IRC logs for #openttd on OFTC at 2020-01-26
        
        
        
            ⏴ go to previous day
00:29:38  *** snail_UES_ has joined #openttd
 
00:47:36  <spnda> I just noticed that since JGR 0.32 the straight bus stop sprites are offset by +4 in SpriteIDs... Weird
 
00:59:29  *** Eddi|zuHause2 is now known as Eddi|zuHause
 
00:59:48  <Eddi|zuHause> spnda: if they added stuff to the baseset, the IDs might shift
 
01:10:24  <spnda> Eddi|zuHause: yeah, that's what I think aswell
 
01:22:17  *** WormnestAndroid has quit IRC
 
01:22:44  *** WormnestAndroid has joined #openttd
 
01:26:29  <spnda> ok wow it took JGRennison like 15 minutes to fix that issue. Very quick lol
 
01:34:52  *** WormnestAndroid has quit IRC
 
01:36:03  *** WormnestAndroid has joined #openttd
 
02:10:14  <_dp_> great, openttd can have sprites with negative dimensions xD
 
02:10:21  * _dp_ was debugging this shit for an hour
 
02:12:05  <_dp_> (7482, 15014, 24) - (7481, 15013, 23)
 
02:16:40  <DorpsGek_III_> [OpenTTD/OpenTTD] JMcKiern updated pull request #7959: Fix #7950: Incorrect setup of Normal Screenshot ViewPort https://git.io/Jvqlk
 
02:24:24  *** WormnestAndroid has quit IRC
 
02:24:48  *** WormnestAndroid has joined #openttd
 
05:07:09  <DorpsGek_III_> [OpenTTD/OpenTTD] glx22 dismissed a review for pull request #7959: Fix #7950: Incorrect setup of Normal Screenshot ViewPort https://git.io/Jvq48
 
06:13:04  *** el3ktr4 has joined #openttd
 
08:55:37  *** Progman has joined #openttd
 
09:36:27  *** sla_ro|master has joined #openttd
 
10:13:39  <peter1138> Original Half-Life playable for free for... 2 months. Nice.
 
10:54:45  *** andythenorth has joined #openttd
 
11:51:36  <Samu> can't multithread sprite sorter?
 
11:54:43  <Samu> coronavirus enters europe via france
 
11:55:31  <andythenorth> pandemics gonna pandemic
 
11:58:56  <nielsm> multithreading will likely have more overhead for tasks that are relatively small like that
 
11:59:04  <nielsm> also distributed sorting algorithms are Hard
 
12:00:19  <Samu> i have no 4k display to test it :(
 
12:00:54  <nielsm> timing a huge screenshot should also work
 
12:02:09  <Samu> that screenshot that writes a gigabyte of .png into a file
 
12:02:42  <_dp_> multithreading can probably help but not much
 
12:02:56  <nielsm> for benchmarking, you could remove the file-writing part, while adding the time measuring part
 
12:03:11  <_dp_> I have more ideas for single thread though xD
 
12:03:39  * andythenorth wonders if it affects mac fps, retina (hidpi) screens 
 
12:03:47  * andythenorth hasn't pulled the PR yet :P
 
12:05:15  <_dp_> andythenorth, no idea how retina works but if there are lot of sprites to draw it should help
 
12:05:56  <Samu> take 2 screenshots, then compare the differences between them
 
12:06:01  <nielsm> I still haven't seen proof that the general performance issues on mac are limited by either simulation time or rendering time
 
12:06:04  <Samu> 1 without optimization, 1 with
 
12:08:08  <andythenorth> nielsm: at least one of the issues is caused by the vehicle window, and might not be mac specific :)
 
12:09:03  <nielsm> maybe it's similar to the town window thing that got looked at recently: sorting the list too often
 
12:09:59  <andythenorth> someone figured out what it was, maybe frosch
 
12:10:09  <andythenorth> I didn't file an issue though, I'd have look in logs
 
12:14:05  <Samu> i just tried pr 7962, got a crash in main menu right away
 
12:14:19  <nielsm> did you make a full rebuild?
 
12:16:08  <Samu> Expression: front() called on empty vector
 
12:16:44  <andythenorth> "looks like the window is invalidated whenever a vehicle loads"
 
12:17:45  <_dp_> Samu, oh, didn't expect that to crash
 
12:18:06  <_dp_> well, I guess crash counts as undefined behaviour xD
 
12:18:49  <Samu> crashed on here ParentSpriteToDraw *s = psdv->front(); // Sprite that we're currently comparing
 
12:19:16  <Samu> gonna rebuild all to make sure
 
12:19:25  <nielsm> sounds like it's not handling an empty input?
 
12:20:00  <nielsm> but yeah rebuild, I've had a bunch of miscompile/mislink issues with vs2019
 
12:20:33  <_dp_> nielsm, yeah, I assumed having ub is fine since I'm not using it anyway
 
12:20:39  <_dp_> just didn't expect a crash xD
 
12:21:16  <Samu> maybe it's because i'm testing in debug build
 
12:21:55  <_dp_> yeah, shouldn't crash anyway tho so I fixed it
 
12:22:32  * andythenorth wonders about 'extended info' window for vehicles
 
12:22:46  <andythenorth> new window that opens, has the vehicle title, rest is blank
 
12:23:02  <andythenorth> populated by a string result from a cb (or prop)
 
12:29:33  <Samu> cool, no crash in main menu
 
12:33:08  <nielsm> so it was that again... :s it seems the safest is to always make a full rebuild after switching branch
 
12:33:34  <peter1138> Ok, my server is empty... what shall I change.
 
12:33:53  *** WormnestAndroid has quit IRC
 
12:34:07  *** WormnestAndroid has joined #openttd
 
12:34:10  <andythenorth> put your OpenTTD patchpack on it peter1138
 
12:34:58  <peter1138> I don't have a patch pack, of course.
 
12:35:21  <peter1138> Acceleration model...
 
12:35:42  <andythenorth> do some grfs or something
 
12:35:57  * andythenorth needs to procrastinate about doing tax return
 
12:38:39  <peter1138> Infrastructure maintenance? Just to kill off planes :p
 
12:41:15  <peter1138> Hmm, toyland? Not sure I'll play that either :p
 
12:41:21  <peter1138> Maybe non-temperate.
 
