IRCloggy #git 2021-10-01

Provider of IRC logs since 2005.

WARNING: As Freenode became unjoinable and lost all warnings in topics, we cannot log channels on Freenode anymore.

	2021-10-01
	← zopsi left	00:00
	← ninjin left	00:00
	→ ninjin joined	00:00
	→ zopsi joined	00:00
	← pyeveryt_ left	00:01
	← jkordani left	00:02
	← Yruama left	00:14
	← christall left	00:15
	→ stats4647 joined	00:19
	← stats4647 left	00:23
	→ chenqisu1 joined	00:25
	chenqisu1 → chenqisu12	00:25
	→ christall joined	00:43
	← tejr left	00:45
	→ roadie`` joined	01:00
	← christall left	01:09
	← durham left	01:14
	→ durham_ joined	01:14
	← roadie`` left	01:14
	← Murr left	01:18
	→ Murr joined	01:19
	→ wender joined	01:19
	← EvilDMP left	01:20
	← computeiro left	01:21
	→ gast0n joined	01:21
	→ ferdna joined	01:21
	→ jimklimov joined	01:22
	→ computeiro joined	01:24
	← wender left	01:26
	← madewokherd left	01:30
	→ emf_ joined	01:32
	← emf left	01:33
	→ madewokherd joined	01:36
	← emf_ left	01:36
	← ChmEarl left	01:40
	← hqdruxn08 left	01:43
	→ hqdruxn08 joined	01:43
	← chenqisu12 left	01:44
	← onizu left	01:44
	→ cryptonector joined	01:44
	→ chenqisu1 joined	01:44
	← chenqisu1 left	01:45
cryptonector	why does this not work: git push upstream HEAD:new-branch-name; but this does: git checkout -b new-branch-name && git push upstream HEAD:new-branch-name ?	01:45
	→ chenqisu1 joined	01:45
cryptonector	I don't understand why I should have to have a local branch name in order to be able to create that in a remote	01:45
	chenqisu1 → chenqisu12	01:46
cryptonector	it's... mildly infuriating :-/ -- I've been making do for a long time, but finally I've decided to ask	01:46
	→ ninjin- joined	01:49
	← ninjin left	01:50
	← cdown left	01:51
	← xx left	01:52
	← elkalamar left	01:54
	← BSaboia left	01:54
	← jimklimov left	01:57
	→ roadie`` joined	02:01
	← Xaldafax left	02:03
	← roadie`` left	02:05
	← durham_ left	02:11
	→ durham joined	02:13
	→ hbautista__ joined	02:13
	← alzgh left	02:22
	→ Samian joined	02:23
	← skapata left	02:27
	← nyah left	02:28
	← betelgeuse left	02:28
	← Samian left	02:30
	← chenqisu12 left	02:31
	→ chenqisu12 joined	02:32
	→ pyeveryt_ joined	02:32
	← pyeveryt_ left	02:37
	← ColdKeyboard left	02:39
	→ jimklimov joined	02:41
	← durham left	02:41
	→ durham joined	02:42
	→ antone joined	02:44
	→ Samian joined	02:52
	← junktext left	02:56
	→ roadie`` joined	03:00
	← FinnElija left	03:01
	← Samian left	03:02
	→ FinnElija joined	03:03
gnoo	cryptonector: it should work with git push -u	03:04
	← roadie`` left	03:05
	→ christall joined	03:06
	← antone left	03:09
	→ antone joined	03:09
	→ roadie`` joined	03:15
	→ ttree joined	03:24
	← dermoth left	03:26
	← roadie`` left	03:29
	→ JibStyle joined	03:32
	→ dermoth joined	03:39
	← keypusher left	03:40
	← christall left	03:44
	← zebrag left	03:45
	← stevenix1 left	03:50
	← durham left	03:53
	→ durham joined	03:55
	← adlternative left	03:55
	→ adlternative joined	03:55
	← antone left	03:56
	→ saroy joined	03:56
	← nightstrike left	03:56
	→ wender joined	04:04
	← bket left	04:05
	← computeiro left	04:06
elibrokeit	cryptonector: as documented by man git push -- it cannot infer what you expect new-branch-name to be, if HEAD doesn't unambiguously resolve to a ref beginning with refs/heads/ or refs/tags/ and new-branch-name doesn't exist on the remote	04:06
gitinfo	cryptonector: the git-push manpage is available at https://gitirc.eu/git-push.html	04:06
	→ bket joined	04:07
elibrokeit	however, if HEAD is a branch, even if that branch is, say, master, you can push HEAD:new-branch-name and it can successfully determine that you mean to push a branch, so it treats new-branch-name as a branch	04:07
	→ computeiro joined	04:10
elibrokeit	gnoo: what does --set-upstream have to do with this?	04:10
	← The_Blode left	04:11
	← wender left	04:11
elibrokeit	cryptonector: incidentally, as I do use detached heads fairly frequently for this sort of thing, I sort of got resigned to specifying git push myfork HEAD^:refs/heads/new-branch-name	04:12
	→ dayday joined	04:14
gnoo	elibrokeit: i thought the question was about different branch names locally and remotely (seeing checkout -b new-branch-name)	04:14
elibrokeit	nope, you can push those just fine :p	04:14
	the question stated "why does push fail", not "why doesn't it set up so that push later works without specifying options"	04:15
	→ The_Blode joined	04:16
dayday	Hi! Curious about some basic workflow stuff here. I got this repo that only I use (commandline), so instead of committing new features and fixes incrementally like I would if I was working with a team, I just rebase back and forth between my main commits like "Add Feature A", "Add Feature B" etc.	04:23
	Unfortunately I often run into merge conflicts (duh), and sometimes I get sloppy and miss a spot, causing commits to contain code that isn't theirs. So I have to rebase, re-rebase etc. -- it always works out in the end but it's extremely time-consuming, and I have without doubt spent more time on my Git history than on my project itself :-)	04:23
	Are there any ways to simplify this workflow? Maybe some UI tool I managed to miss out on?	04:23
elibrokeit	dayday: what does that even mean "rebase back and forth between my main commits"?	04:25
dayday	Say I want to fix something in commit "Add Feature A", then I just rebase back to that commit and apply the fix there and amend the commit. As opposed to making an entirely new commit ie. "Fix this or that in Feature A". Hope that made sense!	04:27
	→ roadie`` joined	04:28
elibrokeit	ah, so basically the same as if you did git commit --fixup=feature-A-commit	04:28
dayday	Really???	04:28
elibrokeit	but you're asking if there is advice on simplifying the process of autosquashing the commits together	04:28
dayday	Oh man :-)	04:28
elibrokeit	git commit --fixup creates a new commit with a magic oneline description beginning with "fixup!"	04:29
	and that tells git rebase -i --autosquash how to automatically sort and mark the commits for squash/fixup	04:29
	actually resolving conflicts is still going to be a challenge, sadly :(	04:30
	← vishal left	04:30
elibrokeit	however -- you do not need to keep going back and forth to amend old commits, you can commit stuff piece by piece and then glue it back together after the fact	04:31
	e.g. typically I don't want to checkout an old commit just so I can work on fixing it up	04:31
	I fix it on the latest commit, commit --fixup, and immediately rebase --autosquash to fold it into the commit it belongs in	04:32
	→ jazzy2 joined	04:32
dayday	Very cool :-) Thank you so much elibrokeit	04:32
	→ vishal joined	04:34
	← jazzy left	04:34
dayday	I've been doing it wrong this whole time! This fixup stuff is brilliant!	04:37
elibrokeit	it is pretty handy :) especially since it means you can now use git add -p to add only part of the current set of edits to an old commit	04:38
	rather than, idk, stashing everything, copying files around, re-creating changes after checking out an old commit, and G-d knows what else	04:38
	← hbautista__ left	04:40
	→ ThorMojito1 joined	04:44
	→ emf joined	04:45
dayday	hehe	04:45
	That's pretty much what I've been doing, yeah :-)	04:46
	← emf left	04:46
dayday	Boy am I glad I joined #git and ran into you tonight elibrokeit!	04:46
elibrokeit	happy to help :)	04:46
dayday	:D	04:47
	← ThorMojito left	04:47
	→ rsrx joined	04:48
cryptonector	gno, elibrokeit: thanks!	04:49
	elibrokeit: yes, I work in detached head state a lot	04:50
	git push myfork HEAD^:refs/heads/new-branch-name -- I'll have to remember it	04:50
elibrokeit	great minds think alike, eh	04:50
cryptonector	elibrokeit: I mean, branches are crutches often	04:51
	→ madewokherd` joined	04:53
	← YuGiOhJCJ left	04:55
elibrokeit	cryptonector: I mainly do this for pushing individual changesets to github	04:55
	← madewokherd left	04:55
elibrokeit	as PR branches	04:55
cryptonector	right, me too	04:56
elibrokeit	elsewhere I might just use one branch called "patchqueue" and send patches via git send-email	04:56
