IRCloggy #git 2024-01-08

Provider of IRC logs since 2005.

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

	2024-01-08
	← pasky left	00:00
	← humanface left	00:00
	→ jklfds joined	00:03
	→ pasky joined	00:05
	← oriba left	00:07
	← FinnElija left	00:08
	← sahil_ left	00:16
	← erectus left	00:31
	← owuser left	00:41
	← stagas left	00:48
	→ stagas joined	00:49
	← m0viefreak left	00:51
	← jmd left	01:00
	← yuljk left	01:01
	← cbreak left	01:05
	→ irc_user joined	01:05
	← qqq left	01:05
	← rgrinberg left	01:08
	← jklfds left	01:08
	→ cbreak joined	01:10
	→ rgrinberg joined	01:11
	→ yuljk joined	01:16
	→ jklfds joined	01:19
	← Lord_of_Life left	01:23
	→ Lord_of_Life joined	01:24
	← rgrinberg left	01:29
	← fflam left	01:31
	→ fflam joined	01:32
	← xx left	01:33
	← stuv left	01:40
	← miigotu left	01:45
	→ miigotu joined	01:46
	→ meritamen joined	01:52
	← navi left	01:59
	← lucasta left	02:04
	← miigotu left	02:05
	→ Gary_Sinise joined	02:23
	← SuperLag left	02:24
	← thiago left	02:24
	→ SuperLag joined	02:24
	← jklfds left	02:25
	← igemnace left	02:26
	→ miigotu joined	02:26
	← meritamen left	02:36
	→ jklfds joined	02:36
	→ schne1der joined	02:37
	← jklfds left	02:41
	← cbreak left	02:42
	→ igemnace joined	02:44
	→ jklfds joined	02:47
	← jklfds left	02:54
	→ cappy joined	02:54
	→ cbreak joined	02:57
	→ meritamen joined	02:57
	← lunatiq left	03:04
	→ JanC_ joined	03:05
	→ hbautista joined	03:05
	JanC → Guest7764	03:05
	← Guest7764 left	03:05
	JanC_ → JanC	03:05
	→ jklfds joined	03:06
	→ lunatiq joined	03:07
	← lunatiq left	03:07
	→ lunatiq joined	03:07
	← stagas left	03:08
	→ stagas joined	03:09
	← jklfds left	03:11
	← tirnanog left	03:11
	← meritamen left	03:12
	→ jklfds joined	03:12
	← irc_user left	03:13
	→ rgrinberg joined	03:17
	← jklfds left	03:17
	→ meritamen joined	03:19
	← bin101 left	03:27
	→ jklfds joined	03:30
	← schne1der left	03:30
	→ gsi_ joined	03:31
	→ bin101 joined	03:33
	← gsi left	03:34
	← jklfds left	03:37
	→ jklfds joined	03:46
	← jklfds left	03:50
	← meritamen left	03:59
	→ fdan joined	04:00
	← rgrinberg left	04:01
	← mrpops2ko left	04:02
	← lester29 left	04:04
	→ jklfds joined	04:05
	← cbreak left	04:06
	← bitoff_ left	04:08
	→ LuckyMan joined	04:10
	→ mrpops2ko joined	04:12
	→ Rashad joined	04:12
	← jklfds left	04:14
	→ jklfds joined	04:17
	← bloody left	04:18
	← reset left	04:18
	← khrbtxyz left	04:19
	→ cbreak joined	04:22
	← jklfds left	04:22
	→ RiFo joined	04:27
	← RiFo left	04:27
	← BrianBlaze left	04:27
	← matt0x6f left	04:29
	← zeka__ left	04:29
	→ zeka_ joined	04:29
	← masoudd_ left	04:30
	→ jklfds joined	04:30
	→ matt0x6f joined	04:30
	← epony left	04:30
	← fdan left	04:33
	→ bitoff_ joined	04:34
	← cpackham left	04:35
	← _qw left	04:37
	→ RiFo joined	04:38
	← eqw left	04:38
	→ lester29 joined	04:38
	← jklfds left	04:42
	→ epony joined	04:49
	→ rgrinberg joined	04:53
	→ jklfds joined	04:54
	→ khrbtxyz joined	04:58
	← gryffus left	04:59
	← praise_the_sun left	05:00
	← jklfds left	05:00
	← ZMarkGC left	05:01
	→ gryffus joined	05:02
	← GNUmoon2 left	05:05
	→ GNUmoon2 joined	05:05
	→ fdan joined	05:07
	→ FinnElija joined	05:10
	← FH_thecat left	05:11
	← cbreak left	05:15
	→ jklfds joined	05:15
	← zeka_ left	05:15
	→ zeka_ joined	05:15
	→ ZMarkGC joined	05:19
	← jklfds left	05:25
	→ jklfds joined	05:33
	→ cbreak joined	05:33
	← jklfds left	05:38
	→ praise_the_sun joined	05:39
	← Rashad left	05:42
	→ jklfds joined	05:42
fdan	Hi there, i have a question	05:44
	we had a customer who had the repo and just a member of our team had an access to their repo	05:44
	our colleague create another repo ( not sure why he didnt fork the repo )	05:44
	whenever there is changes he used to add the changes to the repo he created	05:45
	whenever i tried to do something like this, i lost the commit information	05:45
	could someone tell me how to have the changes from repo1 to repo2	05:46
bookworm	git format patch / git show+ git apply, add the remote and cherry pick.... bunch of ways	05:47
ikke	fdan: git format-patch + git am	05:47
fdan	bookworm how do i add a remote	05:48
bookworm	git remote add	05:48
	what a surprise	05:48
fdan	yes but the remote should be repo1 or repo2	05:48
bookworm	yes and?	05:48
gsish	is any magic even required? is it not just a simple matter of fetch and push with different remotes?	05:49
fdan	i cloned the repo2, when i did git remote -v (it shows the origin in remote for fetch and push )	05:49
	gsish - i have never done that	05:49
ikke	gsish: only if the repos are related	05:49
	ie, they were the same repo but pushed to different projects	05:49
gsish	was to understand that from the above description	05:49
bookworm	fdan: sure, now add a second remote that points to the other	05:49
