IRCloggy #git 2020-01-15

Provider of IRC logs since 2005.

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

	2020-01-15
	← lucasb left	00:00
	→ Codaraxis joined	00:02
	← omnireq left	00:02
	→ orbyt_v7 joined	00:04
	← oatmealraisin left	00:04
	→ net_wayfarer joined	00:07
	← net_wayfarer left	00:07
	→ net_wayfarer joined	00:13
	← net_wayfarer left	00:13
	← TheJollyRoger left	00:14
	→ canton7 joined	00:14
	→ thiago joined	00:15
	→ TheJollyRoger joined	00:17
	→ net_wayfarer joined	00:20
	← net_wayfarer left	00:20
	← H2Q left	00:20
	← Wizek left	00:23
	← jayjo left	00:23
	→ net_wayfarer joined	00:32
	← net_wayfarer left	00:32
	→ net_wayfarer joined	00:37
	← net_wayfarer left	00:37
	← owlscientist left	00:40
	← orbyt_v7 left	00:43
	← orbyt_ left	00:44
	→ net_wayfarer joined	00:44
	← net_wayfarer left	00:44
	→ wrksx_ joined	00:45
	← jhulten left	00:46
	→ Legolars99 joined	00:47
	→ omnireq joined	00:48
	→ net_wayfarer joined	00:52
	← net_wayfarer left	00:52
	← sgn left	00:52
	← Achylles left	00:54
	← danielzgtg left	00:57
	→ net_wayfarer joined	00:58
	← net_wayfarer left	00:58
	→ danielzgtg joined	00:58
	← swalladge left	01:03
	→ net_wayfarer joined	01:05
	← net_wayfarer left	01:05
	→ LordRyan joined	01:06
	← rsrx left	01:06
	← dionysus69 left	01:09
	← Thorn left	01:10
	→ net_wayfarer joined	01:11
	← net_wayfarer left	01:11
	→ Decobus joined	01:11
	← Legolars99 left	01:13
	← m0viefreak left	01:14
	← wrksx_ left	01:15
	→ wrksx_ joined	01:16
	→ causasui joined	01:16
	Disco → Disconsented	01:18
	→ spacefrog joined	01:20
	← jetchisel left	01:21
	→ net_wayfarer joined	01:23
	← net_wayfarer left	01:23
	← causasui left	01:26
	→ orbyt_ joined	01:33
	→ net_wayfarer joined	01:38
	← net_wayfarer left	01:38
energizer	how do i find out which commits have reference to a given tree?	01:41
	answer seems to be 'try them all' https://stackoverflow.com/a/41090798/12055383	01:43
	is there any tool that keeps an index of things like this?	01:43
	→ jayjo joined	01:43
energizer	similarly "find children of this commit" without being all O(n) about it	01:44
	→ net_wayfarer joined	01:50
	← net_wayfarer left	01:50
	← zlogan left	01:51
	→ zlogan2003 joined	01:51
	→ mat001_ joined	01:51
	← Sasazuka left	01:51
	← greatgatsby left	01:53
	← mat001 left	01:54
	→ pandakekok9 joined	01:54
	← IAmNetx left	01:55
	→ fakenerd joined	01:56
	← mat001_ left	01:56
	→ net_wayfarer joined	01:56
Ori_B	energizer: AFAIK, you just need to be O(n) about it.	01:59
	I'm not aware of any particularly common operations that need these to be indexed, so.. eh.	01:59
	I don't think it's done	01:59
	← Dirak left	02:02
	→ maryo joined	02:03
	← crose left	02:10
	→ Onionnion joined	02:16
	← omnireq left	02:23
	→ omnireq joined	02:23
	← maryo left	02:26
	→ Dirak joined	02:28
	→ gpanders_ joined	02:30
	← omnireq left	02:32
	← fahadash left	02:33
	→ omnireq joined	02:33
	→ Elundia joined	02:37
	← Elundia left	02:52
	→ Elundia joined	02:52
	← Elundia left	02:52
	→ donfede joined	02:52
	→ Elundia joined	02:53
	← Elundia left	02:53
	→ Elundia joined	02:53
	← Elundia left	02:53
	→ Elundia joined	02:53
	← Elundia left	02:53
	→ Elundia joined	02:54
	← Elundia left	02:54
donfede	hello - i'm seeking to setup git server side hooks (update); should/can this be commited to the git repo, or am I to edit/copy barerepo.git/hooks/update directly?	02:54
	I've read https://git-scm.com/book/en/v2/Customizing-Git-Git-Hooks , but am not clear on implementation (and am wary of breaking my repo)	02:54
	← fakenerd left	02:55
	→ fakenerd joined	02:56
	← NODE left	02:56
	← gpanders_ left	02:57
	→ NODE joined	02:59
	← NODE left	03:00
	→ gpanders_ joined	03:00
	← F0rTex left	03:01
	← fakenerd left	03:01
	→ NODE joined	03:02
energizer	if i wanted to store an index of children, i'd just put it into .git/foo and it would get pushed?	03:02
	(when i `git push`)	03:02
	→ F0rTex joined	03:03
	← gpanders_ left	03:03
	→ causasui joined	03:03
	→ nickodd joined	03:03
	← Cabanossi left	03:07
	→ gpanders_ joined	03:07
	→ Cabanossi joined	03:11
donfede	@+energizer is your question a response for me, or its own question?	03:12
energizer	donfede: it's my own question	03:12
	→ fakenerd joined	03:13
	← fakenerd left	03:13
	→ fakenerd joined	03:13
gitter1234	Hello! Is there a nice way to undo an old commit and instead move all the changes from that commit into my current working tree?	03:14
	← begui left	03:20
donfede	energizer: ty, pardon my noise	03:20
	answering my own question -- it looks like "Hooks are per-repository and are never pushed." Lily Ballard (though no citation)	03:20
	per https://stackoverflow.com/questions/12222186/are-git-hooks-pushed-to-the-remote-when-i-git-push	03:20
	← shored left	03:21
	→ regedit joined	03:22
	→ begui joined	03:22
	← Brainium left	03:22
	dilfridge → gollum	03:22
	→ igemnace joined	03:23
	← fakenerd left	03:23
	→ fakenerd joined	03:24
	→ _Warl0ck joined	03:26
	← sgen left	03:26
	← zlogan2003 left	03:26
	← fakenerd left	03:28
	lagothrix → Guest61693	03:29
	→ lagothrix joined	03:29
	→ beastDiv_ joined	03:29
kreyren	is there a way to move `.gitmodules` in `.git` directory? it's kinda annoying in the root of repo	03:30
	→ fakenerd joined	03:31
	← Guest61693 left	03:33
	→ im0nde_ joined	03:37
	← Codaraxis left	03:37
	→ swalladge joined	03:37
	← Enphuego left	03:38
	← felipe1982 left	03:38
	← im0nde left	03:39
	→ jetchisel joined	03:41
	→ ShapeShifter499 joined	03:46
ShapeShifter499	hi	03:46
	I'm currently mirroring a large set of repositories using "repo". I have two questions, is it possible to "clone" out of this mirror so that it's a symlink and thus I only need to store one copy of the code?	03:48
	And how do I sync new code into the cloned repository generated from the mirror?	03:49
	← xcm left	03:53
	→ xcm joined	03:53
	→ duderonomy joined	03:56
	→ maryo joined	03:56
	← maryo left	04:01
	← boombatower left	04:04
	← fakenerd left	04:04
	→ fakenerd joined	04:04
	← fakenerd left	04:10
	← jetchisel left	04:10
	→ Dreadnaught joined	04:12
	← Decobus left	04:14
	→ Decobus joined	04:15
	← Kulrak left	04:15
	HOFFMAN → YOU_SUCK_HOFFMAN	04:20
	→ Moongoodgirl joined	04:20
	← gnufr33d0m left	04:25
	← Onionnion left	04:26
	→ gnufr33d0m joined	04:26
	← howdoi left	04:28
	→ Thorn joined	04:29
	→ Goplat joined	04:32
	← g4570n left	04:32
	← FH_thecat left	04:39
	← Dirak left	04:40
	← CodeSlingerPaul left	04:42
	→ omnireq_ joined	04:46
	→ FH_thecat joined	04:47
	→ pi- joined	04:49
	← omnireq left	04:49
	→ ferdna joined	04:53
	→ pks_ joined	05:02
	→ Dirak joined	05:02
	→ felipe1982 joined	05:02
	ghost43_ → ghost43	05:03
	← pks left	05:03
	pks_ → pks	05:03
	→ jetchisel joined	05:18
	→ fakenerd joined	05:23
	← Shentino left	05:23
	← FH_thecat left	05:30
	← fakenerd left	05:30
	→ fakenerd joined	05:30
	→ fphilipe_ joined	05:32
	← xcm left	05:37
	← felipe1982 left	05:37
	→ xcm joined	05:37
	← fphilipe_ left	05:37
	← fakenerd left	05:43
	← orbyt_ left	05:43
	→ fakenerd joined	05:44
	← duderonomy left	05:44
	→ felipe1982 joined	05:45
	← fakenerd left	05:48
	→ durham joined	05:52
	← durham left	05:57
	← cbreak left	05:57
	← gnufr33d0m left	05:57
	→ g00s joined	05:57
	→ cbreak joined	05:58
	← felipe1982 left	05:59
	← begui left	06:04
	→ begui joined	06:04
	← azerus left	06:07
	→ azerus joined	06:08
	← ferdna left	06:18
	→ FH_thecat joined	06:24
	→ fakenerd joined	06:30
	→ Mattiaslndstrm joined	06:30
	← omnireq_ left	06:32
	→ milkt joined	06:32
	→ omnireq_ joined	06:33
	→ sauvin joined	06:33
	→ xco joined	06:35
	← xco left	06:36
	→ TomyWork joined	06:38
	→ maryo joined	06:39
	→ tsdh joined	06:41
	→ Codaraxis joined	06:42
	→ rnmhdn joined	06:44
	← NODE left	06:44
rnmhdn	what is the proper way to mange a vuejs - django website?	06:45
	should I have different git repos for the vuejs frontend and the django backend?	06:45
	→ NODE joined	06:47
bookworm	you can do both, if they are separate split repos make sense	06:48
	← NODE left	06:48
rnmhdn	does it work better with CI/CD if I have separate or combined?	06:48
	gitlab	06:48
bookworm	doesn't matter either for unit tests, for integration tests both are relatively straight forward to setup	06:49
rnmhdn	thanks	06:49
	← rnmhdn left	06:49
	→ NODE joined	06:50
bookworm	I'd split them, makes it easier to reason about changes	06:50
	→ duderonomy joined	06:53
	→ xco joined	06:55
	← duderonomy left	06:58
	← g00s left	07:00
	→ pks_ joined	07:04
	← pks left	07:05
	pks_ → pks	07:05
	← nickodd left	07:06
	← fakenerd left	07:08
	gollum → varys	07:10
	← donofrio left	07:15
	→ donofrio joined	07:15
	← causasui left	07:15
	→ Noti joined	07:18
	→ oxymoron93 joined	07:22
	→ g00s joined	07:23
	varys → dilfridge	07:24
	→ sQVe joined	07:25
	→ chele joined	07:38
	← gxt left	07:43
	→ SteffanW joined	07:46
	← Noti left	07:47
	← Spr0cket left	07:50
	← regedit left	07:50
	→ Noti joined	07:51
	→ Spr0cket joined	07:51
	← SteffanW left	07:52
	← thiago left	07:57
	← jetchisel left	08:02
	→ Jackneill joined	08:05
	← Jackneill left	08:06
	← cd left	08:08
energizer	is there somewhere i can stuff some repo metadata that'll be automatically brought along with a clone or push?	08:09
	→ Jackneill joined	08:09
cbreak	energizer: sure.	08:14
energizer	cbreak: where?	08:15
cbreak	write your meta data into a file and commit it	08:15
	maybe on a separate orphan branch	08:15
	(this is often done for web pages of projects, like gh-pages)	08:15
	if you have very specific requirements, you can even use a tag instead of a branch for this	08:16
	and naturally, instead of creating a commit, you can tag a file directly	08:16
energizer	hmm i dont think that's what i'm looking for	08:16
	i dont want to make commits	08:17
cbreak	as I mentioned, you can tag files directly too	08:19
	← jab416171 left	08:21
	→ Codaraxis_ joined	08:21
	← Codaraxis left	08:21
energizer	maybe i'm just reacting instinctively because it seems like hack	08:22
cbreak	energizer: ... unlike the rest of git? :)	08:22
	→ jab416171 joined	08:25
	← TomyWork left	08:25
	→ dimi1947 joined	08:25
energizer	will only refs/heads/* and refs/tags/* be pushed&cloned automatically?	08:25