cryptonector	for example, if I'm rebasing an OpenJDK branch onto a newer OpenJDK, then I'll get into detached head state, git rebase --onto, and when I'm gone fixing conflicts I'll create a branch and push it	04:57
	it's just, I might want to push to a new remote branch and not keep a local one	04:57
	so, ok, you taught me something	04:57
elibrokeit	I wouldn't be opposed to git adding an option to always assume if it cannot figure out what you are trying to push, assume it is a branch. But it's not really up to me...	04:59
cryptonector	I'd like that; anyways, gtg, thanks	04:59
elibrokeit	(this ambiguity is essentially because "well technically you could want to push it as a lightweight tag")	04:59
	elibrokeit skeptical	05:00
	→ emf joined	05:01
	← ferdna left	05:02
	jazzy2 → jazzy	05:02
	→ nightstrike joined	05:14
nightstrike	Timvde: what is autostash?	05:15
	(sorry, my client disconnected a few hours ago)	05:15
ikke	option to automatically stash changes when you rebabse and apply the stash again after rebase	05:15
	→ emf_ joined	05:16
	← emf_ left	05:17
	← emf left	05:18
	← roadie`` left	05:21
	← chenqisu12 left	05:26
	← davit left	05:28
	← pulse left	05:33
	→ vysn joined	05:34
	→ wender joined	05:34
	← thiago left	05:34
	→ Guest34 joined	05:35
	← computeiro left	05:36
	← gast0n left	05:37
	→ computeiro joined	05:39
	→ christall joined	05:41
	→ davit joined	05:41
	← wender left	05:41
nightstrike	that's convenient	05:46
	→ ThomasD13 joined	05:46
	← rsrx left	05:48
	→ emf joined	05:51
	→ cvmn joined	05:51
	← Guest34 left	05:54
	→ roadie`` joined	06:00
	← thekingofbandit left	06:03
	→ ouzel8___ joined	06:03
	→ thelounge66195 joined	06:04
	← birkoff left	06:06
	← lucerne left	06:06
	← llh left	06:06
	← LiENUS left	06:06
	← shailangsa left	06:06
	← avarab left	06:06
	← wolfshappen left	06:06
	← starfarer left	06:06
	← magic_ninja left	06:06
	← nivag left	06:06
	← thelounge6619 left	06:06
	← ouzel8_ left	06:06
	← wacko left	06:06
	→ Fabricio20_ joined	06:06
	→ wacko_ joined	06:06
	→ vikonen2 joined	06:06
	→ raffaele_ joined	06:06
	← JibStyle left	06:06
	← rtjure left	06:06
	← acidtonic_ left	06:06
	← fjmorazan left	06:06
	← NeatNit left	06:06
	vikonen2 → vikonen	06:07
	← bodiccea_ left	06:07
	← Fabricio20 left	06:07
	→ ilogger2 joined	06:25
	→ wagle_ joined	06:26
	→ shailangsa_ joined	06:26
	→ roadie`` joined	06:26
miikak	elibrokeit: I'm building it in docker container on same machine	06:26
	→ Gustavo6046 joined	06:26
miikak	elibrokeit: can I somehow pass it as a flag to make?	06:26
	→ kenanmarasli joined	06:27
	→ gxt_ joined	06:28
	→ lessless joined	06:29
elibrokeit	docker is by definition not "the same machine"	06:30
	it is a different machine, perhaps with a newer version of glibc ;)	06:31
	→ Vonter joined	06:32
	→ aniruddha joined	06:32
miikak	elibrokeit: yea, on the host I have 2.17 and in container I have 2.28	06:32
	→ xlei joined	06:32
	→ rfuentess joined	06:35
	→ troyt joined	06:36
	→ SpacePlod joined	06:36
	← shailangsa_ left	06:39
	→ cvmn joined	06:41
	→ AnapodoPsalidi joined	06:47
	→ wender joined	06:49
	← cvmn left	06:49
	→ jast joined	06:49
	→ pieguy128_ joined	06:49
	→ glider_ joined	06:50
	→ ChemicalRascal joined	06:51
	→ mikeputnam joined	06:52
	→ Erisa joined	06:53
	→ ChemicalRascal_ joined	06:54
	→ Dragoon joined	06:54
	→ computeiro joined	06:54
	→ theoceaniscool joined	06:55
	← ChemicalRascal_ left	06:56
	← wender left	06:56
	← ChemicalRascal left	06:56
	→ ThorMojito joined	06:56
	→ wender joined	06:59
	→ shailangsa joined	07:00
	← computeiro left	07:01
	→ diverdude joined	07:02
	← Vonter left	07:03
	→ Vonter joined	07:03
	→ computeiro joined	07:05
	→ emf joined	07:05
	← wender left	07:06
	← emf left	07:07
	→ vysn joined	07:10
	→ mcfrd joined	07:11
	→ vikonen4 joined	07:12
	→ matthewcroughan joined	07:12
	→ SuperLag joined	07:12
	vikonen4 → vikonen	07:12
	→ blmt joined	07:13
	→ qsx joined	07:13
	→ avar joined	07:14
	→ mika joined	07:14
	mcfrd → mcfrdy	07:14
	→ henk joined	07:15
	→ Assault joined	07:17
	→ AsenMx joined	07:18
Assault	hi. I already made some commits to develop branch (but did not push the commits to remote). Now realised I should have created a new branch for the commits. Is it possible to somehow move the commits to new branch at this point?	07:18
	→ limon_ joined	07:18
	→ michele_ joined	07:18
	→ i8c joined	07:18
	→ Murr joined	07:18
	→ matthewcroughan_ joined	07:18
	→ blaisebool_ joined	07:19
	→ tmz joined	07:19
	→ OMGOMG_ joined	07:19
	→ diverdud1 joined	07:20
	← matthewcroughan left	07:21
	← diverdude left	07:21
	← AsenMx left	07:21
	← qsx left	07:21
	→ qsx joined	07:21
	→ t3nj1n_ joined	07:21
	→ dskull joined	07:22
	→ Masklin_Gurder joined	07:22
	← jab416171 left	07:22
	→ berndj joined	07:22
	→ barlas joined	07:23
	→ vancz joined	07:23
	→ elkalamar joined	07:23
	→ waldeck_ joined	07:24
	→ eggbean_ joined	07:24
	→ de-vri-es joined	07:25
	→ jab416171 joined	07:25
	→ xkr47_ joined	07:26
	xkr47_ → xkr47	07:27
	→ EPic_ joined	07:27
rudi_s	Assault: Sure, just create a new branch with `git branch new-branch` and then "remove" the commits from the old branch with git reset --hard origin/develop or similar. NOTE: The --hard will remove uncommitted changes in your working tree. So only do this if `git status` shows no pending changes.	07:27
miikak	Assault: I have done it as follows: reset with --soft flag, checkout new branch and commit changes there	07:28
	→ jim87 joined	07:29
	→ Aleksejs_Home joined	07:31
	← blaisebool_ left	07:31
	→ nomicon_ joined	07:32
	→ blaisebool joined	07:32
	→ tomek_ joined	07:32
	→ Henry151_ joined	07:33
	→ phylaz joined	07:34
	← vysn left	07:34
	→ emf joined	07:35
	→ vysn joined	07:35
	→ christall joined	07:35
	→ pyeverything joined	07:36
	→ TomyWork joined	07:37
	→ wender joined	07:39
	← christall left	07:40
	← pyeverything left	07:41
	← computeiro left	07:41
	→ rsx joined	07:42
miikak	hm	07:43
	→ lemontre1 joined	07:44
	→ stoned joined	07:44
	→ computeiro joined	07:45
	stoned → Hash	07:45
	→ justache joined	07:45
	→ SpeakerToMeat joined	07:46
	→ robertparkerx_ joined	07:46
	← wender left	07:46
	→ velix joined	07:46
	→ Industrial[m] joined	07:47
	→ ericnoan joined	07:47
	→ themill joined	07:47
	→ wender joined	07:49
	→ Optimus joined	07:49
	→ drogas joined	07:49
	→ swistak joined	07:49
	← computeiro left	07:51
	→ af joined	07:55
	← theoceaniscool left	07:57
	→ aidalgol joined	07:57
	→ EvilDMP joined	07:58
	← Vonter left	07:59
	→ computeiro joined	08:00
	→ emf_ joined	08:00
	→ Vonter joined	08:00
	→ alkino joined	08:00
	← wender left	08:01
	→ darkstardevx joined	08:02
	← af left	08:02
	→ otisolsen70 joined	08:03
	← emf left	08:04
	→ emf joined	08:04
Assault	got it, thanks!	08:06
	← emf_ left	08:06
	→ af joined	08:08
	→ Guest1 joined	08:08
Guest1	how do i revert to the previous commit as my master?	08:08
	← vysn left	08:11
Guest1	how do i under a hard reset?	08:14
	*undo	08:14
	← alkino left	08:16
	→ Strom- joined	08:16
	→ alkino joined	08:17
	← blaisebool left	08:17
osse	git reseting back to where you started from	08:18
	you can use the reflog to see where that is if you need to	08:18
Guest1	did i just delete a months work of work?	08:18
osse	if it was committed, no.	08:18
Guest1	this was the command i ran: git reset --hard master~1	08:19
osse	were you on master at the time?	08:19
Guest1	i wasn't	08:19
	i was on another branch, but i had committed to it	08:20
osse	do 'git reflog HEAD' and find the commit you were on. Unless you've done some other stuff in the meantime it should be HEAD@{1}	08:20
	→ soifou joined	08:20
osse	or you can do 'git reflog branchnameyoureactuallyon'	08:20
Guest1	do i checkout one of the listed HEADs?	08:21
osse	no	08:22
	you git reset --hard that-thing	08:22
	→ cluelessperson joined	08:24
Guest1	want to make sure i get this right, so this is what git reflog shows:	08:24
	aaaaaaa (HEAD -> feature/skeletons) HEAD@{0}: checkout: moving from master to feature/skeletons	08:24
	bbbbbbb (master) HEAD@{1}: checkout: moving from feature/skeletons to master	08:24
	aaaaaaa (HEAD -> feature/skeletons) HEAD@{2}: reset: moving to master~1	08:24
	ccccccc HEAD@{3}: Branch: renamed refs/heads/feature/refactor to refs/heads/feature/skeletons	08:24