	but as ikke alluded to, you probably only want tp do so if the histories are actually related	05:50
fdan	git remote add [email@hidden.address]	05:50
	says usage: git remote add [<options>] <name> <url>	05:50
ikke	fdan: you need to give it a name	05:50
bookworm	yeah so give ot a name	05:50
fdan	sure	05:51
	that works and what is next	05:52
bookworm	fetch it, look at what you wanna do, do it	05:52
	← jklfds left	05:52
fdan	just git fetch?	05:52
bookworm	no, tell it which remote or use --all	05:52
ikke	git fetch <name-of-remote>	05:52
fdan	alright and then?	05:53
bookworm	do what needs doing	05:53
fdan	how i can add the new changes in the newremote to my branch	05:53
bookworm	I don't see your tree	05:53
gsish	follow your example above and do: some-command some-args	05:53
bookworm	are the histories related?	05:54
	or do you now see two separate graphs?	05:54
	← epony left	05:54
	→ rosco joined	05:54
bookworm	if the first, well, merge will do, else cherry-pick	05:54
fdan	bookworm: how do i see if the histories are related?	05:55
bookworm	look at it	05:55
fdan	when i do git merge feature/mon	05:56
	→ BrianBlaze joined	05:56
fdan	it says merge: feature/mon - not something we can merge	05:56
	→ epony joined	05:56
bookworm	split then	05:56
	→ TheCaptain4 joined	05:56
bookworm	unless you borked the name	05:57
	so use cherry-pick	05:57
fdan	i tried doing git pull --rebase repo1remote feature/mon	05:57
	says there is merge conflict	05:57
bookworm	why would you pull repo1 now?	05:57
	pull is fetch && merge	05:58
fdan	thats what i wanted right.	05:58
	repo1 - customer git repo	05:58
	repo2 - my repo	05:58
bookworm	which side are you trying to get where?	05:58
fdan	i created a test branch in repo2	05:58
	← TheCaptain left	05:58
	TheCaptain4 → TheCaptain	05:58
fdan	and i want to get some changes from some branch in repo1 to repo2	05:58
bookworm	so checkout whatever your branch is from remote 2 and then use cherry-pick to get changes from remote1 then	05:59
	or make sure that the histories are shared so that it's easier and use merge	05:59
	← hamburgler left	05:59
	→ Ox7C5_ joined	06:00
fdan	how do i make sure the histories are shared	06:00
bookworm	you can force git to accept an unrelated graph in a merge, look at the manpage for merge I don't use it like ever so don't remember from the top of my head	06:01
	→ jklfds joined	06:02
	→ liminal joined	06:02
bookworm	but assuming your colleague wasn't a masochist, they probably made it easy on themselves so I'm not sure your repos are really separate.	06:03
	→ clime joined	06:03
	← kyshoc left	06:05
	→ kyshoc joined	06:06
	← kyshoc left	06:06
	→ kyshoc joined	06:06
	← jklfds left	06:07
	→ Rashad joined	06:08
	← rosco left	06:09
	→ jklfds joined	06:10
	← hbautista left	06:10
	← mven9 left	06:11
	→ rosco joined	06:14
	← cbreak left	06:15
	← jklfds left	06:16
	→ jklfds joined	06:18
	→ cbreak joined	06:22
	← Rashad left	06:24
	← jklfds left	06:25
	→ mven9 joined	06:27
	← lunatiq left	06:29
	→ FH_thecat joined	06:29
	← cappy left	06:35
	→ Rashad joined	06:35
	→ hamburgler joined	06:35
	→ TheCaptain6 joined	06:37
	← TheCaptain left	06:39
	TheCaptain6 → TheCaptain	06:39
	→ jklfds joined	06:40
	← mingdao left	06:43
	→ mingdao joined	06:45
	← bindu left	06:45
	← jklfds left	06:46
	→ bindu joined	06:46
	→ jklfds joined	06:49
	→ xx joined	06:50
	→ Chaser joined	06:50
	→ pr100 joined	06:52
	← rgrinberg left	06:53
	← jklfds left	06:54
pr100	How can add everything from one repo in a subdirectory of another keeping all the branches? The examples I can find deal with merging into a single branch (e.g. main or master).	06:56
bookworm	do you want to keep the repo split?	06:59
grawity	that's the only way things work in git; there are practically no "all branch" operations	06:59
bookworm	subtrees come to mind	06:59
	→ rainystorm joined	06:59
bookworm	depends on what you are actually trying to do	06:59
grawity	tbh I started thinking about whether it would be feasible to mass-merge all branches into their correpsonding branches and my head is already spinning	06:59
	→ jklfds joined	07:00
	← bindu left	07:05
	→ bindu joined	07:05
	← jklfds left	07:05
pr100	I don't care about keeping the old repo (well - I will keep it for reference, but the idea is that all new work happens in the new combined repo)	07:06
	← fdan left	07:08
j416	pr100: sounds like filter-branch	07:08
pr100	I can manually create a new branch in the target repo for every branch in the "subrepo" and merge in manually. I guess I can script that	07:08
j416	but filter-branch will of course destroy the original history; I'd probably go with merge if I could	07:09
	or a simple "move to subdir" commit	07:09
	depending on your project	07:09
	→ rgrinberg joined	07:10
	→ _qw joined	07:12
	← LuckyMan left	07:13
