search of the:   for: 
code logs -> 2009 -> Fri, 24 Apr 2009< code.20090423.log - code.20090425.log >
--- Log opened Fri Apr 24 00:00:47 2009
00:46 You're now known as TheWatcher[T-2]
00:48 You're now known as TheWatcher[zZzZ]
01:32 KBot [~karma.bot@Nightstar-29207.neoplus.adsl.tpnet.pl] has joined #Code
01:32 KarmaBot [~karma.bot@Nightstar-28972.neoplus.adsl.tpnet.pl] has quit [Ping Timeout]
01:35 KBot is now known as KarmaBot
03:31 Reiver [~reaverta@Admin.Nightstar.Net] has quit [Ping Timeout]
03:34 Reiver [~reaverta@Admin.Nightstar.Net] has joined #Code
03:34 mode/#code [+o Reiver] by ChanServ
03:59 Vornicus [Vornicus@Admin.Nightstar.Net] has joined #code
03:59 mode/#code [+o Vornicus] by ChanServ
04:36 * Gruber slaps McMartin.
04:36 Gruber is now known as gnolam
04:37
< gnolam>
The kg is never a unit of force. Never.
04:43
<@ToxicFrog>
Oh, gnolam. Now that you're back: SDL_CreateRGBSurface is behaving as documented and as designed; it returns a new blank surface. What did you expect?
04:47
<@McMartin>
gnolam: "gram-force" was in my physics texts.
04:53
< gnolam>
ToxicFrog: sorry, I meant CreateRGBSurfaceFrom.
04:53
< gnolam>
McMartin: then your physics texts are /wrong/.
04:53
< gnolam>
The kilogram and the gram are _strictly_ units of mass.
04:54
< gnolam>
Never units of force.
04:55
<@McMartin>
If you have a standard acceleration, force and mass are in fact convertible.
04:56
<@McMartin>
Which is why, while pounds are also strictly units of force, nobody but archers ever uses them as such.
04:58
<@McMartin>
If you can find me any metric-using country where people give their weight in Newtons I'll even begin to entertain your thesis, but I'll still reject it instantly because the form is always "I weigh 60kg, last I checked."
04:59
<@ToxicFrog>
gnolam: if CreateRBGSurfaceFrom is giving unexpected results, this implies that you got some of the arguments wrong.
04:59
<@ToxicFrog>
That said, since IIRC you're using this as part of an image loader, Seriously, Just Use SDL_Image Already.
05:00
< gnolam>
McMartin: And you never use an accelerating frame of reference.
05:00
<@McMartin>
Meaningfully accelerating, yes.
05:00
< gnolam>
ToxicFrog: which won't help, because it all still depends on what SDL stores in foo->pixels.
05:00
<@McMartin>
It can accelerate as long as nobody cares.
05:00
<@McMartin>
gnolam: Sable. Does. This. It. Fucking. Works.
05:00
<@McMartin>
Stop. Blaming. The. Library.
05:01
< gnolam>
It's got a complete threading library, yet it still never crossed their minds that someone would want an easy way to convert a loaded image to an OpenGL texture. Or even have a fucking /getpixel()/ function.
05:01
< gnolam>
So yeah, I will keep blaming the library.
05:02
<@McMartin>
As it happens, you never *do* want GetPixel.
05:02
<@ToxicFrog>
And if you do, it has a subsidiary library for that, SDL_gfx
05:02
<@McMartin>
Since it's a synchornized operation and requires locking and unlocking a mutex, that must hit the graphics bus if it's "accelerated"
05:02
<@ToxicFrog>
Which covers graphics primitives of all sorts.
05:02
< gnolam>
I might eventually, when I stop being pissed off, rewrite lines like
05:02
< gnolam>
SDL_Surface *SDLblowshobosforbeermoney = SDL_CreateRGBSurface(SDL_SWSURFACE, 1, 1, 32, rmask, gmask, bmask, amask);
05:03
<@McMartin>
Why are you creating a 1x1 image?
05:03
< gnolam>
to something else, but that depends on if I ever actually stop hating on the library.
05:04
<@McMartin>
But, uh, yeah, the "make an OpenGL texture from an SDL image" code in Sable was basically just a direct reference to the "pixels" member.