	like, what if i added refs/foo/bar	08:26
	← Goplat left	08:27
	energizer suddenly realizes there's another way to answer this question than asking here	08:27
cbreak	energizer: you can change that	08:27
energizer	cbreak: how?	08:27
cbreak	for each remote you have something like fetch = +refs/heads/:refs/remotes/origin/ in .git/config	08:28
	you can fetch more than that though	08:29
	→ TomyWork joined	08:29
cbreak	the logic to fetch tags is separate, I think they have special rules	08:29
	← leprechau left	08:30
	→ jas4711 joined	08:30
cbreak	I think you'd have to add to that fetch ref spec after a clone though, there doesn't seem to be much to influence it	08:32
	→ jomofcw joined	08:33
jomofcw	Hello !	08:33
	← ChrisLane_ left	08:34
	→ floppydh joined	08:34
jomofcw	I'm trying to understand the config of a git repository I don't setup myself. It uses hooks. I've read the doc but I still have a doubt : is the post-*receive hook called automatically right after I perform a push on a branch of this repository from my local copy of the repository, please ?	08:35
energizer	cbreak: ok sounds like a branch is the way to go, awkward as it is	08:35
	→ sbeyer joined	08:35
energizer	cbreak: wait..an orphan branch?	08:37
cbreak	yeah	08:37
	one that doesn't have any history in common	08:37
	→ mobidrop joined	08:37
energizer	how do i make a commit that has no history?	08:37
	← figuinhas left	08:38
cbreak	git checkout --orphan newbranchname	08:38
	→ figuinhas joined	08:38
	→ throsturx joined	08:39
	→ CCR5-D32 joined	08:40
	→ ChrisLane joined	08:41
	← figuinhas left	08:44
	→ feriman joined	08:45
	→ zlogan joined	08:47
	→ inkbottle joined	08:47
	← zebrag left	08:47
energizer	cbreak: that's interesting, thanks	08:50
jomofcw	Anyone knows about my question, please ?	08:51
	I think that's my hypothesis is correct, but I would likle to be sure.	08:51
cbreak	jomofcw: man githooks should have detailed descriptions	08:52
gitinfo	jomofcw: the githooks manpage is available at https://gitirc.eu/githooks.html	08:52
jomofcw	[I can be wrong as I'm oftenly wrong concerning GIT :D]	08:53
	→ selckin joined	08:53
	← gitter1234 left	08:53
	→ dartmouthed joined	08:54
jomofcw	cbreak hello ! Thanks for your answer ! Yup, i've read it already. What I understand from it (I'm not a natal english speaker) is that it detect when a push is done for a branch and then it's called after an "automatic pull".	08:54
cbreak	it's not really detecting, git will call the hooks on its own	08:55
	→ enoq joined	08:55
cbreak	pull isn't really involved	08:56
jomofcw	cbreak OK ! Thanks.	08:58
	Here is the hook that was wrote by the previous dev : https://gist.github.com/jomofcw/c7c6b827a212be5fc9d93acf6d56c135	08:58
	I don't understand what "git --work-tree=$FPV_PROD_DIR --git-dir=$GIT_DIR checkout -f master" is there for :/...	08:58
cbreak	to "deploy" the code?	08:59
	!deploy	08:59
gitinfo	Git is not a deployment tool, but you can build one around it (in simple environments) or use it as an object store(for complex ones). Here are some options/ideas to get you started: http://gitolite.com/deploy.html	08:59
jomofcw	I absolutely don't use GIT for deployement curently :s.	08:59
	I don't know what it means in GIT context so... I'll read this, thanks.	09:00
	← feriman left	09:00
	← mobidrop left	09:01
jomofcw	The problem I curently have is that, when I make a push on my master from my local repository, it overwrite files on my PROD server that are covered by rules in my .gitignore... and i seriously can't understand why :/.	09:01
	→ feriman joined	09:02
	→ dre_ joined	09:03
selckin	if you are using this hook then you are using git for deployment, so either remove it, or figure out what it does (like deploy (= overrite files))	09:03
	→ SerialSmile joined	09:04
jomofcw	selckin hello ! That's what I'm curently trying to do "understanding how it works". But seems to be hard to me :/.	09:05
	And the doc doesn't make me more confident as it says : "But it's hard to explain why.".	09:05
selckin	well the checkout cmd, puts the content of master in $FPV_PROD_DIR	09:06
jomofcw	"So if you run into trouble, it becomes a game of trying to dig out of you what exactly you did in what sequence. And if you think that's easy you've never done tech support :)"	09:06
	selckin all the content, including those covered by .gitignore ?	09:06
selckin	so you say you don't use this anymore, since they guy left i assume, so remove the hook	09:06
jomofcw	selckin we still use it as it's the deploying process.	09:07
selckin	jomofcw: whats committed, gitignore only counts for untracked files	09:07
	jomofcw: then you lied	09:07
jomofcw	selckin then I express it the bad way. Anyway lier or not I need to still use it, but I need to understand and fix it :s.	09:07
jast	the line you mentioned seems to be the one that overwrites the files in the production dir	09:08
	→ figuinhas joined	09:08
selckin	if you do 'git --git-dir=$GIT_DIR status' in $FPV_PROD_DIR does it show untracked files or anything ?	09:09
jomofcw	Hello jast ! Yup, with selckin explanation I get it. Now I need to have a solution to kinda get it out of the master tree files.	09:09
selckin	maybe you need a git clean in there to remove stale files	09:09
	or you know, not use git to deploy	09:09
jomofcw	That's the worst thing I've no command line (SSH or whatever) access to the PROD server :/...	09:09
jast	oh, you were asking about .gitignore, right?	09:09
	→ Achylles joined	09:10
jast	.gitignore only ever applies to untracked files	09:10
	so if you have a file that would be ignored in a git commit, it's going to be used anyway	09:10
	actually let me rephrase to make that clearer	09:10
jomofcw	jast OK, that's the problem.	09:10
selckin	then remove them, and commit that	09:10
jast	if a file is already added to git and you have an ignore pattern for it, that pattern doesn't do anything	09:10
jomofcw	selckin yea i'm concidering the fact to stop use that things :s and deploy manually in PROD as I do for my others project :s.	09:11
jast	if you need to exclude a few files from the deploy, you may have to do something more complicated in the script, instead of just using a big 'checkout'	09:11
	← dimi1947 left	09:12
jomofcw	jast I get it. And as i'm not really confident with all those things, I may just stop to use this process imo...	09:13
	Anyway thanks for your help, both of you jast and selckin (and cbreak on start ^^). I understand how it works now and why it's a bad idea to keep using it.	09:15
	→ _ADN_ joined	09:15
	→ makaveli7 joined	09:17
	← Helenah left	09:18
	→ Helenah joined	09:20
Dirak	Hi Git, is there a command for stashing changes on a branch or for naming a stash?	09:20
_ikke_	Dirak: stashing on a branch is just comitting it.	09:21
	← figuinhas left	09:22
jast	there's "git stash branch"	09:22
	← SerialSmile left	09:22
_ikke_	That's for restoring the stash, not creating it	09:24
	→ dviola joined	09:24
jast	whoops :)	09:24
Dirak	ah, good call to just use a commit	09:24
jast	I like to add "[WIP]" in front of the commit message when I do that	09:25
	→ dimi1947 joined	09:25
	→ mra90 joined	09:26
selckin	git reset HEAD^ to undo the last commit	09:27
mra90	how can I change commit hash of several consecutive commits	09:27
	→ Syzygy_ joined	09:27
selckin	why do you want to do that	09:27
mra90	because when I rebase other branch on top of that these commits instead to be the first ones are moved down the list	09:28
	→ settermjd joined	09:29
osse	rebase doesn't care about the hash for that	09:30
selckin	you can manually resort commits with interactive rebase, not sure what you're trying to do	09:30
osse	sounds like you want to rebase in the other direction	09:30
	← maryo left	09:30
mra90	osse: so git takes only the date?	09:31
	in this case I need to change thge date then	09:32
selckin	it uses the content	09:32
	can't change it	09:32
mra90	as I have a stituation i which my newest commits from branch A are moved down the list after the rebase witch branch B which has never commits	09:32
Dirak	The hash is based off the contents of the diff. This side project uses this feature to make git into a blockchain https://github.com/grahamjenson/gitpow	09:32
	← beastDiv_ left	09:32
	← deltab left	09:33
	← kleisli left	09:33
osse	mra90: rebase doesn't care about dates and hashes. it takes commits from one branch and puts then on top of the other. that's all	09:34
	maybe you want to rebase A onto B instead of the other way around	09:34
_ikke_	"to make git into a blockchain": It already is a blockchain :)	09:37
	← settermjd left	09:37
	← Achylles left	09:40
jast	not exactly	09:40
	→ felipe1982 joined	09:41
	→ deltab joined	09:42
jast	and by the way, the hash is not "based off the contents of the diff". the diff isn't even stored anywhere. it's based off the contents of the commit.	09:42
Dirak	isn't a commit represented as a patch aka a diff?	09:43
jast	rebase, however, uses patch IDs which are based on the diff	09:43
	no	09:43
	in terms of data structures, a commit represents the actual file contents, not changes to the files	09:43
	each commit references a tree object, which references other tree objects (for subdirs) and blob objects (file contents)	09:44
_ikke_	so a commit is a complete snapshot	09:44
	→ T_UNIX joined	09:46
jast	part of the trick is that if a file doesn't change from one commit to another, both end up referencing the same blob object	09:47
	but for git to figure out whether a file changed in a commit, it needs to compare that commit's contents to the parent commit's	09:48
	← j7k6 left	09:48
	→ dionysus69 joined	09:49
	← felipe1982 left	09:49
	← Dirak left	09:50
	→ j7k6 joined	09:50
	← j7k6 left	09:51
	→ Achylles joined	09:52
	→ felipe1982 joined	09:52
	→ fphilipe_ joined	09:53
	← dre_ left	09:53
	→ shored joined	09:54
	→ j7k6 joined	09:54
	← Joel left	09:55
	→ crose joined	09:55
	← johnny56 left	09:56
	→ jetchisel joined	09:56
	← j7k6 left	09:58
	→ johnny56 joined	09:58
	→ j7k6 joined	09:58
	← j7k6 left	09:58
