IRC logs for #openttd on OFTC at 2024-05-25
⏴ go to previous day
01:32:23 *** Wormnest has quit IRC (Quit: Leaving)
02:21:58 *** ChanServ sets mode: +v tokai
02:26:01 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
02:38:14 *** debdog has quit IRC (Ping timeout: 480 seconds)
04:27:01 *** jinks has quit IRC (Remote host closed the connection)
04:42:14 <DorpsGek> - Update: Translations from eints (by translators)
06:34:51 *** salut3585 has joined #openttd
06:35:24 <merni> Wait for one of the devs to review it
06:37:03 <merni> Depends on how much free time they have :)
06:38:07 <salut3585> ok, This is the Way
06:39:01 <salut3585> when he does the merge, will it be merged into one commit?
06:39:32 <merni> Depends on whether whoever merges it selects "Squash and merge" or not
06:40:05 <merni> But for reviewing it's easier if you split it into logical commits
06:40:44 <merni> (or any other method idk much about git)
06:42:12 <salut3585> can you watch the video in PR, do you think it will be useful?
06:42:42 <merni> It's not my opinion that matters anyway 😆
06:42:58 <merni> It sounds like a good idea based on the description though
06:43:21 <salut3585> Your opinion is important to me
06:45:26 <merni> Well, it seems to work from the video
06:45:34 <merni> How do you know where to click though
06:48:27 <salut3585> Click first button, select area
06:48:27 <salut3585> then click the second button and you can set the area where the selected area will be pasted.
06:50:54 <merni> Yes, but it seems like in the second step you don't need to select an area that's the same size as the copied area (which is a good thing)
06:51:21 <merni> So in that case, the pasted area is bigger than the area you copied
06:51:55 <merni> Hm, I should try the preview
06:52:25 <salut3585> yes the selected area will be pasted but you can just click or point the direction if you select the area to paste
06:52:25 <salut3585> it can rotate the area
06:54:33 <merni> Also, it is somewhat unexpected that this option is in the Landscaping menu but it only applies to rail tracks
06:54:52 <merni> If it is rail-specific it should be in the Railway Construction menu IMO
06:54:58 <merni> Though ideally it would cover other things too
06:55:09 <merni> And it seems to leave gaps where there are level crossings
06:55:47 <salut3585> it will spread to all roads and change the surface
06:58:25 <salut3585> if you want to rotate, just select an area in a different direction
07:11:47 <andythenorth> I have 2 commits in my brain branch that aren't in my feature branch
07:12:14 <andythenorth> (non-interactive) rebase fails with more merge conflicts than it's worth resolving
07:12:59 <andythenorth> but when I try to cherrypick either of the commits, cherrypick seems to think they're empty
07:13:15 <andythenorth> On branch v5-release-track
07:13:15 <andythenorth> You are currently cherry-picking commit 70fbc27a.
07:13:15 <andythenorth> (all conflicts fixed: run "git cherry-pick --continue")
07:13:15 <andythenorth> (use "git cherry-pick --skip" to skip this patch)
07:13:15 <andythenorth> (use "git cherry-pick --abort" to cancel the cherry-pick operation)
07:13:17 <andythenorth> nothing to commit, working tree clean
07:13:17 <andythenorth> The previous cherry-pick is now empty, possibly due to conflict resolution.
07:13:19 <andythenorth> If you wish to commit it anyway, use:
07:13:19 <andythenorth> git commit --allow-empty
07:13:21 <andythenorth> Otherwise, please use 'git cherry-pick --skip'
07:24:18 <andythenorth> unclear if it's because I've got a branch merge in history that appears to have duplicated a bunch of commits
07:24:26 <andythenorth> not going to try and clean that up, too risky
07:27:04 <peter1138> Life is easier when you don't merge.
07:28:16 *** tokai|noir has joined #openttd
07:28:16 *** ChanServ sets mode: +v tokai|noir
07:32:49 *** tokai has quit IRC (Ping timeout: 480 seconds)
07:33:09 <andythenorth> I can't even do a rebase -i to reword a commit
07:33:18 <andythenorth> this branch is terminally wounded
07:33:21 *** ChanServ sets mode: +v tokai
07:37:49 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
07:50:14 <andythenorth> maybe I can redo it from scratch, one commit at a time
07:58:17 <andythenorth> how TF do individual cherrypicks conflict so much with main
08:02:42 <andythenorth> rebuilding a branch one cp at a time 😛
08:07:59 *** tokai|noir has joined #openttd
08:08:00 *** ChanServ sets mode: +v tokai|noir
08:12:04 *** tokai has quit IRC (Ping timeout: 480 seconds)
08:12:22 *** ChanServ sets mode: +v tokai
08:16:14 *** tokai|noir has quit IRC (Ping timeout: 480 seconds)
08:16:39 *** tokai|noir has joined #openttd
08:16:39 *** ChanServ sets mode: +v tokai|noir
08:20:44 *** tokai has quit IRC (Ping timeout: 480 seconds)
08:22:24 <andythenorth> never understood what triggers a merge commit to be an actual commit
08:45:27 <truebrain> those bugs are the best bugs! (12715)
09:04:52 *** HerzogDeXtEr has joined #openttd
09:50:15 <kuhnovic> Slowly improving the YAPF implementation. Maybe one day it will make sense 😛
10:04:33 *** SigHunter has quit IRC (Remote host closed the connection)
10:05:35 *** SigHunter has joined #openttd
11:01:38 <peter1138> andythenorth: All merge commits are actual commits
11:13:19 <peter1138> Self-fixing issues are the best.
