IRC logs for #openttd on OFTC at 2020-07-08
⏴ go to previous day
00:23:46 *** andythenorth has joined #openttd
00:25:02 *** snail_UES_ has joined #openttd
02:16:38 <DorpsGek_III> [OpenTTD/OpenTTD] techgeeknz commented on pull request #8217: Port GUI rendering improvements from JGRPP https://git.io/JJtdg
02:42:32 <CornsMcGowan[m]> @_oftc_jewo_GRFS:matrix.org: vcpkg, you need to checkout an older tag
02:43:11 <glx> CornsMcGowan[m]: only for non cmake
02:43:53 <CornsMcGowan[m]> oh its v453000 :0
02:58:39 <CornsMcGowan[m]> whats the usual size of network commands? (in bytes)
03:05:03 *** mindlesstux6662 has quit IRC
03:06:11 *** mindlesstux6662 has joined #openttd
04:47:11 <CornsMcGowan[m]> also, i remember there being a document that explains the different between files like uh
04:47:25 <CornsMcGowan[m]> viewport.h, viewport.cpp, viewport_type.h and viewport_func.h
04:47:30 <CornsMcGowan[m]> does anybody know where it is?
04:47:49 <CornsMcGowan[m]> oh theres no viewport.h
05:07:51 <Eddi|zuHause> there was a really old and outdated set of wikipages called "dev black book" or something
05:19:04 <Eddi|zuHause> like i said, outdated
06:25:19 *** Gustavo6046 is now known as Guest2702
06:25:23 *** Gustavo6046 has joined #openttd
06:47:43 <DorpsGek_III> [OpenTTD/OpenTTD] HotelCalifornia commented on issue #8201: Consider upgrading CMake to latest release https://git.io/JfMP8
07:02:50 <CornsMcGowan[m]> any way i can control a loop within OnRealtimeTick() to only execute for a fixed amount of time (say 1millisecond)
07:03:18 <CornsMcGowan[m]> i don't know what timing functions i should use
07:43:40 *** HerzogDeXtEr has joined #openttd
08:01:56 *** sla_ro|master has joined #openttd
09:32:39 <CornsMcGowan[m]> i guess i should use whatever the framerate window uses
09:42:54 <DorpsGek_III> [OpenTTD/OpenTTD] fsimonis commented on issue #8201: Consider upgrading CMake to latest release https://git.io/JfMP8
09:48:14 <nielsm> remember to stick to the C++11 things if you intend to get merged into master
09:49:16 *** andythenorth has joined #openttd
10:10:55 *** gelignite has joined #openttd
11:10:16 <CornsMcGowan[m]> wait what new C++17 features are there
11:10:19 <CornsMcGowan[m]> i should google this
11:36:08 *** arikover has joined #openttd
11:55:11 *** tokai|noir has joined #openttd
11:55:11 *** ChanServ sets mode: +v tokai|noir
12:06:43 <CornsMcGowan[m]> should i be worried about the performance of selecting large areas (eg demolish, landscape or tree tools) in debug build?
12:27:17 <planetmaker> oh how lovely. Someone complaining that we don't distribute sources for the newgrfs under GPL. Told him to check our TOS - and to ask authors of the NewGRFs for source.
12:43:15 *** andythenorth_ has joined #openttd
13:01:44 *** HerzogDeXtEr1 has joined #openttd
13:12:09 <andythenorth_> planetmaker presumably trooper?
13:12:29 <andythenorth_> if so, it's already had quite enough discussion
13:26:23 *** snail_UES_ has joined #openttd
13:32:15 <planetmaker> andythenorth_, dunno... Mikhail A
13:32:40 <planetmaker> no. Going the very broad general way
13:32:53 <planetmaker> "you distribute stuff under GPL, thus you must distribute its sources"
13:33:14 <andythenorth_> I don't want to say person A is person B if is B is someone else
13:33:23 <andythenorth_> but all this was covered on Monday
13:33:50 <andythenorth_> it's what you said already, we don't 'distribute' in the way GPL means
13:34:04 <andythenorth_> we're just plumbing / pipes / media
13:34:36 <andythenorth_> there was some blah blah around it here and in forums relating to CZ set, Pikka seems to have deleted it from forums though
13:35:28 <andythenorth_> 'nothing to see, move on' :)
13:36:38 <planetmaker> well, that guy replied already. And seems to not have understood my answer, that we have a special license as by our TOS, thus GPL requirements do not apply for us...
13:37:23 <andythenorth_> logs from Monday will show TB explaining it
13:37:39 <andythenorth_> or you could leave it alone and get on with life :)
13:38:08 <planetmaker> I shall not reply further :)
13:38:19 <planetmaker> things which needed saying have been said :)
13:39:51 <SpComb> yeah, you have a separate license to distribute via the bananas tos, so the GPL isn't really relevant
13:42:11 <SpComb> did the uploader re-upload it with a different license?
