22:27:32 <kerlo> Is it Indo-European?
22:27:43 <ais523> anyway, I love the word "cwm", because it's normally a giveaway that someone is attempting a pangram
22:28:03 <AnMaster> Deewiant, isn't it the only language in that family?
22:28:20 <fizzie> I freely admit I probably wouldn't be able to differentiate between 'food' /fu:d/ and 'foot' /fʊt/ based on only the wovel; I mean, sure, there's a difference, but in the grand scheme of life, the universe and everything, it's not so big.
22:28:26 <Deewiant> AnMaster: Finnish, Estonian, Hungarian and a bunch of very-tiny languages
22:28:38 <Deewiant> Meänkieli, Saami, Mordva, etc.
22:28:41 <AnMaster> <fizzie> I freely admit I probably wouldn't be able to differentiate between 'food' /fu:d/ and 'foot' /fʊt/ based on only the wovel; I mean, sure, there's a difference, but in the grand scheme of life, the universe and everything, it's not so big. <-- it is rather obvious to me
22:29:01 <Deewiant> AnMaster: Very-tiny, like I said.
22:29:11 <kerlo> I could differentiate between "wide" and "white" based only on the vowel.
22:29:29 * AnMaster hates when you get double "of" in English
22:29:44 <kerlo> Everyone hates when you get double anything in English.
22:30:03 <Deewiant> Is that not a very different vowel
22:30:06 <kerlo> But sometimes we just have to accept that that that is is.
22:30:16 <AnMaster> Deewiant, foot vs. food is more different
22:30:19 <kerlo> Deewiant: those vowels sound pretty much the same to me.
22:30:26 <kerlo> Foot is different from both.
22:30:28 <Deewiant> AnMaster: Yes, but I'm not interested in that.
22:30:31 <AnMaster> <kerlo> But sometimes we just have to accept that that that is is. <-- parser failure on three of them
22:31:13 <oerjan> AnMaster: but but but it's easy
22:31:16 <fizzie> Heh, typical Wikipedia style: "The term Finno-Ugric is somewhat controversial today[citation needed], with many historical linguists[who?] feeling --"
22:31:20 <kerlo> Noun phrase: that that that is is
22:31:23 <kerlo> Sentence: that that is is
22:31:52 <kerlo> Restrictive clause: that is
22:32:48 <kerlo> If you understand the phrase "that that is is", you just stick a complementizer in front of it and get the noun phrase found in "accept that . . .".
22:34:21 <kerlo> Spanish is less ambiguous: Pero a veces sólo tenemos que aceptar que lo que es es.
22:35:22 <oerjan> and norwegian even less: ... akseptere at det som er, er.
22:36:04 -!- puzzlet has quit (Remote closed the connection).
22:36:08 -!- puzzlet has joined.
22:37:31 <Deewiant> http://en.wikipedia.org/wiki/Advanced_%28phonetics%29
22:37:40 <Deewiant> In English, the back vowel /u/ is farther forward than what is normally indicated by the IPA letter <u>. This fronting may be shown explicitly, especially within a narrow transcription: [u̟]. Whether this is as far front as the central vowel [ʉ], or somewhere between [u] and [ʉ], may need to be clarified verbally.
22:37:51 <Deewiant> No shit it's farther forward, it's a mile away :-P
22:38:20 <fizzie> You have one big-ass mouth, then.
22:38:25 <oerjan> those english always sticking out their tongue
22:38:27 <fizzie> (Do not move the dash.)
22:40:03 <Deewiant> fizzie: But seriously, am I the only one who thinks that "food" and fi:"kuu" have completely different vowel sounds
22:40:52 <fizzie> I'm not sure I'd go so far as to say "completely", but different, yes.
22:41:04 <fizzie> Of course my understanding of 'food' might be the wrong.
22:41:10 <Deewiant> I honestly thought they'd be different even in broad transcription
22:41:15 <fizzie> I am reasonably sure I know what fi:kuu sounds like.
22:41:33 <Deewiant> I assume you know what that sounds like, in general
22:42:27 <Deewiant> Although it does obfuscate a bit with the /j/ rolling into it
22:43:21 <oerjan> doncha comfuse yer prunciation, ye haer
22:44:10 <fizzie> I can plausibly believe a "you" that's exactly like fi:juu, with identical wovel to fi:kuu, but I'm not so sure I know what "you" should exactly sound like; I think if I were to say "you" and "food" they'd have somewhat different sounds, anyway.
22:44:30 <fizzie> I mean "plausibly believe that someone, esp. a Finnish someone, would say it like that".
22:44:46 <Deewiant> Certainly a Finnish someone might, but no English one IMO :-P
22:45:13 <pikhq> ... Kuu? For food?
22:45:31 <fizzie> fi:kuu is made out of cheese.