12:43:55  <peter1138> Hmm, but I just set a 1950 start.
 
12:44:42  <peter1138> Autoclean turned off except for companies with no vehicles.
 
12:45:23  <Samu> how do i time this screenshoting thing
 
12:45:59  <Samu> where do I put TICC/TOCC
 
12:47:06  <peter1138> Around the bit you want to time...
 
12:47:44  <Samu> somewhere between zoom out to 4x
 
12:48:07  <nielsm> somewhere in MakeScreenshot in screenshot.cpp I'd say
 
12:48:16  <andythenorth> much industry peter1138 :D
 
12:49:59  <peter1138> Hmm, yes, density is high.
 
12:50:29  <peter1138> Should reduce that next time.
 
12:50:48  <Samu> are u sure about the huge screenshot thing, is that going to do spritesorting?
 
12:51:23  <Samu> but it's at default zoom level i think
 
12:52:16  <peter1138> What does that have to do with sprite sorting?
 
12:52:21  <peter1138> Also, is it breakfast time yet?
 
12:53:09  <nielsm> drawing the world always involves sorting sprites
 
12:53:33  <nielsm> the reason zooming out is slow on large resolutions is because it will draw lots of sprites
 
12:53:50  <nielsm> drawing a full world screenshot will also draw lots of sprites
 
12:58:52  <peter1138> And that may require crunching through the sprite cache.
 
12:59:49  <Samu> tree generation is slow in arctic
 
13:04:55  <nielsm> testing with PSG 201 save in 2048x1536 resolution, debug build
 
13:05:02  <nielsm> definitely seeing big improvements
 
13:05:17  <Samu> it is confirmed that trees are evil
 
13:07:01  <andythenorth> if you run a game long enough though Samu, all the trees disappear from the map
 
13:07:06  <andythenorth> so that's ok no? :)
 
13:07:20  <Samu> i didn't disable trees from respawning
 
13:07:25  <andythenorth> oh it's a setting?
 
13:07:30  <andythenorth> I assumed it was a bug
 
13:07:44  <andythenorth> I noticed in my 100 year game that all trees were gone
 
13:08:10  <andythenorth> oh that's a very weird setting :)
 
13:08:15  <andythenorth> why would that even exist?
 
13:08:17  <nielsm> there is some improvement
 
13:08:19  <DorpsGek_III_> [OpenTTD/OpenTTD] JMcKiern commented on pull request #7959: Fix #7950: Incorrect setup of Normal Screenshot ViewPort https://git.io/JvqHX
 
13:08:21  <nielsm> but I should try with release builds
 
13:08:45  <andythenorth> that save game is intense :)
 
13:09:04  <Samu> where do u find those savegames :(
 
13:09:14  <Samu> im generating a map with huge amounts of trees
 
13:09:42  <LordAro> shall we do RC1 today?
 
13:10:08  <Samu> 4096x4096 map, gonna take huge screenshot with ticc/tocc, be back tomorrow
 
13:10:08  <nielsm> I think we want to know if this sprite sorter patch can be merged before that
 
13:11:36  <nielsm> but I'm pretty sure it's fine
 
13:11:48  <Samu> The screenshot will have a resolution of 262,081 x 131,191 pixels
 
13:11:57  <LordAro> checked that the GRF it was having trouble with before was fine?
 
13:11:58  <nielsm> especially if _dp_ unit tested it against the original
 
13:12:03  <LordAro> MBs ships or something
 
13:12:40  <nielsm> hm? last sprite sorter patch we merged and then reverted showed issues just with default bridges for me
 
13:13:12  <LordAro> i remember a screenshot of some GRF stations, i think
 
13:13:55  *** gelignite has joined #openttd
 
13:15:54  <_dp_> well, I kinda tested it... on one sprite set and some random_shuffle :)
 
13:19:55  <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro approved pull request #7959: Fix #7950: Incorrect setup of Normal Screenshot ViewPort https://git.io/JvqHd
 
13:20:55  <Samu> this is gonna take forever, oh well, i'll let you know when it finishes
 
13:22:24  <_dp_> reverted sorter patch used std::sort so all stuff that relies on implicit ordering could go wrong randomly
 
13:23:31  <LordAro> Samu: have you checked whether anyone wants whatever you're about to spend hours doing?
 
13:24:03  <peter1138> Does anyone want to know whatever he's about to spend hour doing?
 
13:24:23  <_dp_> I usually test implicit ordering with how planes draw around international airport terminal
 
13:25:04  <_dp_> As sorter itself doesn't ensure that planes in front draw first
 
13:26:23  <_dp_> That's also the only case I know of that fails std::sort by center mass xD
 
13:26:44  <nielsm> hm I'm not seeing any huge improvement, it seems sorting time is dwarfed by painting time for me?
 
13:28:35  <TrueBrain> why is my transfer speed with the old binaries server only 4 MB/s ... this is horribly slow ... (I need to download ~100GB)
 
13:33:15  <peter1138> You've got fixes in there anyway.
 
13:33:39  <peter1138> Hmm, unless that's fixing master, heh.
 
13:33:41  <_dp_> nielsm, I just tried with psg201 and it's 20s unzoom lag vs 1s for me
 
13:35:07  <_dp_> it's unzoom that benefits the most, just panning around doesn't do much sorting
 
13:40:44  <nielsm> and my previous flicker.sav does not cause flicker, so at least that configuration is fine
 
13:46:31  <peter1138> Playing Doom II: Hellbound. And I died on the first level, oof.
 