05:04
< gnolam>
Because, surprisingly enough, it's easier than filling an entire SDL_PixelFormat structure from scratch.
05:04
<@McMartin>
Fair enough.
05:04
<@McMartin>
Why you consider that an imposition is beyond me.
05:04
<@McMartin>
(I tend to use the backbuffer for that)
05:05
<@McMartin>
right, so.
05:05
< gnolam>
You use the backbuffer for loading textures? :P
05:05
<@McMartin>
No, for having the PixelFormat I want.
05:06
< gnolam>
And the backbuffer isn't guaranteed to be RGB.
05:06
<@McMartin>
It is when I make it SDL_SWSURFACE and don't set SDL_ANYFORMAT.
05:06
< gnolam>
Which is why you have to go through the 20 LOC ordeal in the first place.
05:06
<@ToxicFrog>
If your complaint is "I can't turn random surfaces into OpenGL textures", SDL_ConvertSurface or SDL_DisplayFormat to a format OGL expects, and load that.
05:06
<@McMartin>
... one line to declare reference.
05:06
<@McMartin>
Then:
05:06
<@McMartin>
 SDL_Surface *base = IMG_Load(global_texture_files[i]);
05:07
<@McMartin>
 surfaces[i] = SDL_ConvertSurface (base, reference->format, SDL_SWSURFACE);
05:07
<@McMartin>
 SDL_FreeSurface (base);
05:07
<@McMartin>
(various OpenGL setup things to Gen/Bind textures)
05:07
< Derakon>
In general, if you find yourself needing a half-dozen lines or more to do something, then you're missing an API or library.
05:07
<@McMartin>
 gluBuild2DMipmaps(GL_TEXTURE_2D, GL_RGBA, 256, 256, GL_RGBA, GL_UNSIGNED_BYTE, surfaces[i]->pixels);
05:07
<@McMartin>
That's three lines.
05:07
< gnolam>
ToxicFrog: And those don't actually work. You have to do the SDL_ConvertSurface thingy with the 10 lines of mask defines just to LOAD A GODDAMNED TEXTURE.
05:07
<@McMartin>
And it fucking works on PNGs.
05:07
<@McMartin>
10?
05:07
< Derakon>
Gnolam, I think it's pretty clear here that You're Doing It Wrong.
05:07
<@McMartin>
That's one line for the reference, and one for Convert.
05:08
<@ToxicFrog>
gnolam: they work for me, they work for McM, they work for every other goddamn SDL-using project on the planet, ergo you are not using the library properly.
05:08
<@McMartin>
I've quoted sufficient code to do it, except for the OpenGL parts that are entirely SDL-independent.
05:08
< gnolam>
Well, explain to me how to load _arbitrary_ images into OpenGL textures without it then.
05:08
<@ToxicFrog>
McM just pasted code that does exactly that.
05:09
<@McMartin>
Here, I'll paste the code from Sable. Note that it uses SDL_image because I vastly prefer my content to be PNGs. Replace IMG_Load with LoadBMP if you insist on terrible file formats.
05:10
< gnolam>
Which, again, requires the base surface to be RGBA... which is NOT GUARANTEED.
05:10
<@McMartin>
http://paste.ubuntu.com/156924/
05:11
<@McMartin>
Then you're either building your reference surface wrong or calling SDL_ConvertSurface wrong.
05:11
<@McMartin>
It appears in this case after refactoring my reference surface is actually going away afterwards, but eh.
05:12
<@McMartin>
You can probably use a smaller reference surface in this case.
05:13 gnolam [lenin@Nightstar-1382.A163.priv.bahnhof.se] has quit [Quit: Z?]
05:14
<@ToxicFrog>
Sigh.
05:14
<@McMartin>
Going with "He thought SDL_ConvertSurface was a mutator"
05:15
<@McMartin>
Since the bug I actually hit developing Sable involved failing to check byte order.
