| --- Log opened Sat Oct 28 00:00:51 2017 |
| 00:10 | | Vornicus [Vorn@Nightstar-1l3nul.res.rr.com] has joined #code |
| 00:10 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 00:14 | | Degi [Degi@Nightstar-rdjb0m.dyn.telefonica.de] has quit [Connection reset by peer] |
| 00:49 | | Kindamoody is now known as Kindamoody[zZz] |
| 00:49 | | Derakon[AFK] is now known as Derakon |
| 01:13 | | Vornlicious [Vorn@Nightstar-7e1eq8.sub-70-197-73.myvzw.com] has joined #code |
| 01:14 | | Vornicus [Vorn@Nightstar-1l3nul.res.rr.com] has quit [Ping timeout: 121 seconds] |
| 01:16 | | Vorntastic [Vorn@Nightstar-1l3nul.res.rr.com] has quit [Ping timeout: 121 seconds] |
| 01:34 | | readerror [khyjclt@Nightstar-nrrjor.prodain.com] has joined #code |
| 01:35 | < readerror> | THE LODE RADIO IS LIVE AGAIN YAY!! |
| 01:35 | < readerror> | call 415-349-5666 |
| 01:36 | < readerror> | live show @ https://www.youtube.com/watch?v=rXWx3lPlwgE |
| 01:36 | < readerror> | or go to #lrh efnet irc for more information. |
| 01:36 | < readerror> | Namegduf bowlich himi Mahal iospace simon_ abudhabi Vornlicious ktemkin VirusJTG Tamber ToxicFrog Orthia Derakon crystalclaw gnolam Alek Kizor PinkFreud Jessikat TheWatcher jeroud Syloq ErikMesoy mac JustBob jerith Pi Kindamoody[zZz] McMartin Attilla Reiver [R] |
| 01:36 | | readerror [khyjclt@Nightstar-nrrjor.prodain.com] has left #code ["call l0de @ 415-349-5666"] |
| 01:43 | < Vornlicious> | Hey two, got another one for you |
| 01:50 | | * Alek facepalms. |
| 02:13 | | Jessikat [Jessikat@Nightstar-bt5k4h.81.in-addr.arpa] has quit [[NS] Quit: Leaving] |
| 03:07 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 03:07 | | mode/#code [+o macdjord] by ChanServ |
| 03:09 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 03:28 | | Vornlicious [Vorn@Nightstar-7e1eq8.sub-70-197-73.myvzw.com] has quit [[NS] Quit: Bye] |
| 03:28 | | Vorntastic [Vorn@Nightstar-1l3nul.res.rr.com] has joined #code |
| 03:31 | | Vornicus [Vorn@Nightstar-1l3nul.res.rr.com] has joined #code |
| 03:31 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 03:56 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 03:56 | | mode/#code [+o mac] by ChanServ |
| 03:59 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 04:18 | <~Vornicus> | https://cstheory.stackexchange.com/questions/11855/how-do-the-state-of-the-art-p athfinding-algorithms-for-changing-graphs-d-d-l stack exchange is my favorite website for a reason |
| 04:54 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 04:54 | | mode/#code [+o macdjord|slep] by ChanServ |
| 04:57 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 05:24 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 05:24 | | mode/#code [+o macdjord] by ChanServ |
| 05:26 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 05:28 | | Derakon is now known as Derakon[AFK] |
| 07:40 | <&McMartin> | Wow. That is amazing |
| 07:40 | <&McMartin> | I assume I need to blame Debian for this one |
| 07:40 | <&McMartin> | Installing libpng16-dev forces an uninstallation of libgtk-3-dev and refuses to allow a reinstall until I uninstall libpng16 and put libpng12 back |
| 07:40 | <&McMartin> | I guess they can't coexist despite having different names! |
| 07:40 | <&McMartin> | I'm so glad we made them separate packages |
| 07:41 | <&McMartin> | This is technically an Ubuntu system but this sounds like part they'd just take from the standard repos |
| 07:42 | <~Vornicus> | what the drugs |
| 07:42 | <&McMartin> | 1.2 and 1.6 are of course separate packages because Debian firmly believes that an autoupdate system should never actually autoupdate anything because that's just asking for everything to break |
| 07:42 | <&McMartin> | This is the year, man |
| 07:42 | <&McMartin> | Linux on the Desktop |
| 07:44 | <&McMartin> | I also happen to know for a fact that libpng16 and gtk3 coexist just fine on my Fedora system |
| 07:45 | <~Vornicus> | i mean okay py2 vs py3 but |
| 07:46 | <&McMartin> | Debian divides py2 by minor version, IIRC |
| 07:47 | | Kindamoody[zZz] is now known as Kindamoody |
| 07:47 | <&McMartin> | yep, the package is indeed python2.7 |
| 07:52 | <~Vornicus> | bs |
| 07:54 | <&McMartin> | I have complained about this many a time in the past and have always met with angry insistence that this is the only correct way to do things |
| 07:54 | <&McMartin> | I remain unconvinced, which is one of several reasons my primary Linux machine is Fedora-based. |
