IRCloggy #git 2021-09-02

Provider of IRC logs since 2005.

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

	2021-09-02
	→ igemnace joined	00:01
	← arcatech left	00:01
	→ arcatech joined	00:02
	← jstein left	00:03
	→ christall joined	00:04
	← jwillikers left	00:04
	← pretty_dumm_guy left	00:15
	← christall left	00:18
	← vysn left	00:19
	← Tobbi_ left	00:19
	→ Brain joined	00:20
Brain	can someone please give me a command to forcibly, non-interactively, rebase ALL commits on master into one commit? I know this sounds harmful an destructive -- it is! i want to use it to periodically clean up automatic commits made to my github pages repository from an autobuild script.	00:21
	ive done some searching but most rebase examples seem to be interactive or know that you want to rebase the last 'n' commits	00:21
	→ rkta joined	00:23
	→ Xenguy joined	00:29
	← arcatech left	00:32
	→ pyeveryt_ joined	00:32
	→ christall joined	00:36
	→ muscularduckling joined	00:37
	→ arcatech joined	00:37
	→ audiocat joined	00:38
	← christall left	00:40
	← XV8 left	00:45
	← arcatech left	00:45
	→ cation joined	00:46
	→ XV8 joined	00:47
	← hubble88 left	00:47
	→ CommunistWolf joined	00:49
circuitbone	https://github.com/stevenjack/cig	00:50
	please post any extra tools you find good value	00:51
	← XV8 left	00:51
	← reset left	00:53
	→ christall joined	00:55
	→ arcatech joined	00:58
	← arcatech left	00:58
	→ arcatech joined	01:02
	← muscularduckling left	01:05
	← pyeveryt_ left	01:05
	→ pyeveryt_ joined	01:05
	← pyeveryt_ left	01:10
	← christall left	01:13
nedbat	Brain: you could create a branch from your first commit, then when you want to squash, "git reset THATBRANCH" then "git add"; "git commit"	01:17
	← arktnld left	01:20
	← travaldo left	01:30
	← arcatech left	01:31
	← mei left	01:31
	← nyah left	01:32
	→ vysn joined	01:41
	→ muscularduckling joined	01:47
	← Murr left	01:49
	→ Murr joined	01:50
	→ pyeveryt_ joined	01:51
	← qpdb left	01:52
	→ qpdb joined	01:53
	→ ferdna joined	02:02
	→ magic_ninja joined	02:05
	→ vicfred joined	02:07
	→ arcatech joined	02:09
	→ gast0n joined	02:10
	← dsrt^ left	02:10
	→ Stanley joined	02:11
	← thiago left	02:12
	← FinnElija left	02:12
Stanley	Hey all. Solo dev working on my own project. I made a bunch of commits and pushed them. Then I decided to revert, so I reverted back to the last commit before the currenet batch of changes. When I tried to git push it said that I was behind and had to git pull, but I didn't want to pull back the changes again, so I did a git push --force	02:12
	it went through fine but now my deployment system is asking who I am and telling me to run git config --global user.name "whatewver", did I break something>	02:13
	→ FinnElija joined	02:14
Stanley	if I log into the server and run git status it says	02:17
	Your branch and 'origin/master' have diverged, and have 1 and 1 different commits each, respectively.	02:17
twb	Stanley: if you don't care about the version on the server you can "git push --force".	02:29
	Stanley: this is usually considered Bad Practice, but for a solo dev it's less important	02:30
	"which repos/branches can I push -f to" is a big part of multi-dev workflow design	02:30
	Stanley: you can also do something like "git diff main origin/main" to check how they differ before you push	02:31
	→ thiago joined	02:31
	→ gabin joined	02:32
	→ psyc0ld joined	02:33
Stanley	thanks1	02:44
	!	02:44
	← arcatech left	02:47
	→ arcatech joined	02:49
	← Xaldafax left	02:50
	→ euouae joined	02:56
	→ kyshoc joined	02:57
	← kyshoc left	02:57
	→ kyshoc joined	02:57
	→ cdown_ joined	02:58
	cdown_ → cdown	03:02
	← ChmEarl left	03:03
	← jazzy left	03:03
	→ jazzy joined	03:04
	→ ttree joined	03:05
	← arcatech left	03:06
	→ arcatech joined	03:07
	→ christall joined	03:10
	→ pulse joined	03:22