	→ shady_traveller joined	10:00
	→ j7k6 joined	10:00
	← j7k6 left	10:02
	→ mobidrop joined	10:02
	→ Shentino joined	10:02
	→ j7k6 joined	10:02
	← fphilipe_ left	10:05
	← dimi1947 left	10:05
	→ Dirak joined	10:06
	← j7k6 left	10:06
	→ Joel joined	10:07
	→ beastDiv_ joined	10:08
	← Dirak left	10:11
	→ dimi1947 joined	10:12
	→ palooka-jock joined	10:13
	→ j7k6 joined	10:14
	→ fphilipe_ joined	10:18
	← Trieste left	10:23
	← fphilipe_ left	10:23
	→ Trieste joined	10:24
	→ barteks2x joined	10:25
	pandakekok9 → pandakekok9\|away	10:25
	← tsdh left	10:27
	→ Aquazi joined	10:27
	→ Brainium joined	10:29
	← jab416171 left	10:33
	← theorangeone left	10:33
	→ theorangeone joined	10:34
	← Achylles left	10:37
	→ jab416171 joined	10:38
	← beastDiv_ left	10:39
	→ leprechau joined	10:40
	← leprechau left	10:43
	pandakekok9\|away → pandakekok9	10:44
	→ sz0 joined	10:47
	→ figuinhas joined	10:53
	→ felipe1982__ joined	10:56
	← j7k6 left	10:57
	→ j7k6 joined	10:57
	← jaziz left	10:58
	← felipe1982 left	11:00
	← _xor left	11:00
	← felipe1982__ left	11:01
	→ royal_screwup21 joined	11:02
	→ _xor joined	11:06
	→ gloomy joined	11:08
	→ Achylles joined	11:11
	← wrksx_ left	11:18
	→ wrksx_ joined	11:18
	← palooka-jock left	11:19
	→ palooka-jock joined	11:19
	← palooka-jock left	11:21
	→ gnufr33d0m joined	11:23
dzho	the diffs that are displayed via, for example, git diff, are computed on the fly, not presented as already-available stored items	11:23
	← Trieste left	11:23
dzho	oh, gone, and I'm very late, but my understanding is out there to be critiqued so I guess it's still a win, of sorts :P	11:24
	→ palooka-jock joined	11:25
_ikke_	dzho: it's completely correct fyi	11:28
dzho	:)	11:31
	← hussam left	11:37
	→ Trieste joined	11:37
	← lacrymology left	11:46
	← Trieste left	11:46
	← iz16 left	11:50
	← palooka-jock left	11:51
	→ palooka-jock joined	11:52
	→ Trieste joined	11:53
	← misuto left	11:53
	→ iz16 joined	11:54
	← dviola left	11:54
	← bvdw left	11:55
	→ bvdw joined	11:55
	← NODE left	11:56
	→ NODE joined	11:59
	← Achylles left	11:59
	← NODE left	12:00
	→ NODE joined	12:02
	← NODE left	12:02
	→ NODE joined	12:05
	← NODE left	12:06
	→ NODE joined	12:08
	→ Achylles joined	12:11
	← milkt left	12:14
	→ milkt joined	12:14
	← pandakekok9 left	12:16
	← xco left	12:19
	→ pandakekok9 joined	12:21
	← sbeyer left	12:24
	→ zebrag joined	12:25
	← inkbottle left	12:26
	← makaveli7 left	12:32
	→ narcos joined	12:33
	→ xco joined	12:36
narcos	Hi all. I've forked a project, made a pull request, and made loads of commits after various requests for changes. During my testing I accidentally committed a bunch of .min.js files. I'd like to remove/revert these. What's the best way to do this? Thank	12:36
	← xco left	12:37
	← akemhp left	12:37
	→ lacrymology joined	12:37
	→ akemhp joined	12:39
_ikke_	narcos: interactive rebase	12:40
	!fixup	12:40
gitinfo	So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!	12:40
narcos	Thanks	12:41
	← gloomy left	12:41
	→ xco joined	12:41
	← xco left	12:42
oxymoron93	lazar	12:42
	that is not password ^ ...luckily	12:43
	→ xco joined	12:43
narcos	Hmm looks like git checkout <goodhash> <filename> might be the solution	12:43
	Or "git reset <goodhash> <filename>"	12:43
	← crose left	12:43
	← narcos left	12:53
	← pandakekok9 left	12:55
	→ pandakekok9 joined	12:57
	→ jchevalier joined	12:58
	← mra90 left	13:00
	→ oatmealraisin joined	13:02
	← pandakekok9 left	13:05
	→ mowcat joined	13:06
	← xco left	13:09
	→ pandakekok9 joined	13:09
	→ xco joined	13:15
	→ sgen joined	13:23
	→ shabius_ joined	13:24
	← shabius left	13:26
	→ CodeSlingerPaul joined	13:27
	← pandakekok9 left	13:29
	← yena left	13:32
	→ yena joined	13:32
	← feriman left	13:32
	→ hussam joined	13:33
	← royal_screwup21 left	13:34
	→ beastDiv_ joined	13:34
	← beastDiv_ left	13:34
	← figuinhas left	13:42
	→ figuinhas joined	13:42
	← jchevalier left	13:46
	→ leprechau joined	13:49
	→ rgr joined	13:54
	← hahuang65 left	13:55
	→ hahuang65 joined	13:56
	← Foxboron left	14:00
	→ jackedin joined	14:01
	→ Janni joined	14:02
Janni	Hello! I would like to set up a system for notifying all developers (for example of important API changes).	14:03
	→ mra90 joined	14:03
mra90	I have ap roblem with rebase of branch A and B	14:04
Janni	The scenario is rather simple. There are a number of people working on the same project. Sometimes a change occurs of which everybody should take notice. (But no-one wants to send around emails for this purpose).	14:04
mra90	branch A is my branch on which I have like 5 commits I want to upstream	14:04
	but beforee I do so I need to rebase branch A with branch B	14:04
	← gnufr33d0m left	14:05
mra90	Now if I do rebase B on A my changes are down the list of commits co I can not upstream them separately	14:05
Janni	I'm thinking of a CHANGES file which is partially displayed whenever does a checkout or a pull, plus some logic that remembers what has already been shown.	14:05
	I was hoping that all this could be implemented in a simple hook. Anybody has ever attempted something like this?	14:06
	Alternatively (preferably) I'd like to mark certain commits of which developers should take notice.	14:06
bremner	some kind of out-of-band notification whether email / irc / slack seems like the obvious solution	14:07
	← rgr left	14:08
jast	mra90: !repro	14:08
gitinfo	mra90: [!transcript] Please paste (using https://gist.github.com/ or similar) a transcript ( https://git.io/viMGr ) of your terminal session so we can see exactly what you see	14:08
Janni	bremner: Right. That might be the right solution for some people. Not for me. I need something more automatic.	14:08
pandem1	Janni: do you use lab or hub or something similar?	14:10
Janni	pandem1: github	14:10
pandem1	i don't think git hooks work on github repos	14:10
	→ lucasb joined	14:11
jast	checkout hooks would be local, but then you still have to install them in each clone	14:11
Janni	pandem1: I was thinking of a pure git + scripting solution. Something entirely local.	14:11
	→ fphilipe_ joined	14:11
	← JackMc left	14:11
	→ JackMc joined	14:11
jast	(hooks never install themselves automatically)	14:12
Janni	jast: I'll try out what I can do with post-checkout. (We already use hooks, so that's fine.)	14:13
mra90	I think it is happening because these commits has been reverted some time ago	14:13
	and it looks like git things there are the same, even though I have chnaged the comit hash	14:14
	s/there.they	14:14
jast	rebase doesn't look at commit hashes	14:14
	essentially it compares the diffs	14:14
	so if you have old commits that have the same diffs, rebase will tend to make the new ones disappear	14:14
	though I think that really only should happen if those 'old commits' are part of what the rebase brings in	14:15
	→ jackedin_ joined	14:15
mra90	and what is the diff?	14:16
	they have been applied to branch B but then rewerted	14:16
	now when I rebase B on branch A these commits get covered (down the list) aftert he revbase	14:17
	→ prepend joined	14:17
	← jackedin left	14:17
jast	what you say 'rebase B on A', which command are you running and which branch is checked out at the time?	14:18
	→ localhorse joined	14:18
mra90	so I am checked out on branch A and go with "rebase B"	14:18
	← CCR5-D32 left	14:18
mra90	afterwards my newset commits from branch A (those i wanted to upstream) gets "covered" by these form branch B	14:19
	and this is not what I wanted	14:19
jast	all right, that's what I thought... the 'official' way to say that is "rebase A against B", and I'm not completely sure what people mean when they use another way of saying it :)	14:19
	← figuinhas left	14:20
	← gjnoonan left	14:20
mra90	so what can I do about it?	14:20
jast	but yeah, as I said, rebase compares the changes in the commits to see which commits it doesn't need to apply again	14:20
	→ gjnoonan joined	14:20
	→ figuinhas joined	14:20
jast	in this case, the reverts are immaterial since git looks at each commit individually... you know that that commit got reverted, but git doesn't	14:21
	or at least it doesn't know what your sequence of commits means	14:21
	← jackedin_ left	14:21
mra90	yes it looks like the case	14:21
jast	so I guess the easiest way would be to do the same thing rebase does, but manually	14:21
mra90	rebase -i is what you suggest?	14:22
jast	A has nothing new other than those commits you're talking about?	14:22
	→ Foxboron joined	14:22
mra90	Yes only these reverted commits are there	14:22
jast	all right, then one thing you could do is this:	14:22
	note down the ID of the base commit (immediate parent of your set of commits) and the topmost commit	14:22
	'git reset --hard B'	14:23
	'git cherry-pick base..topmost'	14:23
	← bluezone left	14:24
	→ bluezone joined	14:24
	→ jackedin joined	14:24
mra90	but I should not alter B	14:25
	as this is master	14:25
jast	yeah, all of this is done on A	14:27
	← tcsc left	14:28
	→ tcsc joined	14:28
	→ jetpack_joe joined	14:28
mra90	ok I will give it a try	14:29
	← Janni left	14:30
	→ Anticom joined	14:30
	← enoq left	14:32
	→ enoq joined	14:35
	→ crose joined	14:35
mra90	how can I skip hooks while cherry picking	14:40
	I am thinking about --no-verify equivalent	14:40
	← oatmealraisin left	14:43
	→ Xiti` joined	14:47
	← Stoot left	14:48
	← Xiti left	14:51
mra90	ok I manualy turn pre commit hooks	14:53
	off	14:53
Abdullah	if you have multiple origins, is it normal, I don't get commits I have made localy in git status?	14:53
osse	no	14:53
Abdullah	or I'm missing something in gitconfig?	14:53
	in a different repo I have only one origin and it shows me	14:54
osse	what do you mean exactly?	14:54
	→ feriman joined	14:54
Abdullah	osse: I have multiple origins in one repo. when I do commits localy and check it with git status, I don't get like commits made and need to pushed (I mean the count)	14:55
osse	that is not related to that	14:55
	what branch do you want to compare the current one with?	14:55
	→ thiago joined	14:56
	← Trieste left	14:57
	→ Trieste joined	14:57
Abdullah	osse: https://i.imgur.com/5HOoTDk.png	14:58
