IRC logs for #openttd on OFTC at 2019-01-24
⏴ go to previous day
00:03:30 *** sushibear has joined #openttd
00:09:19 <LordAro> seemed more appropriate
00:10:42 <Eddi|zuHause> "66% complete" seems like a nonsense number :p
00:37:31 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on issue #6896: game crashes on first autosave if compiled without lzo but with lzma https://git.io/fhVj5
00:45:21 <Samu> i had reported a crash with lzo some years ago that got fixed
00:45:37 <Samu> but maybe it broke something else
01:14:56 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro requested changes for pull request #7085: Fix: Remove duplicate French and Slovak town names https://git.io/fhwfW
01:28:00 *** supermop_work_ has quit IRC
01:29:56 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas updated pull request #7085: Fix: Remove duplicate French and Slovak town names https://git.io/fh2Af
01:30:24 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas commented on pull request #7085: Fix: Remove duplicate French and Slovak town names https://git.io/fhwJ8
01:31:51 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on pull request #7080: Change: Cache stations and links for whole map for linkgraph GUI to eliminate delay when scrolling or zooming https://git.io/fhwJ0
01:37:24 <Samu> still testing windows closures
01:41:07 <m1cr0man> What happens when the passenger distribution is set to manual?
02:19:04 *** Wormnest has joined #openttd
02:29:56 <Samu> makes them unload at the next station that accepts passengers, instead of the station they want to go
02:57:25 <m1cr0man> right cool. I just found the big wiki page on cargodist too that cleared it all up
02:58:18 <m1cr0man> I was running asymmetric distribution for my last game and it got..weird (5000 people to a city with 100 people), so I'm going to do 75% symmetric for the next run
03:37:39 <Samu> aha, 1024 is the correct size
04:27:31 <Samu> i'm getting a crash only in debug mode
04:28:28 <Samu> in release it doesn't crash. Regardless, it's a bug, a difficult one to reproduce
05:04:57 *** snail_UES_ has joined #openttd
06:03:41 *** HeyCitizen has joined #openttd
07:12:30 *** HerzogDeXtEr has joined #openttd
07:19:47 <DorpsGek_II> [OpenTTD/OpenTTD] btzy updated pull request #7080: Change: Cache stations and links for whole map for linkgraph GUI to eliminate delay when scrolling or zooming https://git.io/fh2LL
07:25:15 <DorpsGek_II> [OpenTTD/OpenTTD] btzy updated pull request #7080: Change: Cache stations and links for whole map for linkgraph GUI to eliminate delay when scrolling or zooming https://git.io/fh2LL
07:26:41 <DorpsGek_II> [OpenTTD/OpenTTD] btzy commented on pull request #7080: Change: Cache stations and links for whole map for linkgraph GUI to eliminate delay when scrolling or zooming https://git.io/fhwGy
08:09:11 *** andythenorth has joined #openttd
08:17:46 *** sla_ro|master has joined #openttd
08:29:02 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on pull request #7083: Codechange: [Linkgraph GUI] Replace line visibility detection algorithm https://git.io/fhwcv
09:26:54 *** HeyCitizen_ has joined #openttd
09:58:42 *** andythenorth has joined #openttd
10:28:26 *** ChanServ sets mode: +v tokai
10:43:30 *** Smedles has joined #openttd
11:02:34 *** andythenorth has joined #openttd
11:13:18 *** Smedles has joined #openttd
11:53:07 <andythenorth> yeah, I'm inclined to agree
12:06:55 <Borg> is Company Statue is not counted for Oil Rigs?
12:07:05 <Borg> hmm right! it should NOT
12:07:10 <Borg> because Oil Rig is neutral
12:17:54 <andythenorth> can't believe it's lunchtime already :o
12:21:11 <peter1138> Might be cuppa time, though.
12:25:11 <peter1138> A cup of macaroni cheese, eh?
12:25:45 <andythenorth> if a cup is also a bowl, yes
12:27:27 <peter1138> It's good enough for a cup-a-soup...
12:29:24 <Gabda> is there a code editor that shows the documentation texts of the functions, when as mouseover text when I select a use of that function somewhere in the code?
12:32:56 <Gabda> there are format strings in the documentatios, so I assume something display those texts
12:35:22 <Eddi|zuHause> most IDEs nowaday should be able to do that
12:39:13 <peter1138> Not sure what IDEs people use for OpenTTD.
12:39:19 <peter1138> I'm still using vim.
12:40:19 <andythenorth> "what's an IDE" :P
12:40:41 <Borg> guys.. help. GRF.. arytmetic... its normal standard x86 one? right?
12:40:58 <Borg> for callback results.. I see info that sign is bit14.. but its just informative?
12:40:59 <peter1138> Borg, er, what's the actual question?
12:41:17 *** WWacko1976-work has quit IRC
12:41:27 <Borg> still -1 ix 0xFFFF (16bit) and not 0x4001 ?
12:42:20 <peter1138> You can't return 0xffff.
12:42:36 <Borg> callbacks results are limited to 14 bits
12:42:50 <Borg> 15 bit.. is reserved for.. callback result or jump
12:43:04 <Borg> but now.. I want to return negative values
12:43:25 <Borg> and I see note in callback results.. that sign is 14 bit...
12:43:33 <Borg> I know architectures.. where there is ACTUAL sign bit
12:43:47 <Borg> too hard question and I should go RTFS?
12:44:01 <peter1138> bit 15 (counting from 0) leaves you with 15 bits for the result, not 14.
12:44:13 <peter1138> (Caveat that grfv7 has some reserved values)
12:44:37 <Borg> I want to use nvar=0 actually
12:44:39 <peter1138> In a 15 bit value, 7fff would be -1
12:44:50 <peter1138> Why would you think it's 4001?
12:45:05 <Borg> because if you have sign bit.. its completly different
12:45:10 <peter1138> normal two's complement
12:45:23 <peter1138> the sign bit is in the value
12:45:32 <peter1138> if it's set, it's negative.
12:45:41 <Borg> there is no sign bit in x86 :)
12:46:24 <Borg> okey. anyway. problem solved. thx
12:46:26 <peter1138> i think you are wrong there
12:47:36 <peter1138> but that's not really relevant
12:49:13 <Borg> ih.. its ages when I done x86 assembly
12:49:38 <Borg> x86 values does not have sign
12:49:47 <peter1138> That's literally the sign flag, set in the CPU's status register depending on the last op.
12:50:06 <peter1138> Sure they do. It's bit 15 in a 16 bit value, for example.
12:50:11 <Borg> but.. its completly optional
12:50:23 <Borg> the beauty of x86 is that is irravelant
12:50:35 <peter1138> If bit 15 is set, then the sign flag will be set.
12:50:38 <Borg> thays why 0x80 can be 128 or -128
12:50:59 <peter1138> That depends if you are treating your value as signed or unsigned.
12:51:09 <peter1138> Maybe you are not using the right terminology.
12:51:10 <Borg> thats why x86 doesnt have sign..
12:51:22 <Borg> some architectures actually can reserve bit for sign :)
12:51:32 <peter1138> Whether it is signed or not depends how your code uses it, not on the CPU.
12:51:49 <peter1138> That must be pretty rare and obsolete.
12:52:07 <Borg> but ok.. I probably went too far away w/ it..
12:52:11 <peter1138> Two's complement has been the standard since... forever.
12:52:13 <Borg> TTD was developed in DOS for x86
12:52:23 <Borg> and I should use it as reference
12:52:35 <peter1138> No, you should use the specs as your reference.
12:53:27 <Eddi|zuHause> two's complement has been the standard for so long, they considered putting it in the C standard instead of leaving it undefined
12:53:36 <Borg> thats why wording in newgrf worring me
12:53:48 <Eddi|zuHause> i've never ever seen an architecture not using two's complement
12:54:30 <peter1138> In a 16 bit signed value, bit 15 is the sign bit.
12:54:39 <peter1138> I don't understand what's confusing :p
12:55:23 <peter1138> "x86 doesn't have a sign bit" ... it's that register at that moment that has the sign bit.
12:55:29 <Borg> read Signed Magnitude representation
12:55:48 <Borg> thats what I was talking about
12:55:58 <peter1138> We're not using calculator's visual output, here...
12:56:40 <peter1138> I guess two's complement is so standard, nobody ever considered documenting in the specs that it is still two's complement :p
12:57:11 <peter1138> We don't use binary coded decimal either.
12:57:41 <Borg> so.. as I said.. doing x86 as reference is good enough
12:57:52 <Borg> anyway.. back to grf coding..
12:58:02 <peter1138> (Except that, conveniently, 0x50 = 80km/h = 50mph)
12:58:08 <Eddi|zuHause> no, but we use "4-byte strings as uint"
12:59:54 <Gabda> is there a reason that in sortlist_type the sortable list (GUIList) has GUI in it?
13:00:28 <peter1138> It was written for GUI code originally, I suppose.
13:00:48 <Gabda> can it be used for other pusposes than displaying a sorted list on the screen?
13:01:44 *** andythenorth has left #openttd
13:01:44 <peter1138> Maybe. It was written with GUI in mind such that the list be cached when possible.
13:01:52 <Gabda> or there is a hidden restriction?
13:03:00 <Gabda> ok, if there is no obvious restriction, I will try to use it
13:05:09 *** WWacko1976-work has joined #openttd
13:05:53 <peter1138> Yeah, probably better to use the existing function (even if it's misnamed?) than inventing more code.
13:06:57 <Gabda> abd the sorter behind it seems nice
13:30:01 <Borg> ! *00 0B 01 0B 00 01 02 03 04 05 06 07 08 09 0A 1A 02
13:30:35 <Borg> if I specify more that 1 ? how that should to work.. this one doesnt work
13:30:58 <Borg> I want to change 1 property of 11 cargos
13:31:11 <Borg> 1A 02 is custom rating calc CB
13:32:01 <peter1138> Id "The Vehicle ID of the first vehicle or station to change. If num-info is greater than one, this vehicle/station and the following vehicles/stations will be changed. "
13:32:59 <Borg> but what about props? I should do num-id * num-props?
13:33:48 <peter1138> "Each property byte is followed by <num-info> new-info sections"
13:34:18 <Eddi|zuHause> num-info > 1 goes through each consecutive ID
13:34:53 <Borg> so its easier to actually split it
13:34:59 <Eddi|zuHause> for (i=id; i<id+num_info; i++)
13:35:23 <peter1138> It's easier to use nml ;)
13:35:25 <Borg> because I still dont get it.. why this one does not work
13:35:42 <Eddi|zuHause> i wasn't going to open that can of worms :p
13:35:44 <Borg> peter1138: haha ;) nah.. im fine w/ my own NFO system.. works pretty well
13:36:04 <peter1138> I don't know what "! *" means in your input.
13:36:18 <Borg> I have ruby script that does sprite number and size calculations
13:36:49 <Borg> also.. rewriting it to NML would be nightmare..
13:36:59 <Borg> I already rewritten it to support opcode mnemonics
13:37:00 <Eddi|zuHause> you know there's nforenum?
13:37:22 <Borg> oh.. nope..but IRC is better imo? :)
13:37:27 <Borg> is there nfo channel? ;)
13:37:27 <peter1138> So you wrote: action 00, feature 0B, num-props 01, num-info 0B, Starting at ID 00, changing property 01, value for first ID 02, value for second ID 03 ... etc, etc...
13:38:11 <Borg> at least its how I understood newgrf specs..
13:38:25 <Borg> u specifify n IDs.. and then props they will be changed for..
13:38:26 <peter1138> 12:31 < Borg> 1A 02 is custom rating calc CB
13:38:43 <peter1138> Borg, that's not what I wrote.