| 07:54 | <@himi> | Dependencies are frequently not very nice |
| 07:55 | <&McMartin> | (The Xenial machine has been sticking around both because working on Fedora doesn't mean it works on Ubuntu and vice versa, and because Steam only runs properly on Ubuntu) |
| 07:55 | <@himi> | To get into this situation you have packages that depend on 1.2 and break on 1.6, or vice-versa - hence you split them out into different packages |
| 07:56 | <&McMartin> | Which is fine, but then you put them in different places |
| 07:56 | <&McMartin> | You do not uninstall the entire GNOME dev stack and refuse to reinstall it |
| 07:56 | <@himi> | And then you have someone specify the dependencies and conflicts a little too zealously, meaning you end up not being able to have both at once |
| 07:57 | <&McMartin> | Yes, but that should restrict itself to "within the same program" |
| 07:57 | <&McMartin> | Not "on the same system". Otherwise why have them be different packages at all? |
| 07:57 | <@himi> | Debian tends to be fairly careful about that, but given the size of the system and the complexity of the dependencies you've got a decent chance of hitting this problem |
| 07:57 | <&McMartin> | PNG and GTK are not exactly wacky corner-case packages. |
| 07:57 | <@himi> | No, that's not the issue - someone created a package that says "I need 1.2 and I conflict with 1.6" |
| 07:58 | <&McMartin> | No, that's not the issue |
| 07:58 | <@himi> | It sounds like it from your initial description . . . |
| 07:58 | <&McMartin> | The issue is "I refuse to let you install libgtk-3-dev because that would require installing libcairo2-dev and that requires libpng12-dev and I can't install tha tbecause libpng16-dev is installed. Dying." |
| 07:58 | <&McMartin> | Except it wasn't that until after step 1 |
| 07:59 | <&McMartin> | Step 1 was "OK, I need to install libpng16-dev, I'll do that now" |
| 07:59 | | * himi isn't actually saying this is the right way to do it, by the way |
| 07:59 | <&McMartin> | "OK. *installs libpng16-dev* *uninstalls entire GTK3 dev stack*" |
| 07:59 | <&McMartin> | If sudo apt-get install libgtk-3-dev is not the way to do it something has gone tremendousely wrong |
| 07:59 | <@himi> | Yeah - this is driven by the depends/conflicts resolution handling |
| 07:59 | <&McMartin> | OK, but |
| 08:00 | <&McMartin> | In order for that to make sense, then despite being an entirely different package, libpng16 has to prevent libpng12 from existing |
| 08:00 | <&McMartin> | If you're going to do that why are they different packages |
| 08:00 | <&McMartin> | And again, this is libpng |
| 08:01 | <&McMartin> | My solution is actually "fuck it, I only needed it for a test program, I'll just run that test on Fedora systems and revisit Ubuntu only at major milestones" |
| 08:01 | <@himi> | No, some other package in the collection of dependencies which is implied by your command "install libgtk-3-dev" is saying "there can be only one!" |
| 08:01 | <@himi> | Oooh, wait - it's the /dev/ packages |
| 08:02 | <&McMartin> | Yes |
| 08:02 | <@himi> | They /regularly/ have that kind of thing, because they often install files in the same locations |
| 08:02 | <~Vornicus> | why the hell are they doing that that is foolish |
| 08:02 | <@himi> | Sometimes that can be worked around, but not always |
| 08:02 | <&McMartin> | Yeah, my workaround is "just do the work on Fedora |
| 08:03 | <&McMartin> | I only need libpng to load test data into a rendering program anyway, it won't be in the final product. |
| 08:03 | <@himi> | Vornicus: sometimes that's the only way to package things |
| 08:03 | <&McMartin> | Isn't this literally what "alternatives" and pkg-config are for |
| 08:03 | <@himi> | At least, not without requiring changes all over the place, which may not be considered worth it |
| 08:04 | <@himi> | Yes, but alternatives are system-wide (i.e. you can only have one libpng dev files diversion active at any given time), and pkg-config isn't actually universally supported by the libraries . . . |
| 08:05 | <@himi> | I'm surprised that something like libpng would have this kind of problem, but it's not unheard of . . . |
| 08:06 | <@himi> | libjpeg can be a right pain in the arse, for example, despite being about as standardised as you can possibly imagine |
| 08:06 | | * himi had to deal with some of this crap when trying to get stuff working with libjpeg-turbo way back in the day |