13:45:33 <LordAro> definitely sounds like trooper
13:53:35 <SpComb> the uploader might be breaking something by attaching a GPL license to something that they are not the sole author of, and the other authors clearly didn't intend it to be under the GPL. If I was one of those other authors, I would certainly complain that my works are being improperly distributed
13:54:54 <planetmaker> which were actually banned from bananannas
13:55:27 <planetmaker> I couldn't be arsed to file a DMCA to github about it, though
13:55:56 <andythenorth_> TB was quite clear, we get a complaint, we take it down
13:56:00 <andythenorth_> no complaint, no action
13:56:10 <andythenorth_> complaint = from copyright owner
13:56:14 <planetmaker> but a valid complaint...^^^
13:56:31 <andythenorth_> has to be valid yes +1
13:56:48 <planetmaker> that has always been our position. And it's the only one you can take as content distributor as we are in this case
13:57:29 <planetmaker> especially as we ask the uploader to ensure that s/he is allowed to upload what is being uploaded... so... yes
13:58:45 <SpComb> downloading something with a GPL license attached doesn't really give you any special rights. If you don't have the source, then you cannot comply with the license and benefit from being able to redistribute. Whoever distributed it to you may be breaking the original author's copyright, unless they are the copyright holders or have a separate license to distribute
13:59:58 <dP> there can be some notification on bananas or something if invalid license was used though
14:00:18 <dP> just so someone else who uses that seemingly gpled content doesn't get complaints out of the blue
14:00:18 <planetmaker> strictly speaking, that's problematic for any set which is developed by more than one person. It would need permission by each contributor
14:00:54 <planetmaker> however, those who use this license in OpenTTD context mean to use it in the way it is being used...
14:01:20 <planetmaker> @dP, what makes the license invalid and how do you tell?
14:02:06 <planetmaker> seems totally impractical. You basically... ask for an upload filter and manual review of each upload. Which basically means the service ceases to exist.
14:02:16 <dP> planetmaker, well, author of said grf admitting he used the wrong license should be enough to tell :p
14:02:21 <planetmaker> Luckily we are not there yet
14:03:48 <andythenorth_> if we *know* the license is invalid we would ideally take it down, as we shouldn't be distributing poisoned GPL things, it makes a mess for people who reuse
14:03:54 <andythenorth_> but our ToS don't say that
14:03:59 <andythenorth_> it is what it is
16:01:08 *** Speeder has joined #openttd
16:25:03 *** WormnestAndroid has quit IRC
16:25:16 *** WormnestAndroid has joined #openttd
16:42:25 *** gelignite has joined #openttd
17:09:20 <TrueBrain> <planetmaker> andythenorth_, dunno... Mikhail A <- without reading most of the rest of the conversation, yes, that is trooper; see the forum for that evidence :)
17:09:25 *** Wormnest has joined #openttd
17:09:56 <andythenorth_> pikka has deleted it from the forum
17:10:07 <TrueBrain> ah; well, then believe me with my blue eyes :)
17:10:11 <TrueBrain> which I do not have :P
17:10:13 <andythenorth_> I believe you :)
17:10:17 <planetmaker> but from what remains,that's still conclusive
17:10:44 <planetmaker> however *I* do have a copyright case for relasing the source of the CZTR :P
17:11:11 <TrueBrain> you want me to reply on his mail btw?
17:11:53 <planetmaker> I replied to the first. The 2nd proves he didn't read/understand my 1st
17:11:55 <TrueBrain> your reply on the first was nice btw :)
17:12:11 <TrueBrain> yeah, so we can reply to him: we told you this on IRC too already, you misunderstand licensing
17:12:19 <TrueBrain> but we can also just not reply :) all the same to me!
17:13:09 <TrueBrain> I love how he things licenses stack :) The dual-license concept is completely lost on him :)
17:13:36 <TrueBrain> but good to see he still hasn't took our advise to heart: learn to communicate in a polite way
17:14:57 <TrueBrain> "your opinion on how licenses work is not relevant; please contact the original author to figure this out" :D
17:15:39 <andythenorth_> so much words :)
17:15:46 <andythenorth_> did anyone sort out my IAM identities yet?
17:15:52 <andythenorth_> what? Only I can do that?
17:16:45 <TrueBrain> it just saddens me that people can get in this mode of "harm was done to me, so now everything has to burn with me" :( This is an Open Source game, run for free .. give it some slack .. meh
17:17:23 <andythenorth_> don't be sad Truebrain!
17:17:28 <andythenorth_> it's only cartoon tanks!