12:18:20 <andythenorth> we have a workflow at work where we merge feature branches frequently, but I never encounter merge commits
12:18:25 <andythenorth> possibly I just didn't notice them
12:32:04 *** SigHunter has quit IRC (Ping timeout: 480 seconds)
12:45:17 *** SigHunter has joined #openttd
13:42:29 <andythenorth> nearly fixed my branch 😛
13:42:37 <andythenorth> manual cherrypick of 370 commits not so bad
13:42:49 <andythenorth> been able to squash and clean
13:48:24 *** Flygon has quit IRC (Quit: A toaster's basically a soldering iron designed to toast bread)
13:49:31 <peter1138> ICQ has a month left.
13:59:27 <andythenorth> how do I get these 2 commits into my v5 release branch?
13:59:52 <andythenorth> cherry pick says they're empty (they were already cherry picked into the release branch)
14:00:17 <andythenorth> github claims it can do a PR and merge for them, so I did, but after the merge it still reports they're not in v5 release track
14:01:09 <peter1138> ```git checkout v5-release-track
14:01:09 <peter1138> git cherry-pick a57473b
14:01:09 <peter1138> git cherry-pick 70fbc27```
14:01:40 <andythenorth> yeah they're empty
14:01:52 <andythenorth> and I can force them as empty, but it's a zero diff
14:02:01 <andythenorth> and github still thinks they're behind when I do that
14:03:45 <andythenorth> not sure I care enough at this point 😛
14:03:56 <andythenorth> 4 hours of manual cherry pick, and I still can't rebase to main
14:03:59 <andythenorth> but at least it's cleaner
14:04:08 <peter1138> Those changes are already in v5-release-track
14:04:43 <peter1138> ec2bea236 is the jquery update
14:06:30 <peter1138> The polar fox changes are spread out in other commits.
14:09:00 <peter1138> You can't cherry pick the changes because the changes are already present. They don't conflict as they are the same, so they end up empty.
14:09:58 <andythenorth> but I can't seem to tell github to understand that
14:11:22 <peter1138> What operation are you trying to do?
14:12:18 <andythenorth> make github UI stop telling me the v5-release branch is behind
14:12:25 <andythenorth> hmm I think I just deleted all my work
14:12:44 <andythenorth> I tried merging main into v5-release-track and everything I did today is now gone
14:13:00 <andythenorth> that's interesting
14:14:01 <andythenorth> yeah that's deleted local branches
14:14:13 <andythenorth> since when does a simple merge delete local feature branches?
14:15:25 <peter1138> main is a completely different tree to v5-release?
14:15:28 <andythenorth> branch state before and after 'git merge main' to v5-release-track
14:15:46 <andythenorth> maybe I need a tree visualiser
14:15:51 <peter1138> They diverge in Oct 2022
14:17:21 <peter1138> But then again, those jquery and polar fox updates are the only commits to main since they diverge.
14:17:27 <andythenorth> the merge of main deleted this branch locally `v5-release-track.old2` but it's present on origin
14:17:39 <andythenorth> I have never seen a merge do destructive branch removal before
14:17:53 <peter1138> It didn't delete it.
14:18:28 <andythenorth> maybe the branch names are in some git reflog that the merge to main has changed
14:18:36 <_glx_> git kinda expect you know what you are doing 😉
14:19:01 <peter1138> git expects you to know what you're doing, but unless you do something bad, it won't just delete branches.
14:19:01 <andythenorth> I know I don't know, I mean 😛
14:19:17 <peter1138> If you're using the reflog then you're generally doing it wrong.
14:19:36 <_glx_> but indeed cherry picking a commit already in the branch typically is blank
14:19:41 <peter1138> Do you want to make main be v5-release-track?
14:20:03 <andythenorth> I want github UI to stop being angry about v5-release-track being behind main
14:20:10 <andythenorth> so I can make a v5 beta release
14:21:21 <_glx_> branches often diverge for a valid reason
14:21:24 *** michi_cc has joined #openttd
14:21:24 <michi_cc> If two branches aren't completely identical, it will always says x ahead/behind.
14:21:26 <peter1138> You could just do `git merge main v5-release-track` if you want to make it up to date with main.
14:22:23 <andythenorth> that's what caused it to delete the branch locally
14:22:25 <peter1138> It won't actually add any changes as all the changes in main are already in v5-release-branch.