| 08:07 | <@himi> | Short of large patch sets that make the problematic libraries play nicely with (say) pkg-config, it's hard to come up with a solution that doesn't break pathologically somewhere or other |
| 08:08 | <@himi> | . . . libgkt-3 development is an unusually big thing to break, mind, but sometimes shit happens |
| 08:08 | <&McMartin> | Right, I'm less patient for this because everything works out of the box on Fedora using only core repo libraries |
| 08:09 | <&McMartin> | While Fedora *is* the "home distro" of the GNOME project, this should not require that level of home team advantage |
| 08:09 | <@himi> | And again I'll note that I don't think this is The Right Way To Do It(tm), it's just that it's hard to know of a better way to handle it in the general case |
| 08:09 | <@himi> | Yeah, Fedora package libgkt3-devel in a way that's more robust in your use case |
| 08:10 | <@himi> | Probably because, yes, GNOME is more interested in Fedora than they are in Debian |
| 08:10 | <&McMartin> | They also, for what it's worth, don't leave libpng1.2 around. |
| 08:10 | <@himi> | On the other hand, OpenStack packaging on Ubuntu is /miles/ ahead of the packaging on Fedora/RHEL - because most of the development happens on Ubuntu |
| 08:11 | <@himi> | This is one of the other reasons Debian can get into this kind of thing more often - they keep a /lot/ of older stuff around |
| 08:11 | <&McMartin> | Fedora's packaging tends towards a much smaller repository but fully self-consistent |
| 08:11 | <&McMartin> | Yeah |
| 08:11 | <&McMartin> | OTOH, I've taken a dim view of that ever since our UQM packager vanished and has been unable to be reasonably replaced |
| 08:12 | <@himi> | Fedora has far less stuff in general, and they package things with far less granularity |
| 08:12 | <&McMartin> | Which is why it's like 8 years behind because nobody would make the required one-line change to the build options to build 0.7 on later GCCs |
| 08:12 | <&McMartin> | gtk3-devel does still have a squillion dependencies but that's because the gnome footprint is enormous |
| 08:12 | | * himi nods |
| 08:13 | <&McMartin> | (And unlike Qt, the cuts are all horizontal instead of mostly vertical, so doing anything in GTK requires basically the entire GNOME stack except for the IPC stuff) |
| 08:13 | <&McMartin> | (Whereas in Qt4, at least, you would have 20 libraries and any given application would need maybe four of them.) |
| 08:14 | <@himi> | It has advantages, but it also has disadvantages, and it ultimately gets down to the core packaging philosophy of the distro, and in some ways the core philosophies underlying the distro (if any) |
| 08:14 | <&McMartin> | (Two of which were always QtCore and QtGui, but those were the only "horizontal" cuts... and QtGui wasn't even that.) |
| 08:14 | <&McMartin> | Yeah. |
| 08:15 | <@himi> | I'd file a bug, though, given the breadth of pain provided by this (if you actually care enough) |
| 08:15 | <&McMartin> | The GTK team is not exactly aligned with the Debian philosophy on anything to begin with, and the sudden design change requirements introduced by Wayland becoming standard faster than everyone expected made things a great deal worse. |
| 08:15 | <@himi> | Why did you need both libpng1.2-dev and libpng1.6-dev? |
| 08:15 | <&McMartin> | I didn't. I needed libpng1.6-dev and libgtk-3-dev, and the latter insisted ultimately on libpng1.2-dev. |
| 08:15 | <&McMartin> | A thing the Fedora repos do not do because it only has 1.6 |
| 08:16 | <@himi> | Huh |
| 08:16 | <&McMartin> | I needed png1.6 because 1.6 introduces the non-insane API for loading images into RAM buffers and I was testing the usage and compatibility of the GtkGlArea widget. |
| 08:16 | <&McMartin> | And packaging a 256x256 32-bit color XPM in a header file did not appeal |
| 08:16 | <@himi> | If you care enough file a bug - I'd be surprised if libgtk3 /really/ depends on 1.2, so it's entirely possible it's an unnecessary issue |
| 08:17 | <&McMartin> | I know for a fact that it does not, but it probably needs to be built not to |
| 08:17 | <&McMartin> | This is only worth filing if it's still true in April. |
| 08:17 | <@himi> | Yeah - hence a packaging bug (with Ubuntu or Debian) rather than GNOME or libpng |
| 08:17 | <&McMartin> | Right. The reason I'd hold off is because Xenial came out at an extremely awkward time |
| 08:18 | <&McMartin> | In that it ships with GTK3 3.18 and will not leave that minor revision |