	ccccccc HEAD@{5}: commit: started skeletons, not working	08:24
	so i run 'git reset --hard HEAD@{0}'?	08:24
	→ rfuentess_ joined	08:25
	→ jimklimov1 joined	08:26
	← rfuentess left	08:26
osse	(by the way, it would make more sense to censor the branch names than the sha1s :P)	08:26
	→ jancoow2 joined	08:26
	→ LewisCawte joined	08:26
	→ itok joined	08:26
	→ hqdruxn08_ joined	08:26
	LewisCawte → Lcawte	08:27
	→ brw joined	08:27
	→ fxrs_ joined	08:27
	→ pizdets_ joined	08:28
	→ sandipan joined	08:28
	→ ztrawhcse joined	08:28
osse	Anyway, we can see that aaaaaaa (aka. HEAD@{2}) is the commit that you reset to, and the commit immediately before that is ccccccc (aka. HEAD@{3})	08:28
	just to make sure, you can 'git reflog feature/skeletons' (i assume that's the branch you're on) and verify that feature/skeletons@{1} is the same sha1 as HEAD@{3} from that output	08:28
	jancoow2 → jancoow	08:28
osse	that's the sha1 you'll want to reset --hard to, to get back to where you started	08:28
Guest1	osse: next time xD	08:28
	→ Techcable joined	08:29
	→ Duckle joined	08:29
	→ roxlu joined	08:29
	→ sahilister joined	08:29
	→ hnOsmium0001 joined	08:29
	→ bencevans joined	08:29
	→ retro_ joined	08:29
	← aniruddha left	08:29
	→ aniruddha joined	08:29
	→ LordLion joined	08:30
	→ JayDoubleu joined	08:30
	→ DrowningElysium joined	08:30
osse	where did 0 come from?	08:30
Guest1	so what happened is i was on my feature/skeletons branch, realized i didn't want to keep the work there, and tried to revert back to the previous commit, so i ran that reset command, then checkout'd master, and then checkout'd feature/skeletons again	08:31
	i think the command i used for that was either "git reflog HEAD" or "git reflog" (while on master perhaps)?	08:33
	→ unmanbearpig joined	08:33
	→ rahl joined	08:33
osse	run them again ,then	08:34
	← unmanbearpig left	08:34
	→ unmanbearpig joined	08:34
	← unmanbearpig left	08:35
Guest1	`git reflog HEAD` while in the feature/skeletons branch?	08:35
	→ unmanbearpig joined	08:35
	← unmanbearpig left	08:36
	→ unmanbearpig joined	08:36
Guest1	osse: ok, that paste was the output of running `git reflog HEAD` while in the feature/skeletons branch, i just checked	08:36
	← Optimus left	08:37
Guest1	osse: ok and HEAD@{3} and HEAD@{1} also match up	08:37
	← unmanbearpig left	08:38
osse	that makes sense, because you moed back and forth	08:38
	→ unmanbearpig joined	08:38
Guest1	so then i just do `git reset --hard [HEAD@{3}.sha]`?	08:38
	→ makara joined	08:40
	← shailangsa left	08:41
Guest1	osse: i.e. ccccccc	08:41
	`git reset --hard [cccccc]`	08:41
makara	here's a puzzle. I kept a branch locally that was already committed to master because I needed to see the list of unique commits it made	08:41
	That was a while ago. Today I rebased the branch with master by accident, but without changing to it first. So the reflog doesn't show the commit it was on before rebasing. How do I find that commit?	08:41
Guest1	and do i run that command while in master or in feature/skeletons?	08:42
	→ tchan1 joined	08:42
	← rahl left	08:42
	← mika left	08:42
	→ frobnic_ joined	08:42
	→ mika joined	08:42
	→ cambrian_invader joined	08:42
	→ nyuszika7h joined	08:42
	→ Pent joined	08:42
	→ AlexKalopsia4 joined	08:43
	→ eqw joined	08:43
	→ daemon joined	08:43
	→ rhe joined	08:43
	→ kanzure joined	08:43
	→ rahl joined	08:43
	→ vysn joined	08:43
	→ mackerman joined	08:43
	→ strudl joined	08:43
	→ haymawk joined	08:43
	→ synthmeat joined	08:43
	→ daurnimator joined	08:43
	→ NightMonkey joined	08:44
	→ derfj joined	08:44
osse	Guest1: the branch you want to modify	08:44
	→ Bilge joined	08:44
osse	ro rather, undo the modification you did	08:44
makara	oh, i see i did check it out. All good	08:44
osse	which is feature/skeletons	08:44
	← emf left	08:45
	← ztrawhcse left	08:45
Guest1	osse: ah awesome, it's back, thank you man! is it possible to go back a single commit in this branch too?	08:45
	that's what i originally wanted to do, but didn't know how	08:46
	→ emf joined	08:46
osse	that would be git reset --hard feature/skeletons~1	08:46
	or simply HEAD~1	08:46
	→ elibrokeit joined	08:46
	← xkr47 left	08:47
	→ xkr47 joined	08:47
	← emf left	08:47
	← hnOsmium0001 left	08:49
Guest1	osse: thank you so much dude, i thought i'd lost like a months worth of work for a moment there	08:49
osse	as long as it's committed at some point it's almost impossible to lose	08:50
	might be a bitch to get back, but it's in there somewehere	08:50
Guest1	great to know, wont need to panic next time it happens then	08:51
	→ natrys joined	08:52
	→ supernovah joined	08:59
	→ shailangsa joined	09:03
	→ roman joined	09:05
	→ bencevans3 joined	09:05
	← roadie`` left	09:06
	→ unluckyshrubbery joined	09:06
	→ blmt4 joined	09:07
	← nomicon_ left	09:07
	← mika left	09:07
	← shailangsa left	09:07
	← bencevans left	09:07
	bencevans3 → bencevans	09:07
	← Masklin_Gurder left	09:07
	→ vikonen6 joined	09:08
	→ mika joined	09:08
	→ _xor joined	09:08
	← brw left	09:08
	← matthewcroughan_ left	09:08
	→ matthewcroughan joined	09:08
	← Guest1 left	09:08
	→ Masklin_Gurder joined	09:09
	→ alkino_ joined	09:10
	→ fflam_ joined	09:10
	→ berndj-blackout joined	09:11
	→ avarab joined	09:11
	→ kanzure_ joined	09:12
	→ nomicon joined	09:12
	← vancz left	09:14
	← alkino left	09:14
	← blmt left	09:14
	blmt4 → blmt	09:14
	← vikonen left	09:14
	← avar left	09:14
	vikonen6 → vikonen	09:14
	← makara left	09:14
	← kanzure left	09:14
	← berndj left	09:14
	← mika left	09:14
	→ makara joined	09:14
	→ mika joined	09:14
	→ vancz joined	09:15
	berndj-blackout → berndj	09:15
	→ chronon joined	09:16
	← vysn left	09:18
	→ otisolsen70_ joined	09:19
	→ vladoski joined	09:19
	← darkstardevx left	09:20
	← rfuentess_ left	09:20
	← computeiro left	09:20
	→ rfuentess__ joined	09:20
	→ darkstardevx joined	09:20
	→ computeiro joined	09:20
	→ rond_ joined	09:22
	← otisolsen70 left	09:22
	← cluelessperson left	09:24
	→ cluelessperson joined	09:25
	→ shailangsa_ joined	09:25
	→ wender joined	09:27
	→ roadie`` joined	09:30
	← computeiro left	09:30
	← roadie`` left	09:35
	→ roadie`` joined	09:35
	→ christall joined	09:36
	← elkalamar left	09:43
	← natrys left	09:47
	→ BSaboia joined	09:49
	→ pyeverything joined	09:49
	EPic_ → APic	09:53
	← pyeverything left	09:54
	→ Guest57 joined	09:56
	← Guest57 left	09:56
	→ cdown joined	10:03
	← roadie`` left	10:06
	wagle_ → wagle	10:10
	→ roadie`` joined	10:14
	→ masber joined	10:17
	→ ino joined	10:18
	← ino left	10:19
	← roadie`` left	10:19
	→ Yruama joined	10:20
	→ stats4647 joined	10:21
	← christall left	10:23
	→ casaca joined	10:24
	← stats4647 left	10:25
	← rond_ left	10:28
	→ Misotauros joined	10:31
	→ roadie`` joined	10:32
	→ vysn joined	10:36
	→ alzgh joined	10:45
	→ darkbit joined	10:47
	→ dviola joined	10:48
	→ BUSY joined	10:49
	← otisolsen70_ left	10:50
	→ merethan joined	10:56
merethan	electronics	10:56
	oops	10:57
osse	kitchenware	10:58
	oops	10:58
	← rsx left	11:05
	← alkino_ left	11:05
	→ pulse joined	11:09
	← roadie`` left	11:12
	→ roadie`` joined	11:12
	→ daoudr joined	11:14
	→ audiobirb joined	11:15
	← roadie`` left	11:17
	← Murr left	11:18
	→ Murr joined	11:18
	→ xx joined	11:20
	→ roadie`` joined	11:28
	→ alkino joined	11:30
merethan	bdsm equipment	11:30
	oops	11:30
	osse ;)	11:30
	I ended up in ##electronics. There´s no channel for kitchenware though.	11:32
	→ Guest7843 joined	11:32
	← roadie`` left	11:33
Guest7843	I have a patch with full absolute path and git tells me 'Invalid path' while the path is really valid, file exists there. How can I apply such path with full absolute path in it?	11:33
	git 2.33	11:33
osse	look at the -p option	11:33
Guest7843	osse: Does it matter where the *.patch file itself located?	11:34
osse	no	11:34
	→ jwillikers joined	11:38
	→ speckz joined	11:41
	← darkbit left	11:42
	→ roadie`` joined	11:45
lantech19446	damn merethan i was actually interested in your conversation til i foubd out you were joking and there is no equipment	11:47
	← jwillikers left	11:47
	→ lechner joined	11:49
merethan	Well there´s no channel for kitchenware, but the other two have one lantech19446 ;)	11:51