13:48:37  <andythenorth> Torpedo Wagons eh
 
13:54:15  <TrueBrain> hmmm .. sshd is at 100% CPU ... why is that ... it is only doing 3 MB/s ... I expect more :P
 
13:54:51  <peter1138> Is it a CPU from 15 years ago? :p
 
13:55:41  <DorpsGek_III_> [OpenTTD/OpenTTD] nielsmh approved pull request #7859: Feature #7756: Allow server to supply a reason to kicked/banned clients https://git.io/JvqQE
 
13:56:24  <TrueBrain> well, the OS is around that age, so that might have something to do with it for sure :)
 
13:56:34  <TrueBrain> @calc 100*1024 / 3 / 3600
 
13:56:34  <DorpsGek> TrueBrain: 9.48148148148
 
13:56:40  <TrueBrain> 10 hours of downloading at this rate
 
13:57:06  <peter1138> nielsm, compile errors :-)
 
13:58:21  <peter1138> /home/petern/Projects/OpenTTD/src/table/../newgrf_roadtype.h:35:31: error: 'GetScope' overrides a member function but is not marked 'override' [-Werror,-Winconsistent-missing-override]
 
14:03:33  <LordAro> really need Werror on the CF really
 
14:03:43  <LordAro> just that irritating clang false positive...
 
14:04:19  <LordAro> nielsm: merge seems fine :p
 
14:11:20  <peter1138> Quick! Undo it! Break everything!
 
14:11:45  <peter1138> Mmm, breakfast salad :D
 
14:15:02  <peter1138> Mouse look in Doom. Is it acceptable?
 
14:17:46  *** el3ktr4 has joined #openttd
 
14:18:58  <nielsm> I think generally only for wads designed for it
 
14:21:48  <peter1138> Hmm. It wasn't there in original Doom only for technical reasons, really.
 
14:22:12  <peter1138> Jumping is an obvious big no-no of course.
 
14:23:29  <peter1138> Mouse look is generally benign but can give advantages in some places, indeed.
 
14:24:01  *** frosch123 has joined #openttd
 
14:41:50  <peter1138> Oof, chocolate doom is a bit unpleasant now.
 
14:46:53  *** ChanServ sets mode: +v tokai
 
15:05:43  <peter1138> Rage quitting at Hellbound.
 
15:06:00  <peter1138> Maybe I should just play an easier difficulty?
 
15:08:58  <peter1138> Oh, my git hooks no longer work :/
 
15:17:13  <peter1138> As we use that variable to determine if we're being run as a git hook, rather than for the location, it seems like the test could be removed as well.
 
15:18:57  <peter1138> Mmm, that doesn't remove the redundant tests, but thanks for the pointer.
 
15:24:45  *** el3ktr4 has joined #openttd
 
15:33:50  <TrueBrain> 26GB done ... 70-ish to go .. ugh .. this is slow :P
 
15:33:59  <TrueBrain> why did I want this to be pretty and clean?
 
15:34:06  <DorpsGek_III_> [OpenTTD/OpenTTD-git-hooks] PeterN commented on pull request #11: Fix: GIT_DIR is no longer set in git >= 2.18.0 https://git.io/Jvq5a
 
15:35:16  <TrueBrain> peter1138: I think the problem is more that the safety check no longer does what it should be doing
 
15:35:19  <TrueBrain> more than being redundant
 
15:35:33  <TrueBrain> (it is legit to check if it is being run from a commit-hook or not .. it now just always work :P)
 
15:35:56  <TrueBrain> I have no clue if that is a bad thing btw ..
 
15:36:05  <TrueBrain> trying to stay as far away as possible from commit-hooks :D
 
15:36:28  <peter1138> My comment on the PR doesn't suggest removal, just that the test is wrong :-)
 
15:36:42  <TrueBrain> sorry, didn't read your comment on the PR yet :P Was reading IRC :D
 
15:37:20  <TrueBrain> other not less important news: I AM HUNGRY! And have to wait 2 more hours before food-time
 
15:38:46  <peter1138> So I ate lunch 1h30 ago, and I've picked on food since. Fail :(
 
15:39:36  <TrueBrain> 220K    tmp/releases/0.1.1
 
15:39:37  <TrueBrain> 268M    tmp/releases/1.4.4
 
15:39:43  <TrueBrain> 200M    openttd-releases/1.9.3
 
15:39:47  <TrueBrain> 258M    openttd-releases/1.10.0-beta2
 
15:40:04  <TrueBrain> that 0.1.1 release size ... and the difference between 1.9 and 1.10 :P
 
15:40:25  <LordAro> ...is there actually anything in 0.1.1?
 
15:40:38  <TrueBrain> it is a working version
 
15:40:58  <TrueBrain> 1.10 is 0.3MB bigger for the win64 version
 
15:41:21  <peter1138> Secret trick to get me back into PRing :p
 
15:41:36  <TrueBrain> well, now you are creating PRs, can you do <enters list here>
 
15:42:03  <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro commented on pull request #7963: Fix #7868: Missing override attribute. https://git.io/Jvq5P
 
15:42:08  <TrueBrain> LordAro: sadly, the source for 0.1.1, 0.1.2, and 0.1.3 is lost
 
15:42:16  <TrueBrain> the first source we have is from 0.1.4
 
15:42:30  <TrueBrain> 400K Mar 26  2004 openttd-0.1.4-windows-win32.zip
 
15:42:36  <TrueBrain> 581K Mar 26  2004 openttd-0.1.4-source.zip
 
15:43:03  <LordAro> weren't those all within the space of a couple of weeks though?
 
15:43:05  <TrueBrain> first binary next to win32 was mac :)
 
15:43:25  <TrueBrain> 11 days between 0.1.1 and0.1.4
 
15:43:45  <TrueBrain> after mac, the next platform was BeOS :)
 
15:43:47  <LordAro> if only we could do releases that fast these days :p
 
15:43:54  <peter1138> LordAro, dunno, they seem to exist in many places in the code, so...
 
15:43:59  <LordAro> how long between 0.1 & 0.1.1 ?
 
15:44:08  <TrueBrain> we have no record of 0.1
 
15:44:32  <peter1138> LordAro, oh. "/* virtual */"
 
15:45:27  <TrueBrain> I do not know; let me cross-reference the content
 
15:46:27  <TrueBrain> it for sure pre-dated 0.1.4
 
15:46:36  <TrueBrain> but as I have no older sources, not sure which 0.1.N this is
 
15:46:52  <TrueBrain> I guess that is the reason it never got added to the repo
 
15:46:52  <LordAro> maybe orudge remembers?
 
15:47:01  <LordAro> i was under the impression it was the original ludde release
 
15:47:03  <TrueBrain> I remember he collected the 0.1.N series for us to start with :)
 
15:47:28  <peter1138> Grr, got a spot on my cheek :(
 
15:47:38  <peter1138> Must be a teenager again.
 