13:38:57 <peter1138> You only specify 1 ID.
13:39:04 <peter1138> You specify the value multiple times.
13:39:49 <peter1138> Ah, sorry assumed the value was a byte, if it's two bytes, then you've not got enough values, and the first value would be 0x0302
13:40:14 <Borg> its the only thing I dont get in NFO :D
13:40:26 <Borg> I never managed to set properties for multiple objects.. it never worked
13:40:38 <peter1138> value. the value of the property you want to set
13:40:53 <Borg> okey.. but lets read it again
13:41:00 <Borg> Id: The Vehicle ID of the first vehicle or station to change. If num-info is greater than one, this vehicle/station and the following vehicles/stations will be changed.
13:41:07 <Borg> its all right.. when num-info is 1
13:41:32 <Borg> to my understanding.. if you have num-info > 1 u have to specify multiple IDs.. like list..
13:41:45 <peter1138> No, you always specify one ID, the rest are sequential from there.
13:41:47 <peter1138> 12:37 <@peter1138> So you wrote: action 00, feature 0B, num-props 01, num-info 0B, Starting at ID 00, changing property 01, value for first ID 02, value for second ID 03 ... etc, etc...
13:41:53 <peter1138> As I wrote but you didn't seem to understand either.
13:42:46 <peter1138> Feature 0B (cargo) property 0B is "TextID for 1 unit of cargo"... so where are you getting custom rating calc CB?
13:42:48 <Borg> so what it means if num-info > 1 following stations/vehicles will be changed?
13:43:16 <peter1138> If you specify ID 00, and num-info is 3, then you will be setting the value for IDs 00, 01, and 02.
13:43:21 <Borg> where that folling IDs are specifed.. or they are calculated on fly?
13:43:28 <peter1138> They are SEQUENTIAL.
13:43:46 <peter1138> The IDs following the first ID.
13:43:47 <Borg> someone need add comment there.. that its calculated on fly!
13:43:55 <Borg> now is all clear :) lets try
13:44:25 <peter1138> It's specified that you provide the first ID only.
13:44:43 <Borg> yeah.. but adding (calculated on fly). would be good too :D
13:45:10 <peter1138> Oh yes, ignore my text thing, I was misinterpreting too :p
13:46:04 <Borg> yep! it doesnt work either :(
13:46:09 <peter1138> You need something like 00 0B 01 0B 00 1A 02 02 02 02 02 02 02 02 02 02 02
13:46:30 <peter1138> And you also need nforenum.
13:46:41 <Borg> what is 02 02 02 at the end?
13:46:46 <peter1138> That's the value you are setting.
13:46:57 <Borg> dammit.. its more complicated that I tought
13:47:13 <peter1138> "Each property byte is followed by <num-info> new-info sections. "
13:47:19 <Pikka> reconsider changing to nml? :P
13:47:42 <Borg> Pikka: no need to :) really.. this is last thing to be sorted out in NFO..
13:47:54 <Borg> as I said.. I can always do copy pasta
13:48:03 <Borg> and I will do it! since this thing is completly illogical to me
13:48:35 <Borg> Pikka: part of NFO code for rating calc:
13:48:38 <peter1138> There's a reason nml was developed.
13:48:38 <Borg> ! *02 0B 01 gw pcl B0 avs FF 00 add pcl B1 avs FF 00 add
13:48:45 <Eddi|zuHause> that's why people don't code in NFO anymore
13:48:54 <peter1138> NFO is very low level, based on TTDPatch internals.
13:49:03 <Pikka> who doesn't code in nfo any more, Eddi|zuHause?
13:49:12 <peter1138> Anyway, thanks for that diversion, my lunch is now at a more reasonable time ;)
13:49:20 <Eddi|zuHause> you know, people
13:49:34 <Eddi|zuHause> no sane people code in nfo
13:49:55 <peter1138> Looks like giberish to me :-)
13:50:02 <Pikka> that explains a lot, eddi
13:50:14 <Eddi|zuHause> well, it is gibberish, that's why it doesn't work :p
13:50:27 <Borg> and mnemonics are easy to remember
13:50:42 <Borg> gw -is generic word access type
13:50:48 <Borg> avs is advanced variable shift :)
13:51:21 <Eddi|zuHause> for now there is only one single person in the world for whom it is "easy"
13:52:33 <peter1138> I'd go with NML, yes.
13:53:44 <Borg> okey.. I did copy&paste.. GRF loads
13:56:00 <Borg> Pikka: so? you switched to NML too?
13:57:45 <Borg> Pikka: so you might be interested in my grfix.rb ;)
13:58:27 <Eddi|zuHause> because he knows about nforenum
13:58:42 <Borg> sure.. if Pikka remember all hex opcodes :) then no problem
13:59:04 <Eddi|zuHause> there's escape sequences, you know?
13:59:20 <Borg> Eddi|zuHause: yeah.. and except \d \w they are ugly as fuck :D
13:59:38 <peter1138> NFO is ugly as fuck.
14:00:01 <Borg> well.. its not pretties too.. but aint that bad
14:02:19 <planetmaker> it reads to me like we have s/o here who re-invents nforenum ;)
14:02:21 <Borg> Eddi|zuHause: also.. escape sequences doesnt cover everything
14:05:03 <Borg> it also does minimal sanity checks for act2
14:07:30 *** WWacko1976-work has joined #openttd
14:48:16 *** snail_UES_ has joined #openttd
14:56:19 <Borg> my rating calculations does NOT work
14:56:24 <Borg> I always get good rating..
14:56:29 <Borg> I did RTFS.. and found this
14:56:31 <Borg> if (HasBit(callback, 14)) rating -= 0x4000;
15:00:49 <milek7> looks like fancy way to clear 14 bit
15:01:43 <Borg> % ruby -e 'puts "%08X" % (-1&0x7FFF-0x4000)'
15:03:25 <Borg> in CB 154 I am supposed to return negative value..
15:03:25 *** WWacko1976-work has quit IRC
15:04:36 *** WWacko1976-work has joined #openttd
15:06:21 <Borg> it looks like that HasBit check should NOT happen at all
15:06:31 <peter1138> I guess it is 0x3fff
15:06:42 <Borg> thats why I argued peter1138 about sign
15:06:58 <Borg> 3fff is maximal signed value for 15 bits
15:07:05 <peter1138> It's still two's complement.
15:07:27 <Samu> WC_AI_SETTINGS, WC_GAME_OPTIONS, WC_AI_DEBUG, WC_AI_LIST, WC_QUERY_STRING, WC_TEXTFILE
15:07:33 <peter1138> I didn't read the spec :p
15:07:49 <Samu> dropdown lists aren't a window?
15:07:51 <Borg> peter1138: specs say: s long as you stay in the range -16384..16383).
15:08:28 <peter1138> So it is a 15 bit value or a 14 bit value...
15:08:39 <Borg> 0x8000 means callback have result
15:09:06 <Borg> so.. .. its like result&0x7FFF|0x8000)
15:09:20 <Borg> and u are supposed to get 15 bits from result
15:10:30 <Borg> rating = GB(callback, 0, 14);
15:11:01 <peter1138> Ah, you mean CB 145
15:11:12 <Borg> % ruby -e 'puts "%08X" % (-1&0x3FFF-0x4000)'
15:12:02 <Eddi|zuHause> Borg: internally everything is treated as unsigned, so you need to sign-extend it, if you want to store it in a 16bit or 32bit int
15:13:21 <Samu> peter1138 help me fix this
15:13:28 <Eddi|zuHause> sign-extend means: if (highest original bit set) {set all higher bits} else {unset all higher bits}
15:13:48 <Samu> i want to close WC_QUERY_STRING on ai parameters that just became invisible
15:15:10 <Borg> okey.. so calc looks about right...
15:15:26 <peter1138> I mean, it's possible it's been wrong since forever, but...
15:16:10 <Borg> return (x >> s) & (((T)1U << n) - 1);
15:16:15 <Eddi|zuHause> i think the specs were never really meant to work for signed
15:16:20 <Borg> I wonder if -1 is right here
15:16:47 <Eddi|zuHause> Borg: test again for positive numbers
15:17:01 <Borg> Eddi|zuHause: ok.. I will return 0
15:17:07 <peter1138> Eddi|zuHause, this is existing code, so it should work.
15:17:09 <Borg> instead of negative in my CB
15:17:18 <peter1138> Not something Borg is adding.
15:17:27 <Borg> yeah.. im coding GRF.. CB 145
15:17:56 <peter1138> CB 145 doesn't take a signed value?
15:18:11 <peter1138> Or I haven't read it enough.
15:18:14 <peter1138> It's quite verbose.
15:18:18 <Borg> CB 145 takes 2 params... 18 in form ssaaaatt
15:18:22 <peter1138> The callback allows you to replace the first three components (days since last pickup, amount waiting and max. speed of last vehicle) in the calculation above. If your callback succeeds, the first three components are skipped and the returned value is used instead their results. Bit 14 is considered the sign bit, so you can return negative numbers (and need not to worry about calculated callback
15:18:28 <peter1138> results yielding a negative result, as long as you stay in the range -16384..16383). During the callback, variable 18 has the value ssaaaatt, where
15:18:40 <peter1138> So yeah, it's meant to be a signed value.
15:18:47 <peter1138> 14:16 < Eddi|zuHause> i think the specs were never really meant to work for signed
15:18:50 <Borg> I wonder why my rating are so damn high then
15:18:52 <peter1138> Is clearly wrong :D
15:20:33 <Borg> okey, lets add shortcut..
15:20:38 <Borg> to see if bug is in my calc
15:21:42 <Samu> aha, i knew there was another window
15:22:38 <Eddi|zuHause> /* Simulate a 15 bit signed value */
15:22:39 <Eddi|zuHause> if (HasBit(callback, 14)) rating -= 0x4000;
15:23:16 <Eddi|zuHause> so what is the problem again?
15:23:42 <Borg> Eddi|zuHause: no problem.. all works as it should
15:23:51 <Borg> too much binary operations
15:24:19 <Borg> % ruby -e 'puts "%08X" % (-50&0x3FFF-0x4000)'
15:26:08 <Eddi|zuHause> "Bit 14 is considered the sign bit" is a bit ambiguous
15:26:27 <Eddi|zuHause> it should mention 2-complement there
15:26:27 <Borg> hmm I just also realised that there will be problem w/ -1 value
15:27:22 <Borg> static const uint CALLBACK_FAILED = 0xFFFF;
15:27:31 <Borg> so return -1 as signed value.. means.. callback failed?
15:27:40 <Eddi|zuHause> because -1 is 0x3FFF
15:27:56 <Eddi|zuHause> 0x3FFF != 0xFFFF
15:28:07 <peter1138> Callback failed is checked at a different stage.
15:28:10 <Eddi|zuHause> that's why callbacks are only 15 bits, the 16th bit means "callback failed"
15:28:28 <Eddi|zuHause> there's a separate "if (callback != CALLBACK_FAILED) {"
15:29:03 <Borg> so there is somewhere result &= ~0x8000 in code?
15:29:25 <Eddi|zuHause> the values between 0x3FFF and 0xFFFF can never be returned
15:29:57 <Borg> Eddi|zuHause: noo?!! stop bullshitng me :D
15:29:59 <Borg> this->result &= ~0x8000;
15:30:16 <Borg> src/newgrf_spritegroup.h: this->result &= ~0x8000;
15:31:29 <Borg> yeah.. so return -1 is no problem
15:32:12 <Borg> okey.. back to my GRF.. all seems good in OpenTTD code.. there must be a bug in my GRF then
15:32:18 <planetmaker> @base 2 16 010000000000000
15:32:27 <Eddi|zuHause> Borg: i don't think that code is relevant to your problem
15:32:34 <planetmaker> @base 2 16 0100000000000000
15:32:39 <Borg> Eddi|zuHause: thats.. completly different story :)
15:33:27 *** supermop_work has joined #openttd
15:33:33 <Eddi|zuHause> Borg: dumb question: did you set the callback flag so the callback is called at all?
15:33:43 <Borg> I returned 80 FF in callback.. and now my rating is stuck at 60%
15:33:53 <Borg> yes it is.. because station rating is manipulated
15:34:19 <Samu> ... that face when i can crash openttd in debug mode but not in release mode...
15:34:20 <planetmaker> 80FF > 3FFF, thus undefined or failed?
15:34:48 <Borg> planetmaker: it should not matter.. there should be no cmp in callbacks!! only bit operations
15:35:20 <Borg> okey.. lets turn 0... 0x8000
15:35:33 <planetmaker> @calc 0x80ff & 0x8000
15:35:33 <DorpsGek> planetmaker: Error: Something in there wasn't a valid number.
15:35:50 *** supermop_work_ has joined #openttd
15:35:55 <peter1138> Did I hear that someone was going to get OpenTTD on Steam?
15:36:17 <Eddi|zuHause> orudge or it didn't happen.
15:36:30 <Borg> ok.. confirmed.. something is broken in OpenTTD CB result handling
15:36:37 <Borg> hope its only for CB 145
15:36:48 <Borg> when I return 0 (0x8000)
15:36:55 <Borg> station rating drops flat...
15:37:23 <Borg> when I return -128 (0x8080)
15:37:29 <Borg> station rating stuck at 60%
15:37:41 <planetmaker> I guess you might have heard that
15:38:38 <Borg> now.. lets do RTFS again.. why this might happen
15:38:50 <Borg> let me paste GRF code as well
15:39:43 <Eddi|zuHause> i don't think anyone gets what you're doing
15:39:56 <Eddi|zuHause> that includes yourself, most likely
15:40:05 <planetmaker> Borg, -128 must be 0x4080, IIRC
15:40:18 <planetmaker> the first bit is the CB_FAILED flag
15:40:37 <planetmaker> (and that's why 0x3FFF is max
15:40:49 <Borg> planetmaker: how did you came w/ 0x4080 ?
15:41:11 <planetmaker> 15 bit... 15th is sign. Not 16th
15:41:34 <Borg> planetmaker: no.. u are wrong
15:41:40 <Borg> first... bits are from 0.. ok?
15:41:52 <planetmaker> then decrement my statement by one
15:41:54 <Borg> so.. 16bit value we have 0 15. but 15 bit.. means result (not jump to another VA2)
15:42:13 <Borg> so.. to return negative values... I need to use bits 0 14.. + 15bit set as result
15:42:31 <Borg> now please elaborate.. why not
15:43:04 <planetmaker> you have 15bit signed value. You do not touch the 16th bit (or 15 if you start counting with 0)
15:43:20 <planetmaker> and 0x8... touches the highest bit
15:43:24 <planetmaker> you must not do that
15:43:27 <Borg> but I need to set it!!!!
15:43:51 <Borg> not to do another chained VarAct2
15:44:04 <Eddi|zuHause> Borg: really, nobody should code in NFO
15:44:50 <Borg> Eddi|zuHause: first.. if NFO spec is fucked up how u can be sure that NML compiler actually works?
15:45:01 <Borg> 99% time in OpenTTD we operate on unsigned values
15:45:15 <Borg> this is first time in my GRF I need signed values
15:45:25 <Borg> and boom! got slapped by them
15:45:38 <Borg> I use PCLs... CB results.. chaining a lot..
15:45:53 <Borg> I always do result|0x8000
15:46:28 <Borg> so.. now I try to use signed stuff.. and it does not work
15:46:35 <Borg> either I do it wrong.. or there is a bug
15:46:38 <Eddi|zuHause> but you said "80 FF" earlier, which is... which value?
15:46:42 <Borg> hence../ we have discussion
15:47:13 <planetmaker> 0x8080 is not -128. It is 128. -128 is 0x4080 in 15 bit
15:47:37 <Eddi|zuHause> it's 0xFF80 because of little endian
15:47:54 <planetmaker> oh, and that mess, yes
15:48:02 <Eddi|zuHause> highest bit is discarded because of callback result
15:48:19 <Eddi|zuHause> and then extended back, because of hacky sign treatment
15:49:18 <Borg> yes.. OpenTTD is little endian
15:49:39 <Eddi|zuHause> Borg: so, have you tried printf to see what "rating" is immediately after the callback handling?
15:49:56 <Borg> Eddi|zuHause: if callback results is 00 80
15:50:10 <Borg> because rating is clampled 0,255
15:50:45 <Borg> w/ is wrong.. but I understand why
15:50:54 <Eddi|zuHause> in station_cmd.cpp:3326, add a printf?
15:50:55 <Borg> 80 is excaly 50% rating.. in 0,255
15:51:06 <Borg> Eddi|zuHause: I cant compile OpenTTD myself :/
15:51:38 <Eddi|zuHause> Borg: have test grf?
15:51:43 <Borg> im trying to do planetmaker calc to prove or not prove that -128 in 15% is 0x4080
15:52:07 <Borg> Eddi|zuHause: sure I can give you one
15:53:16 <Borg> Eddi|zuHause: ftp://ds-1.ovh.uu3.net/home/borg/tmp/X2025.grf
15:53:20 <planetmaker> Borg, for what it's worth: the station rating callback seems to work with nml for FIRS ;)
15:53:31 <Borg> I return 80 FF right now
15:54:06 <Borg> well.. I can have work around.. like clamp myself o 0 255 before returning
15:54:07 <planetmaker> at least it presumably does what andy wants it to do
15:54:46 <Eddi|zuHause> man, i haven't started openttd in so long, "b[pgup]" doesn't result in "bin/openttd &" anymore
15:56:00 <Samu> how do i explain in a bug report that a crash only happens in debug mode?
15:56:13 <peter1138> "this crash only happens in debug mode"
15:56:32 <Samu> and that in release mode there's no crash
15:56:35 <Borg> Eddi|zuHause: dont bother.. problem solved!!!!
15:56:44 <planetmaker> "it doesn't happen in release mode" ;)
15:56:45 <Samu> but it still sets the value
15:57:14 <Borg> * Old style callback results (only valid for version < 8) have the highest byte 0xFF so signify it is a callback result.
15:57:17 <Borg> * New style ones only have the highest bit set (allows 15-bit results, instead of just 8)
15:57:27 <Borg> so I am limited to 8bit only!
15:57:40 *** supermop_work_ has quit IRC
15:57:46 <Samu> ok, gonna try fill in a report
15:57:53 <Samu> it's gonna be a weird one
15:57:55 <Eddi|zuHause> uhm, how old do you think old-style is?
15:57:56 <Borg> this one was hard to nail :D
15:58:27 <Borg> Eddi|zuHause: I use grf version 6
15:58:29 <planetmaker> we kinda should... hide the old style stuff
15:58:37 <planetmaker> !!! 6?!!! are you nuts?
15:58:52 <Borg> dunno.. I found it esier
15:58:56 <Borg> no cargo translation tables.
15:58:58 <planetmaker> well... grf v8 and nfo v32
15:59:00 <Eddi|zuHause> nobody should ever need to use grf version 6
15:59:09 <Borg> and some other things were easier..
15:59:24 <peter1138> easier ... less features ... less compatibility ...
15:59:24 <planetmaker> no support for anything <7.
15:59:57 <Eddi|zuHause> does grfv6 then even have callbacks >100?
16:00:09 <planetmaker> we should really start deprecating grf versions
16:00:30 <planetmaker> so we can get rid of old cruft. And make maintenance and extensions easier
16:00:30 <Borg> I use quite a lot >FF CBs here and there
16:00:36 <Eddi|zuHause> planetmaker: but... DBSet... (non-XL)...
16:00:45 <planetmaker> you know my stance
16:01:02 <planetmaker> old grf, old openttd ;)
16:01:12 <planetmaker> or better ... ttd
16:01:42 <Borg> planetmaker: btw.. newest GRF version is 8
16:01:55 <Borg> so 6 is not that old to me :D
16:02:14 <planetmaker> it is ... one to two *decades* old
16:02:19 <Eddi|zuHause> Borg: grfv6 was already old when i started, about 12 years ago
16:02:21 <planetmaker> more two than one
16:02:34 <Borg> damn... time fly fast.. ;)
16:02:36 <Borg> anyway. thx for help guys
16:02:44 <Borg> myster solved! I will clamp values myself
16:03:02 <planetmaker> Eddi|zuHause, exactly that xkcd :)
16:03:35 <Borg> haha I love that xkcd :D
16:04:03 <Eddi|zuHause> google doesn't have any results for "grfv6", except one that mentions "OpenTTD 0.3.3 Problem"
16:04:06 <planetmaker> indeed I've never had to use grf v6 and I'm in grf business for more than a decade...
16:04:49 <Eddi|zuHause> and that is from 2004?
16:05:43 <peter1138> I recommend not using grfv6, Borg.
16:06:01 <peter1138> You don't, for instance, HAVE to use cargo translation tables, just because they are there...
16:06:02 <planetmaker> git log | grep 'v7'
16:06:02 <planetmaker> (svn r13855) -Fix [FS#2157]: Cargo type lookup was incorrect for GRFv7 files without a translation table.
16:06:53 <Eddi|zuHause> @openttd log 13855
16:06:57 <Eddi|zuHause> does that even still work?
16:07:03 <Eddi|zuHause> i guess it's trying to access the old svn
16:07:13 <Eddi|zuHause> which isn't where it expects it to be
16:08:06 <planetmaker> Date: Mon Jul 28 06:16:34 2008 +0000
16:08:16 <Borg> peter1138: well, for now I will stick w GRFv6.. I bet many things will break when I will switch to GRFv8
16:08:24 <planetmaker> by a certain peter...n
16:08:48 <Eddi|zuHause> planetmaker: tried "grep -C3"?
16:08:56 <Borg> I see a lot of GRFv7 comments in newgrf specs
16:09:00 <planetmaker> tried git log --grep
16:09:11 <peter1138> Must be old if it was still me doing NewGRF development...
16:11:17 <Eddi|zuHause> these decades might start to get annoying
16:12:27 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick opened issue #7091: Dropdown menu can set value to a parameter of a different AI https://git.io/fhwhr
16:13:01 <Eddi|zuHause> Borg: anyway, negative values with 8bit callback results will never work, as the code that tries to sign-extend cannot know whether the result was 8 or 15 bits, and the place that would know, doesn't know it should be treated as signed
16:14:25 <Borg> Eddi|zuHause: yep I know :) I clamp stuff in GRF directly now
16:14:41 <Borg> pcl B2 avs FF 00 max aim \w0 min im \w255 00
16:14:57 <Eddi|zuHause> Borg: thus for 8-bit results, the range is 0..255, not -16k..16k
16:15:10 <Eddi|zuHause> and also not -128..127
16:15:19 <Borg> yeah.. now when mistery is solved.. its all easy again :D
16:15:52 <peter1138> You can't have a negative result, right?
16:16:08 <Borg> peter1138: yeah.. thats why I used signed min and max
16:16:10 <peter1138> So you can't use grfv6 to do what you want.
16:16:16 <Samu> Expression: can't dereference out of range vector iterator
16:16:26 <Eddi|zuHause> "80 FF" is +128 if treated as 8bit
16:16:40 <Eddi|zuHause> and -128 if treated as 15bit
16:16:55 <Borg> Eddi|zuHause: exacly! thats why I had station rating stuck at 60% :) 50% rating + 10% from statue
16:17:17 <peter1138> I don't understand why you wouldn't just use the current version :/
16:17:25 <Borg> peter1138: I can.. CB 145 can accept negative values betwen -16384 and 16383.. but it will clamp it 0..255
16:17:27 <peter1138> Are you using Windows 95 by any chance?
16:17:35 <Borg> if I provide clamped result.. its all right
16:17:40 <Borg> peter1138: close :) im on Win2003, why?
16:18:03 <Eddi|zuHause> i knew i shouldn't have unignored people
16:18:04 <peter1138> OpenTTD 0.6 I suppose.
16:18:08 <Borg> peter1138: because I will have to rewrite my GRF again.. I might do this.. but NOT today
16:18:16 <Borg> peter1138: no.. im using 1,8.0
16:18:30 <Borg> peter1138: okey.. I see you have grudge to me.. so lets clarify :)
16:18:49 <Borg> I am not enemy of progress :) but progress must bring me benefits.. not only costs of upgrade and relearning
16:18:52 <peter1138> CB 145 does not clamp from 0..255.
16:18:52 <nielsm> nobody wants to support 16 year old operating systems
16:19:04 <Borg> peter1138: really? let me RTFS again
16:19:15 <Eddi|zuHause> i'm pretty sure OpenTTD 1.9.0 will not run on win2k3
16:19:34 <peter1138> To be clear, it clamps *after* it has calculated the result.
16:19:53 <peter1138> Negative values in the callback reduce the rating.
16:20:02 <Eddi|zuHause> "clamping" is the wrong word
16:20:16 *** supermop_work has joined #openttd
16:20:16 <Borg> peter1138: yes.. its score
16:20:18 <peter1138> Eddi|zuHause, the function name is literally Clamp()
16:20:31 <Borg> valid range is 0..255 to feed station rating
16:20:35 <Eddi|zuHause> peter1138: yeah, but it doesn't even arrive at that point
16:20:46 <peter1138> No, valid range is -16384 ... 16383
16:21:07 <peter1138> (-16384 does not make much sense though)
16:21:38 <Borg> peter1138: okey, lets start from the begining
16:21:49 <Borg> uint16 callback = GetCargoCallback(CBID_CARGO_STATION_RATING_CALC, var10, var18, cs);
16:21:57 <Eddi|zuHause> peter1138: because he is using stupid version, valid range is 0..255 but the values are not CLAMPED, but masked to 8 bit
16:22:05 <Borg> if callback is not failed.. we get result
16:22:21 <Borg> so calculation is skiped.. we then reach next blok
16:22:27 <Eddi|zuHause> Borg: that's aleady not "the beginning"
16:22:29 <peter1138> Eddi|zuHause, i'm talking about grfv7 case.
16:22:40 <Eddi|zuHause> peter1138: yeah, but... stupid
16:22:48 <peter1138> The callback expects -16384..16383.
16:23:03 <Borg> but I can provide 0..255
16:23:11 <Borg> and it will work too.. not fully tho
16:23:23 <Eddi|zuHause> ah fuck it, nothing good can ever come out of this discussion... back to /ignore
16:23:40 <Borg> yeah.. calm down guys.. I know whats going on now
16:24:04 <Borg> I wonder why I missed info that CB results are only 8 bit in GRFv6
16:24:24 <peter1138> You can't provide a negative value, which is what you had originally intended.
16:24:31 <milek7> win2k3 doesn't run any modern browser build, right?
16:24:32 <Borg> lets read newgrf specs.. if migrating to GRFv8 will not be hard
16:24:57 <Borg> milek7: yep.. im stuck w/ last FF ESR
16:25:11 <Borg> milek7: basicaly.. its ending.. true..
16:25:23 <Borg> but I cant find anything good enough that can replace it
16:26:30 <Eddi|zuHause> if i had op rights, i would permaban for "don't waste our time"
16:27:23 <Borg> busy? dont get into conversation
16:27:31 <Borg> use /ignore.. thatever..
16:27:35 <Borg> dont decide for everyone
16:33:03 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on issue #7090: Query box sets value to hidden parameters https://git.io/fhrvB
16:35:00 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick commented on issue #7091: Dropdown menu can set value to a parameter of a different AI https://git.io/fhrvw
16:35:46 <Samu> peter1138, fix my bugs. hehehe
16:36:20 *** Thedarkb-T60 has joined #openttd
16:36:26 <Samu> I have an idea how to fix them, but it's a lazy approach
16:39:56 <Samu> close oth WC_QUERY_STRING and WC_DROPDOWN_MENU
16:41:14 <Samu> problem is, i don't know how to properly detect whether I'm closing the right WC_QUERY_STRING / WC_DROPDOWN_MENU
16:42:17 <Samu> have to know if they're opened via WC_AI_SETTINGS, and not from some other window
16:42:33 <Samu> WC_AI_SETTINGS is the parent window
16:43:11 *** Wormnest has joined #openttd
16:44:21 <Borg> peter1138: hmm I changed my grf from v6 to v8
16:44:26 <Borg> and it didnt exploded ;) interesting
16:46:09 *** supermop_work_ has joined #openttd
16:59:33 <Samu> HideDropDownMenu(WC_AI_SETTINGS)
16:59:41 <Samu> wow, don't tell me this was it
16:59:47 <Samu> sounds too easy to be true
17:05:11 <Samu> what about WC_QUERY_STRING
17:05:28 <Samu> how do i close this one?
17:08:27 *** synchris has joined #openttd
17:09:50 <Samu> gonna try DeleteChildWindows(WC_QUERY_STRING);
17:11:18 *** WWacko1976-work has quit IRC
17:15:01 *** WWacko1976-work has joined #openttd
17:22:19 *** supermop_work has joined #openttd
17:22:25 *** supermop_work_ has quit IRC
17:23:39 *** supermop_work_ has joined #openttd
17:41:49 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick opened pull request #7092: Fix #7091: Close dropdown menu windows after rebuilding AI/GS settings https://git.io/fhrIN
17:46:26 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick opened pull request #7093: Fix #7090: Close Query String window after rebuilding AI/GS settings https://git.io/fhrLG
18:01:11 <DorpsGek_II> [OpenTTD/OpenTTD] SamuXarick opened pull request #7094: Fix #7088: Retrieve an appropriate name for a non-existant AI/GS when… https://git.io/fhrqY
18:21:50 *** HerzogDeXtEr has joined #openttd
18:31:30 <DorpsGek_II> [OpenTTD/OpenTTD] planetmaker approved pull request #7085: Fix: Remove duplicate French and Slovak town names https://git.io/fhrO8
18:33:42 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas commented on pull request #7085: Fix: Remove duplicate French and Slovak town names https://git.io/fhrOd
18:36:16 *** andythenorth has joined #openttd
18:36:18 <DorpsGek_II> [OpenTTD/OpenTTD] nikolas updated pull request #7086: Change #6173: Update SDL driver to use SDL 2.0 https://git.io/fhamZ
18:38:54 <DorpsGek_II> [OpenTTD/OpenTTD] JGRennison updated pull request #7083: Codechange: [Linkgraph GUI] Replace line visibility detection algorithm https://git.io/fh2PV
18:42:39 *** Progman has joined #openttd
18:42:45 *** Gustavo6046 has joined #openttd
18:45:13 <Samu> ShowAIDebugWindowIfAIError hmm this function is being misused
18:46:01 <Samu> when an AI dies, it doesn't call this function
18:46:17 <Samu> it calls ShowAIDebugWindow directly
19:04:07 <supermop_work_> i want an AI that rage-quits if it can't get it's desired color
19:06:30 <Samu> is it intended that the debug window can be spawned in main menu?
19:06:37 <acklen> I wrote an AI which would create mocking signs around my son's headquarters
19:06:46 <acklen> it quickly went haywire
19:07:38 <Samu> there is some weird saveload stuff happening during openttd launch
19:07:41 *** smoke_fumus has joined #openttd
19:07:51 <Samu> and it can call the AI Debug window if there's a crashed AI
19:08:12 <Samu> do you really want the ai debug there?
19:10:30 <glx> it's just loading the intro game
19:12:22 <Samu> the intro game doesn't have AIs, usually
19:12:47 <Samu> but if it does, it's current behaviour would be to display the ai debug window if it crashed
19:13:24 <glx> it should not have, same for newgrf
19:14:30 <Samu> void ShowAIDebugWindowIfAIError(CompanyID show_company)
19:14:30 <Samu> if (_game_mode != GM_NORMAL) return;
19:17:27 <m1cr0man> anyone got a rec for a road vehicle pack that works with FIRS 3 to complement a grf set including av8 + dutch trains? I'm just looking for a few real vehicle based variations. I've found the german vehicles set but it seems hella old
19:18:50 <nielsm> I'd say not worth the effort
19:37:15 *** frosch123 has joined #openttd
19:47:41 <Samu> hmm, the reload ai button should be disabled inside scenario editor
19:48:48 <Samu> or maybe... maybe i could do something else
19:53:48 <Samu> yeah, reload button still enabled inside scenario editor
19:54:02 <Samu> making the start stop button also enabled
19:55:35 <Samu> i have doubts about starting AIs in the scenario editor
19:55:45 <Samu> it doesn't quite make sense
19:56:33 <Samu> it's not consistent with the console
19:56:45 <Samu> startai command doesn't work there
19:56:50 <Samu> but the reload button does
19:57:39 <Samu> guess i better be consistent
19:57:48 <Samu> gonna disable reload ai button
20:01:57 *** andythenorth has joined #openttd
20:03:50 * andythenorth declares authorship
20:04:48 <nielsm> close that bit twiddling issue, or let someone else make some tests too?
20:08:56 <m1cr0man> andythenorth, This looks amazing :D thanks man! Is this listed on the newgrf wiki page? Because if so I overlooked it. That was my primary source before I came here.
20:12:23 <andythenorth> feel free to add :D
20:12:38 <DorpsGek> andythenorth: TrueBrain was last seen in #openttd 22 hours, 11 minutes, and 38 seconds ago: <TrueBrain> tomorrow it will be available around 20:15 CE(S)T :D
20:14:34 <andythenorth> that's a regression
20:14:40 <andythenorth> and 38 PRs also a regression :(
20:15:34 <DorpsGek> andythenorth: 0.342105263158
20:17:24 <andythenorth> on the one hand, you have PRs about trying to stop docks blocking ships
20:17:37 <andythenorth> and on the other you have a PR that increases the chances of dock construction blocking ships
20:17:51 <andythenorth> there doesn't seem to be a guiding concept here :)
20:18:12 <andythenorth> I don't want to be discouraging, you do a lot of work
20:18:15 <DorpsGek> TrueBrain: andythenorth was last seen in #openttd 2 seconds ago: <andythenorth> I don't want to be discouraging, you do a lot of work
20:18:23 <andythenorth> that was me TB :P
20:18:43 <TrueBrain> (what are we talking about?)
20:18:52 <andythenorth> I am talking about samu has 34% of open PRs
20:18:59 <andythenorth> but isn't it also binary o'clock?
20:19:23 <andythenorth> beer for me then
20:19:30 <TrueBrain> 12 minutes to build it
20:19:41 <TrueBrain> 4 minutes to rebuild the website
20:19:51 <andythenorth> I really think it's bad for the project to have a lot of open PRs going nowhere
20:19:53 <TrueBrain> and a lot more minutes waiting for resources
20:20:00 *** supermop_work_ has quit IRC
20:20:02 <andythenorth> it discourages drive-by contributors
20:20:08 <TrueBrain> in general things that are not going anywhere are not useful :D
20:20:34 <TrueBrain> owh bah, someone queued other things ..
20:20:40 *** smoke_fumus has joined #openttd
20:20:40 <TrueBrain> that explains why it is taking so darn long :P
20:21:21 *** supermop_work has joined #openttd
20:22:37 *** supermop_work_ has joined #openttd
20:26:58 <andythenorth> supermop_work: have you figured out industrial trains for Horse yet? o_O
20:27:25 <TrueBrain> finally the website is being published .. damn .. took a while
20:27:28 <supermop_work_> DO THEY SHUNT?
20:27:59 <andythenorth> supermop_work: do you even lift?
20:28:14 <andythenorth> wait, wat, there's a new website? :o
20:29:50 <TrueBrain> so within 30 minutes, good enough
20:32:07 <andythenorth> OpenTTD is dying
20:32:58 <nielsm> killed by fork-wielding gits
20:35:05 * andythenorth wants to kill issues
20:35:51 <supermop_work_> WHOA THERE IS A NEW SCREENSHOT
20:37:10 <andythenorth> supermop_work_: is it of steel mill trains?
20:37:17 * andythenorth needs to teddybear set design :P
20:38:50 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on issue #5451: Do not overlap vehicle view window with child windows (orders, timetable, details) https://git.io/fhrR7
20:38:51 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed issue #5451: Do not overlap vehicle view window with child windows (orders, timetable, details) https://git.io/fhrRd
20:39:14 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on issue #5679: Allow characters as hotkey instead of assuming everyone uses a Latin keyboard https://git.io/fhrRN
20:39:15 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed issue #5679: Allow characters as hotkey instead of assuming everyone uses a Latin keyboard https://git.io/fhrRA
20:41:08 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed issue #6413: Signals: PBS Signal vs 'Skip Signal' Train on Platform https://git.io/fhr0k
20:41:09 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on issue #6413: Signals: PBS Signal vs 'Skip Signal' Train on Platform https://git.io/fhr0I
20:42:02 <andythenorth> keeps it fresh eh?
20:42:07 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh commented on issue #6857: Linux SDL video driver runs slightly too fast https://git.io/fhr0q
20:42:20 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on issue #6398: Path remains reserved when extending train station https://git.io/fhr0Y
20:42:21 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed issue #6398: Path remains reserved when extending train station https://git.io/fhr0O
20:43:05 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on issue #6418: bitmask_vehicle_info is not specified for all the engines https://git.io/fhr0c
20:43:06 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed issue #6418: bitmask_vehicle_info is not specified for all the engines https://git.io/fhr0C
20:46:45 <LordAro> nielsm: i tried the sdl thing yesterday, it seemed to be running pretty consistently at just over 34ms
20:52:48 <nielsm> and win32 is consistently stuck just below 33 fps for me
20:55:23 <andythenorth> TrueBrain: at least 99 I reckon
20:56:02 <andythenorth> how many issues can I close before we start to complain?
20:56:43 <LordAro> though i should point out this is sdl on windows/mingw
20:57:13 <LordAro> windows itself seems to spend most of its time at 32.9xfps, with occasional spikes to 33.0x
20:57:52 <andythenorth> does the mac port use SDL?
20:57:58 <nielsm> maybe the framerate algorithm should be unified between video drivers
20:58:10 <nielsm> no it uses an apple api
20:58:10 <LordAro> andythenorth: if only there was a nightly someone could download to check
20:58:30 <andythenorth> mac version is hovering just above 34 fps
20:58:42 <andythenorth> occasionally drops to 33.x
20:59:31 <andythenorth> if I turn on full animation, and do nothing in game, it drops to around 32.x
21:01:14 <andythenorth> moving the cursor around slams it down further
21:02:09 <LordAro> interesting, it's presumably not on the limit of being able to keep up?
21:03:51 <TrueBrain> andythenorth: stalebot didn't find anything to close in PullRequests :P
21:04:24 <andythenorth> make it more aggressive :P
21:04:29 *** Vasatko has joined #openttd
21:04:39 <andythenorth> LordAro: dunno, the mac performance is pretty dire
21:04:50 <andythenorth> this is with a new map, no vehicles etc
21:05:00 <TrueBrain> andythenorth: yeah ... 2 days for PRs? :P
21:05:07 <andythenorth> every time I upgrade my mac, graphics performance gets worse
21:06:27 <andythenorth> I suspect the recent FFWD improvements remove the worst symptoms though
21:06:44 <andythenorth> it's usually only noticeably slow during FFWD
21:06:55 <andythenorth> or in cases like having a few vehicle windows
21:07:08 <andythenorth> or a lot of the cost / income animations on screen
21:08:39 <andythenorth> how many reviews can a PR have before it becomes 'no hope'?
21:09:08 <LordAro> depends on what the review is :p
21:10:58 <TrueBrain> the 26 "stale" issues, can be automatically closed if they expire?
21:11:06 <TrueBrain> (roughly 30 days left)
21:12:14 <andythenorth> some really look legit
21:12:34 <andythenorth> nobody is doing anything about them, and empiricism suggests they won't :P
21:12:42 <andythenorth> so let them go free :D
21:15:22 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on issue #6583: Scrolling down/right out of map on a 255 high flat map bounces viewport back to edge of map https://git.io/fhrzq
21:15:41 <andythenorth> and will close 6583
21:15:45 <andythenorth> can someone merge?
21:17:20 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh merged pull request #7061: Fix #6583: Rework Tileheight handling (patch by adf88) https://git.io/fhnDK
21:17:38 <nielsm> hope it breaks nothing :D
21:17:45 <TrueBrain> andythenorth: fixed
21:18:06 <DorpsGek_II> [OpenTTD/OpenTTD] nielsmh closed issue #6583: Scrolling down/right out of map on a 255 high flat map bounces viewport back to edge of map https://git.io/fhrzi
21:18:06 <LordAro> irritatingly, "Fix #nnnn" in the PR title doesn't close issues
21:18:12 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain opened pull request #7096: Add: [GitHub] enable probot/stale, to flush out stale issues/pull req… https://git.io/fhrzX
21:18:38 <TrueBrain> please check my PR request on both if you agree with the time, text, and if I wrote correct english :D
21:18:50 <TrueBrain> well, I just opened 20 stale.yml, and cherry-picked what I liked
21:19:14 <nielsm> andythenorth, no, it uses sprite fonts
21:19:39 <glx> but it's an old RC version
21:19:41 <andythenorth> there's limited value to preserving these crashes
21:19:50 <andythenorth> OpenTTD must crash hundreds of times a day
21:19:53 <andythenorth> in the player base
21:20:18 <andythenorth> this is like trying to suck the ocean up through a straw
21:20:56 <glx> there are more important issues, like the actual deadlock with "-n something"
21:21:24 <glx> and I don't see how to prevent it
21:22:52 <andythenorth> TrueBrain: my wording too cheese?
21:22:53 <andythenorth> "Thanks for this. There's been no activity on this for some time, and as it stands, it doesn't look likely that it will go any further. I'm closing it as we try to keep the issue count low for OpenTTD, it helps us focus on things that are important and fun. Feel free to discuss in irc or request re-opening if you disagree. Thanks for contributing!"
21:23:02 <glx> hmm segfault, ukrainian, let's blame ICU ;)
21:23:29 <TrueBrain> andythenorth: like I said, I picked what other projects used :) They all use the same first and last two lines
21:23:44 <TrueBrain> I liked the second line from a few other projects
21:23:51 <TrueBrain> thought it might be good to have it more like other projects
21:23:57 <TrueBrain> but I am up for changing it completely
21:24:16 <glx> and useless stack trace, andythenorth just close 6572, nothing usable there
21:24:30 <TrueBrain> glx: you close it :P
21:24:49 <LordAro> TrueBrain: what's "closeComment: false" mean?
21:25:00 <TrueBrain> LordAro: that on close, it doesn't show any comments
21:25:12 <TrueBrain> you can also make it give a comment when it closes it after 7 days
21:25:16 <TrueBrain> just more fuzz, basically
21:25:39 <TrueBrain> as the stale bot was the last in the thread anyway, why say something again, most projects seems to say :)
21:26:25 <TrueBrain> as you can see, "security", "good first issue" and "regression" are never closed. They should always be fixed :P
21:26:35 <TrueBrain> well, the second because we need some tickets that are easy to get into OpenTTD with :)
21:26:38 <TrueBrain> no matter how old they are
21:30:08 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on issue #6621: Saving game then loading one of the same name via console, all in the same directory, fails. https://git.io/fhr2Y
21:30:09 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed issue #6621: Saving game then loading one of the same name via console, all in the same directory, fails. https://git.io/fhr2O
21:32:04 <TrueBrain> hmm, when I merge my PR, it will close all current 'stale' marked tickets, it seems :)
21:33:00 <LordAro> TrueBrain: perhaps exemptMilestones/exemptAssignees might be worthwhile?
21:33:11 <TrueBrain> LordAro: I considered it. But no other project uses it
21:33:24 <TrueBrain> so I was unsure about that
21:34:20 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain commented on issue #6916: Re-implement building binaries via compile farm https://git.io/fhr26
21:34:54 <andythenorth> TrueBrain: I can take 'stale' off some issues
21:35:11 <andythenorth> I think some are legit, but I added stale to indicate they're just rotting
21:35:22 <LordAro> i'm still concerned that any sort of stalebot is just going to end up closing issues before anyone can get to them
21:35:23 <TrueBrain> andythenorth: possibly we should remove them from all
21:35:28 <TrueBrain> and let the normal stale do its thing
21:35:39 <LordAro> then it's just going to keep bouncing between getting reopened (if it gets noticed) and closed by the bot
21:35:42 <TrueBrain> LordAro: 60 days to get to them .. and 1 comment to keep it another 60 :)
21:35:54 <TrueBrain> but ... in 120 days we can look back on that, see if you are right :)
21:35:56 <andythenorth> tree, falling, forest
21:36:01 <TrueBrain> we should not be afraid to try things out
21:36:08 <andythenorth> if nobody cares about the issue, does it matter if it was closed? o_O
21:36:09 <TrueBrain> just because we might have an unpleasent outcome :)
21:36:52 <andythenorth> also we seem to have an inexhaustible supply of PRs from just one contributor
21:36:53 <LordAro> i'd probably be happy with exemptMilestone & exemptAssignee, if only because i'm the only one that uses them :p
21:36:59 <frosch123> what about the "pinned" label?
21:36:59 <andythenorth> so we won't be short of admin work to do :P
21:37:04 <frosch123> does it also exist for pr?
21:37:15 <TrueBrain> LordAro: all the current ones will be exempt anyway :)
21:37:27 <TrueBrain> frosch123: labels are over issues and pull requests :)
21:37:53 <frosch123> stale.yml does not list exceptions for prs though
21:38:04 <andythenorth> ok I knock a bunch of labels off
21:38:16 <TrueBrain> frosch123: good point!
21:39:13 <TrueBrain> frosch123: only 'pinned' enough you think?
21:39:34 <frosch123> you can always set "pinned"
21:39:48 <TrueBrain> good first issue makes no sense
21:39:49 <frosch123> so, technically a single label is enough
21:39:54 <TrueBrain> security/regression PRs are unlikely
21:40:26 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain updated pull request #7096: Add: [GitHub] enable probot/stale, to flush out stale issues/pull requests https://git.io/fhrzX
21:40:27 <TrueBrain> there we go, should fix both the issue for frosch123 and LordAro
21:40:40 <andythenorth> oof these issues are dull :)
21:42:30 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed issue #6592: Missing symbols while linking with MinGW and LTO https://git.io/fhraW
21:42:31 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on issue #6592: Missing symbols while linking with MinGW and LTO https://git.io/fhral
21:44:10 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on issue #6337: Mapgen: no tropical trees and no snow with default world generation settings https://git.io/fhraV
21:44:28 <andythenorth> ok removed all stale
21:45:19 <LordAro> dang it, i meant to comment on 6592
21:45:27 <LordAro> this is what i'm worried about :p
21:45:49 <TrueBrain> LordAro: that you are getting old, or what?
21:46:06 <andythenorth> just comment and reopen, no kittens died, and now you got a prompt :)
21:47:26 <TrueBrain> LordAro: we no longer support MingW, do we?
21:47:52 <LordAro> TrueBrain: me & michi have put effort into making it work :p
21:47:53 <frosch123> isn't glx using mingw64 all the time?
21:48:08 <TrueBrain> didn't you say it doesn't support x11?
21:48:11 <LordAro> but, MSYS is probably worth deprecating, MSYS2 works fine
21:48:30 <TrueBrain> ah; so we should add that to the CF too
21:48:32 <frosch123> TrueBrain: we do not support mingw32, and thus do not support win9x
21:48:54 <TrueBrain> is mingw64 installed on the Windows Agents ..
21:49:16 <glx> it is if the github wiki is right
21:49:53 <orudge> Hmm, is there a reason the OS X build now seems to require 10.12? I've only got 10.11 (max my hardware can run), and the underlying openttd binary seems to run fine, but the package is marked as 10.12+ only
21:50:14 <glx> ha no it's the outdated one indeed
21:50:18 <TrueBrain> orudge: we said a while back that we are only "supporting" N-2
21:50:27 <TrueBrain> possibly it runs on older ones
21:50:32 <TrueBrain> but we are not building for it, etc
21:50:35 <andythenorth> matches Apple support policy afaict
21:50:40 <TrueBrain> (we only "support" what Apple does)
21:50:52 <TrueBrain> but yeah, it might even run on 10.8 :P
21:50:59 <orudge> Anyone want to donate a new Mac then? :D
21:50:59 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on issue #6592: Missing symbols while linking with MinGW and LTO https://git.io/fhrVL
21:51:00 <TrueBrain> we just know it runs on 10.12 :D
21:51:02 <glx> anyway as I said if someone really wants win9x, a PR will work :)
21:51:16 <planetmaker> where do I find the new nightlies / builds anyway?
21:51:23 * LordAro wonders when the last time someone tried compiling with --enable-lto on Linux
21:51:24 <TrueBrain> don't you read?! :'( :P
21:51:32 <glx> not from me, from the someone :)
21:51:45 <TrueBrain> LordAro: I would like to have a azure-pipelines.yml which tests all possible combinations
21:52:12 <andythenorth> orudge: I considered it briefly :P
21:52:13 <TrueBrain> but I would like a mingw64 in CompileFarm on the CI so we can validate we don't break mingw64 :)
21:52:35 <TrueBrain> LordAro: we can run that pipeline every week or so :P
21:52:47 <TrueBrain> anyway .. stalebot PR .. anyone dares to approve?
21:53:51 <andythenorth> you know I will if nobody else does :P
21:54:00 <andythenorth> but I'd have to actually read the rules properly
21:54:04 <LordAro> TrueBrain: remember to turn off github notifications before you set it to work :p
21:54:13 <planetmaker> ty. And no, I didn't read back hours of chatty IRC.
21:54:17 <LordAro> or it'll be a bit noisy in here
21:54:33 <TrueBrain> LordAro: andythenorth makes more noise :P
21:54:34 <andythenorth> I've removed the stale label from all open issues
21:54:40 <glx> TrueBrain: vcpkg uses msys2/mingw64
21:54:42 <planetmaker> hm. Why turning off github notifications?
21:55:10 <TrueBrain> glx: hmm .. can we also use it?
21:55:21 <TrueBrain> as that might be a quicker way to add it to azure-pipelines
21:55:28 *** gelignite has joined #openttd
21:55:42 <glx> but I don't know if it provides pacman
21:55:46 <planetmaker> no nightly linux builds yet?
21:56:01 <TrueBrain> planetmaker: 'yet', as in: till someone finds a way to build those
21:56:32 <TrueBrain> planetmaker: releases will have debian and ubuntu deb-files, but we have no 'generic' linux
21:56:40 <TrueBrain> someone has to invent them :D
21:56:47 <DorpsGek_II> [OpenTTD/OpenTTD] frosch123 approved pull request #7096: Add: [GitHub] enable probot/stale, to flush out stale issues/pull requests https://git.io/fhrVF
21:57:09 <DorpsGek_II> [OpenTTD/OpenTTD] TrueBrain merged pull request #7096: Add: [GitHub] enable probot/stale, to flush out stale issues/pull requests https://git.io/fhrzX
21:57:40 <TrueBrain> planetmaker: the version I had was a bit too modern for most systems, so that was useless :P
21:57:48 <TrueBrain> someone keeps telling me you have to keep the libc in mind or something like that
21:57:57 <TrueBrain> and than there were issues with what is compiled static
21:57:59 <TrueBrain> and ICU was being stupid
21:58:05 <TrueBrain> and more of that shit
21:58:17 <TrueBrain> someone who understands what they are doing should make a Docker that produces a generic linux binary :)
21:58:24 <planetmaker> meh. sounds nasty. But it used to work...
21:58:30 <TrueBrain> but .. honestly .. the amount of downloads we had on it, were .. low :)
21:58:40 <TrueBrain> "used to work" was more: nobody ever used them, so nobody noticed :P
21:59:13 <LordAro> linux users can use git & make :p
21:59:27 <TrueBrain> let me check the stats on it ...
21:59:43 <glx> hmm or it's just git shell
22:00:13 <TrueBrain> linux-generic-amd64: 101510
22:00:19 <peter1138> Hrmm, my watch says it's time to sleep :/
22:00:22 <peter1138> I've not eaten yet.
22:00:29 <TrueBrain> @calc 101510 / (3268067 + 3094301 + 438740)
22:00:29 <DorpsGek> TrueBrain: 0.0149255091964
22:00:32 <TrueBrain> @calc 101510 / (3268067 + 3094301 + 438740) * 100
22:00:32 <DorpsGek> TrueBrain: 1.49255091964
22:00:48 <TrueBrain> more people download the source, than the generic linux :D
22:00:51 <frosch123> TrueBrain: at some point some docker also built ai/gs docs. are they not built, or does the website not know them?
22:00:52 <TrueBrain> like ..... 3 times more people :P
22:01:01 <peter1138> People do want Linux builds.
22:01:02 <TrueBrain> frosch123: 'docs' should be there
22:01:17 <peter1138> Would be nice to build them, even if they are "too new"
22:01:19 <TrueBrain> as many people that want win9x :D
22:01:25 <planetmaker> TrueBrain, ŵhat time period?
22:01:31 <LordAro> @calc 101510 / (3268067 + 3094301 + 438740 + 101510 + 90867)
22:01:31 <DorpsGek> LordAro: 0.0145149378314
22:01:36 <planetmaker> I agree with peter
22:01:37 <glx> people usually prefer to get them via their distrib
22:01:55 <TrueBrain> planetmaker: the exact numbers hav elittle meaning
22:02:03 <peter1138> What happens to the Linux builds the CI makes?
22:02:06 <peter1138> Are they not runnable on anything?
22:02:10 <TrueBrain> and just to be clear: we talk about NIGHTLIES. not releases.
22:02:21 <LordAro> TrueBrain: if the numbers are post-april 2018, they're a little less useful
22:02:22 <frosch123> the farm hast changed so often, i do not recognise anything :p
22:02:25 <TrueBrain> peter1138: not really. They are not static, and depend on very specific installed versions
22:03:01 <planetmaker> can't one just build a static on the debian or ubuntu one and that's it?
22:03:17 <LordAro> old style linux generic is probably runnable on ubuntu versions (12.04+?), assuming they have the right dependencies
22:03:18 <TrueBrain> at least, I did the same as we used to do :)
22:03:35 <TrueBrain> planetmaker: I have not found a way to do that; after spending 10+ hours on it, I was like: fuck this, not worth it
22:03:37 <glx> linux generic were static
22:03:45 <TrueBrain> but I would really appreciate it if anyone did find a good way to make it work :)
22:03:52 <glx> so theorically no depenencies
22:03:52 <peter1138> I heard something about Steam the other day.
22:04:14 <peter1138> Can we target just whatever Steam on Linux targets?
22:04:19 <TrueBrain> all the code is available, and all the repositories welcome any PR :)
22:04:46 <planetmaker> that probably is a good idea with "steam's target"
22:05:13 <planetmaker> glx, yet it still depends on a certain glib version
22:05:17 <DorpsGek_II> [OpenTTD/OpenTTD] stale[bot] commented on issue #6922: Replace ICU ParagraphLayout with something else? https://git.io/fhrwX
22:05:19 <DorpsGek_II> [OpenTTD/OpenTTD] stale[bot] commented on issue #6907: Cargo capacity should be recalculated on TRIGGER_VEHICLE_NEW_LOAD https://git.io/fhrw1
22:05:23 <DorpsGek_II> [OpenTTD/OpenTTD] stale[bot] commented on issue #6322: AI can allocate more memory than the system has, crashing the game https://git.io/fhrwD
22:05:25 <DorpsGek_II> [OpenTTD/OpenTTD] stale[bot] commented on issue #6884: Installer reports success even if not all files were copied https://git.io/fhrwy
22:05:29 <DorpsGek_II> [OpenTTD/OpenTTD] stale[bot] commented on issue #6682: menu bars disappear if enable windows screen scaling on high dpi monitors https://git.io/fhrwS
22:05:35 <DorpsGek_II> [OpenTTD/OpenTTD] stale[bot] commented on issue #6800: [OSX] NSScrollWheel event handling for 2D scrolling should use scrollingDeltaX/Y https://git.io/fhrwH
22:05:37 <DorpsGek_II> [OpenTTD/OpenTTD] stale[bot] commented on issue #6712: Startup ticks don't allow multiple DoCommands https://git.io/fhrwQ
22:05:38 <andythenorth> easier than me doing it
22:05:41 <DorpsGek_II> [OpenTTD/OpenTTD] stale[bot] commented on issue #6493: Helicopters do not autorenew/autoreplace while in routes consisting of heliports https://git.io/fhrw7
22:05:43 <DorpsGek_II> [OpenTTD/OpenTTD] stale[bot] commented on issue #6233: Lock price for when built on sloped river and bare land tiles https://git.io/fhrw5
22:05:47 <DorpsGek_II> [OpenTTD/OpenTTD] stale[bot] commented on issue #6574: Heliport "hangar" can cause stuck helicopters if it replaced a helidepot/helistation. https://git.io/fhrwF
22:05:49 <DorpsGek_II> [OpenTTD/OpenTTD] stale[bot] commented on issue #6603: Insert XDG_DATA_DIRS into data loading path https://git.io/fhrwb
22:06:00 <andythenorth> you have automated my game :(
22:06:13 * andythenorth has to play 2048 instead now :P
22:06:24 <TrueBrain> guess that is no longer the same :D
22:07:09 <frosch123> forgotten, or not wanted?
22:07:13 <TrueBrain> ah! More hidden subdomains :D
22:07:19 <TrueBrain> but one does not exclude the other :)
22:07:33 <TrueBrain> I believe those were only updated on release
22:08:00 <LordAro> it's occasionally useful
22:08:06 <TrueBrain> I guess we should extend the 'release-docs' to also do the noai and nogs
22:08:15 <TrueBrain> I will make an issue out of it; hopefully someone wants to help out and fix it :)
22:08:45 <glx> and intellisense still don't like doxygen comments
22:09:17 <TrueBrain> ah, now I see. We didn't publish NoAI/NoGS for nightlies at all
22:09:33 <glx> no for stable only, and that makes sense :)
22:10:03 <frosch123> i consider script docs more useful then source docs
22:10:15 <TrueBrain> yeah; lets just publish them too with nightlies
22:12:58 <andythenorth> urgh why did I play 2048? :(
22:13:05 <TrueBrain> because it is fun :D
22:14:05 <TrueBrain> hmm, guess the NoAI / NoGS docs should be a Docker to run on the k8s too
22:14:14 <TrueBrain> that should be easy .. if I have the content :D
22:15:01 <TrueBrain> hmm .. seems even for releases there are no NoAI / NoGS tarballs
22:16:55 <LordAro> i don't see a particular need to provide that as a separate download
22:17:22 <TrueBrain> it makes it difficult to create a Docker containing the correct APIs :P
22:17:34 <TrueBrain> it is now uploaded to a storage, and that is it :D
22:18:08 <TrueBrain> I guess I could put it on the CDN .. hmm
22:18:27 <TrueBrain> that reminds me, I should make sure the CDN has a backup :P
22:19:05 <TrueBrain> LordAro: do you see any particular reason to have docs as a download in general? :D
22:19:27 <LordAro> if they're online somewhere
22:19:31 <TrueBrain> we publish NoAI and NoGS, but not generic docs, I think ..
22:20:05 <TrueBrain> holy crap, that is terrible
22:20:26 <LordAro> yeah, no one ever bothered making a doxygen title page
22:20:42 <TrueBrain> welll ..... when was this last generated ....
22:21:07 <TrueBrain> 2018-04-06, so it is updated
22:21:26 <TrueBrain> it is just horrible
22:21:42 <LordAro> the other tabs are a bit more useful
22:22:00 <TrueBrain> so okay ... I guess we should either publish all 3 docs as a binary to download
22:22:05 <TrueBrain> but 1 of the 3 is silly :D
22:22:54 <TrueBrain> okay, putting it on the CDN is easy and possibly the best pick here
22:23:01 <TrueBrain> meaning we can have the nightly overwrite every night
22:23:05 <TrueBrain> and have a "most recent" version
22:23:10 <TrueBrain> together with all the stables
22:23:16 <TrueBrain> that should be doable
22:24:13 <planetmaker> most important for the docs is that they are online readable. Download... is a nice bonus
22:24:57 <LordAro> hmm, do i milestone #6922 for 1.9?
22:26:15 <frosch123> i definitely linked people to the script apis, when they were asking for some specific function
22:26:33 <andythenorth> NoGo docs are pretty essential :)
22:26:33 <frosch123> can't remember doing the same for the source docs :p
22:26:34 <TrueBrain> there, a ticket for someone to pick up
22:27:19 <peter1138> My issue with submitting to the compilefarm... I have no idea how to test stuff, if that's even possible.
22:27:47 <TrueBrain> peter1138: did you even try to read the README? Or did you just assume things here?
22:28:03 <peter1138> Of course not, if I did I would know.
22:28:43 <frosch123> hmm, tb's forum signature...
22:29:34 <TrueBrain> I think I had the one about good men do nothing, right?
22:29:51 <TrueBrain> frosch123: yeah, some things haven't changed, sadly
22:30:56 <TrueBrain> okay ... seems we are pretty close to be able to release a beta :D
22:31:34 <peter1138> Thanks for just saying "please read the readme" rather than being condescending.
22:32:10 <TrueBrain> tnx for taking an effort, instead of seeing only issues :)
22:33:41 <andythenorth> should I comment, or wait for review :P
22:34:09 <LordAro> i think that can be fixed with a squash :p
22:34:25 <LordAro> i don't like that solution though, it's too brittle
22:34:50 <andythenorth> I tried to read what it does
22:35:07 <andythenorth> there seem to be a lot of crashes due to changes with certain windows open
22:35:18 <andythenorth> can't we just enforce window modality? :P
22:35:36 <andythenorth> trying to maintain window state when underlying state changes....
22:35:39 <andythenorth> is a fool's errand
22:35:49 <peter1138> It's not the right way to do what he's trying to do.
22:35:54 <LordAro> andythenorth: exactly
22:38:31 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro requested changes for pull request #7094: Fix #7088: Retrieve an appropriate name for a non-existant AI/GS when… https://git.io/fhr6S
22:39:24 <LordAro> TrueBrain: thought: is there an option to prevent an issue from being made stale/closed if it's referred to in a(n open) PR?
22:39:48 <TrueBrain> probot/stale might have an answer? Dunno
22:40:05 <TrueBrain> I think it only scans for 'updated'
22:40:12 <TrueBrain> but a PR might update that counter
22:43:04 <LordAro> doesn't seem like it, given #6574 was "updated" 12 days ago
22:44:43 <TrueBrain> so indeed, not supported, they accept PRs :D
22:44:59 <LordAro> "Javascript 100.0%" eh.
22:46:13 <peter1138> LordAro, #7085, your changes request is blocking it being merged.
22:46:53 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro dismissed a review for pull request #7085: Fix: Remove duplicate French and Slovak town names https://git.io/fhwfW
22:47:15 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro merged pull request #7085: Fix: Remove duplicate French and Slovak town names https://git.io/fh2Af
22:50:54 <supermop_work_> i like the smart town name request someone wrote
22:51:35 <supermop_work_> would be nice not to get those x-on-sea town in the middle of a prairie
22:51:50 <glx> but it will it work with newgrf townnames ?
22:51:59 <andythenorth> it could be made too
22:52:03 <andythenorth> with a spec extension
22:52:09 <supermop_work_> presumably it would only work with newgrf names
22:52:16 <andythenorth> also some towns should favour forests planting around them
22:52:19 <supermop_work_> and only those that made use of the feature
22:52:22 <andythenorth> oh did I just derail the conversation :P
22:52:38 <TrueBrain> andythenorth: we are used to you doing that :P
22:52:45 <andythenorth> anyone want to close this as stale?
22:52:57 <peter1138> It bases its decision on the string content of the name.
22:53:06 <supermop_work_> andythenorth: my brother and i build the chez albert restaurant this christmas
22:53:11 <peter1138> Which is fine but feels a bit hardcoded.
22:53:43 <TrueBrain> LordAro: do you know if Squirrel is still a maintained language? (looking at the stale you removed)
22:53:45 <peter1138> Also a mix of return types.
22:53:51 <glx> yeah for newgrf it could pass the flags and let the newgrf do its work
22:54:10 <peter1138> That would need ... infrastructure.
22:54:28 <LordAro> TrueBrain: iirc, it depends what you mean by "maintained" :p OTTD uses Squirrel2, whereas Squirrel3 is a thing these days
22:54:41 <TrueBrain> are the languages compatible?
22:54:42 <LordAro> but our copy of Squirrel is fairly significantly modified anyway
22:55:00 <supermop_work_> put nrt in 1.9 so i will feel compelled to work on grfs again
22:55:01 <nielsm> upgrading squirrel in ottd is probably not realistic
22:55:10 <nielsm> may as well switch to a wholly different language then
22:55:22 <TrueBrain> nielsm: that requires AIs/GSs to change too :P
22:55:36 <TrueBrain> if the language it compatible, upgrading the engine is "just work"
22:55:54 <nielsm> at least if it follows my experience with lua 5.0 to 5.1 which are not compatible :)
22:55:56 <TrueBrain> or, we could write a conversion script! :P
22:56:12 <TrueBrain> does lua support any decent form of OO by now?
22:56:23 <TrueBrain> or is it still this annoying to do any sane complex thing in?
22:56:30 <glx> how silly to break compatibility with a minor release
22:56:38 <nielsm> haven't followed it for many years
22:56:41 <TrueBrain> (btw, the most prominent reason Squirrel is in OpenTTD, and not lua. OO :P)
22:56:45 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on pull request #6925: Fix #6574 #6636 #5405 #6493: Aircraft hangar issues https://git.io/fhrPg
22:56:46 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed pull request #6925: Fix #6574 #6636 #5405 #6493: Aircraft hangar issues https://git.io/fpJfr
22:56:51 <frosch123> TrueBrain: lua is still terrible with custom types
22:57:45 <TrueBrain> I would expect them to have that in order by now
22:57:50 <TrueBrain> given ... many games use it :P
22:57:52 <frosch123> like the standard library putting essential dict/array functions into the same library as "load shared library from search path"
22:58:15 <glx> hmm right squirrel vm has no direct access to OTTD code
22:58:19 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on pull request #6926: Change: Allow dock to be constructed in more locations https://git.io/fhrPy
22:58:20 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed pull request #6926: Change: Allow dock to be constructed in more locations https://git.io/fhnCx
22:58:26 <TrueBrain> hmm .. I should have finished NAILs after all .. the BOLTS language was similar to Squirrel, but a bit less weird :P
22:58:36 <frosch123> at work i decided to disable the entire standard library, and reimplement the 3 useful methods
22:58:42 <TrueBrain> the good old days, where writing your own language was an option :D
22:59:30 <nielsm> use javascript (and get a barrage of "javascript is a terrible language"), or use something else (and get a barrage of "you should use de-facto standard javascript instead")
22:59:30 <TrueBrain> omg, GPMI .. that is old :P
22:59:40 <TrueBrain> that was the moment I realised that most script languages suck balls
22:59:52 <TrueBrain> nielsm: add NodeJS support? :D
23:00:01 <frosch123> nielsm: i read left-padding in js is nice
23:00:30 <orudge> VBScript is the way forward
23:00:54 <glx> at least vbs is nice with case
23:00:54 <nielsm> just use c#, import mono and be over with it
23:01:36 <TrueBrain> honestly, if we add NodeJS support, we might get more AIs :P
23:01:41 <TrueBrain> they will all suck more than the next, but ....
23:01:51 <andythenorth> can we import all the node vulns too?
23:01:56 <andythenorth> then we can lose all our bitcoin
23:02:18 <nielsm> make an ai that builds a bitcoin miner with logic trains
23:03:15 <frosch123> can we agree to reject prs that want to add digital currencies?
23:03:19 <andythenorth> LordAro nielsm I made a decision about some PRs you've been reviewing
23:03:26 <TrueBrain> frosch123: no, lets argue about that first
23:03:50 <DorpsGek_II> [OpenTTD/OpenTTD] PeterN commented on issue #6896: game crashes on first autosave if compiled without lzo but with lzma https://git.io/fhrXB
23:05:26 <nielsm> add TycoonCoin as currency option in game, if you use it you will be forced to play with inflation on, and running costs increase at the square of inflation
23:05:54 <TrueBrain> no no; first, you make it worth your while
23:06:04 <TrueBrain> for 1 year or something, everything is incredible cheap
23:06:09 <TrueBrain> you can build the biggest infrastructures
23:06:20 <TrueBrain> after that ..... CRASH :P
23:06:48 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on pull request #6927: Change: Forbid dock placement on docking area https://git.io/fhrXD
23:06:49 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed pull request #6927: Change: Forbid dock placement on docking area https://git.io/fhnCh
23:06:50 <TrueBrain> what ever happened with the game rebalancing ideas ....
23:06:56 <frosch123> how about: the game is slowed the more money you stockpile
23:07:10 <frosch123> so you have to spend it to have stuff moving
23:07:17 <andythenorth> well the coin miner will use more CPU
23:07:21 <andythenorth> so that will slow the game down
23:07:29 <andythenorth> we get 50% of the mined coins, right?
23:07:35 <andythenorth> to pay for servers :P
23:07:42 <TrueBrain> anyway .. 2 more days before we bring the new website online in production .. please let me know if you find any additional issues: https://www.staging.openttd.org/ (only known issue currently is that there are no 'noai/nogs' docs)
23:07:58 <andythenorth> TrueBrain: have you tested the coin miner in it is working? o_O
23:08:16 <nielsm> TrueBrain: should I resurrect my cargo-tiles patch? that rates vehicles by cargo over distance delivered rather than by profit
23:08:35 <andythenorth> nielsm: sounds...interesting :)
23:09:03 <andythenorth> for 2.0 in 2020, can we delete the stupid original high score goals?
23:09:06 <frosch123> are the translations real? or are they fake?
23:09:08 *** Simons_Mith has joined #openttd
23:09:22 <andythenorth> child #1 drove me crazy trying to get 100% of the bars complete
23:09:24 <nielsm> and replaces the "min yearly profit" element of the score goals with one based on that instead
23:09:28 <TrueBrain> frosch123: damn, I forgot about those
23:09:30 <andythenorth> especially the min. vehicle profit
23:09:33 <nielsm> frosch123: it actually measures distance traveled
23:09:41 <TrueBrain> andythenorth: what did we say about the translation stats?
23:09:47 <frosch123> TrueBrain: well, now that the screenshot is different, they can remain the same?
23:09:49 <andythenorth> we said bin them
23:09:52 <andythenorth> and rely on eints
23:10:07 <TrueBrain> I kinda agree, lets remove it. Not sure it ever helped finding people tbh
23:10:08 <andythenorth> we were going to replace that block, or remove it
23:10:12 <nielsm> so a curvy route counts as longer distance traveled, but the vehicle is rated on the manhattan distance the cargo was delivered
23:10:33 <nielsm> so vehicles traveling less direct routes are rated worse
23:10:58 <TrueBrain> andythenorth: why did it pick a screenshot from 0.4.7? Didnt we tell it to pick the latest screenshot?
23:11:37 <nielsm> why aren't there any screenshots from recent versions even?
23:12:09 <TrueBrain> nielsm: because it was not easy to upload new screenshots
23:12:12 <TrueBrain> so people stopped doing it
23:12:44 <TrueBrain> andythenorth: {% assign latest_screenshot = site.screenshots | first %}
23:12:53 <andythenorth> now it's a PR to add screenshots, right?
23:13:16 <TrueBrain> "site.screenshots | group_by_exp: "screenshot", "screenshot.path"
23:13:41 <andythenorth> "The story goes that one programmer, who had to write the code to calculate the height of a line of text, simply wrote “return 12;” and waited for the bug report to come in about how his function is not always correct. "
23:14:12 <andythenorth> TrueBrain: are you fixing, or am I? Or issue for later?
23:14:26 <TrueBrain> I am about to go to bed; I was creating an issue
23:14:30 <TrueBrain> how do I get inline code snippets
23:14:47 <andythenorth> issue works, I am bed also
23:15:50 <andythenorth> there's some kind of utilisation feature added to vehicle groups recently
23:15:53 <andythenorth> dunno what it is
23:15:55 <TrueBrain> frosch123: keep the bugs coming :)
23:16:19 <nielsm> andythenorth yeah, it measures current average load percentage
23:17:39 * LordAro currently looking at the diff between a clean squirrel 2.2.5 and the modified version in ottd
23:17:45 <LordAro> the changes are quite extensive
23:17:45 <andythenorth> nielsm: I could try your patch another day :)
23:17:48 <andythenorth> it's in your fork?
23:17:59 <nielsm> yeah, but haven't updated it lately
23:18:06 <nielsm> probably won't rebase to master
23:18:39 <Samu> how to save the config constructor yada what he said
23:18:51 <frosch123> LordAro: mostly about suspending scripts?
23:18:53 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on pull request #6928: Fix #5713: Use pathfinder to find closest ship depot https://git.io/fhrML
23:18:54 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed pull request #6928: Fix #5713: Use pathfinder to find closest ship depot https://git.io/fhZzm
23:19:10 <LordAro> frosch123: as far as the actual changes go, yes
23:19:51 <LordAro> there's a lot of mechanical changes, things like removing a define that adds extern to a load of functions, and makes all error strings wchars
23:19:58 <LordAro> Samu: you could ask me :p
23:20:35 <andythenorth> I am about to close the PR, just want to be accurate
23:21:11 <Samu> watching my stuff get closed is depressive
23:21:48 <LordAro> Samu: it's nothing personal, they're just changes that don't fit into OTTD into one way or another
23:22:03 <andythenorth> Samu: if it helps, I have deleted entire companies
23:22:15 <frosch123> it changes too much in one thing, and i did not agree with the individual thigns split apart
23:22:28 <LordAro> luckily, there's plenty of other issues you can work on :)
23:22:29 <andythenorth> deleting one company involved about 10 people losing their jobs
23:22:38 <andythenorth> but otherwise I would have gone bankrupt
23:22:51 <andythenorth> all code dies, just some PRs die early
23:23:06 <Samu> ahm that PR was about the bug report
23:23:22 <TrueBrain> DELETE FROM companies LIMIT 0,1
23:23:35 <Samu> found some inconsistencies, then came up with making it more consistent
23:23:44 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on pull request #6933: Fix #6233: Improve lock pricing and infrastructure counting, to achieve better consistency https://git.io/fhrM6
23:23:45 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth closed pull request #6933: Fix #6233: Improve lock pricing and infrastructure counting, to achieve better consistency https://git.io/fhnWe
23:24:12 <andythenorth> I agree that locks and canals cost too much to build
23:24:15 <andythenorth> but I fixed it in newgrf
23:24:29 <andythenorth> it's in the game goals, everything that can be fixed with content, should be fixed with content
23:24:31 <Samu> but the bug... goes unfixed
23:24:48 <LordAro> andythenorth: base game should still be sane though
23:24:55 <nielsm> I don't think you've yet convinced anyone else there is a bug at all?
23:25:27 <andythenorth> I am convinced that some costs for building on water make no sense
23:25:32 <Samu> i tried explaining, requires using base costs, to exploit a flaw
23:25:39 <andythenorth> but I am also convinced it doesn't really matter
23:25:57 <TrueBrain> try convincing that yourself
23:27:23 <andythenorth> but it's very hard to test reliably
23:27:41 <andythenorth> it is silly when town expansion breaks a ship route
23:28:25 <Samu> 6931 is waiting review atm
23:28:33 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on issue #6712: Startup ticks don't allow multiple DoCommands https://git.io/fhrDn
23:29:51 <Samu> for my part, there's nothing to do
23:31:22 <andythenorth> isn't that same as retina etc?
23:31:23 <LordAro> pretty sure there's been some work in the high dpi area since these issues were raised
23:31:45 <andythenorth> I could install windows 7?
23:32:37 <peter1138> Hmm, my simulated high-res modes have disappeared.
23:32:58 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro commented on issue #6682: menu bars disappear if enable windows screen scaling on high dpi monitors https://git.io/fhrDH
23:33:45 <andythenorth> rail / road bridges don't raise a corner
23:33:50 <andythenorth> they have foundations
23:34:20 <andythenorth> I see no rationale for having different behaviour for aqueducts
23:34:23 <LordAro> maybe the issue is that aqueducts don't have foundations :p
23:34:37 <andythenorth> there's no reason to stop canals having foundations
23:34:46 <andythenorth> unless it's an implementation issue
23:35:15 <peter1138> Ok, I set up 4x DSR...
23:35:48 <LordAro> andythenorth: your turn to comment as such :)
23:36:11 <peter1138> Nothing disappears, but I'm using Windows 10.
23:36:34 <LordAro> try checking out some old version?
23:36:55 <peter1138> Too much hassle with VS2017
23:37:02 <LordAro> just downloading 1.8 would do :p
23:37:11 <glx> I think high dpi is ok in master (not testable for me)
23:37:27 <peter1138> No need, already installed...
23:37:39 <peter1138> Simulated high DPI though.
23:37:46 <DorpsGek_II> [OpenTTD/OpenTTD] andythenorth commented on pull request #7078: Feature: Raise a corner of land automatically when building aqueducts https://git.io/fhryg
23:37:53 <glx> I mean something was added in the manifest for that
23:38:09 <LordAro> that's what i was thinking of
23:38:26 <LordAro> presumably only makes a difference for MSVC-compiled versions, rather than MSYS
23:39:13 <glx> hmm MSYS can embed the manifest too I think
23:40:12 <orudge> The manifest ultimately is just an entry in the resource file
23:40:20 <orudge> It may just require adding to the .rc
23:40:32 <frosch123> andythenorth: you can't build canals on foundations
23:40:34 <milek7> maybe system dpi settings could modify openttd interface size setting?
23:40:52 <frosch123> so, to some extent all water construction requires landscaping
23:40:55 <glx> yes system zooming have an effect
23:41:09 <andythenorth> frosch123: for implementation reasons, or realism reasons?
23:41:19 <glx> or scaling whatever it's called
23:42:03 <frosch123> andythenorth: the former may be the case historically, but would be fixable if it applies. so, mostly the latter?
23:42:34 <andythenorth> you should see the canals in wales :P
23:43:02 <andythenorth> I have been along one that is halfway up a hillside, with foundations :P
23:43:33 <frosch123> hmm, actually.. i guess all vehicle types fall of the cliff if the track continues on the lower side
23:43:44 <frosch123> we also forbid placing roads like that
23:43:49 <glx> hmm seems the dpi_aware.manifest is no longer in use btw as it was for vs100
23:44:10 <frosch123> so, maybe it would not be that easy to allow ships on foundations
23:44:47 <frosch123> no other track type can bridge canyons via foundations
23:45:13 <andythenorth> was my comment clear? :)
23:45:23 <andythenorth> I wasn't proposing embankment over canyon :)
23:48:46 <Samu> erm, aqueducts have no foundations
23:49:04 <Samu> they build much like a tunnel in reverse
23:49:28 <Samu> tunnels lower a corner of land, aqueducts raise a corner of land
23:50:17 <andythenorth> and rail bridges? o_O
23:50:35 <Samu> needs drag & drop to build
23:51:40 <andythenorth> ok I see the point about tunnels
23:51:59 <frosch123> i always wanted to add tunnel foundations
23:51:59 <andythenorth> aqueducts are weird :P
23:52:10 <frosch123> and tunnels with slopes at the end
23:53:44 <frosch123> at some point you always run into visibility problems on north-facing slopes
23:56:23 <frosch123> are sure? it looks like kid 1 built the road, and kid 2 wanted their boat to pass
23:56:34 <andythenorth> no, it's all one :P
23:56:54 <andythenorth> MP would be too much trouble :P
23:57:32 <andythenorth> at the top right, the road seems to be in a town zone?
23:57:33 <frosch123> make a PR for the website?
23:57:43 <andythenorth> note the pavements :P
23:57:46 <frosch123> someone wanted more screenshots
23:57:55 <frosch123> yeah, pavements are weird
23:58:15 <andythenorth> town zone for water tiles :P
23:58:18 <frosch123> also the road is a dead end in the south
23:58:29 <andythenorth> canal cheat is the best thing ever
23:58:49 <andythenorth> is it even a cheat?
23:58:49 <frosch123> i think it was voted "not a cheat"
23:59:02 <Samu> save script object in constructor
23:59:02 <andythenorth> "unintended outcome"
23:59:05 <frosch123> i think there was a fix, and it was rejected
23:59:28 <andythenorth> I encourage anything that adds to free-form construction :P
23:59:42 <andythenorth> there is nothing gained by limiting what can be built where
23:59:59 <Samu> what is LordAro trying to tell
continue to next day ⏵