	→ jwillikers joined	11:51
	→ computeiro joined	11:54
	← wender left	11:56
	→ wender joined	11:59
	→ natrys joined	12:00
	← computeiro left	12:01
	→ rewrit3 joined	12:06
	→ nyah joined	12:06
	← makara left	12:08
	→ gh34 joined	12:19
	→ christall joined	12:20
	← tchan1 left	12:27
	→ tchan joined	12:27
	← aidalgol left	12:34
	← vladoski left	12:35
	→ vladoski joined	12:36
	→ aidalgol joined	12:37
	→ mat001 joined	12:40
	→ joes joined	12:43
	→ kyshoc joined	12:46
	← kyshoc left	12:46
	→ kyshoc joined	12:46
	→ a6a45081-2b83 joined	12:47
	← joes left	12:48
	→ BH23 joined	12:48
	← christall left	12:51
	→ christall joined	12:51
	← robertparkerx_ left	12:52
	→ robertparkerx joined	12:53
	← roadie`` left	12:53
	→ roadie`` joined	12:54
	→ vlado joined	12:58
	← roadie`` left	12:59
	→ roadie`` joined	12:59
	← vladoski left	13:00
	← daoudr left	13:00
	→ zebrag joined	13:10
	→ tureba joined	13:15
	→ vladoski joined	13:17
	← Murr left	13:18
	→ Murr joined	13:18
	→ ApostleInTriumph joined	13:22
	→ alvesadrian joined	13:23
	← EvilDMP left	13:24
	→ madewokherd joined	13:30
	← Assault left	13:31
	→ stats4647 joined	13:34
	→ cousteau joined	13:34
	→ Anticom joined	13:34
cousteau	What resources can I use to learn (more) git? Is the Git Book good enough for all levels of expertise, or maybe too lengthy / shallow / outdated in some aspects?	13:35
Anticom	Hi folks. When having a submodule and I'm updating its ref, is there any (convenient) way to view the log for the range of that ref update? (Similar to how `git commit -v` will display those changes)	13:35
cousteau	For example, I recently learned about worktrees, realized they are a really simple thing, and wondered why I didn't learn about them earlier	13:36
osse	Anticom: git diff --submodule=log perhaps?	13:36
Anticom	cousteau: IMHO a good start would be to learn more about the internals of git. You wouldn't have to keep all this knowledge on hand every time but it's good to have read about it. E.g. https://www.freecodecamp.org/news/git-internals-objects-branches-create-repo/	13:36
cousteau	(maybe I should learn git lsf...)	13:36
	Hm, might do that. I think I have a pretty solid idea of how git internals work, but let me check	13:37
Anticom	osse: tried it, didn't work. This SO post said it only works with -p but that didn't yield the desired result as well :/ https://stackoverflow.com/a/11787169/966530	13:37
	← Vonter left	13:38
Anticom	cousteau: For example took me some time to understand the difference between refs and objects but fortunately osse is a very patient human being :)	13:38
	Well and what they are in the first place of course	13:38
	← stats4647 left	13:38
Anticom	cousteau: But this vastly helped me understanding why certain things are the way they are	13:38
	→ Vonter joined	13:40
cousteau	"What is stored between commits? Is it just a diff between the current and previous commit? Or is an entire snapshot of the repo stored each time?" -- this right here is a very good question. I think 100% of people who have ever learned Git have assumed a commit is some sort of diff rather than a snapshot	13:41
	← roadie`` left	13:42
cousteau	I call this "the git lie" or "the diff fallacy"	13:42
	→ roadie`` joined	13:42
Anticom	cousteau: no no, it's just the differences. I've picked the article by random, just skippnig over it to see whether it's not complete bongus	13:42
cousteau	Uh, no. A commit is an entire snapshot of the code.	13:43
Anticom	Wut	13:43
cousteau	Or rather, a list to references to files in the version of that snapshot	13:43
osse	pro-tip: in one terminal run watch -C tree -C .git , in another terminal run 'git init' and play around	13:43
	→ betelgeuse joined	13:44
	→ gnoo joined	13:44
cousteau	It doesn't store a single copy of every file in each commit, but it does contain references to the whole tree in each commit	13:44
Anticom	You learn every day...	13:45
	← masber left	13:45
Anticom	Also it's been a while since I've last devoted myself wih learning about internals... maybe I should do another one of those sessions as well	13:45
cousteau	So basically you wouldn't need to know what the parents of a commit are in order to get the files in the commit. That would be very inefficient! Whenever you change to another branch you'd have to apply all the diffs one after the other	13:46
osse	Anticom: git diff will usually print the old and new shas, so you can do cd submodule; git log old..new	13:46
Anticom	osse: that works of course, was just curious since `git commit -v` is giving me a nice list of commits for the submodule I thought maybe there's another way to get this info without having to jump into the commit msg editor :D	13:47
cousteau	So it instead saves snapshots of the whole tree, and uses diff between commits only for certain operations, like merge, rebase, cherry-pick...	13:47
Anticom	cousteau: https://git-scm.com/book/en/v2/Git-Internals-Git-Objects last image on this article near bottom illustrates this fairly well	13:47
	← roadie`` left	13:47
osse	Anticom: commit -v you say... did you combine the diff --submodule=log with --staged?	13:47
Anticom	osse: I've tried it with the submodule change unstaged as well as staged but never with the --staged flag	13:48
	→ roadie`` joined	13:48
cousteau	Anticom: indeed! Pretty cool	13:48
	→ jensen joined	13:49
Anticom	osse: I've just double-checked my global git config. There's nothing tinkering with commit / log other than my alias `ci = commit -v` which i'm using to commit (and another unrelated alilas).	13:50
osse	Anticom: need for info. are the stages changed? is this right before commit? what does git diff print?	13:51
jensen	How do I `git remote add origin `<repo stored on my system locally>`? `git remote add origin "$HOME/repos/dotfiles.git" on my server results in "fatal: not a git repository (or any of the parent directories): .git". I can clone from this fine	13:52
Anticom	osse: OMG a regular `git status` is giving me exactly what i wanted lol	13:54
cousteau	What do you mean "on your server"? Is this a remote machine?	13:55
Anticom	osse: my alias for status is adding `-s` so because of that the log is left out	13:55
cousteau	Or is everything in the same machine?	13:55
Anticom	osse: but to answer your question: https://gist.github.com/Anticom/d57b389f5783a7cd234e0704f8bdd423	13:57
osse	Anticom: and what if you add --submodule=log ?	13:59
Anticom	osse: then I'm getting the entire diff for the modules	13:59
	I thought this was what --submodule=diff was for	13:59
	→ computeiro joined	14:00
Anticom	Oh no wait, sec	14:00
	→ otisolsen70 joined	14:00
Anticom	Yes, this gives the log. Sorry I changed another file in the meantime and the submodule logs come at the end so i got confused	14:01
osse	rekt	14:01
	← wender left	14:01
Anticom	Didn't now `git diff` also had the --submodule flag. Was trying it with `git log` (as proposed by the above reference SO post) and this didn't work of course	14:01
cousteau	Oh I got jensen's error	14:02
	← vlado left	14:02
cousteau	jensen: did you run git init before trying to call git remote add?	14:02
	→ vlado joined	14:02
	→ wender joined	14:04
Anticom	I think a `git clone /path/to/another/local/repo` might work as well (?)	14:05
cousteau	Yeah, that too. Might be easier	14:05
	← computeiro left	14:06
cousteau	In fact, I did that to know for sure how to clone, and then tried jensen's remote add method	14:06
	but yeah, cloning a local repository will do what you want	14:07
osse	They even said so. "I can clone from this fine"	14:07
cousteau	Oh	14:07
	git clone local1/ local2/ is another example of command that ends up generating relative refs that maybe should use absolute refs	14:08
	Er, generates absolute refs, should generate relative refs	14:08
	→ mei joined	14:09
Anticom	How could it?	14:10
	git clone still produces a working copy	14:10
cousteau	Instead of having `origin /home/user/project/repo`, having `origin ../repo`	14:11
	→ keypusher joined	14:13