05:42 Vornicus [Vornicus@Admin.Nightstar.Net] has quit [Quit: ]
05:42 Vornicus [Vornicus@Admin.Nightstar.Net] has joined #code
05:42 mode/#code [+o Vornicus] by ChanServ
05:43 Reiver [~reaverta@Admin.Nightstar.Net] has quit [Quit: I ATENT'T DEAD]
05:46 Vornicus [Vornicus@Admin.Nightstar.Net] has quit [Quit: ]
05:47 Vornicus [Vornicus@Admin.Nightstar.Net] has joined #code
05:48 Reiver [~reaverta@Admin.Nightstar.Net] has joined #Code
05:48 mode/#code [+o Reiver] by ChanServ
06:02 Syloqs-AFH [Syloq@ServicesAdmin.Nightstar.Net] has quit [Connection reset by peer]
06:14 AnnoDomini [AnnoDomini@Nightstar-29207.neoplus.adsl.tpnet.pl] has joined #Code
06:15 mode/#code [+o AnnoDomini] by ChanServ
06:24 Alek [~omegaboot@Nightstar-5734.dsl.emhril.sbcglobal.net] has quit [Ping Timeout]
06:25 Alek [~omegaboot@Nightstar-5734.dsl.emhril.sbcglobal.net] has joined #code
06:35 * Vornicus is... very frightened of Euler 84.
07:12 somnolence [~somnolenc@Nightstar-3790.hsd1.ca.comcast.net] has quit [Operation timed out]
07:20 Kazuma [~Kazuma@Nightstar-19318.fudgeman.org] has quit [Ping Timeout]
07:45 Derakon is now known as Derakon[AFK]
07:48 Vornicus [Vornicus@ServicesOp.Nightstar.Net] has quit [Quit: ]
08:15 Namegduf [namegduf@86.29.157.ns-3965] has quit [Quit: Heading out.]
09:05 You're now known as TheWatcher
10:09 Consul [~consul@Nightstar-1864.dsl.sfldmi.ameritech.net] has joined #code
10:09 mode/#code [+o Consul] by ChanServ
10:12
<@Consul>
Well, playing Flash video in a browser for a length of time still locks my computer up completely, even though I'm now running Ubuntu 9.04. It has to be a hardware problem. Why just Flash video, though?
10:14
<@TheWatcher>
Are you sure it's "completely"?
10:14
<@Consul>
Sometimes, the mouse pointer can still move, and usually, what's in the audio buffer will start looping.
10:15
<@TheWatcher>
Can you ssh in from another box, Ctrl+Alt+F2 to get to a vtty? Alt+SysRq+i to kill everything but init?
10:16
<@Consul>
The keyboard will not work at all.
10:16
<@Consul>
I haven't tried SSHing in from elsewhere. I suppose I can give that a shot next time it happens.
10:17
<@TheWatcher>
I'd be kinda inclined to suspect it's a driver problem first
10:17
<@Consul>
Well, this did kinda all start with the new soundcard.
10:18 * TheWatcher nod
10:18
<@TheWatcher>
that could do it
10:18
<@Consul>
But nobody else has any of the same issues with this card.
10:19
<@Consul>
I've already sent the damn thing back once for service.
10:20
<@Consul>
I'm still in the warranty period. I might demand a new card and see if that helps.
10:20
<@Consul>
Because I really don't have the money to go for yet another card. :-/
10:22
<@Consul>
Well, no "really" about it. I don't have the money at all.
10:24
<@TheWatcher>
one thing you might want to do is cat /proc/interrupts to see if the sound card is being forced to share an interrupt. If it is, you might want to try and wrangle it to get its own, might help.
10:25
<@Consul>
Is that something that would affect only Flash video? I'll take a look...
10:26
<@Consul>
Doesn't look like it...
10:27
<@Consul>
http://paste.ubuntu.com/157074/
10:29
<@TheWatcher>
Mia is the sound card driver, I guess?
10:29
<@Consul>
Yes
10:29
<@Consul>
Echo MiaMIDI
10:30
<@Consul>
I should point out that this is not the first time I've had issues with Echo cards. Some years ago (twelve?), my brother and I sent back one of their 8 I/O systems for hardware issues (we had outputs cutting in and out).
10:32
<@Consul>
But I need to go shower and get ready and all that. Be back in a bit. Thank you for your help. This has been incredibly frustrating lately.
10:32
<@TheWatcher>
Hm. Is this during playback, or immediately after?
10:32
<@Consul>
During playback...
10:32
<@Consul>
BBIAB
10:33
<@TheWatcher>
(I ask because I've run into some reports of various midi cards spontaneously locking up if they get hit by a lot of midi instructions and once, and this can happen if the driver clears a lot of controller states after playback)
11:08
<@Consul>
Really?