| 08:18 | <&McMartin> | However, thanks ultimately to changes in what assumptions you get to make when you need to support X11 and Wayland independently, GTK 3.22 deprecates at least part of nearly all code prior to it. |
| 08:19 | <&McMartin> | 3.22 is what's in Stretch, what's on Fedora, Arch, Slackware, everyone that isn't LTS. |
| 08:19 | <&McMartin> | It's also in non-LTS Ubuntu |
| 08:19 | <&McMartin> | April 2018 the next LTS Ubuntu lands, has Wayland by default, and has a sufficiently advanced version of GTK3 to make that be sensible |
| 08:19 | <&McMartin> | The team for this project has explicitly decided it doesn't care about any distro that isn't running at least 3.22 |
| 08:20 | <&McMartin> | But it's nice to run on the latest LTS/stable of everything major if possible, so there's a couple holes carved out for 3.18 now with comments saying "Remove when Ubuntu B is released 18.04" |
| 08:27 | | macdjord is now known as macdjord|slep |
| 08:54 | <@himi> | Yeah, Xenial is very much /the/ thing to target as far as Ubuntu goes, despite being a bit of a pain with things like Unity dying and so forth |
| 08:55 | <@himi> | It'll still be a pain for you even when the next LTS release is out, though - Trusty is still around and still common, even 18 months after Xenial came out |
| 08:55 | <@himi> | 18? No, 16 months |
| 09:15 | <&McMartin> | That's less of an issue for us; those older systems will never take the update we're building now. |
| 09:16 | <&McMartin> | They'll just use the ancient GTK2-based system. |
| 09:17 | <&McMartin> | The thing is, the old UNIX builds all assumed X11 and made direct toolkit calls and such as a result |
| 09:17 | <&McMartin> | That won't fly on a Wayland system unless you trigger emulation modes |
| 09:17 | <&McMartin> | So the goal is to build a GTK version with no X11 intrinsic shenanigans and as a happy side effect have that work on most other major OSes too. |
| 09:17 | <&McMartin> | That will drop the number of ports from, like, 20 down to two. |
| 09:18 | <&McMartin> | (GTK3 and SDL) |
| 09:18 | <&McMartin> | I am personally not convinced that the Win32 port can be completely retired, but Windows is a lot more stable an API than Linux desktop these days, which means it can fairly sensibly just not take updates. |
| 09:19 | <&McMartin> | It is also definitely the case that the native Win32 port is full of garbage though and needs a sharp eye and a ruthless editor |
| 09:20 | <&McMartin> | Which I was originally volunteering to be, but they needed people who were better at GTK and OpenGL, and, well, I'm all right with those too |
| 09:20 | <~Vornicus> | this is your work with vice? |
| 09:20 | <&McMartin> | Replacing our Cairo code with OpenGL code when possible is my latest project. |
| 09:20 | <&McMartin> | Yeah. |
| 09:22 | <&McMartin> | But the test program wasn't the VICE core, it was a side test program so I could test out the widgets I wanted to integrate. |
| 09:22 | <@himi> | Huh - is that really viable, even with OpenGL code? |
| 09:23 | <@himi> | The cross platform stuff, that is |
| 09:23 | <&McMartin> | OpenGL is the *most* compatible option. |
| 09:23 | <&McMartin> | The only part that needs to vary is the GLX/AGL/WGL/EGL stuff |
| 09:23 | <@himi> | Even with all the platform specific setup stuff? |
| 09:23 | <&McMartin> | That's almost always invisible to the programmer. The widget toolkit handles it for you. |
| 09:24 | <&McMartin> | GDK covers all that as part of realizing a GdkGlArea, which is what powers the GtkGlArea. |
| 09:24 | <@himi> | Hm |
| 09:24 | <&McMartin> | Qt does the same, as does SDL. |
| 09:24 | <@himi> | Different from how it worked with earlier stuff, I think |
| 09:24 | <&McMartin> | And GLUT, and the new GLUT whose name escapes me (GLSW or something?) |
| 09:24 | <&McMartin> | You can still do it by hand if you really want to |
| 09:25 | <&McMartin> | But we can't use GLX for this anyway because GLX doesn't work on Wayland. |
| 09:25 | <@himi> | Or maybe not - it's a long while since I tried getting my GTK/OpenGL stuff running on Windows |
| 09:25 | <&McMartin> | Windows spent a fairly long span of time as The Best OpenGL Platform. |
| 09:25 | <&McMartin> | Linux only recently caught up. |
| 09:25 | <&McMartin> | Mac is still like four revisions behind. |
| 09:26 | <@himi> | I had all sorts of issues with code running happily on Linux, but failing on Windows - this was gtk2 code, pushing ten years ago |
| 09:26 | <&McMartin> | That's not a huge deal for us - we're only using stuff in OpenGL 3.2 core, and we're only using stuff that late because it's the minimum version GTK3 will permit you to use |