cousteau	The thing is, if you have a whole superproject with submodules, worktrees, etc in a folder, you can't move that folder to another path without breaking absolute refs. The whole thing is static, anchored to that path forever. It would be nicer if it were self-contained.	14:14
	← alkino left	14:14
Anticom	cousteau: Maybe could be achived with `git remote set-url orign <relative-path>` (dunno whether you need file:// scheme there or not. maybe check the remote before updating it)	14:15
jensen	Ok, here's all I did: git clone --bare "$HOME/repos/dotfiles.git" "$HOME/.dotfiles.git ; git --git-dir="$HOME/.dotfiles.git" --work-tree="$HOME" checkout". I am successful up to this point	14:16
	Ok nevermind	14:18
	I probably ran git without the --git-dir="$HOME/.dotfiles.git" --work-tree="$HOME arguments earlier... `d remote -v` shows it's linked to local repo	14:18
cousteau	Why is .dotfiles.git hidden? I mean it can, but it doesn't have to	14:18
	← kenanmarasli left	14:18
	→ gast0n joined	14:19
cousteau	Ah wait never mind, you're creating a hidden repo dir from a non-hidden repo dir for some (probably unrelated) reason	14:20
	→ CSWookie joined	14:21
	→ stats4647 joined	14:21
jensen	Oh yea, I basically follow the approach here: https://www.atlassian.com/git/tutorials/dotfiles. On my desktop machines, I have ~/.dotfiles.git. I was setting up for my server that's hosting the repo to copy the same setup. I guess cloning $HOME/repos/dotfiles.git to ~/.dotfiles.git isn't necessary and I can just use the actual repo since it's just a bare repo	14:23
	← Vonter left	14:24
jensen	One thing I'm still not sure about is how to manage my server dotfiles. Following the approach linked to manage my desktop machines is straightforward because most config files are shared by them and the contents are mostly the same. For the server, I want it to track only a few files already tracked from my desktop machines	14:25
	→ Vonter joined	14:25
imMute	jensen: I have my dotfiles repo cloned to ~/.dots/ on every machine. then I have folders in there that are per-machine (some symlinked to others where I want to share everything between hosts) and a script to setup symlinks in ~ pointing into that ~/.dots/ tree.	14:26
jensen	So I'm thinking of creating a branch for the server from master, do an initially cleaning up of removing all irrelevant files (git rm) and then only cherrypick commits from master here and tehere	14:26
imMute	jensen: I think using branches for different machines is overkill. since to share things (like the script that sets up my command prompt) I'd have to merge, but only merge certain files - some files shouldn't ever be merged. using plain folders is easier (for me) to keep track of	14:27
	← pulse left	14:28
	→ phoenixh joined	14:28
jensen	imMute I haven't branches to be a big deal yet, maybe in the future. I have 95% of the same tracked in master and maybe the content itself is like 10% different between each other	14:33
	imMute I first started using stow	14:33
	imMute Now I'm looking for a way to track config files in /etc. I heard of etckeeper but I'm skeptical of using tools like yadm/etckeeper which add another level of complexity	14:34
	→ Wuzzy joined	14:35
jensen	I'm thinking of using stow or symlinks having etc files in $HOME and then linking them to appropriate places	14:35
	But IIRC, there was a comment in one of the etc files that said "this file shouldn't be a symlink" or something along the lines of avoid using symlinks for that file	14:36
	→ AbleBacon joined	14:36
	→ vvvvv joined	14:36
	→ ChmEarl joined	14:40
	← BSaboia left	14:41
	→ Rhvs joined	14:42
	→ Xenguy joined	14:42
	← phoenixh left	14:44
	→ cmbengue joined	14:44
mackerman	jensen: Read etckeeper's metadata hooks to appreciate what git does not store	14:44
	→ sudoforge joined	14:47
	→ Assault joined	14:47
	→ BSaboia joined	14:50
	← Assault left	14:52
Anticom	jensen: btw i'm currently using dotbot (https://github.com/anishathalye/dotbot). <offtopic>But eventually I'll change to puppet manifests and not only specify my dot files but all my desired software and config</offtopic>	14:54
	← vladoski left	14:54
	→ vladoski joined	14:54
	→ computeiro joined	14:55
	← wender left	14:56
	← madewokherd left	14:56
	← CSWookie left	14:58
	→ thiago joined	14:58
	→ wender joined	14:59
	→ jkordani joined	14:59
cousteau	Anticom: I'm reading that article you linked. I think it doesn't teach me anything new (I already knew how git worked internally, more or less) but it's an interesting way to "demystify" git	14:59
	Basically they grab the magic show that is a git repo, open it and show you the internals, while yelling "See? It's not magic! It's all smoke and mirrors in the inside!"	15:00
	And then they teach you how to build the mirrors and the smoke yourself	15:01
	← computeiro left	15:01
imMute	you could say that about any non-trivial thing that you don't yet understand.	15:01
Anticom	cousteau: Kind of yea. It always helps me to at least roughly understand how things work under the hood to get a better understanding about the capabilities and limitations of a tool	15:01
	imMute: +1	15:02
	→ pulse joined	15:05
	→ rtjure__ joined	15:07
	→ elkalamar joined	15:09
	→ computeiro joined	15:09
	← stats4647 left	15:10
	← roadie`` left	15:10
	→ stats4647 joined	15:10
	← wender left	15:11
	→ wender joined	15:14
	→ hnOsmium0001 joined	15:14
	← computeiro left	15:16
	← Anticom left	15:18
	→ Assault joined	15:19
	→ rsx joined	15:21
	← a6a45081-2b83 left	15:22
	→ sgn joined	15:23
	→ madewokherd joined	15:29
	→ computeiro joined	15:35
	← jkordani left	15:35
	← wender left	15:36
	→ Null_A joined	15:36
	→ wender joined	15:39
	→ cq-work joined	15:41
	← computeiro left	15:41
cq-work	Alright, so I am rebasing a big commit. In this big commit, I want to edit it by removing some files from that commit. So I've run `git restore --staged <files>` on the files I don't want in the commit.	15:41
	However, `git rebase --continue` fails telling me I need to add the fixed conflicts. What do I do	15:42
	← vladoski left	15:42
cq-work	Frankly, the top SO post for this tells me to `git rebase --skip` which I'm like 90% sure would kill my changes.	15:42
	Starting to get a bit scary.	15:42
	→ vladoski joined	15:43
imMute	cq-work: --skip is not what you want. you still need to stage the files before doing --continue though.	15:43
	← Assault left	15:44
cq-work	So I have to stage them and remove them outside of a rebase? I can't remove files in a commit while editing the commit in a rebase?	15:44
imMute	you can.	15:44
	it's been a while, but I think when you mark the commit as "edit" in the rebase, it'll stop just after making that commit, so you can amend it however you please.	15:45
	→ elastic_dog joined	15:45
imMute	I typically use git-gui to do that since it makes it really easy to remove stuff from the previous commit but there's a CLI way to do it as well	15:45
cq-work	Yeah, maybe I should start using a GUI lol	15:48
	→ MrMattsmind joined	15:48
	→ fredi joined	15:48
	← vladoski left	15:48
	← MrMattsmind left	15:49
fredi	hi, I'm basically a layman using git but had to, as per my Advisor orders some time ago	15:49
	basically, I git pulled my data to a server, created about 750 problem instances, and run tests for a solver and my own solution, which means there's at least 750 x 3 = 2250 new files	15:51
	4h ago, I ran "git add ." it's still running.	15:51
	→ grantbow joined	15:53
fredi	did I cross some line...? the problem instances especially are quite large, despite my being oblivious to how that matters, and if.	15:53
	← rfuentess__ left	15:53
	→ kenanmarasli joined	15:54
	← vysn left	15:55
osse	fredi: when you git add a file then git will sha1 hash it and zlib compress it	15:59
	that's probably that's taking time	15:59
	→ jkordani joined	15:59
imMute	cq-work: I'm on the "use CLI for everything" side but git-gui for committing and amending (and most importantly reviewing what you're about to commit) is way more powerful IMO	15:59
	→ jezebel joined	15:59
	→ computeiro joined	16:00
fredi	osse I understand, as long as it does finish I'm OK	16:00
jezebel	is the git remote url format for ssh urls described anywhere?	16:00