	← ferdna left	03:31
	← mw_ left	03:33
	← christall left	03:33
	→ aidalgol joined	03:33
	→ mw_ joined	03:37
	← keypusher left	03:42
	← arcatech left	03:43
	→ eschwartz joined	03:47
	← psyc0ld left	03:49
	← zebrag left	03:49
eschwartz	is there a git plumbing command to list submodules by path? The only thing I can think of is parsing the output of git config --file .gitmodules -l	03:52
	(and once there I can pass submodule paths to git diff-index --quiet to see if they have changed in any way)	03:53
	→ arunkumar413 joined	03:53
	← pyeveryt_ left	03:56
eschwartz	the general goal here is to write a program that uses git archive HEAD to export a dist tarball of a project, and then merge it with git archive foreach submodule too, warning the user along the way if there are uncommitted changes	03:56
	→ gnoo joined	03:56
	← gabin left	03:57
	← gnoo left	03:58
	→ gabin joined	03:58
	← gabin left	03:58
	→ gnoo joined	03:59
	→ gabin joined	04:02
	← stkrdknmibalz left	04:02
	← gast0n left	04:03
	→ gast0n joined	04:04
	→ FH_thecat joined	04:04
thiago	there is one: git submodule	04:05
	as for that script, I have it somewhere	04:06
twb	git submodule \| sort -k2 ?	04:06
	-k2, I suppose	04:06
thiago	remind me when I am in the office tomorrow (that is, in 12 hours) and I'll get it for you	04:06
	sorting is unnecessary	04:06
twb	thiago: they asked for "by path"; I misread that as "sorted"	04:07
thiago	they don't need the paths. They need to run git archive in each submodule. So... what they really need is git submodule foreach	04:07
twb	eschwartz: SO MANY people want git archive + git submodule :/	04:07
	thiago: yeah I agree, I hadn't read properly	04:07
	← gabin left	04:08
thiago	unfortunately, GNU tar isn't a format that is concatenable via pipes	04:08
twb	Actually I was wondering about that	04:08
	→ gabin joined	04:09
twb	https://gitlab.mister-muffin.de/josch/mmdebstrap/src/branch/main/tarfilter isn't quite that	04:09
eschwartz	I'm one of the maintainers of the mesonbuild.com build system, written in python. The current janky implementation uses "git clone the project into a temporary directory, then init and update all submodules, then delete all files/directories beginning with .git* and tar up the results"	04:10
thiago	my solution is to create a .tar file and then use tar -A (--concatenate) into it multiple times	04:10
twb	And https://gitlab.mister-muffin.de/josch/mmdebstrap/src/branch/main/mmdebstrap#L5483	04:10
thiago	so, no pipe streaming	04:10
twb	thiago: yeah that's what I resorted to but it's not scalable because it needs temporary disk space	04:11
eschwartz	this has several flaws, including not supporting submodules with relative paths e.g. ../otherproject.git which should clone from the original github remoteor some such, and not supporting gitattributes	04:11
thiago	twb: indeed	04:11
	I don't see a way out	04:11
eschwartz	I would like to replace that by git archive, but handling submodules is a complicated topic	04:11
	→ saroy joined	04:12
thiago	git archive --submodules should just DTRT	04:12
	but that isn't implemented	04:12
twb	So #L5610 you can see it generates a dev.tar stream by hand from raw bits, THEN runs "tar" program to append additional flies	04:12
	This might be relying on the consumer being GNU tar compatible	04:12
	{ tar c x; tar c y z; } \| tar t # definitely doesn't work	04:13
thiago	--format=pax	04:14
twb	{ tar --format=pax -c x; tar --format=pax -c y z; } \| tar --format=pax -t	04:14
	...also doesn't DTRT	04:14
thiago	no, not concatenable	04:15
	just tried that too	04:15
twb	I thought maybe mmdebstrap was writing the opening header itself and then telling tar "don't write an opening header"	04:15
eschwartz	it is acceptable to untar them all into a staging directory and tar the final results back up	04:15
