code logs -> 2019 -> Sun, 08 Dec 2019< code.20191207.log - code.20191209.log >
--- Log opened Sun Dec 08 00:00:39 2019
00:39 Omega [Alek@Nightstar-o723m2.cicril.sbcglobal.net] has joined #code
00:39 mode/#code [+o Omega] by ChanServ
00:39 Alek [Alek@Nightstar-o723m2.cicril.sbcglobal.net] has quit [NickServ (RECOVER command used by Omega)]
00:39 Omega is now known as Alek
01:11
<@gnolam>
TheWatcher: that doesn't sound like Blender.
01:11
<@gnolam>
I will assume it's some temporary anomaly.
01:20 Kindamoody is now known as Kindamoody[zZz]
01:45 Degi [Degi@Nightstar-qorofn.dyn.telefonica.de] has quit [Ping timeout: 121 seconds]
01:55 Degi [Degi@Nightstar-ekqdcu.dyn.telefonica.de] has joined #code
02:40 Pinkhair [user1@Nightstar-g7hdo5.dyn.optonline.net] has joined #code
02:42 Pink` [user1@Nightstar-g7hdo5.dyn.optonline.net] has joined #code
02:43 Pink [user1@Nightstar-g7hdo5.dyn.optonline.net] has quit [Ping timeout: 121 seconds]
02:44 Pinkhair [user1@Nightstar-g7hdo5.dyn.optonline.net] has quit [Ping timeout: 121 seconds]
04:32 catalyst [catalyst@Nightstar-v6lb30.cable.virginm.net] has joined #code
04:37
< catalyst>
I've been doing advent of code and it's been fun and confidence building :)
04:42
<&jeroud>
catalyst: That's why I look forward to it every year. :-D
04:42
< catalyst>
^^
04:42
< catalyst>
I honestly didn't want to start it in case I couldn't do it
04:42
< catalyst>
but I might be the only one I know who would have thought I couldn't hack it
04:43
<&jeroud>
There has been exactly one problem ever that I've been unable to solve.
04:48
<@Alek>
Fermat's Last Theorem? :P
04:51
<&jerith>
2015 day 19. I only discovered AoC in 2016, but 2015's been the one I go back to when I've run out of current-year problems and am bored.
04:52
<&jerith>
ToxicFrog, anyone else who wants to join: 148552-220e29ae is the leaderboard code.
04:54
< catalyst>
I joined it :)
04:55
<&jerith>
\o/
04:56
<~Vornicus>
argh I have four days of them to do now
04:56
< catalyst>
I'm up to day 4
05:07 Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed]
05:11 Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has joined #code
05:11 mode/#code [+qo Vorntastic Vorntastic] by ChanServ
05:52
<&jerith>
Whee! Finished both parts before taking my meds.
05:54
<&jerith>
Also, I want to complain but minor spoilers.
05:55
<&jerith>
But: "There's one every year, and I kind of wish there weren't because it messes with my system" should be vague enough.
05:58 * McMartin checks 2015 day 19
05:58
<&McMartin>
... I'm a little surprised that one blocked you, unless it's just a matter of not getting around to it.
05:59
<&jeroud>
Part 2 is prohibitively computationally expensive unless you exploit some non-general property of the input.
06:00
<&jeroud>
Some non-obvious non-general property.
06:02
<&jeroud>
I think it was a poorly chosen problem, and the fact that I haven't seen anything similar since indicates it's a mistake that they've learned from.
06:02
<&McMartin>
My solution does not appear to so exploit, but I cannot time it until I translate it into Python 3.
06:04
<&McMartin>
It is possible I thoughtlessly relied on something that is not an invariant in the general case, but it sure doesn't look like it.
06:04 * jeroud thinks.
06:05
<&McMartin>
It is also possible that this is not the script I used to get the solution.
06:05
<&jeroud>
I'm pretty sure I've solved it by getting lucky with path selection or something.
06:06
<&McMartin>
My solution is extremely aggressive about backtracking and it also seems lackadaisical about best-first searching and searches longer to make sure.
06:07
<&McMartin>
That said, it also looks like it was not designed to give an answer _quickly_ in any way.
06:07
<&jeroud>
Yeah.
06:08
<&jeroud>
If it were "sit and wait for a few dozen seconds" I wouldn't have as much of an issue with it.
06:09
<&jeroud>
In other news, I think today I shall swap out my test runner.
06:10
<&jeroud>
I chose the one with the display I didn't like because it runs tests in parallel.
06:10
<&McMartin>
That *also* said, if you're hitting actual infeasibility
06:11
<&McMartin>
Then there's a place you should be turning back or halting the search where you are not
06:11
<&jeroud>
But since I have already exploited the ease with which Clojure enables concurrency, that is less of an advantage.
06:11
<&jeroud>
So prettier output will likely be better.
06:12
<&jeroud>
Yeah, it's definitely some issue with the details of my search algorithm.
06:13
<&jeroud>
But I've tried several variations, and eventually got bored-frustrated enough that I gave up.
06:14
<&jeroud>
Aside from that, no AoC problem has ever taken me more than a day and a half or something.
06:14
<&jeroud>
(There are some later 2015 problems that I may not have tried, though.)
06:17
<&McMartin>
I did have the occasional tendency to do things like run a solution overnight if I didn't get a faster one in time.
06:18
<&McMartin>
It occurs to me that there _is_ a nongeneral property that I'm not exploiting.