15:47:57  <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro merged pull request #7961: Fix #7960: use the same method as findversion.sh to determine tag https://git.io/JvqRr
 
15:48:28  <TrueBrain> -       if (v->vehstatus & 0x80) ormod = 0x3248000;
 
15:48:29  <TrueBrain> +       if (v->vehstatus & VS_CRASHED) ormod = 0x3248000;
 
15:48:38  <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro closed issue #7950: Normal screenshot mode may incorrectly show a huge screenshot warning dialog https://git.io/JvIMx
 
15:48:38  <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro merged pull request #7959: Fix #7950: Incorrect setup of Normal Screenshot ViewPort https://git.io/Jvqlk
 
15:48:40  <TrueBrain> so clearly it predated 0.1.4 .. hard to guess what version it really is
 
15:49:17  <LordAro> those are useful for determining version :p
 
15:49:21  <TrueBrain> indeed, that seems to be 0.1.0 source
 
15:50:43  <TrueBrain> we took quiet some effort to recover the 0.1.N series, funny that this was not added :)
 
15:50:46  <TrueBrain> (we even corrected changelogs etc)
 
15:51:14  <TrueBrain> I wonder if we can redistribute that source package .. I assume we can ..
 
15:52:13  <TrueBrain> do we have a date for this zip .. hmm
 
15:53:09  <TrueBrain>  71 Feb  6  2004 luddemusic.dat
 
15:53:34  <TrueBrain> 316K Aug 14  2003 ttd.exe
 
15:53:39  <TrueBrain> so either of these two dates, I guess
 
15:57:31  <TrueBrain> okay, added it to the archive, as if it was released Feb 7th, 23:00 UTC
 
15:58:29  <TrueBrain> what always was funny to me, that ludde left a few weeks after his release :P
 
15:58:51  <LordAro> "When's your birthday?" "Feb 7th" "What year?" "Every year"
 
15:59:07  <TrueBrain> 2004, ofc, that was obvious you silly
 
15:59:18  <DorpsGek> LordAro: ludde was last seen in #openttd 7 years, 28 weeks, 4 days, 7 hours, 55 minutes, and 56 seconds ago: <ludde> hej
 
15:59:27  <TrueBrain> but the "source" package is both the binary and the source
 
15:59:32  <TrueBrain> I wonder if I should split it :P
 
15:59:50  <LordAro> probably not worth it
 
15:59:57  <LordAro> could you even run that binary these days?
 
16:00:04  <TrueBrain> there is even a crashlog in the zip :)
 
16:00:22  <TrueBrain> honestly, it should run fine
 
16:01:56  <TrueBrain> 0.1.1 already had a grf file :)
 
16:02:36  <TrueBrain> IDA scripts, lol :)
 
16:04:44  <TrueBrain> okay, had to clean up the zip a bit before we can put it on our CDN
 
16:04:55  <TrueBrain> there are some .. euh .. files in there, which we are not allowed to redistribute :)
 
16:05:33  <LordAro> well if orudge has been hosting it for 16 years...
 
16:05:49  <LordAro> but yes, probably for the best
 
16:05:53  <TrueBrain> as that goes .. I don't care what others do; I can only be responsible for what I do :)
 
16:06:39  <_dp_> Oops I did it again ... cut another 15% off sprite sorter xD
 
16:06:56  <_dp_> So... should I add a second commit or just force push a new version?
 
16:07:31  <LordAro> _dp_: probably a second commit
 
16:07:37  <LordAro> we can squash if necessary
 
16:09:12  <_dp_> ok, first I'll try to optimize this new idea to the max though
 
16:10:28  <TrueBrain> first release of 0.1 is set to 14th of March .. but .. that is 0.1.1
 
16:14:42  <LordAro> are there forum threads that could tell you one way or the other?
 
16:16:07  <LordAro> first crash report was 80 minutes after the first release :)
 
16:37:14  <TrueBrain> tnx LordAro, that was fun to add o the archive :)
 
16:43:01  <TrueBrain> right, time to upload everything except the pre-2019 nightlies .. 40GB of archive .. lets go!
 
16:43:13  <TrueBrain> @calc 40 * 1024 / 60 / 3600
 
16:43:13  <DorpsGek> TrueBrain: 0.18962962963
 
16:43:16  <TrueBrain> @calc 40 * 1024 / 60 / 60
 
16:43:16  <DorpsGek> TrueBrain: 11.3777777778
 
16:45:56  <Samu> should I abort screenshot capture, already 8 GB and not finished
 
17:03:20  <TrueBrain> lovely .. we released catcodec 1.0.4 and 1.0.5, but there are no binaries, only source packages :) Lol
 
17:11:14  <TrueBrain> yeah, we should make a decent release there ... no clue what went wrong with the last 2 versions
 
17:14:20  <TrueBrain> but first I am going to get this CDN online :D Not trivial, sadly ...
 
17:14:24  <LordAro> right, wiki updated to current versions
 
17:14:34  <LordAro> i got a couple of years behind :)
 
17:14:41  <TrueBrain> as example, for 2 files the checksums were missing .. for some unknown reason
 
17:15:32  *** gelignite has joined #openttd
 
17:19:38  <frosch123> TrueBrain: i have a tarball of ottd 0.1
 
17:19:52  *** Wormnest has joined #openttd
 
17:20:03  <TrueBrain> bit late to the party, but what 0.1? And what tarball? Source? Binaries? :D
 
17:20:32  <frosch123> there is a forum post by oru*dge somewhere on the forums
 
17:20:38  <frosch123> and i downloaded that some years ago
 
17:20:41  <TrueBrain> is it the same as the zip LordAro linked?
 