17:17:31 <andythenorth_> oh wait, wrong game
17:34:02 <andythenorth_> lol the problems we have
17:34:08 <andythenorth_> and the problems in Blitz
17:34:12 <andythenorth_> are so different
17:34:22 <andythenorth_> a 5-min MMORG tank game, made in Belarus
17:34:51 <andythenorth_> played by children, pepe teenagers, and russian gopniks
17:34:55 <andythenorth_> and men who've been in the army
17:35:03 <andythenorth_> and men who would never ever survive the army, but wish they were hard
17:35:15 <andythenorth_> and then worst of all, people like me
17:35:50 <andythenorth_> they banned most of the user chat recently, which is both great, and takes some of the fun away
17:37:12 <andythenorth_> lol the people who think 'kid' is this major diss
17:37:24 <andythenorth_> that's how you know they're 13
17:43:42 <CornsMcGowan[m]> not major but i guess its somewhat disrespectful
17:44:58 <andythenorth_> well they also say stuff like "I hope you and all your family die"
17:45:11 <andythenorth_> so ya know, as that goes 'kid' is no drama
17:45:27 <andythenorth_> I never know whether the pepe ones are really alt-right, or just out to trigger the libs
17:45:37 <andythenorth_> I bet they don't quite know either
17:45:46 <LordAro> there's a certain amount of thoughtfulness, they did think of your family as well
17:49:18 <CornsMcGowan[m]> pepe faces get used by everybody nowadays
17:49:29 <CornsMcGowan[m]> regardless of political leaning
18:08:02 *** gelignite has joined #openttd
18:19:28 <CornsMcGowan[m]> what do i use to print to debug?
18:24:27 <LordAro> DEBUG(), funnily enough
18:34:40 <dP> I usually just fprintf(stderr,...)
18:35:40 <dP> works without includes, always know how to use, easy to grep when done
18:38:07 <dP> though for server I have some nice functions with fmt stuff so I don't have to deal with % formatting shit
18:54:47 *** frosch123 has joined #openttd
19:03:43 *** Progman has joined #openttd
19:17:37 <dP> crazy idea of the day: make in-game advisor that suggests good money-making options
19:19:59 <frosch123> are you referring to subsidies or to busy bee gs?
19:21:07 <dP> well, I guess subsidies could be an alternative if they actually suggested anything good
19:21:45 <dP> though advisor could work in mp while subsides don't really go well there
19:23:09 <andythenorth_> crazy idea of the day: draw the missing industries for FIRS v4
19:23:15 * andythenorth_ should do stuff
19:27:59 <Borg> dP: crazy idea indeed... ;) in vanilla OpenTTD making money is damn easy...
19:28:17 <Borg> anyway... 2017... Monorail
19:28:37 <dP> Borg, in any openttd it's easy, but idea is to help new players so they can actually play instead of struggling to survive
19:28:42 <Borg> operating profit... between +330000$ and +87000 yearly...
19:29:09 <Borg> dP: come on.. its a part of a game...
19:29:17 <Borg> few failures.. and boom... you got it..
19:29:28 <Borg> also there are shitload of game tutorials
19:29:43 <dP> Borg, idk, to me it seems 90% of players newer get past that part
19:30:04 <Borg> really? if they do.. its good they gone ;)
19:30:27 <Borg> but. of coz.. go do it :) if you feel its fine...
19:30:47 <Borg> I strugle.. to make game harder and more interesting..
19:31:14 <dP> who said I'm going to do it? it's just a crazy idea :p
19:31:27 <Borg> 2 new GS things got my atention for a while :) auto loan + tax
19:31:36 * dP working on 1mil CB server right now
19:32:08 <Borg> CB = callback server? I was like wtf
19:32:27 <dP> Borg, yeah, you can check it out if you want, it's running on citymania test server
19:32:37 <Borg> I like city builder GS scripts.. kinda nice.. too bad games end so quick ;)
19:32:44 <dP> Borg, only 3 people managed to finish it so far iirc
19:32:55 <Borg> whats goal and limitations?
19:33:02 <Borg> im kinda.. sandbox player... really..
19:33:14 <Borg> for me.. game really start after 2001 :D
19:33:19 <Borg> but I usualy start at 1960
19:34:19 <dP> ehm, "sandbox" and "make game harder" are like polar opposites :p
19:34:38 <Borg> ok.. maybe not harder.. but more.. interesting and chalenging..
19:34:59 <dP> and best score for 1 mil is 9 hours currrently so it's not exactly a short game xD
19:35:11 <dP> though I guess it's doable in like 4 by a good team
19:35:40 <Borg> oh... well.. :) maybe I will take a look. when I will get bored w/ BSPI
19:35:49 <Borg> but I run server for months.... (single game)
19:36:07 <Borg> last game I ended in 2700 or so..