fredi	I used a temp token to git pull to the server machien	16:01
cousteau	fredi: are you sure you want to add those files to git? Usually you'd only add the "source" files, not the files that are generated as a byproduct of those files	16:01
	← rewrit3 left	16:01
	← wender left	16:01
osse	yeah I'd Ctrl-C immediately	16:02
fredi	cousteau they're heavy but yes, perhaps I could have them somehow separated but surely not laid down in our research server	16:02
cousteau	(this is, unless you want to add the results to git as well so that other people don't have to generate the files again)	16:02
	→ meator joined	16:02
fredi	the results themselves are light, the instances are a GB or two	16:03
cousteau	Oh	16:03
fredi	ILP	16:03
	→ wender joined	16:04
cousteau	Are those instances generated from some source, or did you generate them one by one?	16:04
fredi	I made them in a plain format using my own generator, then a conversor to .ampl (which is pretty much translate to Mathprog), then with glpsol I made the .lp from the .ampl	16:05
	(so to run cplex)	16:05
	← ApostleInTriumph left	16:05
cousteau	A GB or two each, or the 750 of them together?	16:05
	← alzgh left	16:06
fredi	tbh, I only see the third smallest of 15 directories and it had 11GB	16:06
	← christall left	16:06
	→ alzgh joined	16:06
	← computeiro left	16:06
fredi	cousteau most surely, the 12 remaining are bigger	16:07
	as the parameters increase	16:07
	→ emf joined	16:07
fredi	but I could tell you right now, I'm just kinda scared of du-ing	16:07
	let me see	16:08
cousteau	Well, if you have a script that calls the conversion utility and glpsol, you could put only the plain files + converter + script that calls the converter and glpsol into your git repository	16:08
	And have other users call the script to generate the things	16:08
fredi	I did that. run a couple of days	16:08
	the results are there, in the server. I was about to upload them to github	16:08
cousteau	Which is more the point of git: to use it to control the stuff you're editing, not the stuff that has been generated	16:08
	Oh... Yeah if it takes a couple of days you probably want to avoid the users having to deal with that every time	16:09
fredi	it's a one time only measurement to take some numbers and put in a paper	16:10
cousteau	My advice would be to `git add` the files one by one, and see how long they take	16:10
fredi	yes, I see your point here	16:10
	cousteau ctrl+c won't corrupt anything btw?	16:11
cousteau	If 1 file takes 40 seconds to be added to git, then 750 files will take, um...	16:11
	About 8 hours	16:11
fredi	holy...?	16:12
cousteau	I'm not sure. osse: it won't, right?	16:12
fredi	while we were talking	16:12
	it finished	16:12
cousteau	fredi: but that was an example	16:12
	Lol	16:12
jezebel	can someone break this remote url for me? [email@hidden.address]	16:12
fredi	but now should I commit and push? I'll du	16:12
cousteau	I'm glad we could help by providing entertainment :)	16:12
jezebel	login@host:user/dir-in-their-home.git ?	16:13
fredi	316G	16:13
cousteau	jezebel: I think it's login@host:whichever/path/the/git/server/uses	16:14
fredi	cousteau I should git push this?	16:14
cousteau	I'm not sure to be honest	16:14
	→ jazzy joined	16:14
fredi	oh-oh	16:14
	→ Narrat joined	16:15
jezebel	hmmm	16:16
cousteau	I would ask your advisor what exactly they expect you to do. If they want you to use git for tracking your own work and changes, then you would just upload that to a different platform, like an ftp server, and use git only to track your source code. I suppose that's what an advisor would want you to do, since that's the value of git	16:16
	← merethan left	16:17
fredi	it won't commit because of my global settings as well... I swear it prompted for my user and pass. and I see your point about git too but still can't do much one-sided	16:17
cousteau	If they want you to use git as a mechanism to share files, well, I don't think that's what you typically use git for, but in that case sure, go ahead	16:17
fredi	cousteau he ordered to create a private repo and add him as collaborator	16:18
cousteau	If they just aren't sure how to share large files, I think this is the kind of thing where you want to use git-lsf	16:18
fredi	these files are something we didn't predict tbh	16:19
	the earlier tests weren't enough	16:19
	← jazzy left	16:19
cousteau	I'd just ask him	16:19
fredi	anyway, I'll have to do something. git is running, supposedly committing, but I'll be back 4h later. cousteau and ditto	16:19
	← fredi left	16:20
cousteau	"hey, the tests I generated are several hundred GB large, do I just push that to git? No, right? What do you recommend?"	16:20
	← stats4647 left	16:23
	→ stats4647 joined	16:24
osse	I suspect that ctrl-c'ing and resuming would cost a bit but not much	16:27
	git would have to hash the files again, but then figure out they're already written. I guess zlib'ing takes longer than hashing	16:28
	← stats4647 left	16:29
	← jim87 left	16:31
	← speckz left	16:32
	→ woink joined	16:35
	→ Assault joined	16:36
	← vlado left	16:38
	→ jstein joined	16:38
	→ christall joined	16:41
	← t3nj1n_ left	16:42
	← alzgh left	16:42
	→ alzgh joined	16:43
	→ jazzy joined	16:48
	→ junktext joined	16:51
	← christall left	16:53
	→ christall joined	16:53
	→ stats4647 joined	16:55
	→ fredi joined	16:56
	→ EvilDMP joined	17:00
	← stats4647 left	17:02
	← cmbengue left	17:05
	← jezebel left	17:13
	→ hbautista__ joined	17:15
	← rsx left	17:19
	← Guest7843 left	17:20
	→ pyeveryt_ joined	17:21
fredi	cousteau professor gave permission to let the big files out. if you could enlighten me on this matter, there is a "git remove" command of sorts, right? I assume I'll need it once the add finished processing with them included	17:23
	→ [SyntaxError] joined	17:24
fredi	that's obviously not a way to learn or use git. I hope I'll grasp it properly in the future (git as a versioning tool)	17:24
	← TomyWork left	17:24
	← junktext left	17:25
	← pyeveryt_ left	17:25
	→ junktext joined	17:26
	→ tirnanog joined	17:28
	← EvilDMP left	17:30
	→ EvilDMP joined	17:31
	← EvilDMP left	17:35
	← madewokherd left	17:38
	→ darkbit joined	17:38
cousteau	fredi: if you haven't committed yet, `git reset` will forget about the files you have added but not committed	17:38
	← aniruddha left	17:39
cousteau	(only forget about them, but not actually delete them from your working dir)	17:41
fredi	it's running the commit as we speak, unfortunately. however, that thing about my global config being not properly set	17:41
	no, I think there's no way out (except risking corruption through inconsistency caused by Ctrl+C)	17:43
cousteau	And if you have committed, `git reset HEAD~` will undo the commit	17:43
	← Wuzzy left	17:43
fredi	now, that's great. thank you, cousteau	17:43
cousteau	So you can just wait for it to finish and then undo it	17:43
fredi	plus, I'll configure global data	17:43
	in the mean time	17:43
	exactly	17:43
	← meator left	17:44
	← supernovah left	17:44
fredi	say, github can be quite troublesome for people sshing if this global config is not properly set...?	17:45
cousteau	You could even wait for it to finish, create a new branch or tag the current commit so that you can fetch that commit later, and then rewind to the previous commit. That'll leave your current branch as it was before the commit, but if at some point you change your mind and decide to add the files you don't have to commit again	17:45
	→ skapata joined	17:46
	← skapata left	17:46
	→ skapata joined	17:46
fredi	I didn't buy their arguments for ending support for git tool auth	17:46
cousteau	(btw, I doubt GitHub would allow you to host those many GB for free...)	17:46
fredi	cousteau branches, mein Freund, are some concept I'll learn. properly. someday	17:46
	ditto	17:46
cousteau	Some googling suggests that as of 2020 the limit is 100MB per file and 10GB per repo	17:47
fredi	wasn't taught to versioning tools in college (git or CVS or what-so-ever)	17:47
	so... end up lost these times	17:47
	→ wolfdale joined	17:48
fredi	100MB per file? I'll have to take some care then	17:48