17:20:46  <frosch123> there is also a bundle of ludde's cvs repository
 
17:24:32  <TrueBrain> yeah, I think those are the same as LordAro linked; I now added that as 0.1.0 to the archive :)
 
17:24:51  <TrueBrain> and now the slow process of validating the checksums .. as ... some md5sum files have absolute paths in them .. *writes code to correct that*
 
17:24:55  <frosch123> the newest file in the zip is 2003-08-14 or so
 
17:25:11  <TrueBrain> yeah, and the ttd.exe is newer, as it the luddemusic.dat or what-ever it is called
 
17:25:24  <TrueBrain> he didn't do much between 2013-08 and 2014-03 :P
 
17:25:28  <frosch123> oh, there is one 2004-02-06, probably by or*dge
 
17:25:39  <TrueBrain> yeah, same archive as I just downloaded :)
 
17:25:55  <TrueBrain> there are ida files in there too :P
 
17:26:01  <TrueBrain> for those doubting how this was created :D
 
17:26:20  <TrueBrain> for i in `ls`; do ( cd $i; md5sum -c *.md5sum; ); done | grep -v OK
 
17:26:23  <TrueBrain> doing that over 30GB is slow
 
17:26:26  <TrueBrain> for some unknown reason :P
 
17:27:03  <peter1138> (Not that it'd affect speed)
 
17:27:34  <TrueBrain> I have the sha1 and sha256 to check too in a bit .. :P
 
17:28:21  <peter1138> Want to borrow the compute resources of my original raspberry pi? Single core 600 MHz woo ;)
 
17:28:42  <TrueBrain> I doubt CPU is the issue here :)
 
17:28:47  <TrueBrain> I think I/O is more of a problem :P
 
17:29:41  <frosch123> TrueBrain: also my notes say 2004-03-06 for ottd 0.1.0, though no idea where i got that info from
 
17:29:52  <TrueBrain> frosch123: you are lagging behind on IRC it seems :)
 
17:30:14  <frosch123> yes, i was catching up :)
 
17:41:38  *** snail_UES_ has joined #openttd
 
17:46:24  <peter1138> Hmm, double-scale UI + double-scale fonts looks so much better with double-scale padding & lines... Oh well.
 
18:02:18  *** el3ktr4 has joined #openttd
 
18:05:56  <andythenorth> I like it cramped :)
 
18:08:22  <Samu> dbg: [misc] [MakeScreenshot] 17483083596 us [avg: 17483083596.0 us]
 
18:08:43  <Samu> how much seconds are here
 
18:09:31  <Eddi|zuHause> peter1138: if only there was a patch for that...
 
18:10:16  <peter1138> Samu, how many microseconds in a second?
 
18:13:55  <Samu> nearly 5 hours taking a world screenshot of a 4096x4096 map
 
18:14:10  <Samu> 10,1 GB (10.942.213.840 bytes)
 
18:14:14  <DorpsGek_III_> [OpenTTD/OpenTTD] nielsmh commented on pull request #7963: Fix #7868: Missing override attribute. https://git.io/JvqAJ
 
18:16:21  <Samu> now there is no program who can open it
 
18:16:24  <LordAro> milek7_: Samu likes finding extreme cases
 
18:16:51  <Samu> windows doesn't preview it
 
18:20:51  <Samu> 262,081 x 131,191 pixels hmm
 
18:21:44  <Samu> must convert to tiff format? wow
 
18:30:11  <Samu> can't allocate memory for whatever
 
18:30:51  <LordAro> Samu: how much memory do you have in your system?
 
18:31:25  <LordAro> well, at least you theoretically have enough
 
18:31:27  <Samu> the png file alone is 10 GB
 
18:32:21  <LordAro> is that used or free?
 
18:32:38  <LordAro> regardless, you're going to have trouble allocating more than 10G more
 
18:34:25  <_dp_> that's 128 gb unpacked so not enough ram probably
 
18:34:45  <peter1138> Damn... and I only have 64GB RAM :/
 
18:35:15  <_dp_> same, never had a reason to buy more xD
 
18:35:49  <peter1138> I have no reason to upgrade from 32GB other than "why not"
 
18:36:08  <peter1138> It does mean my Linux VM gets a bit more free RAM to play with.
 
18:36:31  <_dp_> I'm actually close to 32gb used somehow...
 
18:36:46  <peter1138> Yes, because unused memory is wasted memory.
 
18:37:42  <nielsm> I have 24 GB installed and it's annoying because it's DDR3 and I'll have to replace it all if I want a CPU upgrade
 
18:39:34  <milek7_> yes, huge tiffs are certainly used in mapping software
 
18:39:47  <milek7_> but I don't know if this is smart enough to open huge pngs
 
18:40:33  <glx> I have 8GB DDR3 and it's clearly not enough
 
18:47:25  <milek7_> with 130945x65623 image it uses 1.3GiB of RAM, so maybe it could work for samu image
 
18:48:06  <Samu> i feel bored, so im gonna try install it
 
18:51:09  <Samu> eww the installer is blurred
 
18:56:44  <Samu> i find myself installing something I'm probably gonna use once
 
19:04:28  <Samu> ok installed, how to open png here? this is totally confusing
 
19:04:42  <Samu> i dragged the png into it, hope that is the way
 
19:06:27  *** gelignite_ has joined #openttd
 
19:07:19  <Samu> it's opening the file at 40 MB/s
 
19:07:39  <Samu> or at least disk activity says 40 MB/s reading
 
19:08:54  <Samu> they really want me to buy an SSD
 
19:11:58  <Samu> i dont recall my HDD being this slow
 
19:12:17  <Samu> it should achieve 100 MB/s for huge file reading
 
19:12:28  <nielsm> for most desktop workloads an NVMe SSD is not an advantage over SATA SSD
 
19:12:36  <nielsm> but any SSD is an advantage over spinning rust
 
19:13:47  <nielsm> probably because it's a large file that may be fragmented, and it probably needs to index/decode it in a non-linear fashion, it's likely a platter drive will spend a lot of time seeking
 
19:15:33  <Samu> ah, seems that it's the cpu that is slowing down
 
19:16:05  <Samu> program isn't multithreaded?
 
19:19:54  <milek7_> for exporting into tiles vips works quite well
 
19:23:24  <SpComb> yeah, similar things. qmsk/pngtile includes the leaflet.js UI and renders the tiles on the fly
 
19:23:36  <TrueBrain> for big images you should just use IDL :D
 
19:23:59  <nielsm> really, having giant screenshots saved as tiles in the first place would make sense, that can also avoid having the large black corners :P
 
19:35:42  *** WormnestAndroid has quit IRC
 
19:35:58  *** WormnestAndroid has joined #openttd
 
19:45:28  <Samu> it's taking forever to close t.t
 
19:45:55  <DorpsGek_III_>   - Update: Translations from eints (by translators)
 
19:46:23  <Samu> while it closes, i'm gonna have dinner, :o
 
19:51:19  <peter1138> My dinner's been in the oven for 45 minutes already.
 