19:37:29 <nielsm> transport fever 2 released a major update that looks like it has added extensive game scripting support (lua based), so you can add custom behaviour there now
19:38:12 <Eddi|zuHause> i haven't checked in a while, but are there curved stations meanwhile (again)?
19:38:47 <dP> Borg, well, it's hard to stay focused more than few hours so for competitive game 9 hours is already very long
19:38:48 <nielsm> Eddi|zuHause: the new scripting probably allows you to generate curved stations dynamically too
19:39:06 <nielsm> but I haven't looked in detail on it either yet
19:39:13 <Borg> dP: true.. thats why I prefer to play sandbox.. I play openttd to relax.. not to compete..
19:39:22 <Borg> or.. soft compete.. :) if anything
19:39:23 <Eddi|zuHause> the whole station module system looked good on paper, but when playing it wasn't quite right...
19:39:37 <Borg> oh. I need to update openttd... or rather side install
19:39:38 <Eddi|zuHause> especially the road stations
19:40:34 <dP> Borg, yeah, a lot of people play like that, pretending to compete but not really :)
19:40:35 <nielsm> yeah the trf2 road stations are annoying to build and the modules don't always do what I expect them to... and the modular stations cause huge updates across the entire map it seems, building a single module freezes the game for far too long
19:42:22 <Wolf01> Bah, again want to do nothing... I'm like worried about getting bored, so I get bored
19:44:33 <dP> oh, another crazy idea I forgot: allow several currencies/resources
19:45:47 <dP> one example can be energy "coin"
19:46:05 <dP> produced on power plant, used for... something xD
19:46:22 <dP> like, electric trains running cost
19:47:10 <dP> or to make road and rail pieces a limited resource
19:47:36 <CornsMcGowan[m]> which of steady clock or high resolution clock should i use? framerate window uses high res but several websites suggest steady clock
19:48:05 <Borg> doh dP .... whats going on there?
19:48:08 <CornsMcGowan[m]> although i dont need much accuracy (loop for 2ms)
19:48:10 <Borg> I see gazylions of industries....
19:48:46 <dP> CornsMcGowan[m], if you're still doing those planner tools you probably better not use any
19:49:07 <dP> CornsMcGowan[m], ah, for cutting of pathfinding though... nvm
19:49:30 <nielsm> CornsMcGowan[m]: I think on most implementations high res and steady are the same
19:49:43 <dP> Borg, ehm, stuff xD there is some info in the story book
19:49:44 <CornsMcGowan[m]> mm yeah to only calculate for a certain period per realtime tick
19:49:55 <nielsm> but if you want to measure 2 ms you should use one that has at least microsecond precision
19:50:11 <CornsMcGowan[m]> guess i'll follow the precedent set in the codebase (highres)
19:51:39 <Borg> dP: too dense to my taste..
19:51:46 <Borg> too many cities.. and industries..
19:52:08 <dP> Borg, yeah, my vpn is lagging
19:52:15 <dP> Borg, you need a lot of stuff for 1 mil
19:52:18 <Borg> why the hell you play over vpn? ;)
19:52:43 <Borg> dP: I wonder how it would mix w/ BSPI.. :D
19:52:47 <dP> Borg, also makes picking spots a bit more interesting than just sorting industry list by production
19:53:55 <dP> Borg, i do everything over vpn, also it's usually not that bad
19:54:05 <dP> Borg, I've no idea what BSPI does
19:56:50 <dP> Borg, it's hard to understand gameplay implications by just reading
19:57:14 <Borg> yeah.. I read game mechanics... interesting stuff
19:58:16 <Borg> dP: in minial settings.. BSPI should not change much... except it have stockpiling industries...
19:58:38 <nielsm> BSPI forces you to complete the cargo chains since otherwise the stockpiles overflow and you don't get paid for delivery
19:58:57 <Borg> hence you need to grow ciiities
19:59:12 <Borg> at least to the point they accept goods..
19:59:38 <Borg> also.. full complete is not that necessary.. you can provide minimal amounts.. and they will vanish at minimum 4% production ;)
19:59:44 <Borg> small profits.. but at begining it will do
19:59:44 <dP> well, usually stuff with its own agenda doesn't mix well with goal games unless they're specifically designed for it
19:59:49 <dP> like cargodist for example
20:01:50 <nielsm> BSPI would pair well with a GS that ensures every town with a secondary industry also has a power plant, at the minimum
20:01:54 <dP> I'd like to make some newgrf servers but probably not very soon
20:02:10 <nielsm> and in general something that controls where the secondary industries are placed perhaps
20:02:11 <dP> as imo they should look and feel very different from vanilla and that's a lot of work xD
20:02:59 <Borg> nielsm: I fixed it differently.. first.. my GS preseed industries...