pr100	Actually - it's only a few branches that I really care about - I guess it's not too onerous to just go through and do each one: create new branch in target repo, merge in from corresponding branch in source repo. The only branch that'll be named the same atm is almost certainly master.	07:13
j416	makes sense	07:14
	unless those relate to each other in some way, in which case it might make more sense to merge one and then let the others stem from that	07:14
	as in, if you usually merge one to the other	07:14
pr100	OK, ty	07:22
	git-filter-repo does a similar thing if you just want to split out into a new repo, but it's the stitching together...	07:24
	← skapata left	07:26
	→ jklfds joined	07:28
	→ meritamen joined	07:32
	← hamburgler left	07:38
	← jklfds left	07:38
	→ hamburgler joined	07:40
	→ schne1der joined	07:41
	→ jklfds joined	07:48
	← R2robot left	07:52
	→ jmd joined	07:53
	← jklfds left	07:55
	← rainystorm left	07:55
	→ jklfds joined	08:00
	← jklfds left	08:05
	← Ox7C5 left	08:06
	→ Ox7C5 joined	08:06
	← Ox7C5_ left	08:06
	→ Ox7C5_ joined	08:06
	← cbreak left	08:06
	← fling left	08:09
	→ fling joined	08:12
	→ cykelero joined	08:15
	← cykelero left	08:15
	→ jklfds joined	08:20
	→ cbreak joined	08:23
	→ zestyluna joined	08:36
	→ coot joined	08:40
	← Crispy left	08:44
	← zestyluna left	08:47
	← sentriz left	08:51
	→ gandhibobandhi joined	08:58
	→ xzed joined	08:59
	← bindu left	09:04
	→ bindu joined	09:06
	→ R2robot joined	09:07
	→ MGF_Fabio joined	09:10
	← fling left	09:20
	→ sentriz joined	09:20
	→ rainystorm joined	09:21
	→ fling joined	09:21
	← elastic_dog left	09:23
	← rgrinberg left	09:26
	← Rashad left	09:27
	← xzed left	09:27
	← GNUmoon2 left	09:34
	→ GNUmoon2 joined	09:34
	→ elastic_dog joined	09:35
	← gandhibobandhi left	09:37
	→ kurahaupo_ joined	09:40
	→ KnoP joined	09:42
	← FinnElija left	09:46
	← kurahaupo_ left	09:52
	→ danse-nr3 joined	09:59
	→ FinnElija joined	09:59
	← cbreak left	10:00
	← Gary_Sinise left	10:01
	→ gandhibobandhi joined	10:01
	← bindu left	10:03
	→ bindu joined	10:03
	→ cbreak joined	10:07
	← rosco left	10:10
	← Ox7C5_ left	10:17
	→ Ox7C5_ joined	10:17
	← igemnace left	10:24
	→ kurahaupo_ joined	10:27
	kurahaupo_ → kurahaupo__	10:28
	kurahaupo__ → kurahaupo	10:28
	→ thiago joined	10:32
	→ igemnace joined	10:41
	← pvxe left	10:41
	→ pvxe joined	10:42
	kurahaupo → CurriedHippo	10:42
	CurriedHippo → kurahaupo	10:42
	kurahaupo → Kangaroo	10:46
	Kangaroo → kurahaupo	10:47
	← de-facto left	10:51
	→ de-facto joined	10:52
	← meritamen left	10:57
	→ f451 joined	11:01
	→ meritamen joined	11:02
	← bindu left	11:02
	← jab416171 left	11:02
	→ bindu joined	11:03
	→ jab416171 joined	11:08
	← bindu left	11:09
	→ bindu joined	11:10
	← bindu left	11:10
	→ bindu joined	11:10
	← keypusher left	11:19
	→ rsx joined	11:19
	→ keypusher joined	11:19
	→ xdminsy joined	11:24
	→ cousteau joined	11:24
cousteau	Hi	11:24
	Can I have a tutorial for Windows, Git, and symbolic links? I'm having trouble figuring this out	11:25
	← dsimic left	11:26
ikke	cousteau: what do you want to know? As windows does not support symbolic links (as non-admin user), git stores symbolic links as text files on windows	11:27
	→ dsimic joined	11:28
cousteau	(context: by default, both Git and MSYS for Windows replace symbolic links with hard copies. I just installed Git on a new Win 11 machine and enabled the option to use actual symlinks, but now I need to enable this in MSYS and also my Windows user, I think)	11:28
	I have admin rights on this computer so I could enable symlinks for my user account. I also configured the %MSYS% env var but it seems that the git bash console overrides this setting for some reason	11:30
	→ Rashad joined	11:33
	→ fdan joined	11:38
fdan	hi, when i did git pull --rebase <remotename> feature/mon and then did a git commit	11:38
	it asked me to amend the commits i believe, so for some reason the timestamp and date is reported wrongly in the logs	11:39
	how do i fix it	11:39
cousteau	Alternatively... MSYS has an option to handle .lnk files intrinsically as if they were symbolic links. Can Git do that as well? Last I checked it couldn't; it would be nice if it could.	11:39
j416	fdan: the timestamps are correct; there are two timestamps for a commit, the author time the commit time. Your rebase updated only the latter.	11:43
	→ q3lont joined	11:43
fdan	j416 when i do the git log - i dont see the correct log with right timestamp	11:43
osse	try git log --pretty=fuller	11:44
j416	fdan: if you want to update the former as well, an easy way is to amend the commit again and add the flag --reset-author	11:44
	fdan: the log is correct.	11:44
fdan	osse thanks	11:44
	its right	11:44
j416	the author time doesn't change throughout the lifetime of a commit unless you explicitly specify that it should	11:45