| 09:26 | <&McMartin> | Gtk2 didn't directly integrate OpenGL and required extensions. |
| 09:26 | <&McMartin> | Gtk2 also never worked very well at all off of Linux. |
| 09:26 | | * himi nods |
| 09:26 | <&McMartin> | GTK3 is... notionally better at this, but it's really, really obviously alien. |
| 09:27 | <&McMartin> | Having to support Wayland and X11 forced a lot of issues, though. |
| 09:27 | <@himi> | So target gtk3 past 3.22, I guess . . . |
| 09:27 | <&McMartin> | Yep, that's the plan |
| 09:27 | <&McMartin> | We need to anyway; we have popup menus and that whole API either doesn't exist prior to 3.22 or is deprecated *in* 3.22 |
| 09:28 | <&McMartin> | And the deprecated APIs are incredibly obnoxious, so no huge loss~ |
| 09:29 | <@himi> | I really need to bring my code up to more recent stuff like this |
| 09:29 | <&McMartin> | Yeah, this particular code was a horrible mess of stacked legacy code, because it dates back to the 1990s |
| 09:29 | <@himi> | I could see gtk3 being a lot better way back when, but it was still fairly new |
| 09:30 | <&McMartin> | Yeah, it is much more stable at this point, to the extent that there may not be any additional non-patchlevel revisions. |
| 09:30 | <&McMartin> | GTK team now admits it's worth using for non-GNOME projects, now that they're moving their focus to GTK4. |
| 09:30 | <@himi> | As long as it doesn't end up totally stagnated, that's a nice state to be in . . . |
| 09:31 | <&McMartin> | But yeah, this code? Started out as Xaw, was hacked into Gtk1, was hacked from that into Gtk2, with a huge pile of crazy preprocessor stuff to make half the code look right to all three of those with one source base |
| 09:31 | <&McMartin> | So the other part of this project is "burn all that down and get a reasonably clean code base" |
| 09:32 | <@himi> | Ouch |
| 09:32 | <@himi> | Yeah, that's not a great situation to be in |
| 09:33 | <&McMartin> | And of course after 25 years there's very little of the original crew left. |
| 09:33 | <&McMartin> | (VICE started life as x64, which was a base-X11 UNIX program) |
| 09:34 | <@himi> | That's not something I've seen in a /long/ time |
| 09:35 | <&McMartin> | The core is still solid and has only improved over time, and it's independent of all the interaction with it |
| 09:35 | <&McMartin> | So at this point it has no real competitors in that space. |
| 09:35 | <&McMartin> | But, you know |
| 09:35 | <&McMartin> | We really don't need the m68k Amiga port in the build system anymore, etc. |
| 09:36 | <&McMartin> | version 2.4 was still Pretty Good, and that will be the last version that got Literally Every Build |
| 09:36 | <&McMartin> | 3.x will have most of the ones still relevant today, but those will freeze once they can be replaced by GTK3-native. |
| 10:01 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 10:01 | | mode/#code [+o mac] by ChanServ |
| 10:04 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 10:18 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 10:18 | | mode/#code [+o macdjord] by ChanServ |
| 10:19 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 10:19 | | mode/#code [+o macdjord|slep] by ChanServ |
| 10:21 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 10:22 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 11:15 | | Degi [Degi@Nightstar-q6nap3.dyn.telefonica.de] has joined #code |
| 11:16 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 11:16 | | mode/#code [+o macdjord] by ChanServ |
| 11:18 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Operation timed out] |
| 11:19 | | Kindamoody is now known as Kindamoody|afk |
| 11:25 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 11:25 | | mode/#code [+o mac] by ChanServ |
| 11:27 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 13:54 | < Degi> | Hm, this capacitor is as thick as my lower leg. |
| 13:57 | <~Vornicus> | you are in *trouble* if that thing stops working |
| 14:14 | < Degi> | I only need it for discharging |
| 14:14 | < Degi> | Only 1300 joules or so |
| 14:14 | < Degi> | Probably still enough to make it vent when it gets internal breakdown. |
| 14:47 | | Vornicus [Vorn@Nightstar-1l3nul.res.rr.com] has quit [Ping timeout: 121 seconds] |
| 15:07 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 15:07 | | mode/#code [+o macdjord|slep] by ChanServ |
| 15:10 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 15:17 | | Jessikat [Jessikat@Nightstar-bt5k4h.81.in-addr.arpa] has joined #code |