19:52:28  <nielsm> hmm is there any particular reason it's not possible to add/remove a GS from a running game?
 
19:53:18  <nielsm> apart from cheating concerns
 
19:53:42  <frosch123> the gs can run some cycles before the game starts
 
19:54:20  <frosch123> i guess it boils down to: noone knew what gs may do, so let's make it simple for now
 
19:54:39  <frosch123> same reason why you can only have one gs active
 
19:55:02  <frosch123> i kind of expected there would be gs patchpacks at some point, but it never happened
 
19:55:13  <frosch123> everyone made just their own version of city builder
 
19:58:34  <nielsm> ah, you can "hack in" a GS by opening the save in the scenario editor
 
20:18:25  <TrueBrain> this replaces finger just fine
 
20:19:32  <nielsm> should I go ahead and make EternalLove GS?
 
20:21:05  <LordAro> TrueBrain: i'd suggest that (for /latest.yaml) the name should be the key to the object, rather than a standard list
 
20:21:14  <Samu> this thing has no uninstaller? ffs
 
20:21:56  <TrueBrain> LordAro: problem is, there is no key that is unique
 
20:21:57  <Samu> stupid me for installing this
 
20:22:15  <TrueBrain> name "stable" comes by a lot
 
20:22:23  <LordAro> TrueBrain: oh yeah, i didn't spot that before
 
20:22:38  <TrueBrain> otherwise I have to make an hierarcy
 
20:22:42  <TrueBrain> and that is vastly more complex to generate :D
 
20:22:57  <LordAro> Samu: that all seems unlikely
 
20:23:23  <LordAro> does "Add or Remove Programs" from control panel not let you or something?
 
20:23:29  <LordAro> TrueBrain: yeah that's all true
 
20:23:58  <LordAro> unless you use category-name as the key
 
20:24:31  <Samu> nop, it's not on control panel, there's no uninstall
 
20:25:47  <LordAro> Samu: in which case you can probably just delete it
 
20:25:55  *** tokai|noir has joined #openttd
 
20:25:55  *** ChanServ sets mode: +v tokai|noir
 
20:25:57  <LordAro> well, you definitely can
 
20:26:05  <LordAro> but i don't see why windows would worry about it either
 
20:26:32  <LordAro> TrueBrain: incidentally, why are pr6811 & pr7206 still available as builds, other than for testing purposes?
 
20:26:33  <TrueBrain> LordAro: yeah, I considered it .. didn't find a benefit of doing :)
 
20:26:43  <TrueBrain> LordAro: because I haven't cleaned up anything
 
20:26:54  <TrueBrain> the nightlies are completely there back to january 2019
 
20:26:57  <TrueBrain> which is also pointless :D
 
20:27:09  <TrueBrain> (they will soon only have source-tarballs left :P)
 
20:27:24  <TrueBrain> well, I am first downloading the full nightly archive (back to 2004)
 
20:27:30  <TrueBrain> just .. because I fucking can
 
20:27:35  <TrueBrain> (well, also source-tarballs only)
 
20:31:26  <TrueBrain> wrote down how it all works, for those who are curious enough :)
 
20:32:18  <TrueBrain> so, two things left before this can go live ... 1) make Azure Pipelines upload to the S3 bucket, 2) make this generate-everything script more like generate-for-this-release-only :D
 
20:33:31  <LordAro> TrueBrain: seems like all that text should go in a document in a repo somewhere
 
20:33:45  <TrueBrain> will be in a README.md :)
 
20:35:47  <TrueBrain> I completely forgot someone in that ticket claimed XML was the best format in the world .. completely forgot about that :D
 
20:36:00  <LordAro> TrueBrain: probably for the best :p
 
20:38:08  <TrueBrain> on estimate, it will cost 0.0001 dollarcent per release to recalculate these files (S3 usage costs)
 
20:44:20  <peter1138> Hmm, there are 6 GIT_ environment variables set when running as a git hook. Just not GIT_DIR, heh.
 
20:44:53  <peter1138> At least, there are on my PC.
 
20:47:15  <Samu> alright, it's "uninstalled"...
 
20:47:49  <TrueBrain> peter1138: does it really matter if these scripts are executed as non-githook, I wonder?
 
20:54:12  <frosch123> i would insert 3 words: every week, all towns are visided by your mom, and has their records ...
 
20:58:27  <nielsm> hmm... how about adding ctrl+click a town name to open the local authority window directly
 
20:59:29  <frosch123> i thought you would suggest ctrl+click to directly bribe the town :p
 
21:00:16  *** el3ktr4 has joined #openttd
 
21:13:47  <glx> TrueBrain: is the changelog stuff working correctly now ?
 
21:15:23  <_dp_> in cmclient ctrl-click builds a statue xD
 
21:15:24  <peter1138> Okay, the tests are actually still working in the case that you execute outside a git repo.
 
21:15:29  <peter1138> So my comment is wrong :p
 
21:16:28  <nielsm> I should figure out how t upload this to bananas too
 
21:19:48  <nielsm> hm MIT license is missing from bananas? >_>
 
21:22:46  <glx> there's an external tool for uploads
 
21:22:52  <nielsm> also missing versions later than 1.6
 
21:23:12  <peter1138> Yay, found the firmware updater tool for my keyboard and... now it works again ;D
 
21:23:15  <glx> using website is not the recommended way IIRC
 
21:25:16  <LordAro> only not recommended for anything significantly large, i believe
 
21:27:23  <nielsm> I can edit the tags, description, required version etc. later, right?
 
21:27:37  <nielsm> (just not delete the submission or change the file)
 
21:28:02  <glx> where is andythenorth when you need it ?
 
21:30:59  <andythenorth> `where is andythenorth
 
21:31:57  <andythenorth> peter1138: are you restoring 2px window shadows/highlights?
 