	(the space upper limit was expected)	17:49
	→ cmbengue joined	17:49
cousteau	Branches are just pointers to commits, saying on top of which commit you'll add the next commit. They allow you to do things separately or in parallel, and then merge them together	17:49
fredi	thing is, I was reading Git Manual this morning. it is something worth learning, but nearly insane to jump (be thrown) amidst	17:50
	→ junktext_ joined	17:50
fredi	the HEAD tag supposedly points to the last commit, I kinda got it	17:50
cousteau	I never learned CVS and I don't miss it	17:50
tang^	https://docs.github.com/en/repositories/working-with-files/managing-large-files/about-large-files-on-github	17:51
fredi	neither did I but there's one project I'm somewhat interested (FreeBSD) and guess what?	17:51
	→ supernovah joined	17:51
fredi	lucky me they moved to SVN	17:51
	← junktext left	17:51
cousteau	Branches in CVS are an entirely different concept. Branches in Git are a more elegant way to tackle the problem of parallel development	17:52
fredi	I heard CVS is totally outdated	17:52
cousteau	I don't know what's the difference between CVS and SVN but they have the same problem of being centralized	17:52
fredi	and I should stick with SVN and Git	17:52
cousteau	And painful branching and merging	17:53
mspe	I really like that Git branches are only pointers and therefore do not really exist	17:53
cousteau	In my opinion you should just stick with git, and only use SVN for projects already using it	17:53
fredi	it's unlikely for me to work outside the very exception of FreeBSD, but git seems like Haskell	17:54
tang^	Freebsd only recently switched from CVS to Git... it's not _that_ outdated.	17:54
fredi	cousteau too good to be left unlearned	17:54
	→ nightstrike joined	17:55
	← nightstrike left	17:55
cousteau	tang^: projects taking a long time to migrate out of it doesn't mean it's not outdated	17:55
	→ pyeveryt_ joined	17:55
	→ ackyshake joined	17:55
cousteau	Projects are migrating out of it BECAUSE it's outdated. Even despite of how hard it is to migrate from one SCM to another.	17:56
fredi	as FreeBSD itself could show, I do have unusual tastes. but now, tang^ you sure they're on git?	17:56
cousteau	The period from "yo, we should totally migrate out of this, it's painfully outdated" to actually migrating can be several years long	17:57
	→ pyevery__ joined	17:57
fredi	I'm out-of-scene for a year or so but handbook still mentions svn (and devs handbook, CVS)	17:57
tang^	fredi: I'm fairly certain... I've been following them for the past year	17:57
fredi	that's better	17:57
	Henry151_ → Henry151	18:00
	→ przemoc joined	18:00
	← pyeveryt_ left	18:00
	← supernovah left	18:01
	→ supernovah joined	18:03
	← cq-work left	18:05
	→ wootehfoot joined	18:05
	← ThorMojito left	18:07
	→ ThorMojito joined	18:09
	→ Ugrastil joined	18:09
	→ longshanks joined	18:14
	cousteau → ghosteau	18:16
	→ R2robot joined	18:16
longshanks	hi. does mergedown, mergeup have any meaning as I couldn't find any reference to it in any git man pages yet I've heard people using the term at work.	18:17
	← sandipan left	18:17
longshanks	So am guessing it's not a git term per-se.	18:18
	→ j`ey_ joined	18:18
ghosteau	Never heard of it. Context?	18:19
j416	haven't heard.	18:20
	← BH23 left	18:21
longshanks	ghosteau: i think when merging according to git-flow paradigm. i don't really use git flow as plain old topics suit me just fine.	18:21
	← alvesadrian left	18:21
	→ alvesadrian_ joined	18:22
longshanks	i guess it can only really make sense if there's some context to teh merge direction.. a hierarchy? but yeah, dunno. thought I'd ask here :)	18:22
	→ jess joined	18:22
ghosteau	Did you read this in a tutorial or something?	18:24
longshanks	ghosteau: nope just overheard someone at work. I'll ask them Monday.	18:25
	← j`ey_ left	18:25
longshanks	Did google it and found a couple references , but mainly nada.	18:25
	→ alvesadrian joined	18:25
	→ madewokherd joined	18:25
ghosteau	Oh	18:25
wootehfoot	longshanks, ask him whether he's heard of mergesideways, because it's way better	18:26
longshanks	what about orthogonal ?	18:26
wootehfoot	yes	18:26
	← alvesadrian_ left	18:27
ghosteau	https://community.adobe.com/t5/photoshop-ecosystem-discussions/what-is-the-difference-between-merge-down-merge-visible-and-flatten-image/m-p/4626571 this mentions the term but seems unrelated to git	18:28
	← SuperLag left	18:30
	→ joakimk joined	18:30
longshanks	What about https://www.reddit.com/r/git/comments/hshfzq/could_someone_explain_the_concepts_of_downmerge/	18:30
	i wonder if it's a term used by people from non-git systems. couldn't find anything anyting in the man pages or the git pro book. oh well.	18:31
ghosteau	https://www.atlassian.com/continuous-delivery/principles/workflows-with-feature-branching-and-gitflow they seem to use the terms here	18:32
longshanks	ghosteau: ah thanks. i'll go read it	18:32
ghosteau	No idea what they mean by it though. They don't seem to clarify it	18:32
longshanks	ghosteau: well glad I'm not the only one :)	18:34
	→ forgotmynick joined	18:36
	← af left	18:37
	→ af joined	18:37
ghosteau	https://softwareengineering.stackexchange.com/questions/222696/should-i-merge-from-the-same-branch-or-from-a-merge seems to suggest merge up/down has to do with merging into master / from master. No idea which is which.	18:38
	← junktext_ left	18:38
	→ junktext joined	18:39
ghosteau	Following a tree analogy, I would guess merging down merges "towards the trunk", and merging up mergers "towards a branch". The terminology probably comes from another SCM, as you suggested	18:39
longshanks	ghosteau: i think it's always better to be explicit about which branch or branches you're merging into which. I do think of merge into and rebase onto as it helps with the visualisation but up and down. yikes!	18:39
	maybe see-saw scm.	18:40
ghosteau	The thing is that in git there's no trunk and branches, it's all separate branches, so it doesn't make as much sense	18:40
	But there's often a main/master/stable branch that just sits there waiting for PRs, and multiple development branches that implement the experimental features	18:42
ikke	In practice you do have canonical primary / long-lived branches	18:42
	← mat001 left	18:44
	→ mat001 joined	18:44
ghosteau	https://www.plasticscm.com/book/ This PlasticSCM book uses the terms "merge up" and "merge down" extensively	18:46
longshanks	what is plasticSCM ?	18:47
	→ t3nj1n joined	18:48
ghosteau	Apparently it was the other way around. When you create a new branch for some fix, you draw it below the master or "parent" branch. So merging down is merging from master to the development branch, or "syncing with master", and merging up is merging the devel branch back into master	18:48
ikke	"Plastic SCM is a full version control stack designed for branching and merging. It offers speed, visualization, and flexibility."	18:48
ghosteau	PlasticSCM is probably another SCM	18:48
	Like git or mercurial or SVN	18:49
	I recall having heard of it before	18:49
	← mat001 left	18:49
ghosteau	Commercial, not sure if distributed, the Wikipedia page looks like an advertisement	18:51
	→ joakimk31 joined	18:52
	← joakimk left	18:54
	→ Torr joined	18:55
Torr	Hey	18:55
	Is it possible to merge a branch ammending its commit?	18:55
	← longshanks left	18:55
another\|	git merge --edit	18:56
ikke	That's just the commit message, and it's the default	18:57
another\|	is it? I rarely merge by hand	18:57
ikke	Yes, after a rant from Linus triggered by people running git pull randomly	18:58
ghosteau	I think you can instruct git merge to not create a commit yet	18:59
ikke	--no-commit :-)	18:59
	→ durham joined	18:59
ghosteau	With --no-commit perform the merge and stop just before creating a merge commit, to give the user a chance to inspect and further tweak the merge result before committing.	18:59
	→ computeiro joined	18:59
ghosteau	Yep, that was it	19:00
	→ longshanks joined	19:00
	tomek_ → Eustachy_Kapusta	19:00
Torr	ikke: But that'd still create a new commit.	19:00
	I want to ammend.	19:00
	← fredi left	19:00
Torr	amend*	19:01