| 15:26 | | gnolam [lenin@Nightstar-ego6cb.cust.bahnhof.se] has quit [[NS] Quit: Reboot] |
| 15:27 | | gnolam [lenin@Nightstar-ego6cb.cust.bahnhof.se] has joined #code |
| 15:27 | | mode/#code [+o gnolam] by ChanServ |
| 15:58 | | celticminstrel [celticminst@Nightstar-kroq0e.dsl.bell.ca] has joined #code |
| 15:58 | | mode/#code [+o celticminstrel] by ChanServ |
| 16:13 | | [R] [rstamer@genoce.org] has quit [Ping timeout: 121 seconds] |
| 16:15 | | [R] [rstamer@Nightstar-d7h8ki.org] has joined #code |
| 16:15 | | mode/#code [+ao [R] [R]] by ChanServ |
| 16:56 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 16:56 | | mode/#code [+o mac] by ChanServ |
| 16:59 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 17:00 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 17:00 | | mode/#code [+o macdjord|slep] by ChanServ |
| 17:03 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 17:15 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 17:15 | | mode/#code [+o mac] by ChanServ |
| 17:18 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 17:30 | | Kindamoody|afk is now known as Kindamoody |
| 18:12 | | RchrdB [RchrdB@Nightstar-qe9.aug.187.81.IP] has joined #code |
| 19:02 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 19:02 | | mode/#code [+o macdjord] by ChanServ |
| 19:03 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 19:32 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 19:32 | | mode/#code [+o mac] by ChanServ |
| 19:35 | | macdjord [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 19:35 | <&jerith> | Curious Lipstick is... interesting. |
| 19:35 | | macdjord|slep [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has joined #code |
| 19:35 | | mode/#code [+o macdjord|slep] by ChanServ |
| 19:36 | <&jerith> | It's also the level I've most wanted a fast forward button. |
| 19:38 | | mac [macdjord@Nightstar-a1fj2k.mc.videotron.ca] has quit [Ping timeout: 121 seconds] |
| 19:46 | <&jerith> | That said, Universal Solvent promises to be at least as fiddly, if not quite so elaborate. |
| 19:51 | | * jerith goes back to optimise some earlier solutions instead. |
| 19:56 | | Degi [Degi@Nightstar-q6nap3.dyn.telefonica.de] has quit [[NS] Quit: Leaving] |
| 19:56 | | Jessikat [Jessikat@Nightstar-bt5k4h.81.in-addr.arpa] has quit [Connection closed] |
| 19:57 | | Degi [Degi@Nightstar-q6nap3.dyn.telefonica.de] has joined #code |
| 20:58 | <&McMartin> | You should prioritize Universal Solvent for at least the near future. |
| 20:59 | <&jerith> | I'll get back to it once I have the energy again. |
| 21:00 | <&jerith> | For now the simplicity of the early levels combined with the joy of beating your scores is far more relaxing.~ |
| 21:05 | <&McMartin> | As usual, my claim to fame is "time to 100% completion"~ |
| 21:05 | | * jerith grins. |
| 21:05 | <&McMartin> | Less usual, I ended up leading the pack on my friendslist *anyway* this time, albeit not on speed. |
| 21:06 | <&McMartin> | So I guess my other warning is "past a certain point, optimization becomes actively un-fun" |
| 21:07 | <&McMartin> | I believe my scores for Cost and Area on Purified Gold can only be equaled, not beaten |
| 21:07 | <&McMartin> | And that was 45 minutes of entering sigils, even with access to the undocumented copy-paste operation. |
| 21:08 | <&jerith> | There are several puzzles where my first solution beats yours handily on most or all of the axes, which is a good indication that you haven't done much optimisation on them. |
| 21:08 | <&McMartin> | I only did any optimization at all on some things in Chapter 1 and on Purified Gold, which burned me out. |
| 21:09 | <&McMartin> | That also said, you're talented at this. |
| 21:09 | <&jerith> | I don't plan to do much optimisation of the later levels. |
| 21:09 | <@Alek> | different people have different thresholds for how much effort they'll put in to optimize while having fun. |
| 21:09 | <&McMartin> | The core trick for getting high-speed solutions is IMO spectacularly tedious and I don't like it |
| 21:09 | <@Alek> | for that matter, any given person's thresholds can vary over time, depending on other circumstances. |
| 21:09 | <&jerith> | I'm pleased that it avoids the things that caused me to throw SpaceChem off a mountain. |
| 21:10 | <&McMartin> | That turns out to have a blade on both handles. |
| 21:10 | <@Alek> | I've spent weeks on single levels in some games, but wouldn't bother with them now probably. |