21:32:03  <andythenorth> or is that some zoom level I don't know about
 
21:33:03  <nielsm> well uploaded it to bananas now, I hope it's possible to download too :D
 
21:34:50  *** ChanServ sets mode: +v tokai
 
21:37:47  <TrueBrain> that moment you cannot do your escaping correct .. no matter what, it complains .. sometimes I hate shells ...
 
21:37:56  <TrueBrain> that moment you cannot do your escaping correct .. no matter what, it complains .. sometimes I hate shells ...
 
21:38:09  *** gelignite has joined #openttd
 
21:39:13  <LordAro> add more backslashes until it works
 
21:39:14  <TrueBrain> worst part is, it is always your own fault :P
 
21:44:21  <TrueBrain> okay .. I have to extract all archives and repack them to correct a mistake I made 12 years ago :P
 
21:44:27  <TrueBrain> but I finally got to doing it :D
 
21:45:01  <TrueBrain> compression with -6 instead of -9 :)
 
21:45:24  <LordAro> how much difference does it make?
 
21:45:28  <TrueBrain> renaming folders while at it .. it is now: openttd-20070615-trunk-r10166
 
21:45:37  <TrueBrain> on 40GB around 2 to 4 GB, on estimate :)
 
21:45:54  <TrueBrain> 'make bundles' should honestly also be fixed to use -9 :)
 
21:46:30  <TrueBrain> I am mostly happy the date is now in the folder name :)
 
21:46:58  <TrueBrain> and more like current versioning :)
 
21:47:29  <TrueBrain> and I am now putting nightlies in a folder by year .. as the dirlisting became .. annoying :D
 
21:47:40  <TrueBrain> over 4500 entries :D
 
21:48:38  <TrueBrain> except for those from 2020
 
21:49:04  <TrueBrain> okay ... and I want to fix the changelogs for 2019 .. :D
 
21:49:24  <TrueBrain> only 8 changelog.txt are empty up till 2019 .. so ... I cannot handle this :P
 
21:52:24  <TrueBrain> okay, that is not difficult to correct, I guess ..
 
21:53:10  <LordAro> should be able to script it relatively easily, i should think
 
21:56:07  <TrueBrain> seems the changelog script is wrong
 
21:56:19  <TrueBrain> but 'git log' doesn't work with UTC times
 
21:56:31  <TrueBrain> so minor difference exist
 
21:56:41  <TrueBrain> not sure why there is a '-u' there
 
21:57:36  <TrueBrain> okay, changelog.txt generated :)
 
22:00:31  <DorpsGek_III_> [OpenTTD/OpenTTD] TrueBrain opened pull request #7964: Fix: [AzurePipelines] the changelog-generation-script was mixing UTC and non-UTC https://git.io/JvqjN
 
22:00:41  <TrueBrain> fixing bugs that are important! :P
 
22:03:06  <LordAro> TZ=UTC git log --date=iso-local appears to work
 
22:03:24  <TrueBrain> that would work too
 
22:03:51  <LordAro> or whichever other date format you want
 
22:06:16  <TrueBrain> -AuthorDate: Thu Jan 10 19:56:41 2019 -0500
 
22:06:16  <TrueBrain> +AuthorDate: 2019-01-11 00:56:41 +0000
 
22:06:20  <TrueBrain> guess the latter is a bit cleaner :)
 
22:06:28  <TrueBrain> it now had the TZ of the author :)
 
22:06:32  *** y2kboy23 has joined #openttd
 
22:06:44  <LordAro> "local" if you want the same format as previous
 
22:06:55  <TrueBrain> I like YYYY-MM-DD a lot better :P
 
22:06:57  <TrueBrain> for obvious reasons :D
 
22:07:13  <TrueBrain> but okay, strictly seen this is not needed
 
22:07:30  <TrueBrain> the TZ=UTC and 'date -u' should work together
 
22:07:35  <TrueBrain> both give the same result
 
22:07:45  <LordAro> neither is probably better then
 
22:07:45  <TrueBrain> the --date just changes how the already correct information (in both cases) is presented
 
22:08:17  <TrueBrain> and using the TZ of the Author / Commiter might be weird
 
22:08:19  <TrueBrain> but it is what git is
 
22:08:23  <TrueBrain> so not sure setting --date is a good idea
 
22:08:51  <LordAro> definitely should use committer rather than author
 
22:09:01  <LordAro> author commit might have sat around for months :)
 
22:09:30  <TrueBrain> for example the last one, you see that the commiter and author were not in the same TZ
 
22:09:53  <TrueBrain> well, I don't really care which format we use tbh :)
 
22:10:00  <TrueBrain> either "iso-local" or like the URL shows
 
22:10:31  <LordAro> not really, current is better for readability
 
22:10:43  <LordAro> iso is better for parsibility
 
22:11:00  <TrueBrain> lucky this is for humans, not for machines :)
 
22:11:05  <TrueBrain> as -7days is as random as it can be :P
 
22:11:23  <TrueBrain> so okay, my PR is the right fix in that case :D
 
22:11:43  <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro approved pull request #7964: Fix: [AzurePipelines] the changelog-generation-script was mixing UTC and non-UTC https://git.io/JvmeU
 
22:11:50  <DorpsGek_III_> [OpenTTD/OpenTTD] LordAro merged pull request #7964: Fix: [AzurePipelines] the changelog-generation-script was mixing UTC and non-UTC https://git.io/JvqjN
 
22:12:35  <TrueBrain> 300 / 4000 folders converted ... ugh, this is going to take a while ..
 
22:12:54  <TrueBrain> owh well :) After that, the CDN is ready .. it only needs a GitHub Actions workflow script :)
 
22:13:23  <TrueBrain> the new CDN is incredible fast, which is nice :)
 
22:19:26  <TrueBrain> what I did not solve yet, are the "scenarios", and the "installer"
 
22:19:30  <TrueBrain> the first one .. I dont even know what it is
 
22:19:35  <TrueBrain> the second is a magic folder with symlinks
 
22:19:38  <TrueBrain> that no longer really works
 
22:19:43  <TrueBrain> not sure what I can do about that ..
 