11:09
<@Consul>
Well, it's always during playback, as I've mentioned that whatever audio is in the buffer will go into a loop on most occasions.
11:11
<@Consul>
The system seems perfectly stable as long as I don't use Flash video in a browser, though. I'm trying to think of why that might be...
11:12
<@Consul>
And it really sucks because I like watching Leo Laporte and friends a few times a week.
11:13
<@Consul>
And that's the funny thing, because even with this card, this didn't start happening until a few weeks ago. I got the card several months back.
11:14
<@Consul>
7 or 8 months, actually.
11:15
<@TheWatcher>
kernel upgrade around the time it started, maybe?
11:15
<@Consul>
Don't think so. I was having tons of other trouble with Fedora 10 at the time, so something might have slipped in.
11:16
<@TheWatcher>
Have you tried viewing flash outside of the browser?
11:16
<@Consul>
I have a number of FLV files that will play back fine in Totem.
11:16
<@Consul>
A couple of other interesting things to note:
11:17
<@Consul>
1) Lockups happen more on higher bitrate streams and if I go full-screen.
11:17
<@Consul>
In other words, SQ on Youtube doesn't seem to cause a problem.
11:18
<@Consul>
and 2) This was really stupid on my part, but I managed to not put the MB standoff in under the PCI cards, so I had some trouble with the card seating.
11:18
<@Consul>
I really should go in and fix that, but it would be a lot of work.
11:19
<@Consul>
I don't see why the card being slightly unseated would cause issues only with flash plugin video playback, but...
11:19
<@Consul>
Well, stranger things have happened/
11:19
<@TheWatcher>
indeed. That horse becoming pope, for example.
11:19
<@TheWatcher>
*cough*
11:19
<@TheWatcher>
anyway, hm.
11:20
<@Consul>
Because sound is otherwise completely stable.
11:20
<@Consul>
But you mentioned some cards locking if they get hit with tons of control data...
11:21
<@Consul>
And I really have no way of knowing what Flash might actually be doing...
11:21
<@TheWatcher>
Hm, which browser, idly? Firefox?
11:21
<@Consul>
The latest Firefox.
11:22
<@Consul>
One thing I might try is to see if Bitgravity offers a stream for third-party programs.
11:22
<@TheWatcher>
Because I vaguely wonder if this isn't somehow related: http://ubuntuforums.org/archive/index.php/t-667314.html
11:23
<@Consul>
Some of those stories seem awfully familiar...
11:25
<@Consul>
But nobody's saying anything about entire system freezes.
11:25
<@TheWatcher>
Nope
11:26
<@TheWatcher>
It's very curious
11:31
<@Consul>
The thing that jumps out at me, though, is that their problems were also sound card specific.
11:35
<@Consul>
Well, thanks for letting me vent a little, anyway. :-)
11:35
<@Consul>
I need to eat and head off to school.
11:46 Rhamphoryncus [~rhamph@Nightstar-7184.ed.shawcable.net] has quit [Quit: Rhamphoryncus]
11:47
<@Consul>
Hey, turns out I can eat and type at the same time.
11:47
<@Consul>
Not that there's much more to say.
11:55
<@Consul>
I think I'll try watching Leo from Windows next time he has something on I want to watch. That will at least isolate whether it's a hardware or software issue.
11:56
<@Consul>
On this same machine, I mean. I have a rarely-used Windows partition.
12:03
<@Consul>
Oh well, time for school.
13:33 gnolam [lenin@Nightstar-1382.A163.priv.bahnhof.se] has joined #Code
13:33 mode/#code [+o gnolam] by ChanServ
16:04 Syloqs_AFH [Syloq@Admin.Nightstar.Net] has joined #code
16:06 Syloqs_AFH is now known as Syloqs-AFH
16:12 Namegduf [namegduf@86.29.157.ns-3965] has joined #code
16:40 Derakon[AFK] is now known as Derakon
16:41 PinkFreud [~WhyNot@ServicesAdmin.Nightstar.Net] has quit [Operation timed out]
16:44
<@ToxicFrog>
Terminology sanity check:
16:45
<@ToxicFrog>
- does "array" imply a homogeneity constraint to you, or just the usual performance constraints?