	← fphilipe_ left	14:59
osse	yes I know what you mean	14:59
	What do you want the second git status to say?	14:59
	"Your branch is ahead of '__________' by 1 commit" fill in the blanks	15:00
Abdullah	in the left one, status should tell me I have made commits	15:00
	osse: yeah that is what I want.	15:00
osse	But what is _________ ?	15:00
Abdullah	master	15:01
	origin	15:01
osse	git branch -u origin/master	15:01
	done	15:01
	→ thebope joined	15:02
	← mowcat left	15:02
Abdullah	Thanks osse I think I did not set it after deleting the history by creating orphan branch then renaming it master	15:02
osse	That sounds correct	15:03
	When you create a branch from another then git sets it up automatically	15:03
	But not in this case	15:03
	→ SteffanW joined	15:06
Abdullah	osse is there something wrong in this function? purgegit () { git checkout --orphan new ; git add -A; git commit -am 'initializing it' ; git branch -D master ; git branch -m master ; git push -f origin master }	15:06
	← Noti left	15:07
Abdullah	I wrote it for purgin the history	15:08
osse	the git add -A step seems unecessary, but otherwise it looks good to me	15:09
Abdullah	osse: All those files previously tracked will be pushed without doing git add -A again?	15:09
osse	yes	15:10
	they are all staged	15:10
Abdullah	Thank you.	15:10
osse	git add -A might remove or add more files	15:10
	add -u to git push to set up the tracking	15:11
Abdullah	okay	15:11
	git push -uf origin master	15:11
	and this was the answer for my first question :-)	15:13
	← omnireq_ left	15:15
	→ omnireq_ joined	15:15
	← sQVe left	15:17
	← figuinhas left	15:17
	→ figuinhas joined	15:18
	→ mauz555 joined	15:18
	← xco left	15:22
	YOU_SUCK_HOFFMAN → Hoffman	15:23
	→ xco joined	15:23
	← mauz555 left	15:24
	→ fahadash joined	15:24
	← enoq left	15:24
	→ flaviodesousa joined	15:25
	→ maxzor joined	15:26
	← xco left	15:26
maxzor	Hello, I can "ssh -vT to [email@hidden.address] but not "ssh-vT 'user'@github.com" : any clue?	15:27
	→ xco joined	15:27
	← xco left	15:27
maxzor	And I cannot push without logins...	15:28
	→ xco joined	15:28
	← xco left	15:28
	← dedowsdi left	15:28
	→ xco joined	15:29
	← xco left	15:29
	→ xco joined	15:29
	← xco left	15:30
maxzor	ah nvm, have to put ssh url not http url	15:30
osse	maxzor: you're only supposed to use [email@hidden.address]	15:30
	→ xco joined	15:30
	← xco left	15:30
	→ xco joined	15:31
	← xco left	15:31
	← omnireq_ left	15:31
	→ xco joined	15:32
	← xco left	15:32
	→ xco joined	15:32
	← xco left	15:33
	→ gitter1234 joined	15:33
	→ xco joined	15:34
	← xco left	15:34
	→ xco joined	15:35
	← xco left	15:35
	→ xco joined	15:37
	→ kleisli joined	15:37
	← xco left	15:37
	→ xco joined	15:37
	← xco left	15:37
	→ xco joined	15:38
	← xco left	15:38
	→ xco joined	15:39
	← xco left	15:39
	→ xco joined	15:40
	← xco left	15:40
	→ nerdwoman joined	15:40
	→ xco joined	15:41
	← xco left	15:41
	← Achylles left	15:41
	→ xco joined	15:42
	← xco left	15:42
	← lacrymology left	15:42
	← floppydh left	15:43
	→ xco joined	15:43
	← xco left	15:43
	→ xco joined	15:44
	← xco left	15:44
	→ xco joined	15:44
	← xco left	15:45
	→ xco joined	15:46
	← xco left	15:46
	← shored left	15:47
	→ shored joined	15:47
	→ xco joined	15:47
	← xco left	15:47
	→ xco joined	15:47
	← xco left	15:48
	→ xco joined	15:49
	← xco left	15:49
	← shored left	15:49
	← thiago left	15:50
	→ xco joined	15:50
	← xco left	15:50
	→ shored joined	15:51
	→ xco joined	15:52
	← xco left	15:52
	→ xco joined	15:52
	← xco left	15:52
	→ Achylles joined	15:53
	→ xco joined	15:57
	← xco left	15:57
	→ mat001 joined	15:58
	→ xco joined	15:58
	← xco left	15:58
	→ xco joined	15:59
	← xco left	15:59
	→ xco joined	16:00
	← xco left	16:00
	→ xco joined	16:01
	← xco left	16:01
	→ xco joined	16:02
	← xco left	16:02
	→ greggerz joined	16:02
	→ xco joined	16:03
	← xco left	16:03
	→ xco joined	16:04
	← xco left	16:04
	← ttoocs left	16:04
	← comptroller left	16:05
	→ Stoot joined	16:05
	← brettgilio left	16:05
	→ xco joined	16:05
	← xco left	16:05
	→ xco joined	16:06
	← xco left	16:06
	← Achylles left	16:06
	→ xco joined	16:07
	← xco left	16:07
	→ xco joined	16:07
	← xco left	16:08
	→ xco joined	16:09
	← xco left	16:09
	→ brettgilio joined	16:10
	→ xco joined	16:10
	← xco left	16:10
	→ xco joined	16:11
	← mobidrop left	16:11
	← xco left	16:11
	← sgen left	16:11
	→ xco joined	16:12
	← xco left	16:12
	→ xco joined	16:13
	← xco left	16:13
	→ omnireq joined	16:14
	→ xco joined	16:14
	← xco left	16:14
	→ lacrymology joined	16:14
	← SteffanW left	16:15
	→ duderonomy joined	16:15
	→ g4570n joined	16:15
	→ gschanuel joined	16:15
	→ xco joined	16:15
	← xco left	16:15
	→ fphilipe_ joined	16:16
	→ xco joined	16:16
	← xco left	16:16
	← jackedin left	16:17
	→ xco joined	16:18
	← xco left	16:18
	→ Achylles joined	16:18
	→ xco joined	16:19
	← xco left	16:19
	→ causasui joined	16:19
	→ xco joined	16:19
	← xco left	16:19
_ikke_	maxzor: github recognizes you through your ssh key	16:20
	→ xco joined	16:20
	← xco left	16:20
maxzor	_ikke I'm ok, thanks	16:20
	→ troulouliou_dev joined	16:21
	→ xco joined	16:22
	← xco left	16:22
	→ xco joined	16:22
	← xco left	16:23
	→ comptroller joined	16:23
	→ xco joined	16:23
	← xco left	16:23
	→ xco joined	16:25
	← xco left	16:25
	→ durham joined	16:25
	→ xco joined	16:26
	← xco left	16:26
	→ xco joined	16:26
	← xco left	16:27
	→ xco joined	16:27
	← xco left	16:27
	← fphilipe_ left	16:28
	→ xco joined	16:29
	← xco left	16:29
	→ xco joined	16:30
	← xco left	16:30
	→ xco joined	16:30
	← xco left	16:30
	← fk left	16:31
	← Achylles left	16:32
	→ xco joined	16:33
	← xco left	16:33
	→ xco joined	16:34
	← xco left	16:34
	→ xco joined	16:34
	← xco left	16:34
	→ paws_ joined	16:35
	→ Zicklr joined	16:35
	← xcm left	16:35
	→ xco joined	16:36
Zicklr	Hello	16:36
gitinfo	Zicklr: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.	16:36
	← xco left	16:36
	→ xcm joined	16:36
Zicklr	Hi	16:37
	→ xco joined	16:37
	← xco left	16:37
	→ goh joined	16:38
	← paws_ left	16:39
goh	Hi. I have a bunch of staged changes. Is there a way to interactively commit just some of these changes?	16:39
	→ paws_ joined	16:39
goh	i.e. something like `add -p` except for changes that are already staged?	16:39
	→ thiago joined	16:39
	→ xco joined	16:39
	← xco left	16:39
	→ fk joined	16:39
_ikke_	There is git commit -p, but I'm not sure whether that would ignore what is already staged	16:40
	← zlogan left	16:40
	→ xco joined	16:40
	← xco left	16:40
_ikke_	git commit -- <file> would just commit that file, regardless of what's staged, so the idea exists already	16:40
goh	`git commit -p` seems to just commit all the staged changed :\|	16:41
	→ xco joined	16:41
_ikke_	Yeah, that's what I figured	16:41
	← xco left	16:42
goh	I know I can just unstage the changes and `add -p`, but I'd rather not unstage a large mass of changes.	16:42
	Not having any copy makes me nervous.	16:42
	I guess I can also commit them, then reset.	16:42
	then `add -p` and `commit --amend`	16:43
_ikke_	You could make a copy of .git/index	16:43
	(as a rude method)	16:43
goh	I'm polite, sorry	16:43
_ikke_	s/rude/crude/	16:43
	← maxzor left	16:43
goh	I'm refined, sorry	16:43
	=)	16:43
_ikke_	:D	16:43
	← duderonomy left	16:44
	→ xco joined	16:45
	← xco left	16:46
	← dionysus69 left	16:46
	→ xco joined	16:46
	← xco left	16:46
	→ xco joined	16:47
	← xco left	16:47
	→ xco joined	16:48
	← hsiktas left	16:48
	→ hsiktas joined	16:48
	→ duderonomy joined	16:48
	→ IAmNetx joined	16:50
	← TomyWork left	16:50
	← dimi1947 left	16:50
	← Zicklr left	16:50
	← duderonomy left	16:50
	→ duderonomy joined	16:54
	← shady_traveller left	16:55
	← Brainium left	16:55
	→ zlogan joined	16:56
	← figuinhas left	16:58
	→ boombatower joined	16:58
	→ figuinhas joined	16:58
	← Hello71 left	17:00
	← flaviodesousa left	17:01
	→ troulouliou_div2 joined	17:04
	← sinhue left	17:05
	← oxymoron93 left	17:05
	→ Hello71 joined	17:05
	← tag left	17:08
	→ z\|bandito joined	17:08
	← durham left	17:08
	→ tag joined	17:08
	← omnireq left	17:08
	← troulouliou_div2 left	17:12
	→ howdoi joined	17:12
	→ troulouliou_div2 joined	17:12
	→ omnireq joined	17:13
	← mat001 left	17:13
	← xco left	17:17
	→ mowcat joined	17:17
	→ nickodd joined	17:18
	→ moldybits joined	17:18
	← Khisanth left	17:21
	→ omnireq_ joined	17:23
	→ mat001 joined	17:23
	← omnireq left	17:24
	← Mattiaslndstrm left	17:26
	→ orbyt_ joined	17:26
	→ omnireq joined	17:28
	← omnireq_ left	17:30
	→ orbyt_v7 joined	17:31
	→ Khisanth joined	17:34
	→ sebastian_nielse joined	17:40
	→ humanBird joined	17:40
sebastian_nielse	$ git remote -v	17:41
	← kalbasit left	17:41
humanBird	how do i clone a local repository to a separate branch in a different directory?	17:41
sebastian_nielse	$ git push origin master [email@hidden.address]	17:41
	→ kalbasit joined	17:41
humanBird	i need to have 4 different working directories all branched off of mastre	17:41
sebastian_nielse	Does anyone know why I get a refs error?	17:41
	git push origin master works just fine somehow	17:41
	← shabius_ left	17:41
	← _ADN_ left	17:43
_ikke_	sebastian_nielse: what is the exact error?	17:43
sebastian_nielse	error: src refspec [email@hidden.address] does not match any	17:43
	https://gyazo.com/9b6cc5352c7bb97c3d7e9687c29fc204	17:44
_ikke_	sebastian_nielse: that command makes little sense :)	17:44
	Why do you specify both origin and the actual remote url?	17:44
sebastian_nielse	because I got two rsa set up	17:45
	This screenshot shows that "git push origin master" works	17:45
	https://gyazo.com/aa7e98eaa623ef66f68cb1178b2eda66	17:45
	I got two RSAs, one is for work, the other is for personal stuff.	17:46
	← jomofcw left	17:46
_ikke_	I assume you mean ssh keys	17:46
sebastian_nielse	Yes	17:46
	→ stealth[] joined	17:46
humanBird	hrm only see git clone info for urls/github/gitlab etc. how do you clone a local repository on your hard disk	17:46
sebastian_nielse	I set up a config file, so that I can reference each "user" using [email@hidden.address] or [email@hidden.address] respectively	17:47