22:45:35 <pikhq> That's kinda funny, since in Japanese, "Kuu" is a informal way of saying "to eat"...
22:45:49 <fizzie> (It is the same as en:moon.)
22:50:44 <fizzie> We could switch to a two-sound phonology -- say a Hamming-encoded bitstring formed out of 0 = IPA y, 1 = IPA ɑ, duration-insensitive -- that'd certainly make the acoustics side of speech recognition easier, for one thing.
22:51:02 <fizzie> With distinct stops between bits, of course.
22:51:16 <Deewiant> No, I don't think we could. :-P
22:52:02 <fizzie> Deewiant: Well, y-ɑ-ɑ-ɑ-y-ɑ-y-y-y-y-ɑ-ɑ-y you.
22:54:15 <pikhq> ASCII-ise that question mark?
22:55:42 <pikhq> The IPA thing that's not a y.
22:56:02 <Deewiant> pikhq: http://en.wikipedia.org/wiki/Open_back_unrounded_vowel
22:56:05 <fizzie> It's the a variant without the hook-at-the-top thing.
22:56:52 <fizzie> Actually "latin small letter alpha" seems to be the Unicode name.
22:58:35 -!- olsner has quit (Read error: 104 (Connection reset by peer)).
22:58:56 <pikhq> It's also encoded as A is X-SAMPA.
22:59:46 <fizzie> Those X-SAMPA "b_<" and friends look like botched horizontal-smileys.
23:01:11 <pikhq> Yeah, X-SAMPA is a pain.
23:01:21 <pikhq> What's more a pain is the notable lack of terminals that do Unicode right.
23:01:33 <pikhq> And they're the only programs that don't on my system.
23:03:09 <Deewiant> Clearly you're using the wrong terminals.
23:06:00 <pikhq> rxvt-unicode is misnamed.
23:06:10 <pikhq> Must be rxvt minus unicode.
23:07:48 <ehird> pikhq: urxvt doesn't do unicode?
23:08:31 <pikhq> It sure hasn't for me.
23:23:26 -!- BeholdMyGlory has quit (Remote closed the connection).
23:28:28 -!- FireFly has quit ("Later").
23:30:17 <fizzie> It's been doing unicode very well for me. Well, I guess the combining-character rendering could use some work. But other than that, and even that stuff rudimentarily works.
23:34:24 <fizzie> One might need to give it reasonable fonts, though; I'm not sure what the default font search list looks like. I've been using a "URxvt.font: xft:DejaVu Sans Mono:size=8, xft:Kochi Gothic:size=8" (and identical boldFont) resource line.
23:36:37 <fizzie> Well, and I haven't really tried anything fancy, like switching direction with the right-to-left mark; but certainly it's been reasonably good in rendering fancy characters.
23:37:54 <fizzie> ehird: I started with 10, I think; used that about a year, switched to 9, used that one more year, and recently switched to 8.
23:38:04 <fizzie> It's some sort of adaptation technique, I guess.
23:38:15 <fizzie> Maybe I can graduate to 7 soon.
23:38:47 <coppro> pango's got incredible Unicode rendering
23:39:01 <ehird> coppro: that's what SHE said!
23:40:08 <ais523> ehird: doesn't work in that context
23:40:20 <ehird> ais523: that's the point
23:40:27 <ais523> I'm using size-9 font atm, I deliberately sized it down from the default
23:41:00 <fizzie> Heh, urxvt FAQ; "here is a more complete set of non-standard colors" (to replace the ones that look mostly like a VGA screen) "They have been described (not by me) as 'pretty girly'."
23:42:00 <fizzie> This one is low-ish in the DPI rankings; the reasonably common 94dpi.
23:43:56 <ehird> fizzie: ITYM 96dpi
23:44:16 <ehird> This house has some CRTs, my 100dpi screen, and the ole 84dpi screen
23:49:48 <fizzie> No, 94, officially; pixel pitch of 0.270 mm, translating to (25.4 mm/in) / (0.270 mm/dot) = 94.074... dot/in. I guess it might not be "reasonably common", though.
23:53:43 <fizzie> Hadn't noticed before that they're making "normal" laptops nowadays with 16" 1920x1080 displays (which means ~140dpi) -- I did know about the really high-DPI screens in some VAIO models and other "high-end" things, but that 16" one was a cheapo-laptop, some <1000 eur price.
23:54:24 <ehird> fizzie: macbook pros have 17"
23:54:39 <ehird> iphone has 160dpi display
23:55:21 <GregorR> 9 a i o pb qs rn te whyd xv
23:55:41 <fizzie> Last looked at Apple laptop specs when getting the iBook G4.
23:56:19 <fizzie> (That one was 12" 1024x768, meaning a bit over 100 dots.)
23:56:31 <fizzie> Sleepity-sleep anyway.