20:03:16 <Borg> then.. for 20 years. rebuilds them... after that.. player need to found them constantly..
20:03:36 <Borg> also... w/ autoloan.. you can build your own powerstation.. if you really need it.. costly.. but.. may pay off
20:03:54 <Borg> auto loan can have limit up to 25M gbp
20:04:08 <Borg> but that model is of course fun for sandbox..
20:05:08 <nielsm> hm, can GS set a company's loan to something wild disregarding the game settings max loan?
20:05:13 <dP> funding stuff isn't an issue in competive
20:05:22 <Borg> I do it internally in GS
20:05:25 <dP> even for that 1 mil top players fund a lot
20:06:45 <nielsm> I want to have a subsidy GS where you promise a town to provide a service, get some funds to set up the service (ahead-of-time subsidy), and if you fail to do it then those funds turn into debt
20:08:45 *** b_jonas has joined #openttd
20:09:01 <dP> nielsm, promise a service -> get money -> "local authority refuses to allow this" :p
20:12:29 <dP> also knowing openttd economy best strategy would probably be to get money, spend it elsewhere and pay off the debt xD
20:16:27 <Borg> at one point.. I would like to write some simple city growing script.. for my BSPI games... it will spice games up a bit more...
20:18:23 <dP> hm, I know one reason I'll probably never use bspi on a server... they look vanilla :p
20:19:04 <Borg> they are vanilla... :) I cant do gfx
20:19:32 <dP> Borg, I can do whatever mechanics I want, I just need sprites xD
20:19:35 <Borg> BSPI is just logic.. and honestly? I love vanilla gfx..
20:20:13 <Borg> I just wanted more complicated logic out there.. for long fun games.. hence BSPI has born..
20:21:12 <Borg> dP: sprites are not enough :)
20:21:23 <Borg> I tried to improve things... and gfx failed.. miserably ;)
20:22:37 <supermop_Home> do people still buy drobos or other NAS things?
20:23:16 <andythenorth_> I bought one in 2003
20:23:48 <andythenorth_> if I wanted to have the actual files for a lot of HD movies I'd buy one
20:23:51 <andythenorth_> otherwise....why?
20:24:29 <andythenorth_> not needed for streaming media
20:24:38 <andythenorth_> useless for backups
20:24:49 <supermop_Home> i do have a fair amount of personal work stuff, psds, rhino models etc, old work samples
20:25:12 <supermop_Home> that i worry about but want to poke around with time to time
20:25:20 <nielsm> yeah creative work is a good use case for NAS
20:25:32 <supermop_Home> would be a bit clunky to have that in cloud
20:25:40 <Borg> for creative work.. best is VFS...
20:26:02 <supermop_Home> usually only access it from one computer, but might occasionally from another
20:26:07 <andythenorth_> if it's purely for archival /recovery backups, they should be on one or more disks in one or more fire safes
20:26:13 <andythenorth_> NAS for backup is nuts
20:26:31 <supermop_Home> nas is maybe the wrong word
20:26:32 <nielsm> NAS for primary storage of stuff that doesn't need to be hot
20:27:05 <nielsm> (stuff that really needs to be hot, like high res video for editing, obviously goes on a local ssd)
20:27:18 <supermop_Home> currently every few years i get worried any buy a new usb external drive, back up most of my stuff, then forget about that drive in a box or drawer
20:27:42 <supermop_Home> what i want is a little box that can do some raid flavor
20:27:46 <Timberwolf> I've pondered one, mainly for FLAC copies of stuff I've got on physical media that's not available on streaming (usually mono versions, things which got remastered badly, things where the licencing situation is [shrug] )
20:28:14 <supermop_Home> Timberwolf i also have acquired a decent chunk of flac and wavs
20:28:19 <Timberwolf> Currently doing fine with a Pi and a USB HD though.
20:28:23 <supermop_Home> mostly from records
20:29:27 <supermop_Home> i usually would only use mp3s when playing on my phone or computer speakers, but i'd like to keep those larger files accessible for the future
20:29:42 <supermop_Home> likewise for field recordings
20:30:33 <supermop_Home> I've had various 2.5" usb external drives go bad over the years,
20:31:07 <supermop_Home> so currently I've got like a slowly growing and rotating pile of them
20:31:40 <andythenorth_> they're pretty failure prone, especially 2.5"
20:31:50 <supermop_Home> but a raid box with 4-5 bays seems like it would be less of a headache
20:31:53 <andythenorth_> and the enclosures tend to be dirt cheap and prone to failure too
20:32:38 <nielsm> yeah if you want to have a bunch of reliable storage, a synology or qnap NAS bos is a better choice than USB harddrives
20:32:44 <andythenorth_> I bought an nvme enclosure last year
20:32:55 <andythenorth_> it's quite fast as a backup drive
20:33:15 <supermop_Home> usb 3 would be fine, the NAS part is less necessary
20:33:15 <andythenorth_> but it would be easy to fit a lot of them into safes / or just hide them
20:33:34 <andythenorth_> using SSD as long-term storage is probably a gamble
20:33:45 <andythenorth_> I don't know the failure rates
20:33:52 <nielsm> def. don't use SSD for storage that will be unplugged for months at a time
20:34:13 <supermop_Home> well they will turn to noise if you put them in a safe and forget about them for 2 years
20:34:16 <planetmaker> or stone slabs :P
20:34:24 <nielsm> SSD at modern densities will become difficult to read in 3-6 months and probably unreadable after 12-18
20:34:40 <planetmaker> really, that bad?
20:35:11 <supermop_Home> planetmaker they need regular power
20:35:14 <Borg> nielsm: are you insist that plugged in SSD drive does some kind of refresh?
20:35:26 <b_jonas> nielsm: why is it that bad? SD cards are readable even if they sit for years.
20:35:42 <planetmaker> interesting... I had actually no idea about that property of ssds... good that my backups are magnetic dr4ives
20:36:03 <supermop_Home> don't SD cards have a bigger charge per bit? less dense?
20:36:07 <nielsm> are those SD cards you think of lower capacity cards from 10 years ago, or high capacity cards from the last few years?
20:36:13 <Borg> planetmaker: well.. after very long time.. magnetics drives can fail too.. but its 20+ years..
20:36:23 <Borg> I mean.. HDDs.. not tapes.
20:36:31 <supermop_Home> a bit on a ssd may just be a few electrons that can get bored and wander off
20:36:31 <b_jonas> supermop_Home: how can they less dense when you can have 128 gigabytes on a microSD card that's the size of my fingernail?
20:36:34 <nielsm> magnetic storage is more reliable for long term
20:36:40 <planetmaker> of course they can. And will. thermodynamics cannot be beaten
20:37:11 <Borg> everything will fail... I saw holographic storage.. in some kind of special glass..
20:37:17 <b_jonas> nielsm: low capacity ones, I haven't had high capacity ones years ago
20:37:19 <nielsm> the major risk of tape storage is really that the tape sticks to itself in the spool... which can be averted by just respooling it once in a while
20:37:27 <Borg> but we dont need such storage.. for our personal stuff
20:37:37 <b_jonas> obviously everything will fail, I'm just asking about the specific matters of scale
20:37:54 <supermop_Home> i mostly want my stuff not to fail on a 10 year scale
20:38:09 <Borg> supermop_Home: same here. hence.. I still use spinning rust
20:38:11 <planetmaker> I'd like my stuff to not fail in my lifetime :P Preferentially
20:38:19 <supermop_Home> often i go looking for old work from architecture school, every few years
20:38:22 <planetmaker> but 10 years seems fair
20:38:36 <planetmaker> after 10 years you might have connection issues anyhow...
20:38:37 <b_jonas> Borg: does respooling actually help? doesn't it get sticky because water or some other damaging substance gets into it, which you would just accelerate if you got it out of the cupboard and respooled it?
20:38:38 <supermop_Home> it seems each time i look another few things are lost
20:38:59 <Borg> b_jonas: what are you talk about? that glass thingie?
20:39:18 <b_jonas> nielsm: does respooling actually help? doesn't it get sticky because water or some other damaging substance gets into it, which you would just accelerate if you got it out of the cupboard and respooled it?
20:39:18 <supermop_Home> planetmaker well if it lasts 5 years, and i rememeber to refresh or maintain it every 2 years, i'll be ok
20:39:24 <b_jonas> Borg: magnetic tape storage
20:39:33 <nielsm> anyway yes, flash memory cells store data by being in a non-natural state of kind, and that will slowly normalize if it doesn't get erased and rewritten
20:39:40 <planetmaker> supermop_Home, yes, the point is "if I remember" :)
20:39:59 <supermop_Home> but its pretty easy for me to go 12 months forgetting about one of a handful of drives in shoebox on top of a shelf...
20:40:04 <Borg> b_jonas: tapes last longer that HDDs.. afaik.. less density.. different encoding.. stronger signals..
20:40:26 <planetmaker> nielsm, magnetic tapes somewhat do that, too. Just magnetic order used on discs decays slower
20:40:34 <nielsm> and flash memory used as multilevel (storing 2, 3 or 4 bits per cell) is even more prone to data loss
20:40:38 <supermop_Home> 1 or 2 boxes of dives that will let me know when a drive is going bad, would help a lot
20:40:44 * andythenorth_ read about SSD retention
20:41:06 <andythenorth_> it would be unwise to trust one beyond 2 years unplugged in northern european temperaturs
20:41:14 <andythenorth_> 1 year in southern european temperatures
20:41:21 <andythenorth_> also "don't panic"
20:41:22 <supermop_Home> whether the drives inside are ssd or iron doesn't matter to me if they let me know when they are failing
20:41:25 <Borg> storing data on any storage is one thing.. storing it correctly (right software.. checksums.. validation.. is different)
20:41:42 <supermop_Home> i just want the best box to put the drives in
20:41:49 <nielsm> if you want data to last, rewriting it to new media on the regular is really the only way
20:42:02 <nielsm> and keeping context to understand the data :)
20:42:43 <supermop_Home> sinology or qnap recommened?
20:43:23 <nielsm> I have a synology but don't take that as a recommendation in particular
20:43:49 <nielsm> I'm quite sure QNAP also works fine
20:45:08 <andythenorth_> _most_ things are pretty commodity now
20:45:17 * andythenorth_ gets caught out sometimes when they're not
20:46:44 <supermop_Home> well best buy doesn't have much with 4 holes available for same day pick up here in the city
20:46:58 <supermop_Home> so i might as well order whatever online
20:49:19 <nielsm> beware that synology will probably auto-configure whatever disks you install into it into some RAID-5 or 6 setup for you, make sure the disk configuration is what you want before starting to store data ;)
20:51:17 <nielsm> and keep in mind that math shows that RAID-5 and RAID-6 are not terribly safe for disks larger than 2 TB or so, simply due to reconstruction times when one disk fails likely being longer than the time before the risk another disk fails... or something like that
20:51:44 <nielsm> can't remember the exact argument, but after some particular size it's better to do "something else"
20:52:08 <nielsm> (likely just having multiple backups)
20:52:32 <Borg> or distributed storage ;)
20:52:53 <Borg> always wanted to have such thing.. but for personal usage. its kinda too expensive to have something like MooseFS
20:54:50 <supermop_Home> nielsm nested raid levels?
20:58:53 <Eddi|zuHause> so, i had a "disk full" issue with my SSD where i have some games and the steam workshop on... download size was 1.9GB, of which it already downloaded 400MB. so i made space for 3.6GB, and the disk is now full again at 1.4GB out of 1.9GB
21:02:39 <nielsm> supermop_Home, with 1+0 when one disk fails, the next disk failure could either cause total loss or just another degradation, so it's kind of a gamble too
21:05:21 <Eddi|zuHause> that sounds unlikely
21:05:57 <Eddi|zuHause> it's not a particularly big SSD anyway
21:19:13 *** y2kboy23 has joined #openttd
21:29:58 *** innocenat_ has joined #openttd
21:48:10 *** blathijs has joined #openttd
22:00:57 <CornsMcGowan[m]> hm, code works fine on release build but fails assert in debug build
22:01:48 <frosch123> that's because assertions are disabled in release build
22:02:34 <CornsMcGowan[m]> am i using this vector improperly then
22:05:53 <Eddi|zuHause> now, i've 8GB free, let's try again :p
22:07:40 <CornsMcGowan[m]> is delete(*myVector.end()) dodgy?
22:07:46 <CornsMcGowan[m]> oh is end() a null element
22:08:31 <nielsm> yes end is one-past-last
22:09:05 <nielsm> so when begin()==end() the container is empty
22:09:32 <CornsMcGowan[m]> how should i call delete() over each element in a vector then :0
22:09:55 <nielsm> if you want to clear the vector then use clear()
22:10:36 <CornsMcGowan[m]> i see, but its a vector of pointers, i also want to call delete() on each element first
22:10:39 <nielsm> or wait do you have a vector of pointers and you want to delete the pointed-to objects?
22:11:00 <nielsm> make it a vector of std::unique_ptr
22:11:32 <dP> or just for (auto x : v) delete x;
22:11:48 <frosch123> dP: that's the bad answer
22:12:05 <nielsm> avoid manual memory management when possible
22:12:57 <dP> I'd agree if smart pointers in c++ weren't such a mess
22:13:11 <dP> also it's a weird statement in openttd context :p
22:13:14 <Eddi|zuHause> now, after it finished the rest of the 1.9GB, i now have 6GB free
22:13:35 <Eddi|zuHause> i guess steam workshop works in mysterious ways
22:13:37 <frosch123> Eddi|zuHause: just delete all those "train movies"
22:14:04 <Eddi|zuHause> frosch123: completely unrelated :p
22:22:10 <CornsMcGowan[m]> oh man im too tired to wrap my head around unique_ptr at the moment
22:24:23 <CornsMcGowan[m]> trying to implement it in the context of A* search, to delete nodes when no longer needed, and im tossing up between a vector of all nodes or using the two sets (openset and closedset) to manage the unique_ptr since a node can only be in one set anyway
22:26:08 <nielsm> you use std::move() to take a unique_ptr value and move it to a different unique_ptr
22:27:01 <nielsm> closedset.push_back(std::move(openset[3])); openset.erase(openset.begin()+3);
22:27:43 <dP> CornsMcGowan[m], why do you even need dynamic allocation for that?
22:34:35 *** blathijs has joined #openttd
22:36:38 <supermop_Home> well i just bought another whatever external
22:37:07 <supermop_Home> so if my laptop gets destroyed while i'm travelling this weekend at least i'll have that
22:38:25 <supermop_Home> who uses these bare HDD docks?
22:38:59 <nielsm> techs that need to clone disks and such?
22:39:40 <nielsm> or if you use disks for cold-storage backup medium
22:44:54 <andythenorth_> I have a HDD bare dock
22:45:00 <andythenorth_> amongst other things
22:45:18 <supermop_Home> i have a usb to sata cable somewhere
22:45:21 <andythenorth_> I back up to both SSD and spinning disk sporadically
22:45:54 <andythenorth_> recently my most successful backup strategy has been getting one of my slightly older laptops
22:53:44 <supermop_Home> raid still is limited by smallest disk, right?
22:54:14 <supermop_Home> like if i have 3 4TB drives and one 25 GB drive, it's just going to think its 4 25 GB drives?
22:59:40 <nielsm> with classic hardware raid, yes
23:00:04 <nielsm> with software raid you can typically use the "extra" space on the larger drives for other things
23:02:16 <Eddi|zuHause> i was thinking of getting 2 bigger disks, then combining 2 of my smaller disks as one bigger disk (raid0) and then use those virtual bigger disks along the new disks as raid5
23:02:36 <Eddi|zuHause> but it's probably TMWFTLB
23:03:01 <supermop_Home> maybe i can re use a few of these 1 TB 2.5" disks i have laying around
23:03:44 <supermop_Home> or if there was some kind of raid usb hub i could just plug them all into
23:04:23 <Eddi|zuHause> software raid doesn't care if it's usb or sata or whatever
23:04:23 <supermop_Home> Eddi|zuHause yeah, probably the time to spend on it is worth more that the cost of just buying 1-2 more big disks
23:25:02 *** m1cr0man has joined #openttd
23:31:53 <supermop_Home> the have a RAID F1 now...
23:32:34 <supermop_Home> which is just raid 5 but it rides one drive into the ground so that it fails before the others....
23:39:37 <Eddi|zuHause> that sounds not very intelligent :p
23:42:22 <supermop_Home> i guess the idea is that by favoring use of one drive and having it fail first, at least you likely wont have all your drives failing at the same time
23:42:52 <supermop_Home> just 1 drive fails this year and the other 3 fail at the same time next year?
23:52:50 <SpComb> you just need to build your disk arrays out of different disk models by different manufacturers :P
23:52:54 <Eddi|zuHause> supermop_Home: but if the drive that fails because of manufacturing deficits isn't the same as the one that is over-used, then you've got two problems
23:53:10 <SpComb> or someone could start selling pre-aged disks for use in raid arrays...
23:53:28 <supermop_Home> barrel aged HDD
23:53:31 <Eddi|zuHause> you could just replace the drives before they fail
23:53:51 <SpComb> idea being to ensure failures are non-correlated
23:53:57 <supermop_Home> this HDD was aged in sherry casks
23:54:26 <supermop_Home> "... also it is already failed because it is soaked in sherry"
23:54:28 <SpComb> yeah, one approach would be to pre-emptively rotate out disks
23:55:07 <SpComb> I'm sure some people that take their RAID arrays seriously do that
23:55:19 <SpComb> meanwhile, how often do you actually test your backups...
23:55:34 <supermop_Home> on the order of years
23:56:11 <supermop_Home> I've actually only had one hdd failure in a computer i used in the past 20 years
23:56:29 <supermop_Home> and that was some RAID 0 situation
23:57:33 <supermop_Home> i've had more issues with an old external hard drive i haven't looked at in a few years having a bunch of corrupt files on it
23:57:46 <supermop_Home> but again like 3 times in 20 years,
23:57:53 <b_jonas> I should pay more attention to situations where a bridge works but a tunnel would work better, to build tunnels. I default to bridges most of the time.
23:58:12 <supermop_Home> and usually there is another old hard drive in the shoebox that might have that file on it
23:58:55 <supermop_Home> but i feel like i should be taking a more careful approach
23:59:42 <SpComb> I have a fancy ZFS setup with all kinds of snapshot replication, but uh never got around to actually implementing some kind of active monitoring
23:59:46 <b_jonas> argh, my service to this station is so good that trains keep coming in and it's never empty, and I want to modify its setup but the trains are in the way
continue to next day ⏵