	the commit time changes every time a commit is re-created (by for instance rebase, cherry-pick)	11:45
	→ rumpelsMongol joined	11:47
	→ theoceaniscool joined	11:53
	← omnigoat left	11:53
	→ omnigoat joined	11:54
	← danse-nr3 left	11:55
	← Ox7C5_ left	11:57
	← fdan left	12:01
	→ nivag_2c joined	12:01
	← nivag_2c left	12:06
cousteau	OK, so for some reason %MSYS%=winsymlinks:nativestrict doesn't propagate to git bash, where $MSYS defaults to disable_pcon	12:06
jast	how are you trying to set that?	12:07
kurahaupo	cousteau: do you set it in .bashrc or similar?	12:07
cousteau	Windows environment variables	12:07
jast	huh, weird	12:07
cousteau	kurahaupo: that's plan B	12:07
kurahaupo	I mean, it might be setting the wrong value in .bashrc, overriding the windows envvar	12:08
jast	or possibly the launcher overrides it	12:08
cousteau	Oh, gotcha. I already checked; there's no .bashrc file (yet)	12:12
	Has to be the launcher. Gonna ask around in #msys	12:12
	Maybe #mingw ?	12:13
	Oh, they're in OFTC	12:13
	← YuGiOhJCJ left	12:15
	← rumpelsMongol left	12:18
cousteau	So... Somebody had not clicked OK after editing the environment variables m(	12:20
	OK, now all that's missing is making Windows let me do symlinks	12:22
	← Rashad left	12:24
	→ nivag joined	12:24
	cousteau googled "windows allow symbolic links" and got a link to the documentation of Perforce. Finally, a proper SCM! /s	12:24
	← nivag left	12:29
	← q3lont left	12:30
	→ Ox7C5_ joined	12:35
	→ kucha joined	12:40
	→ nivag joined	12:43
	← nivag left	12:47
	→ navi joined	12:48
	← sentriz left	12:48
	→ a51 joined	12:50
	→ danse-nr3 joined	12:50
	→ reset joined	12:55
	→ sentriz joined	12:57
	← KnoP left	12:57
	→ jfsimon1981_c joined	12:58
	← fling left	12:59
	← jfsimon left	12:59
mjt0k	win has mklink, for a very long time, fwiw	12:59
	← gazalam left	13:00
	→ mane joined	13:00
	→ nivag joined	13:05
	← danse-nr3 left	13:09
	→ m257 joined	13:10
	← m257 left	13:11
	→ fling joined	13:13
	← Arsen left	13:15
	← dmvrtx left	13:15
	→ weedmic joined	13:16
	→ dmvrtx joined	13:17
	→ Arsen joined	13:17
	→ m257 joined	13:17
	← nivag left	13:20
	→ danse-nr3 joined	13:21
	→ TheCaptain1 joined	13:22
	← Ox7C5_ left	13:23
	← TheCaptain left	13:24
	TheCaptain1 → TheCaptain	13:24
cousteau	mjt0k: yeah, I'm trying to figure out how to enable user permissions for that	13:30
	TL;DR Perforce lied	13:30
	← weedmic left	13:32
	← de-facto left	13:35
	→ nivag joined	13:36
	→ de-facto joined	13:36
	daemon → daemon^	13:37
	← pr100 left	13:38
cousteau	OK, I give up. I'll use the silly and incompatible default behavior and be done with it.	13:38
	Not git's fault Windows is weird, anyway.	13:39
kucha	hi. does the <https://git-send-email.io>-style workflow assume a possibility to retain a local copy of sent emails? (given i have no control of my smtp server)	13:40
	← nivag left	13:41
	← praise_the_sun left	13:41
	← m257 left	13:42
osse	kucha: sort of. instead of doing git send-email HEAD^ you can do git format-patch HEAD^ which will create a file, and pass that file to git send-email	13:42
ch	and/or see the git-imap-send man page for suggestions	13:42
	→ m257 joined	13:42
osse	I suppose you can also --cc yourself	13:43
	→ Rashad joined	13:45
	→ praise_the_sun joined	13:47
	← elastic_dog left	13:49
	→ nivag joined	13:54
	→ TheCaptain6 joined	13:56
	← nivag left	13:57
	← TheCaptain left	13:58
	TheCaptain6 → TheCaptain	13:58
	→ Kroko joined	13:59
	← praise_the_sun left	13:59
	← Kroko left	13:59
	→ praise_the_sun joined	14:01
	→ elastic_dog joined	14:01
	← miigotu left	14:04
	→ miigotu joined	14:06
	→ TheCaptain3 joined	14:07
	← gloomy_desktop left	14:09
	← TheCaptain left	14:09
	TheCaptain3 → TheCaptain	14:09
	→ bloody joined	14:27
	← Aminda left	14:28
	→ TheCaptain2 joined	14:29
	← jklfds left	14:29
	← m257 left	14:30
	→ Aminda joined	14:30
	← TheCaptain left	14:32
	TheCaptain2 → TheCaptain	14:32
	← Aminda left	14:33
	→ Aminda joined	14:34
	← JanC left	14:37
	→ JanC joined	14:39
	→ jaythepirate2 joined	14:41
	← praise_the_sun left	14:42
	← jaythepirate left	14:43
	← jaythepirate2 left	14:46
	→ jaythepirate joined	14:50
wyre	hi guys, my git client gets stuck when I do "git clone http://..."	14:54
	it should ask for my credentials ... but it's not asking anything	14:54
grawity	export GIT_CURL_VERBOSE=1	14:54
	→ praise_the_sun joined	14:55
	← praise_the_sun left	14:56
	→ praise_the_sun joined	14:56
	→ praise_the_sun63 joined	14:57
	← praise_the_sun63 left	14:57
	→ praise_the_sun28 joined	14:58
	← praise_the_sun28 left	14:59
	→ praise_the_sun63 joined	14:59
	← praise_the_sun63 left	14:59
	→ liefer396 joined	15:00
	← praise_the_sun left	15:01
	→ lunatiq joined	15:01
	← kurahaupo left	15:14
	← mane left	15:15
	→ jklfds joined	15:15
	← meritamen left	15:17
	→ skapata joined	15:17
	← danse-nr3 left	15:20
	→ danse-nr3 joined	15:21
	→ meritamen joined	15:25
	→ gh34 joined	15:25
	→ praise_the_sun joined	15:31
	→ mane joined	15:34
	← mane left	15:40
	→ mane joined	15:41
	← thiago left	15:41
	← mane left	15:43
	→ thiago joined	15:43
	→ fstd_ joined	15:43
	← fstd left	15:43
	fstd_ → fstd	15:43
	← esc left	15:45
	→ rgrinberg joined	15:45
	← skapata left	15:46
	→ skapata joined	15:47
	→ hbautista joined	15:47
	← meritamen left	15:48
	→ mane joined	15:49
	← mane left	15:50
	→ masoudd joined	15:57
	← a51 left	15:57
	→ z1haze joined	15:59
	← danse-nr3 left	16:00
	→ humanface joined	16:02
	→ m257 joined	16:09
	→ esc joined	16:11
	← foul_owl left	16:16
	← m257 left	16:17
	→ stuv joined	16:18
	→ m257 joined	16:25
	← fflam left	16:26
	→ a51 joined	16:27
	→ fflam joined	16:28
	→ foul_owl joined	16:31
	→ danse-nr3 joined	16:33
	← danse-nr3 left	16:36
	→ danse-nr3 joined	16:37
	← rgrinberg left	16:38
	← m257 left	16:43
	→ m257 joined	16:44
	← MGF_Fabio left	16:44
	← mven9 left	16:54
	→ mven9 joined	16:55
	← cousteau left	16:57
	→ cousteau joined	16:57
	← cousteau left	16:58
	→ cousteau joined	16:58
	← a51 left	16:59
	← hbautista left	17:00
	← z1haze left	17:05
	← m257 left	17:05
	→ m257 joined	17:06
	→ rgrinberg joined	17:07
	← Night0wl3 left	17:08
	→ no_gravity joined	17:10
no_gravity	Is there a way to find the last comment I put into my codebase?	17:10
	← gandhibobandhi left	17:10
	→ kurahaupo joined	17:11
j416	no_gravity: do you know what you wrote?	17:13
p14nd4	no_gravity: do you need a one-liner, or you just want to find it manually?	17:13
j416	no_gravity: or what it looks like?	17:13
no_gravity	j416: No. I rarely write comments. I would like to find out when I wrote the last one.	17:13
j416	no_gravity: git log -S 'your comment' # might be of use	17:13
no_gravity	p14nd4: One-liner would be the coolest of course, but I can also do it manually.	17:14
j416	no_gravity: if you have a known pattern for a comment, you could use that, like a '#' or '//' or what have you	17:14
no_gravity	j416: Yeah, I'm trying -S '// ' at the moment. Not yet sure if that works.	17:14
j416	no_gravity: works here	17:14
no_gravity	Yeah. It's a bit manual work, but it works.	17:15
	So for this project, which is one of my bigger ones, I wrote the last comment in May 2023.	17:16
j416	why manual?	17:16
p14nd4	git log --all --branches --pretty='format:%C(auto)%h %d %C(green)%an %Creset%s %C(cyan)(%ci)%Creset' --abbrev-commit \| grep <your name> \| head -n 1	17:16
no_gravity	j416: It finds all kinds of unrelated stuff like removed comments for example.	17:17
p14nd4	there are certainly more elegant ways, and that might break in some cases with different branches or rebases	17:17
j416	no_gravity: ah. you might be able to give it a filter.	17:17
no_gravity	p14nd4: What does the --all do?	17:18
p14nd4	Pretend as if all the refs in refs/, along with HEAD, are listed on the command line as <commit>.	17:18
no_gravity	p14nd4: That is chinese backward to me :)	17:19
j416	hm maybe not	17:19
	p14nd4: he's looking for the last commit that adds a comment	17:20
p14nd4	no_gravity: tbh, i'm not 100% sure, but i've thrown it in to be thorough and grab more than just your current branch, or locally tracked branches, i think	17:20
	oh, i thought he just wanted his last commit message	17:20
no_gravity	The master branch is enough for me.	17:21
p14nd4	okay, then you can skip --all and --branches	17:21
j416	p14nd4: git log -1 --author=name	17:21
	p14nd4: (side note)	17:21
no_gravity	Ok, thanks! Time for dinner :)	17:23
	← no_gravity left	17:23
j416	can't find a way to filter such that git log -S only looks for added lines; maybe it's possible, maybe not	17:23
	and he left ¯\_(ツ)_/¯	17:23
quesker	what does it mean to block a PR? I am supposed to make a github action that runs test and lint on PR and block the PR if either one fails	17:25
	→ Betal joined	17:26
j416	quesker: that's not really a git question, but a github question	17:27
	quesker: if I were to guess, it means something that prevents that PR from being merged	17:27
quesker	yeah what command is that, that is my question I guess	17:28
	→ z1haze joined	17:28
j416	that's unrelated to git	17:28
quesker	oh ok	17:28
j416	git has no such functionality	17:28
quesker	ah ok I never do any pull request stuff just commit or pull or push	17:29
	a precommit hook is a git thing right? it seems to me that would be a better way to do this anyway	17:33
	but I have never used one	17:33
	← opal left	17:34
rewt	precommit hook would be defined locally, so everyone working on the repo would have to set it up for themselves	17:34
quesker	oh ok	17:34
	→ opal joined	17:34
	← danse-nr3 left	17:39
	← m257 left	17:40
	→ m257 joined	17:41
	→ a51 joined	17:41
	→ Night0wl3 joined	17:43
	← Rashad left	17:46
	← elastic_dog left	17:54
	→ undeadscholar joined	17:55
	← f451 left	17:57
	→ f451 joined	17:57
	→ eqw joined	17:57
	→ n1kx joined	18:02
	← m257 left	18:04
	→ qqq joined	18:04
	→ m257 joined	18:05
	→ masoudd_ joined	18:05
	→ elastic_dog joined	18:06
	→ chymera joined	18:06
	← masoudd left	18:08
	← kus left	18:16
p14nd4	when i rebase something, is there a way to have it also automatically update any branches off of those rebased commits to point to the rebased commits? e.g., a--b--c--d (mybranch), but there's also b--x--y--z (otherbranch); i rebase mybranch so it's a'--b'--c'--d', and at the same time i want otherbranch to get updated to b'--x'--y'--z'	18:17
	← lllh left	18:19
	→ llh joined	18:20
	← igemnace left	18:21
	daemon^ → daemon	18:22
nedbat	p14nd4: --update-refs does some of that, not sure if it does all of it.	18:22
	← praise_the_sun left	18:24
	→ praise_the_sun joined	18:24
	← gryffus left	18:25
	→ gryffus joined	18:25
	← stuv left	18:30
	← theoceaniscool left	18:34
	→ igemnace joined	18:38
	← rsx left	18:40
	← opal left	18:44
	→ opal joined	18:44
	← gryffus left	18:46
	→ gryffus joined	18:47
	→ krushia joined	18:48
	← Chaser left	18:49
	← gryffus left	18:55
	← undeadscholar left	18:56
	→ Gary_Sinise joined	18:56
	← sam113101 left	19:01
	→ sam113101 joined	19:02
	← foul_owl left	19:05
	→ OnlineCop joined	19:14
	← Szilva left	19:17
	→ Szilva joined	19:19
	→ foul_owl joined	19:19
OnlineCop	I have created a git repo under ~/bash_files/ and [hard-]linked several of my ~/.bashrc and ~/.bash_* files into it. When I switch between branches, the hard link is not preserved, so the changes are not reflected back into my home directory. Is this possible (like through a git hook or some `git checkout` flag) so the contents of the file are updated, but the inode is not?	19:21
p14nd4	nedbat: indeed, that looks like a very good candidate. thanks!	19:22
	← m257 left	19:23
	← miigotu left	19:25
	→ miigotu joined	19:26
	→ cpackham joined	19:30
	← Gary_Sinise left	19:31
weaksauc_	OnlineCop would a soft link work here since it's basically pointing to a path instead of a inode	19:32
	i don't recall if bash disallows this	19:32
OnlineCop	Would a soft link be able to save the file contents themselves?	19:32
	As long as I can see that I changed "line 54 of file X", I'm fine storing soft links.	19:33
weaksauc_	oh i think the typical way to do this is the other way around	19:33
	have a folder with all your stuff as a repo and then hard/soft link the other way	19:34
	→ PantelicGR joined	19:34
	← miigotu left	19:41
	→ miigotu joined	19:41
j416	OnlineCop: use a symlink instead	19:43
	OnlineCop: keep the file in the repo, put the link in ~/	19:43
	meh, could have just read the backlog before I typed something >.>	19:44
	but, there's a second opinion at least fwiw.	19:44
	→ thuna` joined	19:49
	→ JordiGH joined	19:49
JordiGH	Is there a quick way to checkout a commit's singular child, when it only has one?	19:50
ikke	There is no direct way to refer to children of a commit	19:50
	even if it's just one	19:51
JordiGH	So how do you move up and down a stack?	19:52
	← rainystorm left	19:52
JordiGH	I gotta manually remember where I was?	19:52
ikke	JordiGH: the reflog keeps track of where you were	19:52
anddam	JordiGH: does "where I was" mean HEAD history?	19:53
	i.e. are you "moving" by checking out?	19:53
JordiGH	anddam: I've got a stack of commits, I wanna move up and down this stack to check how my wdir is at each commit.	19:54
	'Cause y'know, I'm weird like that, trying to make sure every commit is buildable.	19:54
anddam	mm sounds like a work for bisect	19:54
JordiGH	Sounds like work to make bisect work.	19:54
j416	JordiGH: move down, then move back to your branch head	19:55
JordiGH	If you've got intermediate nonsense commits, bisect doesn't work as well.	19:55
osse	JordiGH: Have git log running in a second terminal	19:55
JordiGH	j416: So annoying.	19:55
j416	JordiGH: why?	19:55
JordiGH	j416: Because I wish I could just do "git previous", "git next".	19:55
	Maybe I can write some convoluted alias that parses the reflog in a shell script	19:55
j416	but why do you _need_ both?	19:55
JordiGH	Because I make mistakes and I want to go up and down as I'm working to see where my mistake might have been.	19:56
selckin	use relative to your branch, master^ master^^ master^^^	19:56
anddam	can you approach it differently? start from the leaf commit and use a previous counter?	19:56
j416	for a tool, you'd probably be better off with the log, not the reflog. I.e. keep the list of commits and move along that.	19:56
anddam	selckin: what's the syntax with explicit count? master^3 or master~3 ?	19:56
JordiGH	Yeah, git sucks, so I can work around the suckage. Fine, so what I wanna do can't be done. God I hate git so much. Thanks everyone.	19:57
j416	anddam: man gitrevisions	19:57
gitinfo	anddam: the gitrevisions manpage is available at https://gitirc.eu/gitrevisions.html	19:57
j416	anddam: master~3 == master^^^	19:57
	anddam: master^3 means third parent	19:57
anddam	j416: yep, already ran that in terminal :-)	19:57
JordiGH	gitrevision is awful, have you guys even heard of hg revsets?	19:57
j416	JordiGH: what do you mean by that?	19:57
JordiGH	gitrevisions is like its poor , starved cousin.	19:57
j416	JordiGH: it's a man page	19:58
JordiGH	man hg revisions	19:58
selckin	JordiGH: have you heared of jj	19:58
JordiGH	Yeah, git manpages are notoriously and famously unreadable.	19:58
	selckin: I have, might wanna use that instead of this piece of git dvcs.	19:58
j416	JordiGH: that depends on where you come from	19:58
	JordiGH: I haven't had that problem	19:58
	→ user03 joined	19:59
	← user03 left	19:59
	→ user03 joined	19:59
	user03 → gchound	19:59
	→ m0viefreak joined	20:00
JordiGH	There's a nice line from The Beautiful People that comes to mind, right after "you can't see the forest for the trees".	20:00
b0nn	JordiGH: you can move up and down a commit stack - but you need an alias to do it	20:01
	https://stackoverflow.com/questions/16062358/referring-to-the-previous-next-commit-in-git	20:01
	this is "next" git checkout `git log --reverse --ancestry-path HEAD..master \| head -n 1 \| cut -d \ -f 2`	20:02
j416	it's not that easy	20:02
	← praise_the_sun left	20:03
b0nn	last = log -1 HEAD	20:03
	oh, sorry that's the alias I have for the most recent commit	20:03
	prev = checkout HEAD^1	20:04
	I don't use it very much, but I saw someone give a talk once, and they used these aliases to move through their tree, with each commit being a slide in their talk :)	20:05
JordiGH	b0nn: Ah, thanks, I figured there was a shell script you could write.	20:05
	Well, just a single pipeline here.	20:05
b0nn	Honestly, though git's not hg, and hg isn't git, so complaining that one or the other doesn't have the same feature set is a bit... tired	20:06
JordiGH	Git sucks all on its own. I want to do certain things with it and it makes them difficult or impossible.	20:06
	Moving up and down the stack isn't some bizarro thing that nobody should ever do and if you do it you're wrong.	20:07
b0nn	then read the code and make it so :)	20:07
j416	JordiGH: you're free to not use it	20:07
b0nn	or argue for days on IRC - whichever is a better use of your time	20:07
j416	b0nn does have a point.	20:08
	→ fling_ joined	20:08
	← fling left	20:10
JordiGH	j416: Who the heck is free to not use git? They even make graphic artists use git when they work on video games.	20:10
	→ kus joined	20:11
	→ tirnanog joined	20:11
JordiGH	git is more stnadard than ASCII, HTML, or TCP/IP (actually, fine I don't know how anyone would live without TCP/IP these days)	20:11
	← cousteau left	20:12
	→ cousteau joined	20:12
b0nn	Bring back token ring!	20:12
rewt	ipx/spx 4ever!	20:12
osse	find child: git log --format='%H %P' \| awk '$2 == hash { print $1 }	20:13
	'	20:13
	yaay	20:13
p14nd4	UDP is all the rage now	20:13
	← cousteau left	20:13
	→ cousteau joined	20:14
j416	osse: catch-22	20:14
	← zeka_ left	20:15
	→ zeka_ joined	20:16
	fling_ → fling	20:16
p14nd4	HTTP/3 (nee QUIC), wireguard... and, uh ... this hot new protocol called DNS? :)	20:16
	← OnlineCop left	20:16
	← jklfds left	20:16
JordiGH	osse: I guess you need `git log --all` there?	20:17
	← Spawns_Carpeting left	20:18
	→ Guest95 joined	20:20
	→ kjartan_ joined	20:21
Guest95	hi	20:21
	← Guest95 left	20:22
	→ Spawns_Carpeting joined	20:23
	→ cousteau_ joined	20:24
	← rgrinberg left	20:25
	← cousteau left	20:28
	→ jklfds joined	20:29
	← cousteau_ left	20:29
	→ Tweebuffelsmetee joined	20:34
	→ cousteau joined	20:37
	→ chair1 joined	20:37
	← cousteau left	20:38
	← jklfds left	20:39
JordiGH	b0nn: I think it would help some things to add a revision number to git commits. At the very least, it would drastically reduce the search space for a hypothetical `git next` command.	20:39
	→ MGF_Fabio joined	20:39
JordiGH	But I can't be the first idiot to want to try to convince Junio to add revision numbers, so I wonder why it hasn't happened before. Junio is pretty stubborn about some things.	20:39
	← irrgit_ left	20:40
	→ irrgit_ joined	20:40
JordiGH	To be clear, a revision number would be an extra append-only blob that just sequentially maps numbers to new commit hashes as they are seen. It would probably not be exposed anywhere except for .git/ spelunkers.	20:40
	← raffaele__ left	20:40
chair1	hello. I currently work in only one branch at a time. a branch corresponds to one feature. I often add unrelated features/refactors in my single work branch. should I man up and start working with several branches for individual changes and merge accordingly?	20:41
JordiGH	I know Atlassian at one time wanted revision numbers in git too. I wonder why Atlassian failed to convince Junio.	20:41
osse	JordiGH: Yes, if you're unsure where the child might be	20:41
	→ raffaele__ joined	20:41
j416	chair1: not necessarily	20:43
	← kucha left	20:43
	→ cabazon93 joined	20:45
j416	chair1: when I work on my own projects, I do use branches but I like to keep history linear, so I don't merge back and forth just for the sake of it	20:45
chair1	I heard that when you add too much unrelated refactors and features in one branch, it makes reverting and bisect difficult	20:45
j416	it sounds like you mean one _commit_	20:46
	← Tweebuffelsmetee left	20:46
chair1	oh. if I keep every change in separate commits, then no problem? I know how to cherry-pick (select thunks I mean)	20:46
j416	keeping commits logical and scoped such that they do one thing at a time makes sense in that regard, it'll make the things you mention easier	20:46
	a benefit of having merge commits that join branches is that you can easily find the state that was before that branch was merged in	20:48
	for example, we (at dayjob) like to have one merge commit per production deploy, so that we can easily see exactly what went out in that deploy	20:49
	→ jklfds joined	20:50
j416	you can find both extremes	20:50
	there are projects that use merging quite extensively (the git project itself, for example)	20:51
	← gchound left	20:51
JordiGH	Hm. Hunks are parts of a commit, thunks are delayed computations in lazy languages like Haskell, chunks are parts of a video file.	20:51
	hunk/thunk/chunk	20:51
j416	and there are projects that try hard to keep history linear without any merges (the Go project, for example)	20:51
JordiGH	Don't think I've heard people call parts of a diff "thunks" before.	20:51
	← jklfds left	20:54
chair1	yeah I meant hunks hehe	20:56
	← mjt0k left	20:56
chair1	thanks for the insights	20:56
	→ rgrinberg joined	20:56
	→ jklfds joined	20:59
	→ masoudd__ joined	21:02
	← masoudd_ left	21:05
	← jklfds left	21:07
	→ Grace joined	21:12
	← cabazon93 left	21:14
	→ raffo joined	21:20
	→ jklfds joined	21:22
	← thuna` left	21:22
	← jklfds left	21:27
	→ theoceaniscool joined	21:27
	← zeka_ left	21:33
	→ zeka joined	21:34
	→ jklfds joined	21:35
	→ oriba joined	21:36
	← opal left	21:36
	← humanface left	21:37
	→ opal joined	21:37
	← irrgit_ left	21:40
	← kjartan_ left	21:40
	→ irrgit_ joined	21:40
	← jklfds left	21:44
	← xdminsy left	21:44
	← drocan left	21:50
	→ jklfds joined	21:50
b0nn	FTR this is my .gitconfig the alias section is probably most relevant to the earlier discussion (prev and next) https://gist.github.com/mindfarm/fb5ddcfcdfc239eb88ceea7a5c53981b	21:53
	← raffo left	21:54
	← jklfds left	21:55
	→ raffo joined	21:56
	← fling left	22:01
	→ fling_ joined	22:01
	→ jklfds joined	22:01
	← Night0wl3 left	22:02
	← GingerGeek left	22:04
	← rgrinberg left	22:05
	← jklfds left	22:06
	← JordiGH left	22:06
	← xx left	22:07
	fling_ → fling	22:09
	← theoceaniscool left	22:09
	→ jklfds joined	22:11
	→ drocan joined	22:11
	fling → swing	22:14
	← schne1der left	22:16
	← jklfds left	22:16
	→ jklfds joined	22:18
	→ rgrinberg joined	22:19
	← rgrinberg left	22:29
	← coot left	22:30
	← Darkfoe left	22:38
	← clime left	22:39
	→ Darkfoe joined	22:39
	← jmd left	22:42
	← miigotu left	22:54
	→ miigotu joined	22:55
	← miigotu left	22:55
	→ miigotu joined	22:55
	← MGF_Fabio left	23:04
	← gh34 left	23:11
	← deltab left	23:13
	→ GingerGeek joined	23:19
	→ deltab joined	23:22
	→ Jaxef joined	23:26
	← chair1 left	23:27
	← miigotu left	23:28
	→ miigotu joined	23:28
	← miigotu left	23:29
	← jklfds left	23:29
	→ miigotu joined	23:29
	→ owuser joined	23:34
	→ cappy joined	23:35
	→ jklfds joined	23:36
	→ [codemonkey] joined	23:37
	← [codemonkey] left	23:37
	→ [codemonkey] joined	23:38
	← PantelicGR left	23:39
	← jklfds left	23:40
owuser	I have a long-running, 18-hour rebase going. I want to turn off the computer and resume tomorrow, with 'git rebase --continue'	23:45
	Searching the git-rebase man page with 'interrrupt' etc is not telling me if I can do that.	23:46
	→ pac1 joined	23:46
ch	everything is in files	23:46
owuser	ok	23:47
ch	so if you could run git rebase --continue now, then you can also do that tomorrow	23:48
kjetilho	the question is if pressing ^C will terminate the process in a continuable state. I would guess it does, but I don't know for certain	23:49
	→ rgrinberg joined	23:51
owuser	I won't press ^C unless I am at the commandline.	23:59

Logs Search ←Prev date Next date→

Channels Documentation