22:19:50  <TrueBrain> but that is something for another day or something :P
 
22:20:04  <LordAro> make it inaccessible and see who shouts :p
 
22:20:23  <TrueBrain> I know where and how installer is used .. it is just annoying
 
22:21:09  <LordAro> needs more Inno Setup, anyway
 
22:21:16  <LordAro> because who doesn't love Pascal?
 
22:26:11  <frosch123> TrueBrain: ottd bundles "scenarios" up to 0.4.8 or something
 
22:26:35  <frosch123> that would be my guess
 
22:26:38  <TrueBrain> frosch123: yeah, and I think NSIS also downloaded them .. so I will need to check how those URLs are programmed in
 
22:28:15  <LordAro> i would've hoped that the installer would've been made in such a way that it could handle a missing/broken url
 
22:28:16  <frosch123> which is somehow versioned by being compatible
 
22:28:20  <TrueBrain> yeah, the installer folder I know how it is wired .. it is just annoying how it is done
 
22:28:38  <TrueBrain> the version number is not the release value
 
22:28:44  <TrueBrain> so there are symlinks on disk now
 
22:28:47  <frosch123> so ottd 1.10 downloads the same ogfx as ottd 1.2, because that was the last time we changed it incompatibly
 
22:28:49  <TrueBrain> but in an S3 bucket that is not really possible :D
 
22:29:08  <TrueBrain> but okay, something to figure out another day :)
 
22:29:23  <LordAro> seems like the installer should be able to download directly from bananas?
 
22:29:46  <TrueBrain> for the trick mentioned by frosch123, that is not possible
 
22:29:58  <TrueBrain> that is, BaNaNaS has no URL which says: compatible-with
 
22:30:04  <Samu> openttd.org doesn't have a background image behind the OpenTTD logo?
 
22:30:33  <LordAro> TrueBrain: well, neither does this behaviour, it's just manually hardcoded in
 
22:30:48  <LordAro> Samu: no, it does magic stripes instead now
 
22:31:14  <TrueBrain> LordAro: yup; so we cannot point it to BaNaNaS now, and I need to fix this current hardcoded joking :P
 
22:36:26  <LordAro> works fine for me though
 
22:36:53  <LordAro> probably browser cache related
 
22:41:52  <Samu> i refreshed but still white
 
22:42:02  <Samu> not sure how to clear cache
 
22:42:54  <LordAro> oh no, i think you're right
 
22:43:24  <LordAro> ah, the linear-gradient
 
22:44:07  <Samu> tried opening in IE11, it's even worse
 
22:44:23  <LordAro> IE11 is definitely not supported
 
22:46:03  <andythenorth> IE 11 is really easy to support for most things (except flex)
 
22:46:16  <LordAro> because there is lots of flex usage now :p
 
22:47:43  <LordAro> i'm not sure i even have internet explorer
 
22:51:11  <LordAro> no sign of IE11 on my Win10 machine, even after disabling and reenabling the "feature"
 
22:51:20  <LordAro> must have done something in the past to more permanently remove it
 
22:51:35  <Samu> with edge click on the . . . at top right, then there's an option Open with Internet Explorer
 
22:51:38  <nielsm> not even an iexplore.exe ?
 
22:52:25  <LordAro> nielsm: there's a couple in C:\Windows, but they all seem to be in WinSxS dirs and similar
 
22:52:28  <LordAro> Samu: appears to do nothing
 
22:52:57  <LordAro> anyway, staging should be fixed momentarily
 
22:54:29  <Samu> it's been a long time since I did a  disk cleaup  utility
 
22:54:40  <Samu> took 1 hour cleaning stuff :(
 
22:55:46  <TrueBrain> cert-errors, nice :)
 
22:56:19  <TrueBrain> there are more that are broken than those that are working :(
 
22:56:25  <TrueBrain> it used to work for all of them :'(
 
22:56:30  <TrueBrain> stupid CSS3 / HTML5
 
22:56:53  <TrueBrain> (really, it used to work for all of them .. took weeks to get it there :P)
 
22:56:58  <TrueBrain> but this was 2007 :D
 
22:58:08  <TrueBrain> funny that so far 0 browsers render the site correctly :D
 
22:58:19  <LordAro> it has started with the oldest
 
22:58:49  <TrueBrain> and your fix possibly fixes a few of those :)
 
22:58:51  <LordAro> i think most of the firefox ones would've worked with the above fix
 
23:01:30  <TrueBrain> yeah, there already a few arrive that render it correctly :D
 
23:02:46  <TrueBrain> quality of browsershots did not improve in 10 years, I see :P
 
23:03:23  <TrueBrain> so clearly your fix was not Edge only LordAro, and improves many browsers :) Such a simple fix :P
 
23:13:22  <DorpsGek_III_> [OpenTTD/website] TrueBrain opened pull request #147: Add: switch to openttd-cdn.org for all downloads https://git.io/Jvmvk
 
23:14:05  <DorpsGek_III_> [OpenTTD/website] TrueBrain updated pull request #147: Add: switch to openttd-cdn.org for all downloads https://git.io/Jvmvk
 
23:16:05  *** tokai|noir has joined #openttd
 
23:16:05  *** ChanServ sets mode: +v tokai|noir
 
23:20:17  *** andythenorth has left #openttd
 
23:25:48  <DorpsGek_III_> [OpenTTD/workflows] TrueBrain opened pull request #4: Add: [CDN] code and config to generate required files for CDN https://git.io/Jvmvo
 
23:26:27  <TrueBrain> "70 browsers selected, 38 uploaded, 12 failed, 20 expired"
 
23:26:31  <TrueBrain> so it is just crap these days
 
23:27:52  <TrueBrain> okay, I really have to split up that build.py script in the last draft PR :D
 
23:28:46  <TrueBrain> and nightly-archive-conversion is at 25% ..
 
23:28:52  <TrueBrain> this will be night-work for my computer :)
 
23:30:14  <milek7_> you can't use faster compression? e.g. zstd?
 
23:35:19  *** el3ktr4 has joined #openttd
 
23:35:50  <TrueBrain> Speed is not really an issue. It is also only using a single core .. it is fine, it is a one-off conversion
 
23:58:22  *** supermop_Home has joined #openttd
 
continue to next day ⏵