14:22:45 <peter1138> I just typed exactly that, and it did exactly that.
14:23:23 <andythenorth> is that tree view in your editor?
14:23:43 <andythenorth> I never found one in github
14:23:47 <andythenorth> bitbucket has one
14:24:04 <peter1138> This is Git Graph in VS Code.
14:24:27 <andythenorth> if you fetch now, what does it show?
14:24:35 <andythenorth> as github is now reporting what I want it to
14:25:52 <peter1138> Looks like you rebased on main instead of merging.
14:30:05 <andythenorth> nah there's no succesful rebase on main, just failed ones
14:30:18 <andythenorth> I did a merge of main to v5-release-track again I think
14:30:31 <peter1138> You rebased v5-release-track on top of main.
14:32:24 <andythenorth> not sure how that happened
14:32:46 <andythenorth> rebase and merge
14:32:47 <peter1138> `git rebase main v5-release-track` would do it.
14:33:29 <_glx_> except when working with PRs you want to avoid github UI
14:33:32 <andythenorth> pray and spray commands
14:34:52 <andythenorth> at the cost of some self-inflicted pain
14:34:55 <andythenorth> the history is now better
15:05:58 <merni> time to fast forward a game for 12 years to check a news message :)
15:11:18 <merni> Huh, the warning doesn't even show in 14.0 for a test game... there must be something else that triggers it apart from just becoming old??
15:16:12 <merni> Ahh, autorenew is on in the settings...
15:34:11 *** Wormnest has joined #openttd
15:43:48 <peter1138> Uh oh. Microsoft has its eyes on Valve.
15:51:46 *** dropboxkenshiro has joined #openttd
15:51:48 <merni> as long as it isn't hands it's fine I guess
16:15:08 <andythenorth> Unpicking branch merges by cherrypick from scratch
16:15:09 <andythenorth> And squashing some commits
16:15:10 <andythenorth> If the .grf matches at the end, “probably fine”
16:43:26 *** SigHunter has joined #openttd
18:02:39 <andythenorth> new laptop time?
18:02:43 <andythenorth> how much RAM do I need really?
18:08:46 <truebrain> our webservices can't run on 16GB. At least 32GB, sorry.
18:13:25 <audigex> Depends what you do with your laptop, but I wouldn’t buy less than 16GB in 2024 for sure
18:13:37 <audigex> And if you want it to last I’d be saying 32GB
18:17:01 <andythenorth> macs have an irritating habit where coreaudio crackles if there's not enough physical memory free
18:20:17 <andythenorth> 19% faster in single core, nmlc is single-core
18:20:52 <andythenorth> current laptop is m1 pro
18:22:50 <andythenorth> pff no m3 pro on refurb
18:23:08 <andythenorth> m3 max for £3k, but it's no faster than the m3 pro in single core
18:26:57 <merni> have you considered a windows laptop? :p
18:27:51 <andythenorth> yes, I seriously looked at HP laptops a few years ago, when Apple couldn't make keyboards, and OpenTTD performance on macs was really struggling
18:28:06 <andythenorth> both of those issues are fixed
19:27:00 <kuhnovic> Geez, that YAPF rail cache system is an enigma
19:27:58 <kuhnovic> And I doubt it makes any difference in performance tbh
19:33:29 <kuhnovic> Does anyone know anything about that caching system? Or it is a case of lost knowledge?
19:34:54 <_jgr_> kuhnovic: I wouldn't assume that, TBH
19:35:13 <kuhnovic> I'm not, that's why I'm trying to measure it 🙂
19:39:14 <Eddi|zuHause> rail cache is ... complicated...
19:39:42 <Eddi|zuHause> like, you can only cache paths that are far away, when the dynamic signal/reservation penalties aren't applied
19:40:09 <Eddi|zuHause> and the same segment can still have different weights when you have an electric and a diesel train
19:40:49 <Eddi|zuHause> i don't remember if anyone ever did performance test. but i assume someone did...
19:47:59 <kuhnovic> In earlier measurements I have seen that cost calculation, which is the part that's being cached, is indeed quite a large part of the total pathfinder CPU time (at least for trains, which has the caching). But if I enable debug level yapf=3 I do see sporadic hits with large cache hit percentages (most are in the 30-70% range), but most of them are just 0%. And a crude hold-control-to-disable-
19:47:59 <kuhnovic> caching while running in fast forward didn't seem to have any effect on the time spent in calculating train ticks.
21:46:43 <peter1138> How dare we open files based on user input
22:12:46 *** keikoz has quit IRC (Ping timeout: 480 seconds)
22:54:51 <yiffgirl> i am not a c++ developer so if there are things that should be obvious that i'm missing, that's why
22:55:36 <yiffgirl> my background is mostly d
23:06:25 *** Wolf01 has quit IRC (Quit: Once again the world is quick to bury me.)
23:40:52 *** HerzogDeXtEr has quit IRC (Read error: Connection reset by peer)
continue to next day ⏵