| 21:10 | <&jerith> | But I suspect optimising brings those in. |
| 21:10 | <&McMartin> | I have heard a lot of "This is SpaceChem but _________ so it's much less pleasant" |
| 21:11 | <&jerith> | I could *really* use a fast forward button. |
| 21:11 | <&jerith> | Also breakpoints. |
| 21:11 | <@Alek> | by breakpoints you mean? |
| 21:11 | <&jerith> | Same as in code. |
| 21:11 | <&McMartin> | Like in TIS-100 or SHENZHEN I/O |
| 21:11 | <&McMartin> | That said |
| 21:11 | <&jerith> | "Run to this point, then stop and let me step." |
| 21:11 | <&McMartin> | The way stuff desyncs in the fastest solutions seriously damages the utility of same |
| 21:12 | <&McMartin> | "This point" turns out to be a less meaningful concept than in SpaceChem. |
| 21:12 | <&McMartin> | And *much* less meaningful than Shenzhen, which had it much more cleanly represented. |
| 21:12 | <&jerith> | That would have been really useful in Curious Lipstick, where most of my fiddling was right at the end of the loop. |
| 21:13 | <&McMartin> | Yeah. Curious lipstick is an outlier, and I'll be very curious to see how it gets wrecked by the highest-speed players. |
| 21:14 | <&jerith> | McMartin: I think it would be reasonable to put breakpoints in instruction area. There's already a decent indication of what's running when in there. |
| 21:14 | <&McMartin> | That... implies things about your designs that I know for a fact aren't univeral. |
| 21:15 | | * McMartin has regularly ended up, on his faster solutions, with every row having a different column lighted. |
| 21:15 | <@Alek> | aha |
| 21:16 | <&McMartin> | That was when I then went "Oh. Is that what this is going to be about? Yeah, count me out" |
| 21:16 | | * Alek really prefers designs that are stable and don't desync. D: |
| 21:16 | <&jerith> | All rows have the same period, although the may have different starting points. |
| 21:16 | <&jerith> | *they |
| 21:16 | <&McMartin> | Yes |
| 21:16 | <&McMartin> | To be precise |
| 21:16 | <&McMartin> | In every fast solution I've done or seen, every row has period exactly 4. |
| 21:17 | <&McMartin> | And then scoring is managed by how long it takes to get into its steady state. |
| 21:17 | <@Alek> | having the same period is usually good. or at least with a LCM overall period you can figure out. |
| 21:17 | <&McMartin> | Opus Magnum pads to ensure it's always the same |
| 21:18 | <&McMartin> | That's... not always enough to avoid catstrophic interference |
| 21:18 | <&jerith> | Huh. I haven't seen periods that short. |
| 21:18 | <&jerith> | I'm not sure how that would even be possible for some levels. |
| 21:19 | <&McMartin> | Curious Lipstick is indeed one of the cases where I don't know how that would be done |
| 21:19 | <&McMartin> | It is, not coincidentally, also one of the cases where I have not seen a "fast solution" on par with all the other ones I've seen |
| 21:20 | <&McMartin> | They have a number of extremely characteristic traits IME |
| 21:20 | <&McMartin> | This is one of the things my friends who prefer SpaceChem in fact object to; there are clear "god strats" |
| 21:20 | <&jerith> | Curious Lipstick seems much more "organic chemistry" than everything prior to it. |
| 21:21 | <&McMartin> | But, in case my description was overly gnomic: the period is 4, but that is because there are dozens of products in flight at once. |
| 21:21 | <&McMartin> | Heh. |
| 21:21 | <&McMartin> | It totally is |
| 21:21 | <&McMartin> | It's got a ring-structure as a key component, and it will kill the shit out of you if mishandled even slightly! |
| 21:22 | <&McMartin> | My objection is that the solutions I've made that are actually the prettiest - Mist of Incapacitation was, I think, my favorite - end up not being optimal on any measured metric. |
| 21:23 | <&McMartin> | Though Area is usually the closest, but truly optimizing Area requires incredible patience. |
| 21:23 | <&jerith> | I spent several minutes trying to ohvyq gur evat sebz gur bhgfvqr before I gave up and ohvyg vg va gjb unyirf gung V ernffrzoyrq yngre. |
| 21:25 | <&McMartin> | My solution ohvyg vg sebz gur vafvqr, fnir sbe bar and then that finished just in time for the final assembly |
| 21:25 | <&McMartin> | Slightly faster, in fact; it threw everything into a seemingly useless empty space in the middle of the floorplan and then had to incorporate a delay sigil to prevent it from interfering with the next product. |