humanBird	https://stackoverflow.com/questions/21045061/git-clone-from-another-directory foudn it	17:47
	→ wootehfoot joined	17:48
sebastian_nielse	It did work when I initially tested it out, but, surprisingly, it doesn't work now somehow.	17:48
	I wonder why	17:48
	→ GameGamer43 joined	17:48
	→ tonythomas joined	17:51
	→ xalen7 joined	17:52
sebastian_nielse	Well, I guess I'll figure it out myself when I get to the "internals of git" part of the git pro book, so that I can inspect the refs	17:52
	← bvdw left	17:55
	→ bvdw joined	17:55
	→ spytfyre joined	17:56
	← mat001 left	17:57
	→ dionysus69 joined	17:58
	← bl0m1 left	18:00
	→ bl0m1 joined	18:01
	← zlogan left	18:01
humanBird	wow. git status is not working correctly . Here is the proof: on repository C, push to repository A (bare repository which is master). On repository B which is a clone of repository A, do git status. output is "Your branch is up to date with 'origin/master'"	18:03
	then on repository B do git pull and it then it pulls changes that repository C pushed to A	18:03
	→ shabius joined	18:04
	← figuinhas left	18:04
humanBird	if you're up to date you should not have any unknown changes. You either need updates or you don't. you can't be True and False at the same time. Is this just passive aggressive/anti social user interrface?	18:04
	← chele left	18:05
	→ mat001 joined	18:06
	← maxwilliamson left	18:06
humanBird	wheen it says i'm up to date with origin/master and then i pull and get changes from origin master, this is basically like saying "triangles have 4 corners" a contradiction	18:06
	this is prove-ably incorrect.	18:06
	→ maxwilliamson joined	18:06
humanBird	if we can't get the basics correct like not putting contradictions into a tool, it seems pretty messed up	18:07
	← howdoi left	18:09
	→ howdoi joined	18:09
	← Anticom left	18:10
	← m1dnight_ left	18:11
	→ m1dnight_ joined	18:11
sebastian_nielse	Question: Are "tracking branches" only useful if you work with multiple remotes?	18:12
	→ misuto joined	18:13
_ikke_	No	18:14
	even with a single remote they are usefull	18:14
	→ kpease joined	18:15
_ikke_	otherwise you would have to use FETCH_HEAD all the time after fetching a specific branch)	18:15
	← kpease left	18:15
	← m1dnight_ left	18:18
	→ glook joined	18:18
	→ m1dnight_ joined	18:19
humanBird	https://stackoverflow.com/questions/27828404/why-does-git-status-show-branch-is-up-to-date-when-changes-exist-upstream i'm just gonna say that this is really anti-social output on git's part.	18:20
	of course you are up to date locally all the time	18:20
_ikke_	no, not all the time	18:20
	→ Phylock joined	18:20
_ikke_	If you just fetched	18:20
	then it will show you that you are behind	18:20
	→ nowhere_man joined	18:21
_ikke_	or when you commit, it will show you you are ahead	18:21
humanBird	why even bother with saying "you are up to date"	18:21
	this is completely misleading and bad language	18:21
	up to date with what?	18:21
_ikke_	"What the status is telling you is that you're behind the ref called origin/master which is a local ref in your local repo."	18:21
glook	If I update a .gitignore file on my machine, and then do a pull, is the .gitignore file honored or does it only count for pushes?	18:21
humanBird	and if it's up to date with your local ref then this isn't really useful at all	18:21
_ikke_	glook: neither	18:21
	← thebope left	18:21
_ikke_	it is usefull, but not in the way that you expect it to	18:21
humanBird	it's misleading.	18:22
	up to date with what?	18:22
	if not remote. which is what the user should reasonably expect	18:22
_ikke_	People forget that git is first and formost a decentralized tool	18:22
humanBird	to say that "you don't need the internet" for git is also misleading. what is the percentage of use cases where you need the internet for git	18:23
	→ Soni joined	18:23
humanBird	what is the point of gitlab/github/etc	18:23
Eryn_1983_FL	hey guys i keep trying to switch to a new branch in git repo and i get this error error: unable to unlink old 'application/library/Push.php' (Permission denied)	18:23
Soni	if I have a dirty git repo (with ignore rules) how do I clean it?	18:23
_ikke_	humanBird: those services came only after git was writter	18:23
	written	18:23
	← g00s left	18:24
humanBird	Yes. reality is reality. the message is poor. that's all.	18:24
Soni	git clean -x?	18:24
_ikke_	humanBird: I agree it is confusing	18:24
	But only in a centralized workflow	18:24
humanBird	like even in my above example that doesn't need the internet where you are cloning a bare repo on local where you've elminated the need for the internet,	18:25
	it seems like this message is still uselesss and misleading	18:25
_ikke_	it's not useless	18:25
	after you git fetched, it tells you whether you need to update your local branch to match what you just fetched	18:25
humanBird	it's worse than useless, sorry. it misleads	18:25
	← sebastian_nielse left	18:26
humanBird	not everyone is advanced git +10000000 level	18:26
	who needs to reaed some garbage about deecentralized local ref junk when the message is there in plain site to confuse	18:26
bin_bash	wat	18:26
_ikke_	humanBird: reacting like that does not create any sympathy	18:26
humanBird	i mean the proof was already there and this has been around for years.	18:27
	this is fundamentally broken	18:27
_ikke_	it is not	18:27
	← mra90 left	18:28
humanBird	gonna restate the proof: on repository C, push to repository A (bare repository which is master). On repository B which is a clone of repository A, do git status. output is "Your branch is up to date with 'origin/master'"then on repository B do git pull and it then it pulls changes that repository C pushed to A	18:28
	← thiago left	18:28
humanBird	there is CLEARLY a contradiction	18:28
	"up to date with origin/master" when it ISNT	18:28
	→ thebope joined	18:28
humanBird	and I quote "In Git, "origin" is a shorthand name for the remote repository that a project was originally cloned from."	18:29
	→ massmamba joined	18:29
humanBird	if you're not up to date with origin/master you're not up to date with origin/master. this is a tautology. git says you are when you aren't. a contradiction. i can't prove it further than this.	18:29
	← timwis left	18:29
	← fury left	18:30
	← omnigoat left	18:30
	← Nizumzen left	18:30
	← klow left	18:30
	← sz0 left	18:30
	← jleclanche left	18:30
	← IAmNetx left	18:30
	← jetpack_joe left	18:30
	← Jeebiss left	18:30
	← steveeJ left	18:30
	← kalbasit left	18:30
humanBird	there's no point in discussing it further since it's proven incorrect. whether or not you understand the proof is really irrelevant	18:30
	← OvermindDL1 left	18:30
	← Giant81 left	18:30
	← NeoHamled left	18:30
	← amiti left	18:30
	← cyberpear left	18:30
	← prit left	18:30
	← localhorse left	18:30
	← ToBeCloud left	18:30
	← T_UNIX left	18:30
	← hahuang65 left	18:30
	← spacefrog left	18:30
	← stigz left	18:30
	← mithro left	18:30
	← matti left	18:30
	← lineos left	18:30
	← KnownSyntax left	18:30
	← paws_ left	18:30
	← JackMc left	18:30
	← angular_mike left	18:30
	← mgsk left	18:30
	→ amiti joined	18:30
	← PavelB left	18:30
	← Kaisyu left	18:31
	← rfv left	18:31
	→ prit joined	18:31
_ikke_	"humanBird │ i mean the proof was already there and this has been around for years." This actually proofs it's not fundamentally broken, otherwise it would have been fixed already.	18:31
	← bitonic left	18:31
	← ggherdov left	18:31
	← hussam left	18:31
	← yena left	18:31
	← r3my left	18:31
	→ Giant81 joined	18:31
	→ IAmNetx joined	18:31
	← GameGamer43 left	18:31
	← rougeth left	18:31
	← lolmac left	18:31
	← rknLA left	18:31
	→ CCR5-D32 joined	18:31
	→ spacefrog joined	18:31
nedbat	humanBird: there is a difference between your local "origin/master" and "master" on origin.	18:31
	→ cyberpear joined	18:31
	← afisher left	18:31
	← LasseVK left	18:31
	← altendky left	18:31
	← redrambles left	18:31
	← sebhoss left	18:31
	← rodarmor left	18:31
humanBird	what kind of semantics is this...	18:31
	← howdoi left	18:31
	← Fenhl left	18:31
	← tonythomas left	18:32
humanBird	in any case it's pretty stupid to have this kind of language	18:32
nedbat	humanBird: think of "origin/master" as, the last master I got from origin.	18:32
	→ OvermindDL1 joined	18:32
	→ AfroThundr\|dev joined	18:32
	→ kalbasit joined	18:32
	→ redrambles joined	18:32
	← NODE left	18:33
	→ jleclanche joined	18:33
	← mrbkap left	18:33
	← Karyon left	18:33
	← Telerophon left	18:33
	→ fury joined	18:33
	→ howdoi joined	18:33
humanBird	"otherwise it would have been fixed already." you ever see those 10 year old android bugs that are verified bugs?	18:33
	→ hahuang65 joined	18:33
	→ mithro joined	18:33
	→ steveeJ joined	18:33
	→ zalun joined	18:33
	← lightstalker left	18:33
	→ KnownSyntax joined	18:33
	→ Fenhl joined	18:33
	→ lineos joined	18:33
	→ jetpack_joe joined	18:33
	→ sz0 joined	18:33
	→ yena joined	18:33
	← Zandose left	18:33
	← oh4 left	18:33
	← ec left	18:33
	→ Kaisyu joined	18:34
	→ rfv joined	18:34
	→ Zandose joined	18:34
	→ Karyon joined	18:34
	→ tonythomas joined	18:34
	→ angular_mike joined	18:34
	← mithro left	18:34
	→ altendky joined	18:34
	← howdoi left	18:34
	← sz0 left	18:34
	→ LasseVK joined	18:34
	← Syzygy_ left	18:34
nedbat	humanBird: it is a confusing message. we can help explain	18:34
	← TheJollyRoger left	18:34
	→ thiago joined	18:34
	→ bitonic joined	18:34
	→ disruptivenl joined	18:34
	→ GameGamer43 joined	18:34
	→ ToBeCloud joined	18:34
	→ sz0 joined	18:34
	gitinfo set mode: +v	18:34
humanBird	the message should just be changed	18:35
	→ NODE joined	18:35
nedbat	humanBird: it would be better if the message were, "Your branch is up to date with your conception of origin/master"	18:35
	→ PavelB joined	18:35
	← steveeJ left	18:35
	→ rougeth joined	18:35
	→ rodarmor joined	18:35
_ikke_	humanBird: https://lore.kernel.org/git/xmqq1rs91bo2.fsf@gitster-ct.c.googlers.com/	18:35
	→ afisher joined	18:35
	→ lightstalker joined	18:35
	→ r3my joined	18:35
	← ToBeCloud left	18:35
	→ rknLA joined	18:35
	→ ec joined	18:35
	← sz0 left	18:35
	→ m0viefreak joined	18:35
	← NODE left	18:35
	← misuto left	18:36
	→ Telerophon joined	18:36
	→ Nizumzen joined	18:36
	→ omnigoat joined	18:36
	→ misuto joined	18:36
	→ Jeebiss joined	18:36
	→ paws_ joined	18:37
	← Nizumzen left	18:37
	→ oh4 joined	18:37
	→ ggherdov joined	18:37
	→ mithro joined	18:37
	← omnigoat left	18:37
	→ Newami joined	18:37
humanBird	as such, any reasonable person is going to think "oh i'm up to date with remote" and do something under this assumption	18:37
	→ NODE joined	18:38
humanBird	this is the consequence for new users whether or not you want to argue otherwise.	18:38
	→ TheJollyRoger joined	18:38
_ikke_	such a person might also thing, hey, that command went super fast, did it actually contact the remote?	18:38
	→ omnigoat joined	18:38
	← KnownSyntax left	18:38
humanBird	lol, poeple can perceive network latency and throughput via git commands? that's pretty false in general especially internal git repositories	18:38
	← NODE left	18:39
humanBird	and small repositories	18:39
	← shored left	18:39
	← mat001 left	18:40
	← disruptivenl left	18:40
_ikke_	humanBird: if you want to actually change something, reply to that mail from Junio, preferrably with a patch that suggests what you think is a better message and does not ignore what Junio replied	18:40
	→ mat001 joined	18:40
	← prit left	18:40
	← Giant81 left	18:40
	→ NODE joined	18:41
	→ oatmealraisin joined	18:42
	← r3my left	18:42
	← oh4 left	18:42
	← lineos left	18:42
	← redrambles left	18:42
	← massmamba left	18:43
	← ec left	18:43
	← afisher left	18:43
	← cyberpear left	18:43
	← Newami left	18:43
	← bluezone left	18:44
humanBird	that email thread sounds like "oh we can't change the behavior of an existing bug because people have learnt to rely on the bug"	18:44
_ikke_	It helps to not see it as a bug, but a usability issue for new users	18:45
	But that is not fixed by only taking new users into account, but also existing users	18:46
humanBird	very elitist. guess new users will continue to struggle then.	18:46
_ikke_	No	18:46
nedbat	humanBird: this is a complex thing, it can't be explained in one error message, so it's hard to know the best words	18:46
	humanBird: your ideas will help. Respond to the thread.	18:46
_ikke_	humanBird: A new users stays a new for a short period	18:46
	experienced users will always remain experienced	18:46
	← g4570n left	18:47
_ikke_	humanBird: One feature that git has is advise messages that can be disabled	18:47
	these would be perfect for new users	18:47
	but I'm not sure if such an advise message is warranted for git status (would add a lot more text, which might not be that helpful for new users)	18:48
nedbat	humanBird: frankly, you just seem mad now. There are ways to influence things if you want to.	18:49
humanBird	new users are going to make reasonable assumptions based on that message and do wrong things which is worsee.	18:49
nedbat	humanBird: make that point on the email thread.	18:49
glook	exit	18:50
	:)	18:50
	← glook left	18:50
humanBird	"oh. i don't need to pull or merge since i'm alreeady up to date"	18:50
	3 days later "what the fuck? this merge is too big"	18:51
_ikke_	Just ranting here does not change anything	18:51
	not a bit	18:51
nedbat	humanBird: how can we help?	18:51
	humanBird: we get it: git can be complex and confusing.	18:52
_ikke_	humanBird: propose a change to the mailing list, but make sure that you show that you at least understand the other side as well	18:52
	and the decentralized nature of git	18:52
	→ KnownSyntax joined	18:53
	← xcm left	18:53
	← NODE left	18:54
	← KnownSyntax left	18:54
	→ xcm joined	18:54
_ikke_	just saying "git is stupid, fix this" is not going to get you far (as git is created by humans)	18:54
	→ jhulten joined	18:54
_ikke_	humanBird: I often wonder how a lot of developers would act if they work with physical power tools	18:55
	A lot of these tools are outright dangerous if you don't know how to use them	18:56
	→ NODE joined	18:56
	→ howdoi joined	18:56
_ikke_	(which does not mean I'm against adding usability features)	18:56
	→ KnownSyntax joined	18:56
	→ vicfred joined	18:56
	← NODE left	18:57
	→ plexigras joined	18:57
	→ ec joined	18:57
	→ NODE joined	18:58
	→ cyberpear joined	18:59
	bin_bash does not even get what the complaint is about...	18:59
	← IAmNetx left	19:00
_ikke_	that git status says "Everthing up-to-date" without consulting the remote\	19:00
bin_bash	I guess I just always took it for granted that that's the correct and expected functionality	19:01
	← dartmouthed left	19:01
	← Phylock left	19:01
	→ g00s joined	19:01
humanBird	don't really want to dive into what "everything" means.	19:01
_ikke_	bin_bash: Did you assume that it means it's up-to-date in respect to the remote?	19:01
bin_bash	no, i assumed it meant local	19:01
	← ec left	19:02
bin_bash	until i run a fetch and then it would be against remote (from my understanding)	19:02
	→ Mattiaslndstrm joined	19:02
_ikke_	humanBird: this is what it says for me "Your branch is up to date with 'gitlab/master'."	19:02
	bin_bash: yes, that's correct	19:02
bin_bash	ok just making sure	19:03
_ikke_	The complaint is that new users might assume that it means they have everything from the remote	19:03
bin_bash	if it fetched with the status then it would be very slow and resource-intensive, but status should be quick	19:03
	ahhh	19:03
	is that possibly because people don't separate fetch and merge and just use pull?	19:03
	so they don't know what fetch is/what pull is?	19:04
_ikke_	for example	19:04
	← troulouliou_dev left	19:04
	← thiago left	19:04
_ikke_	And because github and the like are promoting a centralized workflow	19:04
	→ Dirak joined	19:04
	→ Narrat joined	19:05
_ikke_	which clouds the decentralized nature of git (ha, no pun intended)	19:05
humanBird	just think of your manager getting updates from you. "Is that everything?"	19:06
	"you didn't git fetch me. therefore i didn't update you with everything sorry"	19:06
	→ sgen joined	19:06
humanBird	"but you said everything"	19:06
bin_bash	That's an extremely false equivalency	19:06
_ikke_	yes	19:06
humanBird	ok. let's dive into what "everything' mans	19:07
	means* stupid mac keyboard e key	19:07
_ikke_	humanBird: afaict, it's only that git push is saying everything up-to-date	19:07
bin_bash	_ikke_: correct me if I'm wrong, but it's more like if you're in your own isolated room and you're looking at a written checklist in an offline state and you say "Is that all?" But then you go online and download the most recent checklist and it has more items; right?	19:08
	→ fstd_ joined	19:08
_ikke_	Or you have to walk 10 flights of stairs	19:08
humanBird	i'm not even online in my example. i'm cloning from local repository	19:08
_ikke_	humanBird: that does not matter to git\	19:08
	a remote repository is a remote repository	19:09
humanBird	a bare repository on local where you don't need internet	19:09
	→ sz0 joined	19:09
	→ ec joined	19:09
_ikke_	A decentralized workflow is not about the internet	19:09
	→ stigz joined	19:09
	→ klow joined	19:09
	← sz0 left	19:10
_ikke_	It means each repository stand on it's own	19:10
	and can fully function without having to contact any remote repository, weather that's on the internet or not	19:10
spytfyre	humanBird: sorry if I misinterpret your problem but I think reading about remote branches would help https://git-scm.com/book/en/v2/Git-Branching-Remote-Branches	19:10
humanBird	in practice, this concept seems pretty ivory tower.	19:10
_ikke_	No, it's not	19:10
	→ sbeyer joined	19:10
_ikke_	it means that many git operations are fast	19:10
	and it means that you can do your work without interference from others	19:11
	→ T_UNIX joined	19:11
	← stigz left	19:11
	← ec left	19:11
_ikke_	until you are ready to synchronize	19:11
humanBird	you will eventually merge, pull or get changes. isn't it already recommended to get changes often to avoid large merges?	19:11
	→ afisher joined	19:11
humanBird	you are already dependent on other repositories	19:11
_ikke_	humanBird: that's completely up to you	19:11
	no, you are not	19:11
	It's only in a centralized workflow that you are	19:12
	← xcm left	19:12
_ikke_	humanBird: imagine a project like git or linux	19:12
	→ mgsk joined	19:12
	→ moo joined	19:12
	← fstd left	19:12
	fstd_ → fstd	19:12
_ikke_	There are companies that keep complete forks of the linux project	19:13
	→ jamiejackson joined	19:13
	← jamiejackson left	19:13
_ikke_	They are not depending on upstream, except for obtaining the latest linux kernel once in a while	19:13
	← nerdwoman left	19:14
_ikke_	the decentralized nature of git makes that very easy (it does have other challenges of course)	19:14
	→ xcm joined	19:14
	→ jamiejackson joined	19:14
_ikke_	git caters for a lot of workflows	19:15
	so when you want to make changes, you cannot just assume there is only one way to work with git	19:15
	← wootehfoot left	19:15
_ikke_	and yes, git has it's warts	19:16
	people are doing a lot of effort in making git more usable for everyone	19:16
	(most recent example, git splitting out checking into switch and restore)	19:16
	→ thiago joined	19:16
_ikke_	humanBird: On to come back to something I mentioned earlier: do you have an example of git status saying "everything up-to-date"?	19:18
humanBird	that was something else	19:19
	← paws_ left	19:19
	→ stigz joined	19:19
humanBird	Your branch is up to date with 'origin/master' is the message in question	19:19
	→ g4570n joined	19:19
bin_bash	yeah, it is, on your local repo.	19:19
	→ oh4 joined	19:19
humanBird	it seems weird that i have to do a git fetch beforee the git status becomes relevant. then it tells me to pull, which does a fetch and merge, the fetch which i've done already	19:19
bin_bash	not really	19:20
osse	humanBird: feel free to do a merge instead of a pull in that case	19:20
	← Mattiaslndstrm left	19:20
humanBird	the thing is that git status won't tell me what i need to know until i fetch. it really is a problem for new users	19:20
osse	I dislike that message a little bit	19:20
	→ guardian joined	19:21
bin_bash	only if those new users fail to understand that git is decentralized by design, and it's referencing the local repo and not the remote humanBird	19:21
nedbat	bin_bash: humanBird is right that many of those users will fail to understand that. the question is, can the message be improved?	19:23
osse	some GUIs run git fetch periodically. I know SourceTree does.	19:23
	I run git fetch willy nilly pretty often. It never hurts	19:23
_ikke_	osse: it does for git push --force-with-lease	19:23
	(without any explicit lease)	19:24
osse	The few times I've used that I've actually wanted to run a git fetch immediately prior anyway	19:24
	→ subopt joined	19:25
osse	to take last-second stuff into account, if needed	19:25
	but I see your point	19:25
humanBird	that seems like a problem. you are changing state to simply view statee	19:25
	→ Sasazuka joined	19:25
osse	hmm	19:26
	→ Fernando-Basso joined	19:26
osse	so it's a problem that the state is outdated, and it's a problem to change the state... I think you're stuck then :p	19:26
	I don't see it as really changing a state; I see it sort of like pressing Refresh in your browser. The state is out there already	19:27
	← sauvin left	19:27
bin_bash	i like the browser refresh analogy.	19:28
nedbat	humanBird: can you say more about why it's a problem?	19:28
bin_bash	it would be awful if the browser automatically refreshed every time you clicked on an already-loaded tab	19:28
R2robot	mobile browsers refresh when you open it. I hate that	19:29
	← jamiejackson left	19:30
humanBird	in general, viewing state should not involve changing state. also, if git is supposed to be centralized, why fetch so much? you're depending on some other repository already	19:30
nedbat	humanBird: git is decentralized	19:31
	humanBird: you decided to use a remote	19:31
humanBird	i don't see a way around using a remote heh	19:31
_ikke_	git init	19:31
	git commit	19:31
	git commit	19:31
	git commit	19:31
	git branch	19:31
	git commit	19:31
	etc	19:31
nedbat	humanBird: i guess the point is that you can stray arbitrarily from the remote. you decided to stay in close sync.	19:31
_ikke_	no remote required :)	19:31
nedbat	humanBird: you want to fetch so much, because you are focused on keeping the delta different between local and remote.	19:33
humanBird	and to make "git status" have meaning	19:33
nedbat	humanBird: right, we've been trying to get your help with that, but i don't know if you suggested a change	19:33
	humanBird: btw, it has a meaning, you just don't like the meaning.	19:34
	← _salt left	19:35
	← stealth[] left	19:35
	← causasui left	19:36
	← troulouliou_div2 left	19:36
	← throsturx left	19:38
osse	humanBird: git status tells you want branch you're on, whether you have uncommitted changes (if so, which of those are staged) and whether there are untracked files present. all of that gives plenty of meaning (or it should at least) without any remote	19:40
	additionally, it shows the current state when you're merging or rebasing	19:40
humanBird	"with 'origin/master'", let's not forget	19:40
osse	if you're focusing solely on that part of what git status says	19:41
	← guardian left	19:41
osse	you can make it not say that if you want	19:41
	→ redrambles joined	19:41
_ikke_	humanBird: there is nothing wrong with that, except for bad assumptions that that refers directly to what is on origin	19:42
	← nowhere_man left	19:42
bin_bash	humanBird: you just need to get it through your head that it's referring to the local and not the remote, and once you fetch, your local is up-to-date with remote.	19:43
humanBird	origin/master is not a remote?	19:43
	origin seems like a misnomer here then	19:44
_ikke_	origin is the name of a remote	19:44
humanBird	so we arrive at the contradiction	19:44
_ikke_	origin/master is the name of a remote tracking branch, which has the remote it was fetched from prefixed	19:44
	← redrambles left	19:44
R2robot	fork git -> create your magical version -> profit	19:45
nedbat	humanBird: what would you call your notation about the state of origin's master the last time you fetched?	19:45
_ikke_	I think less people would complain if it would say "Your branch is up to date with 'refs/remotes/origin/master'" :D	19:46
	humanBird: origin is a remote, origin/master refers to refs/remotes/origin/master	19:47
	but because people don't want to type refs/remotes/origin/master all the time, it gets shortened to origin/master	19:48
humanBird	it would still be confusing since it has the word "remote" in it	19:48
bin_bash	humanBird: are you a troll?	19:48
R2robot	yes	19:48
	← xelxebar left	19:48
humanBird	and here come the ad hominems.	19:48
	up to date with "something that has the word remote or origin in it" new people will still get confused	19:49
	← misuto left	19:49
	← R2robot left	19:49
humanBird	is it up to date with the "not in my repo because it refers to origin or remote stuff" ?	19:49
	→ misuto joined	19:49
	→ steveeJ joined	19:50
_ikke_	If you wonder if that's the case, than it's a good time to consult the documentation and not assume things :)	19:50
bin_bash	asking if someone is a troll because they're being willfully ignorant is not an ad hominem attack, it's a legitimate question to determine whether it's even worth responding for the entertainment value.	19:50
humanBird	i already understand that git wants me to fetch to find the info I need but the message is still the message that it is	19:51
	it refers to origin/master	19:51
	and this is an emperor has no clothes moment.	19:51
sangy	...emperor has no clothes?	19:51
bin_bash	Yeah, it refers to origin/master on your local repo	19:51
_ikke_	That message is very useful when you fetch to know that you still have to merge something, or when you comitting things, that you still need to push them (or a combination of both)	19:51
bin_bash	Just get it through your thick skull	19:51
_ikke_	that message is not usefull as an indication whether you still need to fetch	19:52
spytfyre	yes it refers to origin/master which is a remote tracking branch so it saying you are up to date with it before fetching is perfectly normal	19:52
	→ lineos joined	19:52
	← justanotheruser left	19:52
_ikke_	humanBird: the problem is that you it's hard to call the remote tracking branch something else when you can have multiple remotes all with the same branch	19:52
	if you have 3 remotes, all with the branch master, then you can have a/master, b/master, c/master	19:53
	You need some way to distinguish them and know from what remote they came	19:53
	so it's quite logical that a remote tracking branch for a branch master that you pulled from a remote origin is called origin/master	19:53
	← xcm left	19:53
_ikke_	and a branch that keeps track of what you fetched from a remote is quite logical called a 'remote tracking branch', but that does not mean that branch is remote	19:55
	There is probably a better way to describe it, but that would also be way more verbose	19:55
	→ xcm joined	19:55
humanBird	so when i "git pull origin", this analogy falls apart.	19:56
_ikke_	it does not	19:56
	it first does a git fetch origin	19:56
	which updates your remote tracking branches	19:56
humanBird	i think the overloading of 'origin' is pretty bad	19:56
_ikke_	then it does git merge	19:56
	humanBird: there is nothing special about the word origin	19:57
nedbat	humanBird: what would you call the "origin/master" branch?	19:57
_ikke_	it's just a default name	19:57
	← T_UNIX left	19:57
_ikke_	If I clone from github, I call my default remote github	19:57
humanBird	nedbat, which one, the remote tracking or the actual one that isn't in your current repo?	19:57
nedbat	humanBird: the remote tracking branch. what would you call it?	19:57
_ikke_	humanBird: origin/master is by default the one that is in your current repo	19:57
	s/by default/per definition	19:58
	← xalen7 left	19:58
humanBird	remote tracking branch is basically a pointer, conceptually speaking. start from there.	19:58
nedbat	humanBird: ok. what would you name it?	19:58
_ikke_	every branch is a pointer	19:58
	← sgen left	19:58
_ikke_	(every branch is a ref, a ref is a pointer)	19:58
humanBird	what do you call the contents of a dereferenced branch, going by this analogy	19:59
_ikke_	humanBird: how would you distinguish what branch you fetched from what remote	19:59
humanBird	the actual changes	19:59
nedbat	humanBird: there are three branches in play here: origin's master, your master, and your origin/master	20:00
	humanBird: it's hard to find names that can express those ideas without being misinterpreted.	20:00
humanBird	that's the problem. origin is ambiguous and the semantic of "origin" seems to imply "not local"	20:00
_ikke_	humanBird: there are no semantics in the name origin	20:00
spytfyre	you are free to change it	20:00
_ikke_	git remote rename origin foobar	20:01
nedbat	humanBird: right. origin isn't local. but you have a local idea of what origin's master was the last time you fetched. what would you call that?	20:01
sangy	I feel every 2/3 months there's somebody walking in going "I FIGURED OUT GIT's CONCEPTS ARE WRONG HEAR ME OUT" and there's a circular discussion for about an hour or two	20:01
nedbat	sangy: we'll pencil in the next one for March.	20:01
humanBird	that means i'm not the only one heh	20:01
	→ causasui joined	20:01
nedbat	humanBird: git is complex. no doubt.	20:01
Dirak	git is a local optima. If I had to bet, Git won't be used 50 years from now	20:02
_ikke_	Dirak: That's the case with most software	20:02
sangy	possibly. I also think that generally the attitude doesn't help with understanding what you find confusing	20:02
nedbat	humanBird: the question isn't whether git is complex, or whether people get confused. the question is: what is an improvement?	20:02
Dirak	The problems with git are small, but they're still there -- monorepo management and dev experience are the two facets I would want to see iterated on	20:02
sangy	interestingly enough, I don't see people with that attitude with go modules or kubernetes heh	20:02
	← gitter1234 left	20:03
_ikke_	Dirak: I think a lot of work is being done to help with monorepos	20:03
	→ erebel55 joined	20:03
erebel55	lo	20:03
Dirak	I don't forsee Go will be used in 50 years from now either	20:03
nedbat	Dirak, _ikke_: maybe i'm lucky: what's a problem with monorepos?	20:03
sangy	erebel55: o/	20:03
_ikke_	It doesn't scale (to MS / google levels)	20:03
erebel55	can anyone tell me how to remove all unstaged files? My repo is in a bad state where I have thousands of untracked files and I can't seem to get back to a clean state. I've tried git clean but it only thinks there are a few untracked files for some reason.	20:04
nedbat	_ikke_: i see, just the size.	20:04
_ikke_	MS wrote gvfs for that	20:04
	erebel55: unstaged or untracked files?	20:04
sangy	erebel55: so try using xdff for git-clean	20:04
	and if they are unstaged you probably want git checkout or git reset	20:05
erebel55	source tree is saying "unstaged files"	20:05
sangy	erebel55: so they are being tracked but you changed them right?	20:05
_ikke_	erebel55: then yes, git clean is not helping with that	20:05
erebel55	I didn't really change them. I was pulling a branch and it was taking forever so I canceled. And now I'm in this broken state where it thinks every file is changed and unstaged	20:06
sangy	you probably canceled a merge/rebase	20:06
	→ justanotheruser joined	20:06
sangy	do you see anything on git status?	20:06
_ikke_	erebel55: then you want git merge --abort	20:06
erebel55	it says there is no merge to abort	20:07
	and git status says no commits yet. nothing added to commit but untracked files present	20:07
	dir	20:08
	woops wrong window	20:08
	→ ethernether joined	20:10
erebel55	shouldn't git status be showing all of these files?	20:11
	← Kaisyu left	20:12
_ikke_	it by default shows only the directory name if everything is untracked	20:12
	→ Nizumzen joined	20:12
_ikke_	in that dir	20:12
sangy	but does it say if it's during a merge/rebase?	20:12
_ikke_	it already sait it was not doing a merge	20:13
	said*	20:13
sangy	well, I don't know if the "woops" wrong window meant "wrong terminal" or so :P	20:13
erebel55	lol that was referring to me typing "dir"	20:14
sangy	ah oh	20:14
erebel55	So what should I run?	20:14
	← spytfyre left	20:15
sangy	erebel55: well, I think what you want is to just git clean -xdff to remove everything that's untracked	20:15
_ikke_	git clean -fdx; git reset --hard HEAD	20:15
	← justanotheruser left	20:16
erebel55	phew that worked :)	20:17
	Thanks guys!	20:17
	oh that deleted my entire repo contents locally	20:18
	← Helenah left	20:22
erebel55	So I don't really understand what happened	20:23
	Why would you think every single file in my repo was changed	20:23
	← teej left	20:23
sangy	not sure. There's many reasons that can happen	20:24
	→ Helenah joined	20:24
erebel55	ah okay, and there wasn't a resolution without deleting everything?	20:25
	→ fphilipe_ joined	20:25
_ikke_	erebel55: without more details, hard to tell	20:25
erebel55	okay, makes sense	20:26
	← comptroller left	20:27
	→ sgen joined	20:27
	← orbyt_ left	20:28
	→ justanotheruser joined	20:29
	← fphilipe_ left	20:30
	← erebel55 left	20:30
	→ matti joined	20:31
	← tonythomas left	20:31
	← thebope left	20:33
	← z\|bandito left	20:35
	→ renatogeh joined	20:36
	← spacefrog left	20:40
	→ comptroller joined	20:41
	→ TheFuzzball joined	20:42
	← m1sdirection left	20:43
	→ stealth[] joined	20:43
	→ m1sdirection joined	20:43
	→ Achylles joined	20:43
	→ ARoxdale joined	20:44
	gitinfo set mode: +v	20:44
	← NODE left	20:44
	→ jaziz joined	20:46
	→ NODE joined	20:46
ARoxdale	Are there any recommend git colour schemes for various terminal colour schemes? I'm using a "green on black" terminal and was wondering if anyone with a similar setup has a recommended colour scheme?	20:47
	→ purplebanana joined	20:48
	→ CpAj0 joined	20:49
purplebanana	Does someone know of a more visual guide of https://git-scm.com/docs/gitworkflows? I'm having a hard time understanding it.	20:50
_ikke_	!gitflow !githubflow	20:50
gitinfo	The description of the gitflow branch workflow model is at http://nvie.com/posts/a-successful-git-branching-model/ while a tool to help implement this workflow is at https://github.com/petervanderdoes/gitflow See http://sethrobertson.github.com/GitBestPractices/#workflow for other workflow suggestions/references	20:50
	[!github_flow] This is the workflow followed by github: http://scottchacon.com/2011/08/31/github-flow.html	20:50
purplebanana	That's not the same thing	20:51
CpAj0	i have two branch: master and my branch, i push to my branch two commit, can i revert this push? without make damage to the master?	20:51
bin_bash	CpAj0: was your branch merged into master?	20:53
CpAj0	bin_bash, no, i have pulled the master then merge the master into my branch made two commit and push to my branch	20:55
bin_bash	if it's not been merged into master then there's no risk to master and there's nothing to do. what do you mean by revert the push?	20:55
humanBird	delete the branch from the remote ?	20:55
	git push -d <remote_name> <branch_name>	20:56
nedbat	purplebanana: that page could definitely use some diagrams	20:57
purplebanana	nedbat: I agree.	20:57
CpAj0	i mean return to the state before the push	20:58
humanBird	git revert or reset	20:58
	depends on if you want to rewrite history	20:58
nedbat	purplebanana: https://www.slideshare.net/ktateish/the-gitworkflows7-illustrated (I haven't looked at it yet)	20:59
	← howdoi left	20:59
purplebanana	nedbat: I'll check it out	21:01
	← TheJollyRoger left	21:02
	← Dirak left	21:04
	← CpAj0 left	21:05
	→ TheJollyRoger joined	21:06
	→ jaziz1 joined	21:06
	← jaziz left	21:10
	jaziz1 → jaziz	21:11
	← nickodd left	21:12
	→ vicfred_ joined	21:12
	← vicfred_ left	21:12
	→ prit joined	21:13
	← subopt left	21:14
purplebanana	nedbat: It might just be me, but I don't find that one really clear	21:14
	← Aquazi left	21:15
	← vicfred left	21:15
nedbat	purplebanana: i looked at the description of pu (which i had never heard of), and it helped me understand.	21:15
	purplebanana: is there a particular question you have?	21:15
	← altendky left	21:16
	← NODE left	21:17
	← stealth[] left	21:17
	→ stealth[] joined	21:18
	→ NODE joined	21:19
	← moo left	21:23
purplebanana	nedbat: I'm new to git and I'm currently reading up on different 'strategies'. I'm trying to understand what should go where and what certain branches are for. Anyways, my time is 10:23 PM so a good night sleep might help ;-). Thanks for your help.	21:24
	→ fphilipe_ joined	21:25
	Squarism → meebo	21:28
	→ mrbkap joined	21:28
	meebo → Squarism	21:28
	→ ferdna joined	21:29
	→ redrambles joined	21:30
	← fphilipe_ left	21:30
	← Codaraxis_ left	21:30
	← prit left	21:33
	← jleclanche left	21:35
	← purplebanana left	21:36
	← solars left	21:37
	← fury left	21:37
	← lucasb left	21:38
	← ARoxdale left	21:39
	← diogenese left	21:42
	← fahadash left	21:43
	← TheFuzzball left	21:47
	→ TheFuzzball joined	21:49
	→ ec joined	21:57
	← ec left	21:58
	← sgen left	22:00
	→ Dirak joined	22:01
	→ diogenese joined	22:01
	← mowcat left	22:02
	→ owlscientist joined	22:06
	← humanBird left	22:07
	← c0c0 left	22:10
	→ courrier joined	22:12
	← bittyx left	22:13
	← xcm left	22:15
	→ xcm joined	22:15
	← hagabaka left	22:16
	← barteks2x left	22:16
	→ hagabaka joined	22:16
	→ donofrio_ joined	22:18
	→ Codaraxis joined	22:18
	→ royal_screwup21 joined	22:21
	← emsjessec left	22:22
	← donofrio left	22:22
Forty-Bot	is there a way to compare commits excluding changes before a certain commit?	22:23
	→ emsjessec joined	22:24
Forty-Bot	I want to view the diff of a patch series where one version has been rebased onto a new version of upstream (along with other changes)	22:24
	← duderonomy left	22:24
Forty-Bot	I tried some of the techniques in https://stackoverflow.com/questions/12465184/comparing-differences-across-a-rebase-in-git, but they don't work in my case	22:25
	→ dviola joined	22:25
Forty-Bot	actually, it looks like range-diff is what I wantr	22:27
	→ orbyt_ joined	22:27
	→ joebobjoe joined	22:29
	← plexigras left	22:29
	← greggerz left	22:29
joebobjoe	why do some people say cherry picking strategy creates artificial conflicts?	22:30
	when cherry picking between the same branches again	22:30
	→ greatgatsby joined	22:32
	→ sgen joined	22:32
	→ duderonomy joined	22:33
	→ bittyx joined	22:33
	→ guardian joined	22:33
	← orbyt_v7 left	22:35
	→ jaziz1 joined	22:35
	→ thebope joined	22:36
	→ zher joined	22:36
	→ orbyt_v7 joined	22:38
	← jaziz left	22:39
	← NODE left	22:42
	→ NODE joined	22:44
	← NODE left	22:45
	← Dirak left	22:45
	← stealth[] left	22:45
	→ NODE joined	22:47
	← NODE left	22:48
	← m1dnight_ left	22:48
	→ alexeyneud joined	22:49
	→ m1dnight_ joined	22:49
	→ NODE joined	22:50
	← Anthaas left	22:52
	→ Anthaas joined	22:52
	→ Dirak joined	22:53
	← Codaraxis left	22:55
	← feriman left	22:57
	→ rafasc joined	23:01
	← foul_owl left	23:01
	→ ToBeCloud joined	23:09
	gitinfo set mode: +v	23:09
	← ToBeCloud left	23:10
	→ NeoHamled joined	23:10
	→ Codaraxis joined	23:10
	← misuto left	23:10
	→ misuto joined	23:13
	← sbeyer left	23:15
	→ Dirak_ joined	23:15
	← alexeyneud left	23:17
	← Dirak left	23:18
	← zher left	23:20
	← donfede left	23:20
	← misuto left	23:20
	→ misuto joined	23:21
	→ Brainium joined	23:22
	→ H2Q joined	23:22
	← greatgatsby left	23:23
	→ Newami joined	23:24
	← Newami left	23:25
	→ dedowsdi joined	23:25
	← thiago left	23:28
	→ foul_owl joined	23:28
	← TheFuzzball left	23:29
	← Fernando-Basso left	23:30
	← steven left	23:37
	← Ragnor left	23:37
	→ Ragnor joined	23:38
	← NODE left	23:38
	→ NODE joined	23:41
	→ steven joined	23:41
	← NODE left	23:42
	→ NODE joined	23:43
	← H2Q left	23:45
	← CodeSlingerPaul left	23:50
	← ColdKeyboard left	23:51
	← ethernether left	23:51
	← royal_screwup21 left	23:54
	← jhulten left	23:54
	← bvdw left	23:55
	→ bvdw joined	23:56
	← thebope left	23:59

Logs Search ←Prev date Next date→

Channels Documentation