06:18
<&jeroud>
I only discovered the non-general property by reading reddit posts later.
06:19
<&jeroud>
(I don't recall what it is, even.)
06:21
<&jeroud>
Instead of pacing up and down the house, I should take my meds and maybe eat something.
06:23
<&jeroud>
Also: boop! https://usercontent.irccloud-cdn.com/file/ZxpeVfPw/IMAG1220_1.jpg
06:27
<&McMartin>
Yeah, I srongly suspect I ran this one overnight originally. That said, checking the leaderboard, it seems like the total number of solutions isn't an outlier on the general completion curve and the fastest solutions were within 20 minutes.
06:32
<&McMartin>
OK, yeah, this was clearly a brute-force affair at first if this is even how I solved it.
06:32
<&McMartin>
I only can say with confidence that, at some point, i did, because there is the answer.
06:33 * McMartin kills the process after 12 minutes and 1GB of allocation.
06:38
<&McMartin>
It also occurs to me that the Python 3 version of this would be doing quite a bit more work because of the differences between strings and bytes.
06:39 * Vorntastic looks that problem up.
06:39
<~Vorntastic>
Oh boy.
06:40 celticminstrel is now known as celmin|sleep
06:40
<~Vorntastic>
What's part B?
06:44
<~Vorntastic>
Though I'm reminded, I need to farm out the Intcode terp to a library so I can use it in multiple projects and I don't know how to do that in C#
06:52
<&McMartin>
Part be is "starting from the string 'e' what is the smallest number of replacements required to produce the target molecule"
06:54
<~Vorntastic>
Oh boy
06:58
< catalyst>
woo I did day 4
06:58
< catalyst>
I just brute forced it because I can
06:58
< catalyst>
I'm sure there's some fancy math
07:02
<~Vorntastic>
As the fancy mathematician
07:03
<~Vorntastic>
I am reasonably certain there's a way to count without enumerating but that's a pain
07:03
<~Vorntastic>
I couldn't find it in reasonable time.
07:04
<~Vorntastic>
I did however make sure to only generate numbers that matched the first criterion, about non-descending; it takes it from 600k to 3k candidates.
07:05
< catalyst>
I wonder if this is the spoiler thing
07:06
< catalyst>
#aodspoilers
07:06
< catalyst>
oops
07:06
< catalyst>
#aocspoilers
07:08
<&jeroud>
Yeah, it's spoilery enough to take there.
07:44
<&jeroud>
So, Rich Hickey talks are great.
07:44
< catalyst>
agreed
07:45
<&jeroud>
catalyst: Have you watched any Kevlin Henney talks? I like his style just as much, although it's completely different.
07:45
< catalyst>
yeah, I enjoy his talks as well
07:45
< catalyst>
and Alexandrescu for basically the opposite reasons
07:45
< catalyst>
Alexandrescu is annoying because he's usually right but he's also mad :P
07:46
<&jeroud>
I don't think I've watched any Alexandrescu.
07:46
< catalyst>
it's usually very mechanical low level stuff
07:46
<&jeroud>
I'm always looking for excellent programming talks.
07:46
< catalyst>
he's very much a language designer and optimiser
07:47
<&jeroud>
Ah. So probably not as useful to me, then.
07:48
<&jeroud>
I like the occasional low-level dive into a language I'm very familiar with, but tend to get lost otherwise.
07:48
< catalyst>
yeah, he's a very in-depth kind of person
07:50
<&jeroud>
Sandi Metz is also good, although I sometimes find her stuff a little less language-agnostic than I prefer.
07:52
<&jeroud>
I like Jose Valim's talks as well, but I don't know how much value they have outside Elixir.
09:01 Kindamoody[zZz] is now known as Kindamoody
11:04 Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has joined #code
15:38 Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code
15:38 mode/#code [+qo Vornicus Vornicus] by ChanServ
16:02 celmin|sleep is now known as celticminstrel
17:31 Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity]
17:37 * TheWatcher eyes MSVS
17:37
<@TheWatcher>
Why do I need to reboot to finish updating you, seriously? WTF is this, 2003?
17:39
<&[R]>
Yes, MS still thinks manditory file locks are a good idea
17:45 VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has quit [Connection closed]
17:45 VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has joined #code
17:45 mode/#code [+ao VirusJTG VirusJTG] by ChanServ
20:51 himi [sjjf@Nightstar-v37cpe.internode.on.net] has quit [Ping timeout: 121 seconds]
22:16 JustBob [justbob@Nightstar.Customer.Dissatisfaction.Administrator] has quit [[NS] Quit: This unit is entering a hibernation state.]
22:20 JustBob [justbob@ServerAdministrator.Nightstar.Net] has joined #code
22:20 mode/#code [+o JustBob] by ChanServ
22:48 Netsplit Deepthought.Nightstar.Net <-> Krikkit.Nightstar.Net quits: @PinkFreud
22:51 Netsplit over, joins: @PinkFreud
22:54 Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has quit [Ping timeout: 121 seconds]
23:09 himi [sjjf@Nightstar-1drtbs.anu.edu.au] has joined #code
23:09 mode/#code [+o himi] by ChanServ
23:16 Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed]
23:50 Kindamoody is now known as Kindamoody[zZz]
--- Log closed Mon Dec 09 00:00:41 2019
code logs -> 2019 -> Sun, 08 Dec 2019< code.20191207.log - code.20191209.log >

[ Latest log file ]