| 21:26 | <&jerith> | I considered that, but it would have required shifting everything else around to make space in the right places. |
| 21:27 | | * Alek giggles. |
| 21:27 | <&jerith> | I really like my Voltaic Coil solution, though. |
| 21:28 | <&McMartin> | There are some sleazy things that can be done to make it decide you finished faster on the infinite-polymer products. |
| 21:28 | <&McMartin> | I did that for Very Dark Thread but none of the others. |
| 21:30 | <&McMartin> | For the record, here is my not-the-best-at-anything-but-still-my-favorite MoI solution: https://www.dropbox.com/s/dvv8mwpi884tguk/Opus%20Magnum%20-%20Mist%20of%20Incapa citation%20-%202017-10-19-23-33-03.gif?dl=0 |
| 21:30 | <&McMartin> | I have that one filed under "Function" instead of "Speed" or "Cost" or "Area". |
| 21:31 | | Derakon[AFK] is now known as Derakon |
| 21:31 | <&McMartin> | It's got several bits of obviously wasted motion, but I like how compact it ends up without falling into the trap of the Actually Best Cost And Area Solutions |
| 21:32 | <&jerith> | Mine is one hex smaller and seven cycles slower, but rather less elegant. |
| 21:36 | <&McMartin> | Yeah |
| 21:37 | <&McMartin> | For the others... it seems like it was hilarious and awesome *once* but the fastest solutions have a bit of the clone brush about them and the smallest and cheapest solutions... well, yeah, as you said, where my fast-forward button ;_; |
| 21:38 | <&jerith> | A fast forward button seems like a thing they could add pretty easily. |
| 21:38 | <&McMartin> | I'll have to check in on you later today to see how you're doing |
| 21:39 | <&jerith> | They added an exit button to the puzzle screen yesterday. |
| 21:39 | <&McMartin> | My 'paid a little attention' stuff led my friendslist for a considerable time. |
| 21:39 | <&McMartin> | Until the guy who Got Speed went to it and got like an eighth the time of everyone else |
| 21:40 | <&jerith> | I know this because the game doesn't notice my touchbar escape button, so I had to pull some trickery with bettertouchtool to male it possible for me to leave puzzles without completing them at all. |
| 21:41 | <&jerith> | *make |
| 21:41 | <&jerith> | Also, I was well into chapter two before I even noticed this was a problem. |
| 21:48 | <&McMartin> | So, yeah. Purified Gold, I have two aggressively optimized solutions. |
| 21:48 | <&McMartin> | One takes 112 cycles. |
| 21:48 | <&McMartin> | The other takes 2,948 and has no repeat sigils. |
| 21:49 | <&McMartin> | I suggest not doing that unless there's something you are desperately trying to not pay attention to elsewhere. |
| 21:56 | | * McMartin heads off to refactor some GTK3 application code. |
| 21:58 | <&jerith> | Okay, then. I didn't manage to match your 36 cycles in Waterproof Sealant, but in my fourth attempt to do better than 46 I managed to get 27. |
| 22:37 | | Kindamoody [Kindamoody@Nightstar-eubaqc.tbcn.telia.com] has quit [Connection closed] |
| 22:38 | | Kindamoody|autojoin [Kindamoody@Nightstar-eubaqc.tbcn.telia.com] has joined #code |
| 22:38 | | mode/#code [+o Kindamoody|autojoin] by ChanServ |
| 22:38 | | Kindamoody|autojoin is now known as Kindamoody |
| 23:32 | | * McMartin gets an idea for MoI that might not be any faster but should look neat. |
| 23:37 | <&McMartin> | jerith: You're still only in 2nd place in my list for Sealant. The record's 25. |
| 23:37 | <&McMartin> | Er |
| 23:37 | <&McMartin> | 26. |
| 23:37 | <&McMartin> | This seems to be the kind of thing where "time to steady state" is what is being competed on |
| 23:37 | | * McMartin is kind of meh about that. |
| 23:38 | <&jerith> | You're the only other person in my leaderboards, thankfully. |
| 23:39 | <&jerith> | For the most part, I don't care about comparing myself to the best in the world. Only the people I can chat to on IRC. ;-) |
| 23:51 | <&McMartin> | Yeah, I have more Zachers on my friendslist than you do~ |
| 23:51 | <&McMartin> | Also, that solution for MoI worked *spectacularly* well |
| 23:51 | | Jessikat [Jessikat@Nightstar-bt5k4h.81.in-addr.arpa] has joined #code |
| 23:51 | <&McMartin> | Dropped me to 30 cycles, then tuning the setup got it down to 28. |
| 23:52 | <&jerith> | \o/ |
| 23:53 | <&McMartin> | It's got some very nice almost-symmetry. |
| 23:53 | <&McMartin> | I like some slightly-broken symmetry in my solutions. |
| 23:59 | | Degi [Degi@Nightstar-q6nap3.dyn.telefonica.de] has quit [Connection reset by peer] |
| --- Log closed Sun Oct 29 00:00:53 2017 |