ghosteau	git commit --amend	19:01
ikke	Torr: what exactly do you want to amend?	19:01
	← wender left	19:01
	→ xsx joined	19:01
ghosteau	You probably don't want to do fancy stuff when merging though	19:01
	→ loki_val joined	19:02
ghosteau	Just combine this with that, and maybe choose this version or that version, and maybe in some extreme cases do some fixing (although you probably want to do that BEFORE merging)	19:02
Torr	ikke: The merge changes.	19:02
ikke	Torr: step 1: run git merge --no-commit	19:03
	step 2: change what you want	19:03
	step 3: run git commit to finnish the merge	19:03
xsx	i want to list all tags that change the major version. so lets say i have tags: v1.0.0, v1.1.0, v2.2.2 the output would be v1.0.0 and v2.2.2, any ideas?	19:03
ghosteau	Step 2.5: git add the changes you made	19:03
Torr	-____	19:03
	← pyevery__ left	19:04
	→ wender joined	19:04
ikke	xsx: that's something outside of gits scope	19:04
	→ m0viefreak joined	19:04
ghosteau	xsx: sounds like a task for some text-processing tool	19:04
	← lechner left	19:04
xsx	yeah, that's what i thought, but was hoping there is a hidden way in git to get ot ;)	19:05
ikke	Create a script that iterates over each tag, keeps track of the current major version, and print the tag if the major version changes, and store the new major version	19:05
ghosteau	git tag v[0-9]* \| sort \| uniq with some parameters so that it only reads up to the first .	19:05
	git doesn't care much about version numbers. I mean each project uses a different version numbering scheme	19:06
	← computeiro left	19:06
ghosteau	Oh but if these are version tags they're probably annotated tags, so maybe print only those	19:06
	← longshanks left	19:07
ghosteau	Actually I'm not sure `git tag v[0-9]*` works	19:09
xsx	they are not annotated :( yeah,i will go with awk, sort and other stuff	19:10
	ghosteau: you ommited -l arg	19:10
ghosteau	Yeah I did.	19:12
	Not sure if it's needed if you use a pattern though?	19:12
	Or maybe that only works without arguments	19:12
	← Torr left	19:14
	→ ZacSharp joined	19:16
	← Murr left	19:17
	→ Murr joined	19:18
	← kenanmarasli left	19:19
ghosteau	git tag --list 'v[0-9]*' \| sort -n \| uniq -w3 will work as long as the major version has no more than 2 digits (and is followed by a . if it has only 1 digit)	19:19
	Because if the tag is "v1.2.3" it'll only compare the first 3 chars, i.e. "v1."	19:20
	← Assault left	19:20
ghosteau	And if it's "v10.0.1" it'll compare the "v10"	19:20
	This is a rather poor solution, but it may work for you if all you want is a quick comparison. For a more advanced one you'll want a proper scripting language.	19:21
	git tag --list 'v[0-9]*' will print all tags that begin with "v" followed by a digit	19:22
	(you can go one step further and do things like 'v[0-9]..*' or whatever)	19:23
xsx	ghosteau: what is the -w for uniq? i don't have this option	19:23
ghosteau	sort -n will sort numerically (10 goes between 9 and 11)	19:24
	And uniq -w3 will only check the first 3 chars	19:24
	xsx: what platform are you in?	19:25
	→ Samian joined	19:26
	← durham left	19:26
	→ durham joined	19:27
	→ lechner joined	19:28
	→ vdamewood joined	19:28
	→ CSWookie joined	19:30
	← gh34 left	19:35
	→ stats4647 joined	19:40
	→ dgw joined	19:43
	← stats4647 left	19:44
	← darkbit left	19:45
	← Samian left	19:47
	→ stats4647 joined	19:49
	← stats4647 left	19:49
	→ stats4647 joined	19:50
	→ TheGuestMovie joined	19:51
	→ pyeverything joined	19:52
	→ computeiro joined	19:55
	← _xor left	19:55
	→ _xor joined	19:55
TheGuestMovie	Just tried to push a local commit (containing file changes + updating the commit used in a submodule) and I got "ref integration:: Error in git rev-list --stdin --objects --not --remotes=origin --: exit status 128 fatal: bad object 99b92e2...". "git show 99b92e2" just repeats that it's a bad object. I fixed it by just undoing the local commit and	19:56
	committing again, but what can cause this?	19:56
	→ Wuzzy joined	19:56
	← wender left	19:56
	← SpeakerToMeat left	19:56
TheGuestMovie	could it be memory issues? Or is this relatively common? first time i run into it myself	19:57
	← stats4647 left	19:57
	→ wender joined	19:59
	← computeiro left	20:01
	← CSWookie left	20:02
	← cmbengue left	20:03
	kanzure_ → kanzure	20:03
	→ zumba_addict joined	20:04
	→ computeiro joined	20:04
	← ghosteau left	20:04
	← wender left	20:06
	→ wender joined	20:09
	→ shored joined	20:11
	← computeiro left	20:11
	→ SpeakerToMeat joined	20:15
	→ Samian joined	20:17
	→ phoenixh joined	20:18
	→ nivag joined	20:20
	← joakimk31 left	20:21
	← pyeverything left	20:29
	← Null_A left	20:29
	← drogas left	20:30
	← jstein left	20:30
	← Vonter left	20:34
	→ Null_A joined	20:35
	→ pyever___ joined	20:41
	← phoenixh left	20:42
jast	TheGuestMovie: it's definitely corruption. whether it's RAM or disk, hard to tell. you could look at the object on disk (.git/objects/99/b92e2...), see if it looks like it contains actual data or got truncated with 0 bytes (the latter typically happens if the system crashes or loses power before the object has been fully committed to disk)	20:43
	← forgotmynick left	20:45
	← jensen left	20:50
	→ mw__ joined	20:52
	→ BinarySavior joined	20:55
	← emf left	20:55
	← pyever___ left	21:02
	→ emf joined	21:02
	→ tar_xvf joined	21:06
	← tar_xvf left	21:08
	→ tar_xvf joined	21:09
	→ Xavier7 joined	21:11
	← ackyshake left	21:13
	← _xor left	21:14
	→ Lord_of_Life_ joined	21:14
	Lord_of_Life_ → Lord_of_Life	21:16
	→ _xor joined	21:17
	→ vysn joined	21:18
	→ ss4 joined	21:18
	← tar_xvf left	21:20
	← wootehfoot left	21:22
	→ causasui joined	21:22
	→ tar_xvf joined	21:25
	← Null_A left	21:27
	→ Null_A joined	21:27
	← tar_xvf left	21:28
	→ tar_xvf joined	21:30
	→ EvilDMP joined	21:31
	← m0viefreak left	21:32
	← tar_xvf left	21:32
	← ZacSharp left	21:34
	→ miquella[m] joined	21:36
	← shailangsa_ left	21:36
	→ tar_xvf joined	21:37
	→ fredi joined	21:37
	→ kawics11 joined	21:39
	→ gordonfish- joined	21:39
	← tar_xvf left	21:40
	← EvilDMP left	21:41
fredi	I'm almost done but couldn't push	21:42
	it says as follows: https://dpaste.org/oFzA	21:42
	gordonfish- → gordonfish	21:42
fredi	any help is welcome. at this moment, local repository has plenty of data left out, but, as far as i understand, in the tree. shall I remove them?	21:43
	I'm afraid git pulling as it says cause some disarray	21:43
	as for global configs, all done	21:44
	→ tar_xvf joined	21:44
	← Null_A left	21:46
	← [SyntaxError] left	21:51
	→ cmbengue joined	21:52
	← xsx left	21:53
	← otisolsen70 left	21:55
	← Narrat left	21:55
	← junktext left	21:58
	→ shailangsa joined	21:58
	← Xavier7 left	21:59
	→ stats4647 joined	22:06
	← christall left	22:07
fredi	ok, now all solved. thanks to the people who helped earlier.	22:08
	← fredi left	22:08
	← jkordani left	22:08
	→ christall joined	22:10
	← Strom- left	22:10
	← stats4647 left	22:10
	→ Strom joined	22:11
	← zumba_addict left	22:12
	← jwillikers left	22:13
	→ SuperLag joined	22:13
	← christall left	22:14
	→ hqdruxn08__ joined	22:22
	← hqdruxn08_ left	22:26
	← rtjure__ left	22:31
	→ ttree joined	22:35
	→ odoood joined	22:37
	→ christall joined	22:41
	→ ozoned joined	22:42
	← Gustavo6046 left	22:43
	← natrys left	22:43
	→ Gustavo6046 joined	22:43
	← odoood left	22:45
	← kawics11 left	22:48
	→ ferdna joined	22:51
	← jimklimov1 left	22:52
	→ chexum joined	22:53
	→ odoood joined	22:56
	← ozoned left	22:57
	← emf left	22:58
	→ ozoned joined	22:58
	← odoood left	23:01
	→ Raguile joined	23:01
	→ odoood__ joined	23:03
	→ pizdets joined	23:03
	← christall left	23:03
	← pizdets_ left	23:03
	← ozoned left	23:06
	→ ozoned joined	23:08
	← odoood__ left	23:11
	← af left	23:14
	→ tlaxkit joined	23:19
	→ odoood__ joined	23:23
	← ozoned left	23:24
	← gordonfish left	23:32
	→ christall joined	23:33
	→ gordonfish joined	23:33
	← christall left	23:37
	→ af joined	23:38
	← nyah left	23:40
	← odoood__ left	23:40
	← Yruama left	23:44
	← xx left	23:51
	→ eschwartz joined	23:53
	← vdamewood left	23:55
	← Wuzzy left	23:57
	← ttree left	23:58

Logs Search ←Prev date Next date→

Channels Documentation