16:45
<@ToxicFrog>
- does "list" imply "linked list" specifically, or just any non-homogenous linear data structure?
16:45
< Derakon>
Usually I see "list" used when there's no homogeneity constraint.
16:45
< Derakon>
Python and Perl "lists" have random access but no homogeneity, for example.
16:46
< Derakon>
In Python, if you say "array" generally you're referring to NumPy or Numeric arrays, which definitely do have a homogeneity constraint.
16:50
< simontwo>
ToxicFrog, a) I generally assume homogeinity unless in an environment made by silly people. b) I think I assume homogeinity with lists, too, and don't think about how they are implemented unless in a specific context.
16:52
<@TheWatcher>
"array" does not imply homogeneity to me, but then I use a lot of perl.
16:58 Attilla [~The.Attil@Nightstar-9147.cdif.cable.ntl.com] has joined #code
16:58 mode/#code [+o Attilla] by ChanServ
17:11
< simontwo>
yup
17:52 * Derakon abuses Monte Carlo simulators for fun and profit.
17:52
< Derakon>
Well, really just for fun.
18:09 Chalain [~chalain@ServicesAdmin.Nightstar.Net] has joined #code
18:09 mode/#code [+o Chalain] by ChanServ
18:21 jerith [~jerith@Nightstar-21563.slipgate.za.net] has joined #code
18:23 You're now known as TheWatcher[afk]
18:48 Attilla [~The.Attil@Nightstar-9147.cdif.cable.ntl.com] has quit [Connection reset by peer]
18:50 Attilla [~The.Attil@Nightstar-9147.cdif.cable.ntl.com] has joined #code
18:50 mode/#code [+o Attilla] by ChanServ
18:59 Attilla [~The.Attil@Nightstar-9147.cdif.cable.ntl.com] has quit [Ping Timeout]
19:08 Attilla [~The.Attil@Nightstar-9147.cdif.cable.ntl.com] has joined #code
19:08 mode/#code [+o Attilla] by ChanServ
19:30 Consul [~consul@Nightstar-1864.dsl.sfldmi.ameritech.net] has quit [Ping Timeout]
20:07 You're now known as TheWatcher
21:04 GeekSoldier [~Rob@Nightstar-8573.midstate.ip.cablemo.net] has quit [Ping Timeout]
21:06 GeekSoldier [~Rob@Nightstar-8573.midstate.ip.cablemo.net] has joined #code
21:06 mode/#code [+o GeekSoldier] by ChanServ
21:52 * TheWatcher eyes php
21:52 AnnoDomini is now known as PHP
21:52 * PHP eyes TheWatcher back.
21:52 PHP is now known as AnnoDomini
21:53 * TheWatcher stabsta... blast, missed
21:53 * AnnoDomini laughs.
21:55
< simontwo>
TheWatcher, careful. when you stare into the abyss...
21:55
< simontwo>
supposedly stabbing into the abyss does solve some problems.
21:56
< Namegduf>
Someone has to watch PHP to make sure it doesn't do anything terrible.
21:58
<@TheWatcher>
Too late, then.
22:00
< Namegduf>
Hmm, yeah.
22:04 Rhamphoryncus [~rhamph@Nightstar-7184.ed.shawcable.net] has joined #code
22:05
<@TheWatcher>
I think whoever came up with this one was special: microtime() gives you the string "msec sec", where msec is the number of microseconds, expressed as seconds (ie: 0.0 to 0.999999).. so why does it 8 decimal places?
22:06
<@TheWatcher>
+have
22:08
< simontwo>
TheWatcher, because floats are pretty?
22:19 You're now known as TheWatcher[T-2]
22:22 You're now known as TheWatcher[zZzZ]
22:58 Vornicus [Vornicus@Admin.Nightstar.Net] has joined #code
22:58 mode/#code [+o Vornicus] by ChanServ
23:43 Consul [~consul@Nightstar-1864.dsl.sfldmi.ameritech.net] has joined #code
23:43 mode/#code [+o Consul] by ChanServ
23:55 gnolam [lenin@Nightstar-1382.A163.priv.bahnhof.se] has quit [Quit: Z?]
--- Log closed Sat Apr 25 00:00:01 2009
code logs -> 2009 -> Fri, 24 Apr 2009< code.20090423.log - code.20090425.log >
 search of the:   for: