IRC logs for #openttd on OFTC at 2018-04-26
⏴ go to previous day
00:10:46 *** Thedarkb1-X40 has joined #openttd
00:24:14 *** sim-al2 has joined #openttd
00:30:08 <peter1138> Anyone know what BitBlt does behind the scenes? :s
00:33:26 <peter1138> Oh, that opengl patch breaks the gdi driver :(
00:38:33 <Eddi|zuHause> 00:38? that's not quite petertime
00:38:51 <peter1138> 11:38pm BST, so... er... maybe
00:56:17 <Eddi|zuHause> only in silly backwards countries that have no 24h clock
01:02:32 <peter1138> Yeah but 23:38 doesn't work either :D
01:09:32 *** chomwitt has joined #openttd
01:21:43 *** chomwitt has joined #openttd
01:45:47 <peter1138> Interesting. Added a call to DwmFlush() and that stops my graphical issue.
01:48:31 <peter1138> Or maybe it was luck :p
01:50:40 <peter1138> Ok, found a "solution".
01:50:55 <peter1138> Run another game in the background, on the same screen.
02:22:23 <Eddi|zuHause> reminds me of the bug with very old games which ran way too fast on "modern" computers (386 with 25MHz), and the "solution" was to run them under windows, because that slowed everything down :p
02:23:46 *** Thedarkb1-X40 has joined #openttd
02:55:44 *** goudy[m] has joined #openttd
04:07:43 *** ChanServ sets mode: +v tokai
04:08:03 *** ZehMatt_ has joined #openttd
04:09:09 *** TrueBrain_ has joined #openttd
04:10:16 *** argoneus_ has joined #openttd
04:10:23 *** berndj-blackout has joined #openttd
04:10:58 *** orudge` has joined #openttd
04:11:18 *** rustyshackleford[m] has quit IRC
04:11:27 *** Rubidium has joined #openttd
04:11:38 *** ChanServ sets mode: +o Rubidium
04:11:49 *** jonty-comp has joined #openttd
04:11:51 *** planetmaker has joined #openttd
04:11:51 *** ChanServ sets mode: +o planetmaker
04:12:20 *** lastmikoi has joined #openttd
04:13:01 *** tokai|noir has joined #openttd
04:13:01 *** Terkhen has joined #openttd
04:13:01 *** ZehMatt has joined #openttd
04:13:01 *** TrueBrain has joined #openttd
04:13:01 *** argoneus has joined #openttd
04:13:01 *** triolus[m] has joined #openttd
04:13:01 *** blathijs has joined #openttd
04:13:01 *** rustyshackleford[m] has joined #openttd
04:13:01 *** UncleCJ has joined #openttd
04:13:01 *** dacia.oftc.net sets mode: +vo tokai|noir Terkhen
04:13:19 *** ChanServ sets mode: +v Terkhen
04:13:54 *** Sheogorath has joined #openttd
04:53:58 *** blathijs has joined #openttd
04:55:13 *** muffindrake1 has joined #openttd
05:02:03 *** ToffeeYogurtPots has quit IRC
05:02:30 *** ToffeeYogurtPots has joined #openttd
05:20:50 *** snail_UES_ has joined #openttd
05:33:21 *** muffindrake1 is now known as muffindrake
06:10:09 *** sim-al2 is now known as Guest945
06:10:10 *** sim-al2 has joined #openttd
06:25:00 *** Wacko1976 has joined #openttd
06:25:23 *** sim-al2 is now known as Guest948
06:25:24 *** sim-al2 has joined #openttd
06:45:52 *** ToffeeYogurtPots has quit IRC
06:56:30 *** sla_ro|master has joined #openttd
07:37:05 *** Wacko1976 has joined #openttd
07:40:41 *** andythenorth has joined #openttd
08:45:34 *** Wacko1976 has joined #openttd
08:45:46 *** sim-al2 has joined #openttd
08:50:40 *** andythenorth has joined #openttd
09:03:03 <Eddi|zuHause> depends on your definition of "is", "it", "done" and possibly "o_O"
09:59:08 *** Wacko1976 has joined #openttd
10:54:08 *** andythenorth has joined #openttd
10:55:45 <andythenorth> I removed lots of sprites, because symmetric trains
10:55:53 <andythenorth> then I drew some of them asymmetric :P
11:11:42 <V453000> I iz making pythony things and it's such a mess
11:14:46 <peter1138> So anyone rendering vehicle sprites but using a good amount of sharpening so it's not blurry?
11:15:14 <peter1138> Also, when do we get 3D models in game :D
11:16:11 <andythenorth> they have x, y, z :P
11:16:21 <andythenorth> V453000: WHAT ARE YOU DOING
11:16:47 <peter1138> andythenorth, i'm sure a modern game that wasn't in C++ could do it easily, without specs.
11:16:58 <andythenorth> peter1138: probably just use a unity plugin
11:17:05 <andythenorth> don't worry about the CVEs
11:17:17 <andythenorth> I could rebuild OpenTTD in Flash if you want
11:17:19 <V453000> making a drop-folder-on-python-script, 32bpp x4, 32bpp x1, 8bpp x4, 8bpp x1 comes out automatically
11:17:25 <andythenorth> although my skills are a bit 15 years ago
11:17:31 <andythenorth> that's done isn't it?
11:17:44 <V453000> making spritesheet, scaling, converting with the other tool
11:17:46 <andythenorth> because even unskilled developers can write highly performant apps
11:18:22 <peter1138> Do you still use spritesheets or individual images?
11:18:39 <V453000> spritesheets for each vehicle in each rotations
11:18:44 <V453000> but not the giant things I had before
11:19:02 <peter1138> Yeah, giant one-image-for-the-whole-set is silly.
11:19:20 <V453000> it wasn't so silly when I didn't have the automatic 8bpp converter
11:19:36 <andythenorth> kind of unwieldy to QA
11:19:53 <andythenorth> V453000: what is this 'drop folder' shit?
11:20:09 <peter1138> I'd probably use individual sprites if I was rendering. Image at 0,0, only thing to set is the offset.
11:20:10 <andythenorth> why do you have to manually intervene?
11:20:10 <andythenorth> if you use a mouse while compiling, that's wrong
11:20:16 <V453000> andythenorth: not having to use any commands to launch the thing
11:20:33 <peter1138> Have one command and double click :D
11:20:36 <V453000> just select folder(s) and it does the thing
11:20:45 <andythenorth> peter1138: that's inefficient to QA
11:20:48 <V453000> hey, it's useful for me :P
11:20:50 <peter1138> That's not particularly automated.
11:20:59 <peter1138> andythenorth, probably but easier to rendering.
11:21:01 <V453000> it's able to run from a command too peter1138
11:21:04 <peter1138> andythenorth, hand-drawn, no.
11:21:32 <peter1138> V453000, then I'd script it so it's repeatable rather than relying on a drag&drop operation that could go wrong with misclicks.
11:21:33 * andythenorth back to automating spritesheets :P
11:21:41 <andythenorth> I am using the word 'gestalt' a lot
11:21:45 <peter1138> Let's just make tools instead of doing the thing.
11:21:59 <andythenorth> it is quite satisfying
11:22:11 <andythenorth> 'boom' look at my automation
11:22:20 <andythenorth> "with one shell command I can script my entire editor"
11:22:22 <V453000> the players love it andy
11:22:35 <andythenorth> they love the automated compile? o_O
11:22:50 <andythenorth> they'll love it if we ever ship anything :P
11:22:54 <andythenorth> remember how we used to ship?
11:22:54 <V453000> it's the n1 feature for them
11:23:04 <andythenorth> now we don't ship, all hype machines
11:23:14 <V453000> yes I had like 4 nuts releases per week :D
11:23:21 <andythenorth> I haven't shipped anything for months
11:23:28 <andythenorth> except three fixes to FIRS
11:23:34 <andythenorth> 2 of the 3 were fixes to the fix
11:23:46 <andythenorth> V453000: it's because it's C++
11:23:50 <peter1138> Let's use minecraft with a custom graphics set to create voxel models and then screenshot them to pngs...
11:23:52 <andythenorth> and the need for specs
11:23:57 <andythenorth> peter1138: plausible
11:24:01 <andythenorth> also we can make MC mods
11:24:06 <V453000> my current plan is to make the prototype with mostly placeholder graphics, let people play it, do updates and make graphics at the same time
11:24:09 <andythenorth> they're compiled to a JIT and super reliable
11:24:10 <V453000> it's basically what I did with NUTS
11:24:21 <andythenorth> and it doesn't matter if they don't work, because change is what players want
11:24:26 <V453000> when I realize how unfinished NUTS 0.0.1 was compared to latest, it's ridiculous
11:24:28 <andythenorth> they don't have to work
11:24:38 <V453000> but of course when designing I am aiming for latest level immediately
11:24:44 <peter1138> (except those bugs, but we ignore those)
11:24:50 <andythenorth> I have a TouchBar mac because Players need New Things
11:24:56 <andythenorth> but that's secondary
11:25:02 <andythenorth> tech bloggers were being mean to Apple
11:25:10 <andythenorth> so now I have a stupid thing that doesn't work
11:25:12 <peter1138> Is it lunch time yet? I'm going with no but my tummy is grumbling.
11:25:23 <andythenorth> the ultimate combo
11:25:32 <peter1138> I don't need Gaviscon, yay.
11:25:32 <V453000> I'm on python now, it's a good trip
11:25:40 <andythenorth> V453000 you should make a grf called 'drugs'
11:25:47 <andythenorth> I can't let my kids play though
11:25:53 <peter1138> I was on it for about a year, pointlessly, because it was my bloody gall-bladder that was being a cunt.
11:26:03 <andythenorth> I just eat it for fun
11:26:15 <peter1138> They whipped it out in hospital last year.
11:26:28 <andythenorth> do you have trouble processing fatty foods now?
11:26:31 <peter1138> And Gaviscon tastes a bit nasty.
11:26:41 <andythenorth> someone I know had gall bladder out
11:26:58 <andythenorth> digressing, will Net Nanny stuff ban kids from FIRS?
11:27:12 <andythenorth> someone else than you
11:27:21 <peter1138> But fattiest stuff I eat anyway tends to be gourmet hipster style burgers.
11:27:57 <andythenorth> FIRS has Coke, Slag, Acid
11:28:07 <andythenorth> think of the children
11:28:15 <andythenorth> even Fertiliser and Oil
11:28:40 <andythenorth> allegedly, in early internet days, searching for 'Fertiliser' and 'Diesel' got you flagged on Echelon
11:28:49 <andythenorth> due to the desire to blow things up that way
11:29:24 <andythenorth> also there are Reefers in my newgrfs
11:29:37 <V453000> if you had a vehicle which can run on road and rail, how would you expect it's performance to change based on these tracks?
11:29:47 <andythenorth> and there's an engine called Northcock, which is zero euphemism, but someone is inevitably going to think it is
11:29:56 <V453000> road for more tractive effort, rail for more speed? vice versa?
11:30:54 <andythenorth> V453000: this is railtype hax?
11:30:58 <andythenorth> or you patching core? :P
11:31:13 <V453000> well I have rails which look like water, roads are next
11:31:14 <andythenorth> child #1 wants to know when we're doing trains that go on road too
11:31:54 <V453000> this is pretty much that, and it has the bonus feature of diagonal roads :P
11:31:59 <andythenorth> Pikka: shall I streamline my 2-8-2 and please the crowd? o_O
11:41:07 <andythenorth> only if it looks hench
11:41:28 <andythenorth> Pikka: is your pendolino done yet? :P
11:42:03 <andythenorth> am I on a pendolino?
11:42:17 <andythenorth> no I'm on a Super Voyager thing
11:42:39 <andythenorth> or maybe it's not Super
11:42:46 <andythenorth> doesn't seem to tilt
11:43:06 <Pikka> maybe they turned the tilting off
11:43:17 * Pikka tracking tabled Av32, because that was a useful thing to do
11:43:26 <andythenorth> seems to be locked out according to wikipedia
11:43:34 <andythenorth> it's always nice to design a new one
11:43:41 <andythenorth> rather than draw the current one
11:43:58 <andythenorth> I am thinking about NARS3
11:44:08 <andythenorth> GP38, caboose, done?
11:46:29 <Pikka> 4-4-0, Big Boy, F7, GP38, SD40-2, ES44AC. That's probably about right.
11:47:02 <andythenorth> can I ignore the cabbage?
11:47:14 <andythenorth> "Just say no to DVTs"
11:47:45 <andythenorth> that's what happens when the sides are stuck on in the compile
11:47:48 <andythenorth> and they're not drawn yet
11:51:39 <peter1138> Skeleton Horses in Minecraft were pretty scary.
11:52:41 *** FLHerne has joined #openttd
11:59:31 <Pikka> mmm, minecraft. now there's a game that gets it right when it comes to modern scripting language, half-finished features with no coherent design, and not caring about backwards compatibility
11:59:41 <andythenorth> Sacro: it's UKRSn+1 now
11:59:56 <andythenorth> Pikka: but look at the community!!
12:00:04 <andythenorth> there are only two industry sets for OpenTTD
12:00:11 <andythenorth> one is dead, and the other has only made tiny improvements
12:00:26 <Pikka> never mind the fact that it's only active because every time the game is updated, all the mods need to be rewritten
12:00:37 <andythenorth> and as child #1 keeps pointing out, there are no mods on Minecraft PE, despite it being most popular platforn
12:01:05 <andythenorth> it's change that matters
12:01:12 <andythenorth> like bad therapy
12:01:36 <andythenorth> the problem is that out attitudes are from 30 years ago
12:01:37 <Sacro> I'll stick with Factorio
12:01:41 <andythenorth> and I can't keep up
12:01:51 <Sacro> WHERE ARE THE PONG MODS
12:02:00 <andythenorth> I learnt everything about software when I was 10
12:02:08 <andythenorth> also I played loads of TTD in 1988
12:02:51 <andythenorth> V453000: just bitching
12:03:22 <andythenorth> I would actually like to have sensible debates with the relevant OP every year when the latest version of the thread is born again
12:03:32 <andythenorth> like I think KK has some sensible points
12:03:51 <andythenorth> but it's all just expressed in so much wah wah
12:03:57 <andythenorth> and so much wrong
12:04:24 * andythenorth back to making trains not appear the wrong way round in -> view
12:05:22 <andythenorth> shall I add a steam engine to Horse for 2020?
12:05:25 <andythenorth> it's total realism
12:06:14 <V453000> steamers are fun, case closed.
12:06:25 <andythenorth> it would be the one from 1930, back again
12:06:27 <andythenorth> upgraded, faster
12:06:46 <andythenorth> it is a conundrum
12:06:52 <andythenorth> player feedback is that realism is better
12:07:03 <andythenorth> but the UK has stopped buying new locomotives and stuff
12:07:10 <andythenorth> so when I just repeat the old ones (like realism)
12:07:12 <andythenorth> they are unhappy
12:07:27 <andythenorth> FLHerne: I blame you :D
12:09:18 <andythenorth> maybe I can look out the window for new trains
12:09:26 <andythenorth> and draw them quick
12:10:18 <andythenorth> I can see some now
12:10:22 <andythenorth> all Cargo Sprinters though
12:11:44 <FLHerne> Aw, you renamed the Raven :-(
12:11:59 <FLHerne> I know it was 'realistic' before, but it was also a cool name. The new one is silly :P
12:12:24 <FLHerne> (the new tank names are nice, though)
12:15:30 <andythenorth> I could unrename the Raven
12:15:37 <andythenorth> Raven is a bird, animal names are possible
12:15:46 <andythenorth> it's also named after the designer
12:15:52 <andythenorth> it just didn't sound very hench
12:16:01 <andythenorth> 'Super Raven' :P
12:18:05 <FLHerne> Looking at your revised linear speed bands, I still think they're wrong ;-)
12:18:22 <andythenorth> that's ok, the possibility of a parameter still
12:18:30 <andythenorth> I just make one step at a time, cleaner
12:18:31 <FLHerne> You have 60mph locos from 1860-1900, then nothing, then 1930-1950
12:19:08 <FLHerne> Then 75mph from 1900-1930, nothing, 1960-1990
12:19:16 <andythenorth> that looks like errors
12:20:06 <FLHerne> No, it's because freight and pax are still two bands apart, but you only move up one each generation, so there's no overlap
12:20:16 <FLHerne> (except the first one because it's a bit different)
12:21:15 <andythenorth> branch engines are bugged though
12:21:30 <FLHerne> (90mph 1930-1960, nothing, 1990-2020, etc.)
12:21:55 <FLHerne> Oh, right, those were meant to be pax speed
12:22:45 <FLHerne> I still think the smaller freight loco, and maybe also the branch loco? should be in the middle band that's currently empty in each gen
12:22:56 <FLHerne> Latter might be annoying if they have to share tracks with expresses ever
12:22:57 <andythenorth> k biab, change trains
13:04:31 *** tokai|noir has joined #openttd
13:04:31 *** ChanServ sets mode: +v tokai|noir
13:28:15 *** chomwitt has joined #openttd
13:30:54 *** Wacko1976 has joined #openttd
14:19:44 *** Wacko1976 has joined #openttd
14:43:51 *** sim-al2 has joined #openttd
15:22:46 *** Flygon_ has joined #openttd
15:31:27 *** supermop_work has joined #openttd
15:39:49 *** Wacko1976 has joined #openttd
16:17:16 *** Wormnest has joined #openttd
16:18:07 <nielsm> okay, back from vacation and most major troubles that occurred in the meantime are resolved... time to try to rebase those patchsets I worked on before
16:28:44 *** andythenorth_ has joined #openttd
16:29:19 <andythenorth_> there are trains
16:35:15 <peter1138> I can see... sunshne.
16:35:27 <peter1138> Seems it is going to piss down tomorrow though :(
16:39:05 *** andythenorth_ has joined #openttd
16:39:16 <andythenorth_> even more trains
16:44:33 *** andythenorth_ has joined #openttd
16:52:31 *** andythenorth has joined #openttd
16:56:08 <nielsm> okay managed to rebase one patchset correctly!
17:09:32 *** dustinm` has joined #openttd
17:18:48 *** Thedarkb has joined #openttd
17:41:07 <nielsm> :( jenkins fails to merge my pull request, somehow
17:47:12 *** Wacko1976 has joined #openttd
17:47:24 <LordAro> nielsm: it's very particular about commit messages
17:47:38 <LordAro> and you appear to have merge conflict
17:48:01 <LordAro> (there's also a an svn commit in your branch, somehow)
17:48:28 <nielsm> ugh so I did make a mistake when rebasing after all?
17:48:41 <LordAro> well, you may have made a mistake :p
17:48:49 <LordAro> but just re-rebasing onto HEAD should do the job
17:49:03 <LordAro> the hardest bit will be squashing/moving/merging commits
17:49:18 <nielsm> what I did was rebase onto the commit in the new github repo that matched the one my old branch was based on
17:49:28 <nielsm> then afterwards merge master into the branch
17:49:40 <nielsm> should have rebased onto master instead?
17:51:07 <nielsm> that svn commit that somehow sneaks in is probably from something that happened during merge master
17:52:17 *** TheMask96 has joined #openttd
17:56:51 <LordAro> nielsm: yeah, no merge commits in this repo pls
18:00:55 <LordAro> nielsm: also, force pushing means you don't need to create new pull requests
18:01:36 <nielsm> trying to make sense of things locally atm :)
18:02:35 <nielsm> git UI is annoying, I'm pretty sure I understand the graph and objects stuff behind the scenes correctly, and I know how I want things to look, but I just don't know the commands to make it happen
18:07:25 <LordAro> git rebase master, probably
18:07:37 <LordAro> upstream/master or origin/master, depending on your checkout
18:07:53 <LordAro> and then git rebase -i master to reword/reorganise commits
18:14:19 *** frosch123 has joined #openttd
18:19:17 <nielsm> will have to do that again after changing all commit messages though...
18:20:00 <peter1138> Problem is lots of commits like "WIP" and "now working"
18:20:28 <LordAro> rebasing is love, rebasing is life
18:21:31 <peter1138> "untested" whew, brave.
18:22:30 <peter1138> Who needs reviews :p
18:23:40 <nielsm> commits like "update visual studio project files" should they be squashed into the previous? since that previous actually removes some source files
18:24:31 <peter1138> Don't commit project file changes.
18:24:42 <peter1138> Always edit source.list and run projects/generate
18:24:58 <peter1138> It's made a right mess
18:25:05 <LordAro> well you still need to commit the project file changes :p
18:25:27 <LordAro> ah, i haven't looked at the actual changes
18:25:42 <peter1138> LordAro, and source.list hasn't been touched
18:26:07 <nielsm> I am certainly updated source.list and running generate.vbs
18:26:15 <nielsm> not editing the VS project files by hand
18:26:28 <peter1138> What's with the xml change to every file then?
18:27:03 <peter1138> Maybe generate.vbs doesn't work properly. Hmm.
18:27:09 <nielsm> the source.list change is in the commit that deletes the old source files
18:27:46 <peter1138> So to answer your original question , yes :P
18:27:52 <nielsm> the generate_vs14[01].vcxproj change to xml declaration would seem to be something with UTF8 BOM
18:28:26 <nielsm> but why it changes the order of spanish.txt and spanish_MX.txt around everywhere, no clue
18:28:43 <peter1138> Thing is, you didn't review what was changed in the project file commit.
18:29:31 <nielsm> does the unix projects/generate script also update the VS project files?
18:29:46 <nielsm> I can run that in WSL then
18:30:06 <nielsm> marking that commit as 'edit' then
18:30:12 <peter1138> I don't actually know how to run generate.vbs
18:31:01 <peter1138> cscript wasn't obvious
18:31:10 <peter1138> Ok, Yeah, running that fucks all the proejct files up :p
18:31:22 <peter1138> I don't get the utf-8 bom change
18:31:29 <peter1138> but there is the reordering
18:31:46 <peter1138> Actually the reordering is the only thing.
18:31:52 *** Progman has joined #openttd
18:31:57 <peter1138> So something odd happened with yours for it to mess with the BOM
18:32:23 *** HerzogDeXtEr has joined #openttd
18:32:36 <peter1138> WIPs should be squashed together into something that works :p
18:33:12 <nielsm> then I have this "code style" where I fix up all previous changes to follow the style conventions.... spreading that out will be a major pain :(
18:33:26 <peter1138> SHould've got it right in the first place ;)
18:35:29 <peter1138> Also probably too many things going on in one PR
18:36:01 <nielsm> I tried separating things, I really did
18:36:28 <nielsm> and just kept getting all the changes intertwined
18:36:45 <peter1138> feature to change music set during gameplay
18:37:09 *** Thedarkb has joined #openttd
18:37:14 <peter1138> Nice separate PR, that one.
18:42:31 <nielsm> projects/generate run on WLS on my system also changes the order of the spanish/spanish_MX language files
18:43:08 <Rubidium> sounds like different sorting on file system level between OSes
18:43:48 <nielsm> the generate script should probably do an ASCII sort of filenames before adding them?
18:44:05 <nielsm> instead of depending on filesystem randomness
18:53:52 <nielsm> and I think it is portable across different unices/compatibles?
18:55:56 <glx> I guess it's about mx lang file in projects
18:57:47 <nielsm> made a PR for that generate script fix
18:58:43 <peter1138> Needs more than that.
18:58:57 <peter1138> With sort -d, spanish_MX comes before spanish for me
18:59:04 <peter1138> as it does in our project files
18:59:15 <peter1138> it's generate.vbs that does it the other way
19:00:44 <LordAro> shouldn't be using ls anyway
19:03:13 <glx> generate on mingw changes the order too
19:03:30 <glx> it's more related to the OS
19:03:54 <nielsm> yes it seems to be windows' FS driver that returns files in another order
19:04:24 <nielsm> and of course VBS doesn't have any built-in sorting
19:04:40 <nielsm> so enjoy writing a little bubblesort or whatever for it
19:05:00 <nielsm> of course, comparing the filenames exactly like 'sort -d' would
19:05:16 <peter1138> on linux it seems to sort by alphanumerics
19:05:40 <peter1138> so windows is right :p
19:06:19 <peter1138> spanish.txt comes after spanish_mx on linux cos t > m
19:06:46 <peter1138> english_au comes before english.txt on linux cos a < t
19:07:11 <peter1138> egnlish_us comes after english cos u > t
19:07:23 <peter1138> so it needs the opposite of sort -d
19:07:44 <glx> but it should sort based on what is before the . :)
19:07:50 <Wolf01> If you select a list of files and copy them on windows, it copies first the last one, then back from the first, and that causes a LOT of problems where you need to have files placed sequentially independent of the filename... like my car radio and some dvd players I use to see old photo albums on dvd
19:07:59 <peter1138> glx, yeah. windows does, linux doesn't.
19:08:27 <peter1138> Wolf01, yeah, screw those systems that can't understand filenames :(
19:08:45 <peter1138> Wolf01, there's even a tool specifically for reordering the file list on a dos partition
19:08:46 <Wolf01> They just read in the fs order
19:08:53 <glx> Wolf01: it's also funny when you drag a selection in vlc, you won't know the order in the playlist
19:09:05 <Wolf01> Ah, yes, I noticed that too
19:09:42 <Wolf01> I now use teracopy which does the copy in the right order and so no more problems :P
19:21:25 <peter1138> (Interesting that only spanish shows up differently, though)
19:22:08 <Wolf01> Oh, that, it is a pita every time I regenerate the projects :D
19:22:48 <peter1138> Thanks for reporting it as a bug :p
19:23:42 <peter1138> Actually there probably is one in there, a few years ago ;)
19:30:19 <glx> <peter1138> (Interesting that only spanish shows up differently, though) <-- because the other file that could cause problem is english.txt but it's skipped by generate
19:30:31 *** chomwitt has joined #openttd
19:31:32 <nielsm> looking at generate.vbs now, changing it to sort the files in "linux ordering" would be rather troublesome
19:31:48 <nielsm> since it loops directly on the FileSystemObject data rather on a copy of an array
19:32:21 <nielsm> so it'd have to first copy out all the filenames to a new array, then sort them (using a custom sorting function), then do the current stuff
19:34:45 <peter1138> ls -1 src/lang | sed 's/.txt//g' | sort | sed 's/$/.txt/g'
19:35:14 <nielsm> question is whether you can depend on windows always returning the files in this particular order ;)
19:35:41 <glx> windows tends to be conservative
19:35:44 <nielsm> (in which case changing the shell script to produce that ordering would be the easier)
19:36:17 <glx> because a lot of existing stuff relies on 'defined' behaviour
19:36:26 <peter1138> And also do it for the other ls.
19:46:10 *** supermop_work has joined #openttd
20:11:31 <TrueBrain_> w00p, lets see if I can make a 64bit Windows binary :D SO MUCH HYPE
20:11:54 *** TrueBrain_ is now known as TrueBrain
20:12:24 <TrueBrain> okay, so supybot is no longer really supported I see
20:14:52 <TrueBrain> 3 minute and 45 seconds
20:14:57 <TrueBrain> that is a lot faster than I expected tbfh
20:15:35 <TrueBrain> why is OpenTTD so hugggeeeeeee
20:16:44 <peter1138> Hmm, I've been building 32 bit all this time :p
20:16:53 <TrueBrain> last build it was only 13 MiB .. so why is this so much bigger ..
20:17:10 <peter1138> Is it a debug build?
20:17:31 <TrueBrain> did /p:Configuration=Release
20:17:44 <TrueBrain> I did use static libraries
20:17:50 <TrueBrain> and I see libpng.lib is 2 MiB ..
20:17:58 <TrueBrain> that is a huge number right there
20:18:44 <TrueBrain> ICU is also BIIGGGGG
20:20:24 <TrueBrain> yeah, those libraries should be a lot smaller
20:20:43 <peter1138> So I think you've got a debug build.
20:20:44 <TrueBrain> libpng.lib should be around 400 kib
20:21:03 <TrueBrain> if /p:Configuration=Release doesnt give a release, than yes
20:21:04 <peter1138> Hmm, or debug builds of the libs?
20:21:10 <TrueBrain> otherwise, I think my libs are wrong
20:21:35 <peter1138> libpng.lib is 497KB for me.
20:22:34 <TrueBrain> I think the libraries are static too
20:22:40 <LordAro> pretty sure i read somewhere recently that vcpkg gets 32bit libs by default
20:22:41 <TrueBrain> no clue why the executable is so big
20:22:54 <TrueBrain> nope; vcpkg does what it should do
20:22:57 <TrueBrain> nicely has the targets split
20:23:05 <TrueBrain> but also totally irrelevant; 32bit libraries dont work for a 64bit binary
20:23:06 <TrueBrain> and I have a binary
20:23:52 <TrueBrain> running a Headless windows; no clue how to check that
20:25:00 <TrueBrain> the executable runs fine; pretty smooth even for a VM it runs in
20:27:01 <TrueBrain> guess it is time to try the non-static vcpkg files
20:27:43 *** andythenorth has joined #openttd
20:29:04 <peter1138> God, VS is annoying to use.
20:29:43 <TrueBrain> still no word from AWS .. really think they put me on their ignore list :(
20:30:12 <andythenorth> how does int(foo/2) work then?
20:30:19 <Wolf01> TrueBrain: write at Bezos directly
20:30:20 <andythenorth> seems to round down
20:30:42 <TrueBrain> andythenorth: depending on the language, it tends to round down, yes
20:30:56 <TrueBrain> never do that in Python .... Python2 or Python3?
20:31:24 <TrueBrain> took you long enough :P
20:31:41 <andythenorth> does it reliably round down?
20:31:43 <TrueBrain> andythenorth: 3 / 2 -> 1.5 3 // 2 -> 1
20:31:45 <andythenorth> or should I floor it
20:31:55 <TrueBrain> just never ever use int() ....
20:31:55 <andythenorth> I need the result as an int
20:32:06 <TrueBrain> always returns an int, always rounded down
20:32:14 <Zuu> I actually read the topic when it was quite new, but took me some time to get here yes.
20:32:30 <TrueBrain> andythenorth: internally, // is called 'floordiv'
20:32:44 <TrueBrain> Zuu: the travels on the web is long :D
20:33:05 <TrueBrain> andythenorth: it is very nice of Python to have 2 operators for it :) int / int -> float, int // int -> int
20:33:10 <TrueBrain> makes so much code so much nicer
20:33:31 <peter1138> Hmm, my jenkins has blown up again :(
20:33:57 <TrueBrain> peter1138: no need for (float). what I wrote is correct. No casting. dont cast in Python. Never
20:34:08 <TrueBrain> (well, holds for Python3)
20:34:22 <TrueBrain> int / int -> float in Python 3 .. well, or int
20:34:43 <TrueBrain> if you are casting in Python, you are really doing it wrong
20:34:54 <TrueBrain> cast with (bytes), learn how to do decode/encode
20:35:02 <TrueBrain> cast with int/float, learn the rules :)
20:35:19 <peter1138> TrueBrain, yeah i was refering to C :p
20:35:30 <TrueBrain> yes, as we were talking about C. Sorry. My mistake.
20:35:50 <peter1138> As in "that's the not so much nicer way"
20:36:01 <peter1138> i may have optimised that bit away
20:36:20 <TrueBrain> optmizations give for bad conversations :D
20:36:24 <TrueBrain> we just all talk empty-space and dots
20:36:54 <TrueBrain> your smile is wrong :s
20:37:18 <andythenorth> 7 hours on trains today :P
20:37:27 <TrueBrain> I normally ride in trains
20:37:31 <TrueBrain> but I guess on can also be fun
20:37:34 <TrueBrain> did you go to India?
20:37:58 <TrueBrain> I just imagine andythenorth sitting on a train. His son walks in: DAD! GET OFF THAT TRAIN!
20:38:38 <TrueBrain> or maybe he is not walking on eggs, but he is walking on trains
20:38:48 <TrueBrain> does have another vibe to it ... "I have been walking on trains today"
20:39:05 <TrueBrain> even his connection lost it
20:39:20 <TrueBrain> nielsm: note that none of these commit messages will be accepted by the CI :)
20:39:23 *** andythenorth has joined #openttd
20:39:29 <LordAro> nielsm: see the list of approved prefixes
20:39:32 <nielsm> TrueBrain I think two of them will! :)
20:39:35 <TrueBrain> andythenorth: dont go offline when I am talking trash to you!
20:39:49 <andythenorth> sorry, Apple think it's good to turn off TCP/IP
20:39:52 <nielsm> I'm fixing things bit by bit, shuffling around code between commits is tricky to get right
20:39:54 <andythenorth> it saves battery
20:39:59 <TrueBrain> nielsm: in general we have to think about how to process your PR .. as there is no way we can keep working on it till the whole set gets an approval :)
20:40:05 <andythenorth> so they micro-sleep the network connection constantly
20:40:10 <andythenorth> which is crap for irc
20:40:35 <TrueBrain> yes, as the network negotiation and TCP negotiation are so cheap ..
20:40:51 <nielsm> if I separate the GUI changes and the playlist/title screen logic out into separate patchsets, is it fine to keep the win32 music driver and dos music support as a single PR?
20:40:57 <andythenorth> just waiting for a better alternative...
20:41:04 <nielsm> since those are very closely intertwined
20:41:26 <TrueBrain> yeah, there should be a balance of work vs reward :D
20:41:43 <TrueBrain> but I am mostly interested in the music driver .. as I really dont want to findout how to get dmusic working ...
20:41:47 <TrueBrain> its horrible ......
20:41:56 <TrueBrain> does win64 currently have midi support?
20:42:55 <andythenorth> oops forgot to 'make install'
20:43:04 <andythenorth> that's why my grf still is broken in-game :P
20:43:13 <TrueBrain> I had that many times today
20:43:18 <andythenorth> how do I even tie my shoelaces? :P
20:43:20 <TrueBrain> WHY DOESNT THIS WORK! Wait ... 'make' ....
20:43:25 * andythenorth has shoes without shoelaces
20:45:17 <TrueBrain> real men don't automate!
20:45:30 <nielsm> TrueBrain yes win64 has full midi support
20:45:41 <TrueBrain> then why does win32 need DirectX for?
20:45:42 <nielsm> (professional music software depends on that too)
20:45:49 <TrueBrain> no, I meant OpenTTD :D
20:46:05 <TrueBrain> OpenTTD win32 needs DirectX dmusic shit
20:46:12 <TrueBrain> OpenTTD win64 doesn't ... made me wonder if MIDI worked at all
20:46:15 <nielsm> the current two music drivers in openttd are more properly called directmusic and mci
20:46:19 <nielsm> rather than directmusic and win32
20:46:27 <nielsm> since the "win32" driver uses the ancient MCI interface
20:46:38 <TrueBrain> yes; for which is no win64 from what I understood
20:46:43 <TrueBrain> but so why do we still have it?
20:46:48 <TrueBrain> or does your patch fixes all that?
20:46:48 <nielsm> and for some reason using MCI MIDI on newer versions of windows has massive delays when loading files
20:47:04 <nielsm> 5-7 second pauses between tracks
20:47:11 <nielsm> (or when looping the title song)
20:47:12 <peter1138> Whatever is there works for me.
20:47:12 <Zuu> But it is easy to just remove music support and compile without dmusic. At least that is what I used to do. Now I have to see if I can get VS up and runnig to compille OpenTTD.
20:47:31 <TrueBrain> removing music is never an option :D
20:47:39 <TrueBrain> and I am more worried how I should do the official binaries :P
20:47:42 <peter1138> Zuu, that's not useful for a compile farm :)
20:47:48 <TrueBrain> I really really really do not want to look into dmusic :(
20:48:03 <nielsm> my new driver on windows basically implements a brand new low-level midi file parser and timing control to second it to the system midi output...
20:48:20 <TrueBrain> and removes dmusic?
20:48:43 <TrueBrain> so yeah, if you can split that off from the GUI changes and random README changes etc, would be lovely
20:48:58 <nielsm> dmusic has some neat ideas, honestly, but it was too late and no game developers wanted to use it anyway :D
20:49:33 <TrueBrain> especially if you can offer only the new MIDI driver .. makes reviewing a lot easier :)
20:49:52 <Zuu> peter1138: that put it in a different perspective yes.
20:50:08 <nielsm> just gonna make a force push again...
20:50:13 <TrueBrain> if supybot is no longer the standard .. what IRC bot is (in Python3)?
20:50:22 <TrueBrain> make all the force pushes you need nielsm :)
20:50:25 <TrueBrain> we can handle them! :)
20:51:43 <nielsm> 7f7c2d0 is the commit that contains the entire new music driver in the basic form, but there are some later fixes I haven't cleaned up yet
20:51:56 <nielsm> if you want to test it ;)
20:52:29 <TrueBrain> I wait for the cleanup :D
20:52:50 <TrueBrain> gives me some time to prepare the release binaries a bit more .. maybe I can try to build the PR with it so we can all test it on different systems :)
20:53:13 *** Thedarkb has joined #openttd
20:53:33 <nielsm> I'd appreciate if someone could try out the dos music support on non-windows systems now, before the cleanup is done
20:53:43 <nielsm> so I don't have to go back and fix that later again
20:54:01 <nielsm> it's really simple, just decodes the song, dumps to a temp file, and loads the temp file
20:54:24 <nielsm> but I'm not sure whether it's an appropriate location I write the temp file to, and some other details
20:54:27 <TrueBrain> cheaty :D But MIDI on non-windows always has been a bit of a problem :)
20:54:41 <nielsm> (I don't have any non-windows systems on hand)
20:56:13 <nielsm> oh yeah, also the fix for music looping I've made, can only work with the new win32 driver if the windows-TTD music files are used, since they depend on low level control of playback
20:56:34 <nielsm> (this includes the fix for "can't get there from here" having the long silence at the end)
20:56:39 <TrueBrain> I think it will take a bit of time and a few PRs before we get your stuff in :D
21:00:17 <nielsm> please call it version 2.0 then ;)
21:01:00 <nielsm> (more like 7.0 with all the other features also introduced since 1.0, but version numbers are meaningless anyway)
21:06:40 <glx> I used MidiOut for opendune, but music playing was a lot more low level there than in openttd
21:14:43 <TrueBrain> that is what OpenTTD needs ... auto-updating, and channels! So we can push test versions to a select few people
21:14:51 <TrueBrain> for that we need to get statistics how people use the game
21:14:53 <TrueBrain> so we can target them
21:15:01 <TrueBrain> of course than we can talk with companies to sell that data too
21:15:08 <TrueBrain> (I AM JOKING! Ffs ...)
21:15:09 <peter1138> Yeah...er... no not that.
21:15:18 <peter1138> Auto updating would be good.
21:15:28 <peter1138> I even had people request that.
21:15:37 <TrueBrain> yup; I am surprised that is not in yet :D
21:15:46 <TrueBrain> channels is also nice, for nightly/beta/release
21:15:58 <TrueBrain> statistics how peopleuse the game .... I rather burn the server to the ground
21:16:02 <Zuu> Then I could finally retire ottdau :-)
21:16:31 <peter1138> Hmm, wonder how to support it on non-Windows.
21:16:34 <TrueBrain> so why havent you added this to OpenTTD yet Zuu?! :P
21:16:37 *** ToffeeYogurtPots has joined #openttd
21:16:41 <TrueBrain> apt upgrade openttd? :P
21:16:42 <peter1138> Of course it's possible but package managers and shit.
21:16:42 <glx> so for now dmusic is disabled at compile time for win64, and win64 has music via mci
21:16:55 <Zuu> I don't relly have a compiler for it other than at work, but is to lazy to rewrite it.
21:17:02 <LordAro> some sort of notification is usually good enough
21:17:24 <peter1138> Nah, now that everyone is used to Steam auto updating everything...
21:17:33 <peter1138> (Let's not go on Steam though)
21:17:39 <LordAro> "a newer vsrsion is available at xyz, go get it"
21:18:15 * peter1138 attempts to find out how to make Jenkins parse gendarme reports.
21:18:25 <Zuu> TrueBrain: Good question. Maybe I got drowned to much into AIs and Game Scripts. :-)
21:20:34 * andythenorth would like to collect user data
21:20:50 <andythenorth> we can call it a fair trade for connecting people
21:21:50 <andythenorth> there is no sarcasm emoticon is there
21:21:57 <andythenorth> ^^ this is the kind of thing that came back to haunt Zuck
21:22:10 <andythenorth> except he maybe meant what he said 15 years ago :P
21:31:55 <TrueBrain> right, turns out that after a Plesk update, emails could no longer be recieved
21:32:58 <TrueBrain> and the emails telling me about it ... couldnt be sent :D
21:36:13 <TrueBrain> amazing how you dont really notice if you dont receive any emails .. it is like: hmm .. weird .. but what-evah
21:38:06 <andythenorth> email is the inverse of a good alerting system :P
21:40:17 *** gelignite has joined #openttd
21:42:50 <TrueBrain> now the binary is 4 MiB .. that seems too small
21:43:09 <TrueBrain> now it needs dlls etc
21:43:18 <TrueBrain> so either it is mega-fat .. or too-slim
21:44:43 <TrueBrain> error while loading shared libraries: ?: cannot open shared object file: No such file or directory
21:44:49 <TrueBrain> is ? not a valid file?!
21:46:07 <glx> static is easier, else we need to distribute the dlls
21:46:15 <TrueBrain> but static was 40+ MiB
21:46:21 <TrueBrain> which is a bit weird
21:47:04 <glx> I have 31,7MB for a debug x64 build
21:47:12 <TrueBrain> it was a release build ...
21:47:40 <glx> starting a release x64 build
21:49:04 *** Thedarkb-X40 has joined #openttd
21:49:56 <andythenorth> so we're not adding depot flip for articulated vehicles, right?
21:52:19 <nielsm> .... I _just_ now realized there's the files in media/baseset/ that are used to generate the bin/baseset/ files from with translations for descriptions
21:52:24 <nielsm> that's also a thing I'll have to fix
21:57:20 <andythenorth> we could allow flipping the bit if the vehicle permits it :P
21:57:30 <andythenorth> but not actually reversing the sprites in openttd
21:57:40 <andythenorth> although openttd doesn't seem to reverse the sprites for me anyway
21:57:47 <andythenorth> I had to handle it explicitly
21:58:25 <TrueBrain> glx: yeah, I think my issue is in the deps .. not using openttd-useful :)
22:00:06 <TrueBrain> I was trying to avoid using some random libraries :)
22:00:13 <TrueBrain> vcpkg is a much cleaner way to drag them in
22:00:21 <TrueBrain> (and they are updated!)
22:04:45 <glx> oh yes vcpkg looks nice, I should try it
22:06:03 <glx> and I like the integration option
22:08:47 <TrueBrain> yes; it isa very clean solution; not as dirty as our current :)
22:10:55 <peter1138> Wonder why it's bigger with them. Hmm.
22:10:56 <TrueBrain> ah; despite the configuration settings, it did build a debug after all
22:10:58 <TrueBrain> so what is wrong here ..
22:11:23 <TrueBrain> ah, because this round I couldnt type
22:11:26 <TrueBrain> explains why the linker fails
22:13:34 <TrueBrain> right, now it links again
22:13:51 <TrueBrain> people keep telling me I should do /MT and DEFAULTNOLIB .. no clue how to do that via msbuild :D
22:14:20 <TrueBrain> I see MT is set in the project file anyway
22:14:50 <TrueBrain> 32 MiB is the RELEASE binary (I double checked :P)
22:16:16 <peter1138> How big is the debug binary?
22:16:33 <TrueBrain> not tried to build that yet .. will try next
22:17:03 <glx> or we can compare your .lib sizes vs useful-6.0
22:17:18 <TrueBrain> I assume MSVC also does some optimizations, and prune unneeded stuff?
22:17:53 <orudge`> TrueBrain: I tried using vcpkg for ICU
22:17:58 <orudge`> but we use things in ICU that are no longer present
22:18:04 <TrueBrain> you have to downgrade to 57
22:18:15 *** ChanServ sets mode: +o orudge`
22:18:20 *** orudge` is now known as orudge
22:18:41 <TrueBrain> I wish LordAro hurried up with Pango .. than you no longer have to
22:18:47 <nielsm> the TTO theme sounds bad on the MS GM synth after having heard it on a real roland sc-55 :(
22:18:49 <TrueBrain> (as in 61, even in 59, Layout is removed)
22:19:04 <glx> pango is a pain with deps
22:19:40 <glx> but it's probably better than icu ;)
22:20:14 <orudge> Maybe I should dust off my ancient MP3/Vorbis music patches for OpenTTD (I think they're so ancient they're still in C) - give you a few more libraries to have to support, TrueBrain :D
22:20:47 <orudge> To be fair, when I had the time to play OpenTTD, I tended to play with the various FM recordings up on my web site in Winamp, and mute the OpenTTD MIDI music itself
22:21:27 <TrueBrain> hmm .. this is interesting .. I guess openttd-useful is build with static library, but dynamic crt
22:21:33 <TrueBrain> in vcpkg, static means both library and crt static
22:21:34 <nielsm> do like GZDoom/other modern doom ports and include a bunch of different synthesizers right in the program
22:21:45 <nielsm> afaik they also have some FM synth emulators and more built in
22:21:56 <peter1138> Yeah, including a synth would be a nice option.
22:21:56 <orudge> TrueBrain: no, pretty sure OpenTTD.exe is entirely static CRT, at least the version that gets built on the server
22:22:23 <orudge> A new option for the content downloader, your prefered MIDI synth :)
22:22:40 <orudge> I think the OpenTTD Android port basically includes a copy of timidity
22:22:59 <nielsm> note: if you add softsynth support I'd also have to reverse the FM music from the dos version!
22:23:20 <peter1138> Content Download: fluidsynth patches
22:23:33 <peter1138> Hmm, wonder if this is working.
22:24:10 <TrueBrain> orudge: not what I said at all :) I was talking about the libraries :)
22:24:10 <peter1138> Publishing gendarme results probably shouldn't take longer than actually running gendarme.
22:24:42 <peter1138> Does that mean you have copies of a static CRT for each static lib?
22:25:20 <orudge> TrueBrain: so the libraries are built to use a dynamic CRT, but OpenTTD is linked statically? Or something? Either way it sounds wrong :)
22:25:36 <TrueBrain> not helping, but tnx for the letters :)
22:25:56 <peter1138> Oh, I think I didn't save that palette sprite sheet :(
22:26:09 <peter1138> And my history seems to be poor these days
22:26:10 <glx> my release build is linked against msvcrt.dll
22:26:21 <orudge> glx: that would be a mingw32 build then
22:26:28 <TrueBrain> glx: does the installer install msvcrt?
22:26:37 <glx> no it's an msvc release x64 build
22:26:41 <TrueBrain> glx: which tool to use on windows to look to deps ?
22:27:03 <TrueBrain> I only have a PowerShell :D
22:27:07 <orudge> glx: and it's definitely msvcrt.dll, and not msvcr140.dll?
22:27:13 <TrueBrain> can you check the deps of the libpng.lib in openttd-uesful for me?
22:27:24 <orudge> since msvcrt.dll is the VC6 runtime library
22:27:47 <glx> it's the msvcrt.dll present in system32
22:28:21 <orudge> mingw uses msvcrt.dll, but you have to do some fun hacks to get MSVC 7+ to link against msvcrt.dll I believe
22:28:29 <TrueBrain> ugh, creating a debug build is painful ... vcpkg named the libs differently (and rightfully), but openttd-useful used a special way of writing them .. so they are not compatible :D
22:31:20 <Zuu> In the new git commit style, there is no attribution field. Should I omit the patch author when making a PR for an flysplay patch?
22:31:30 <TrueBrain> okay, building crt:dynamic, lib:static libraries .. see if that changes anything at all
22:31:42 <glx> and it doesn't work with .lib
22:31:51 <TrueBrain> Zuu: you mean the (patch by NN) at the end of commit messages?
22:32:02 <TrueBrain> Zuu: as I still do that. And mention the bug/patch
22:32:10 <TrueBrain> check on of the older commits from last week
22:32:29 <michi_cc> TrueBrain: The openttd-useful libs are also for a static CRT.
22:32:48 <TrueBrain> michi_cc: than why is libpng.lib for me 2MiB, and 500 KiB in openttd-useful?
22:33:57 <TrueBrain> (well, I more tried to type: do you have any clue why :D)
22:34:39 <michi_cc> Is vcpkg defaulting to link time code generation? In that case, the .lib contains intermediate code and not real code. I never enabled that because it is compiler version dependant.
22:35:19 <TrueBrain> I just see my resulting .exe is HUGE
22:35:28 <TrueBrain> so I was hoping I could blame the libraries
22:35:47 <TrueBrain> I wish we used cmake, than things would go a bit more smooth
22:36:07 <michi_cc> The release build could still include symbols in the exe.
22:36:58 <orudge> Visual Studio 2017 is best buddies with cmake now apparently, although I haven't really tried it much myself
22:37:12 <TrueBrain> any suggestions I could try, are welcome michi_cc :) I have no clue what your sentence implies for me :D
22:39:32 <TrueBrain> lol, vcpkg adds *.lib in the linker command
22:39:43 <michi_cc> Are you building the project files or something different. If yes, it has to be caused by the libraries, but I have no idea how vcpkg operates under to hood.
22:40:13 <TrueBrain> are you doing A or B. if yes, .. ? :D
22:40:24 <michi_cc> The openttd-useful libs were always carefully hand-crafted from the original sources :)
22:40:27 <TrueBrain> but I am doing 'msbuild projects\openttd_vs141.sln'
22:40:37 <TrueBrain> yes; hand-crafted, and unmaintained :)
22:41:46 <TrueBrain> I would much rather switch to something upstream :)
22:42:38 <michi_cc> Unless a recent vs2017 update made it produce huge binaries, I'd very much suspect the libs somehow. The sln contains all compiler options for the exe itself.
22:43:10 <TrueBrain> I compile with vs141, and I get warnings during linking of missing vs140.pdb
22:45:01 <michi_cc> A random older vs141 openttd.exe I have around is almost 32MB, and it is a DEBUG build. So double and tripple check that :)
22:45:12 <TrueBrain> did, did it again, and again
22:45:17 <TrueBrain> debug doesn't even compile :)
22:45:23 <TrueBrain> as I am too lazy to rename libs
22:45:52 <TrueBrain> and if you know a way to find out if a library is compiled with release/debug, I would love to hear it. vcpkg says they are release, but .. *shrug*
22:45:53 <TrueBrain> cannot confirm that
22:46:09 <TrueBrain> openttd.exe is also in objs/x64/Release/
22:46:25 <TrueBrain> so if it is not a release, it is not because I didnt yell it to be a release :)
22:47:33 <glx> if it's in x64/Release it's a release
22:47:57 <glx> our project files are correct for that :)
22:48:02 <TrueBrain> some optimization? debug symbols left in there?
22:48:15 <glx> debug stuff is in the pdb
22:49:09 <TrueBrain> the .pdb is slightly bigger than the .exe :)
22:49:22 <michi_cc> I don't think there's any way to see how a lib was build. Especially as you could put the Debug options into Release and vice versa to anger people :)
22:50:06 <michi_cc> glx: There is also CodeView debug symbols that are embedded in the exe/libs, but I can't see any reason why vcpkg would use them by default.
22:50:53 <TrueBrain> the linker command has a /debug in there .. not sure why, or what it does
22:52:37 <glx> opened libpng.lib in notepad++, I see x64\Release Library\pngwutil.obj and stuff like that in it
22:53:10 <glx> maybe you can check your vcpkg libpng.lib in a text editor
22:53:25 <TrueBrain> okay, the /debug just makes the pdb
22:54:19 <andythenorth> I should have PR-ed your repo eh :P
22:54:43 <TrueBrain> what is the powershell command to grep in an output ..
22:56:22 <glx> Get-Content .\doc.txt | Select-String -Pattern (Get-Content .\regex.txt)
22:56:55 *** sim-al2 has joined #openttd
22:57:55 <TrueBrain> hard to tell anything .. my paths are different, but duhh :D
22:58:53 <TrueBrain> ah, the compile flags are in the lib
22:59:09 <TrueBrain> -DWIN32 .. silly libpng :D
22:59:30 <TrueBrain> so yeah, sounds like a release lib :)
23:02:28 <glx> I can't find the flags in .lib
23:03:57 <glx> but openttd-useful is built with vs140 so maybe it's just vs141
23:04:06 <TrueBrain> glx: would you mind installing something like libpng with vcpkg, and see what that does to your binary size of OpenTTD?
23:04:06 <TrueBrain> if any change at all etc?
23:04:06 <TrueBrain> lol; now I get unresolved external symbols :D fdopen, strncpy, etc :D
23:04:06 <TrueBrain> so MT is correct for the origin set I tried ..
23:04:06 <TrueBrain> this is a bit weird
23:04:34 <TrueBrain> could you try locally by any chance?
23:07:07 <TrueBrain> libpng:x64-windows-static
23:07:12 <TrueBrain> to install the static versions
23:10:35 <TrueBrain> and I downloaded openttd-useful now, and put those libraries in there
23:10:38 <TrueBrain> see if that changes anything
23:11:41 <glx> I hate when a tool changes the font size ;)
23:12:38 <glx> and powershell is not smart enough to add the .exe when I just type vcpkg
23:13:03 <TrueBrain> okay, the include-files also changed ..
23:14:05 <TrueBrain> -R, nope; -Recursive, nope ... hmm .. -Recurse? YES!
23:14:06 <glx> A suitable version of git was not found (required v2.16.2). Downloading portable git v2.16.2...
23:14:33 <glx> ie the powershell opened via github
23:15:03 <andythenorth> SPRITE_STACK works on purchase menu, right?
23:15:41 <TrueBrain> trigger-happy-enter-key :(
23:16:56 <TrueBrain> cannot find sprintf
23:17:02 <glx> ah yes git-shell has 2.11.0
23:17:03 <TrueBrain> seems I cannot link against openttd-useful
23:17:31 <TrueBrain> should that work, with vs141?
23:17:37 <glx> ok installed, let's check the sizes
23:18:01 <TrueBrain> is anyone using VS2017? :D
23:18:03 *** supermop_work has joined #openttd
23:18:27 <glx> -a---- 26/04/2018 23:16 2027948 libpng16.lib
23:18:27 <glx> -a---- 26/04/2018 23:15 424940 zlib.lib
23:18:53 <TrueBrain> what does compiling against those do for your openttd.exe?
23:18:56 <TrueBrain> and what VS do you use?
23:20:02 <TrueBrain> well, that at least means you can confirm if it is vcpkg, or 2017 indeed :D
23:20:07 <TrueBrain> so far it seems to be vcpkg :)
23:20:21 <TrueBrain> but yeah .. from what I can tell, openttd-useful, latest, doesn't work with VS2017 ..
23:21:49 <orudge> I wasn't able to get it to work
23:22:01 <orudge> I've been through pretty much what you've been doing tonight
23:22:03 <TrueBrain> legacy_stdio_definitions.lib is needed :D
23:22:24 <orudge> Think I eventually got something building after disabling ICU and installing the other packages with vcpkg
23:23:48 <TrueBrain> I got it to compile; it is just huge when static
23:24:02 <TrueBrain> okay .. I did not install the 140 toolset
23:24:06 <TrueBrain> so I cannot try if that changes anything
23:24:29 <TrueBrain> I could install them; just takes for-ever ..
23:25:32 <TrueBrain> let me see .. what other things did I pick ..
23:25:57 <TrueBrain> .NET 4 (instead of 4.6)
23:28:46 *** berndj-blackout is now known as berndj
23:30:33 *** Stimrol has joined #openttd
23:35:24 <glx> rebuilt openttd and it's still 13.4MB
23:36:10 <TrueBrain> cool; so it is 2017 ..
23:36:16 <TrueBrain> Optimization is set to Full
23:36:22 <TrueBrain> what is the setting in the project for Size?
23:36:43 <glx> but I'm not sure it used the vcpkg build
23:37:01 <TrueBrain> remove the openttd-useful files :P (or rename them ;))
23:38:26 <TrueBrain> manual say /O1 and /O2, but those are invalid values
23:38:29 <Zuu> Oh... I forgot to add to the script changelog. I knew I forgot something. Now how do I fix this in git?
23:38:32 <TrueBrain> it is now Full .. which is also not in the documentation
23:38:54 <TrueBrain> Zuu: git commit --amend and git rebase -i upstream/master, depending
23:39:02 <TrueBrain> some really good guides about this stuff on the interwebz :)
23:39:11 <TrueBrain> amend works for only the last patch
23:39:15 <TrueBrain> rebase -i for any other
23:39:22 <TrueBrain> git push -f, to force push your update
23:39:24 <Zuu> I'm sure there is a solution. I just need to learn how.
23:40:04 <TrueBrain> and of course, git GUIs :)
23:40:37 <Zuu> I also have a paperback git book.
23:40:41 <TrueBrain> glx: any hints what the correct entry for Optimization could be? :D
23:41:07 <Zuu> hm.. no it was an e-book.
23:41:48 <glx> hmm can't find libpng.lib
23:41:59 <glx> so it doesn't use vcpkg one
23:42:21 <peter1138> I'm using 2017, but with the useful pack obvs
23:42:38 <TrueBrain> peter1138: doesn't link for me without modifications
23:42:49 <TrueBrain> sprintf is no longer available, only with legacy settings on it is
23:42:55 <TrueBrain> (and openttd-useful is compiled with them)
23:42:57 <peter1138> Odd, no issues here.
23:43:07 <TrueBrain> anyway, can either of you please tell me what /O1 is in the project file?
23:43:13 <TrueBrain> the Optimization entry?
23:43:48 <TrueBrain> fucking crap documentation :(
23:44:00 <TrueBrain> after guess 200111 I finally guessed right ...
23:44:54 <peter1138> I don't know what O1 is.
23:45:04 <TrueBrain> its in the dropdown on the UI :)
23:45:08 <TrueBrain> but I dont have an UI :(
23:46:28 <TrueBrain> good news about OSX: seems the new build fixes issues people having :D \o/ :)
23:49:35 <TrueBrain> the builds I can now make with the new docker
23:49:49 <peter1138> Are they published somewhere then?
23:50:38 <TrueBrain> okay, MinSpace doesnt change the binary size :P
23:51:41 <Zuu> If I did push my faulty commits, I should not use rebase? Instead I should make a new branch? (I have two commits and want to change the first one)
23:51:57 <TrueBrain> new branch means new PR, which is annoying :D
23:52:13 <TrueBrain> (I did already mention that above :P)
23:52:14 <Zuu> But git help page says I should not rewrite public history
23:52:21 <TrueBrain> yes; they say a lot of things :)
23:52:27 *** FLHerne has joined #openttd
23:52:29 <TrueBrain> and strictly seen, they are correct
23:52:36 <TrueBrain> just lets not assume someone forked your work :P
23:52:43 <TrueBrain> (otherwise they will be pissed I guess)
23:53:06 <TrueBrain> if there is only 1 commit in your PR you are fixing, you can just commit your fixes as new patches
23:53:09 <peter1138> Ish. It's not hard for them to rebase either.
23:53:11 <TrueBrain> but as you had 2 commits in your PR, you have to rebase
23:53:37 <TrueBrain> peter1138: seeing the rants I have gotten for forcing rebase of OpenTTD, I am not sure that is true :P
23:54:02 <peter1138> Yeah, I guess it's tricky if they've been merging bits all over the place.
23:54:28 <Zuu> Now ... this works (rebase -i upstream/master, with the new change as a third commit that I moved and squashed)
23:55:22 <TrueBrain> now a push will tell you: OMG YOU REWROTE HISTORY WHAT ARE YOU DOING
23:55:26 <TrueBrain> so you just tell him: -f :)
23:55:47 <TrueBrain> worked perfectly :)
23:55:55 <peter1138> TrueBrain, yeah, LC_ALL=C didn't solve anything.
23:56:05 <peter1138> Already tried that ;p
23:56:35 <TrueBrain> peter1138: it does solve any locale issue
23:56:39 <TrueBrain> all I was saying :P
23:56:47 <peter1138> Ok. It's not a locale issue :-)
23:57:08 <TrueBrain> okay, disabled tons of stuff in the MSVC project, binary stays HUGE
23:57:32 <Zuu> Oh and I think this whole OpenTTD moving to git is a good thing by the way.
23:57:32 <TrueBrain> it ... just .. doesnt .. want ... to .. shrink ...
23:58:00 <TrueBrain> Zuu: I think so too :D
23:58:07 <TrueBrain> peter1138: yeah .. well, it is not the size itself that worries me
23:58:11 <TrueBrain> just to huge jump of 100%
23:58:23 <TrueBrain> the Mac OS X is also this size
23:59:00 <peter1138> Using useful is out the question I guess.
23:59:11 <peter1138> Hmm, what's this other thing you used then? Maybe I can try it and see what I get.
23:59:28 <TrueBrain> supplied by Microsoft :D
23:59:43 <peter1138> What do I have to do to invoke that?
23:59:53 <peter1138> I'm too lazy to look anything up :p
continue to next day ⏵