twb	Maybe it's not writing a footer and then relying on GNU to write the footer. It's doing a more limited case of a single catenation	04:15
	eschwartz: maybe for you, but I hate it :-)	04:16
thiago	cpio is also not concatenable. I've tried that too	04:16
twb	thiago: and git archive --format=zip is obviously not	04:16
thiago	zip is never a good idea	04:16
eschwartz	... in fact, now that I think about it, for the specific use case of meson it will need to untar and re-tar anyway, because users can add custom post-process scripts that can modify the source tree before the final dist tarball is created	04:17
thiago	if you're producing content for windows, use 7z	04:17
eschwartz	git archive --subprojects would, of course, solve everyone's problems...	04:17
twb	As far as I'm concerned, the goal is to produce a source tarball without writing any intermediary streams to disk or to RAM	04:18
thiago	right	04:18
	I don't know of a solution to that	04:18
twb	Have to start with "git archive" because otherwise you need an intermediary "git checkout"	04:18
eschwartz	that's a good goal, anything else one would want could be added on after that anyway	04:18
thiago	like I said, the best I could come up with was tar --concatenate	04:18
	I can give you the script tomorrow	04:18
twb	Which means the streams begin life as either tar or zip	04:18
eschwartz	is `git submodule` a stable (plumbing-worthy) output format?	04:19
thiago	I think git get-tar-commit-id still works	04:19
	eschwartz: like I said above, you actually want git submodule foreach	04:19
eschwartz	seems like the general format is '{plus-if-modified}{current_commit} {path} ({decorated_name})'	04:20
twb	eschwartz: TIL	04:20
	← AnAverageHuman left	04:22
thiago	git submodule foreach --recursive 'git archive --prefix=$sm_path $sha1 \| tar --concatenaate output.tar -' or something like that	04:22
eschwartz	git submodule foreach git archive --prefix '$sm_path/' HEAD -o '$sm_path'.tar.gz if you don't want to rely on gnu tar, I suppose	04:23
twb	http://ix.io/3xHy	04:23
thiago	I always forget the / too	04:24
	I wonder why it's necessary	04:24
	it's not like I want my configure script name changed...	04:24
	→ christall joined	04:25
eschwartz	--prefix should probably really be --directory-prefix, indeed	04:25
	> $sha1 is the commit as recorded in the immediate superproject	04:25
	hmm, so this is the one committed, not the one manually checked out and marking the index as dirty?	04:25
thiago	yes	04:26
	once concatenated, xz -7	04:26
eschwartz	that conveniently solves the problem of not using HEAD and having to warn on uncommitted changes...	04:27
thiago	-9 if you don't mind the extra time, but in my experience -7 is the sweet spot	04:27
	→ dmc joined	04:28
	← christall left	04:29
	← jazzy left	04:31
twb	thiago: something along these lines http://ix.io/3xHz	04:33
eschwartz	git submodule foreach only affects currently inited submodules, which, well, doesn't seem like something to feasibly work around. logging a warning, at least, is probably a good thing, but this gets back to "how to list submodules" and submodule status doesn't. I'm thinking git config --file .gitmodules might still be the only way to figure out a list of submodules for comparison to the ones that	04:33
	are actually inited	04:33
twb	python and golang have streamable tar; perl doesn't	04:34
thiago	submodule status lists uninited modules	04:35
	→ homura joined	04:35
eschwartz	... right, I'm dumb. foreach won't operate on them, but status will sure list them	04:36
twb	Yeah that's also a huge open problem in my mockup	04:36
	if it's a bare repo how do you access the submodules at all	04:36
	Really my #1 preferred answer here is "don't use submodules"	04:36
thiago	you need to know how to access a local clone of the repository	04:37
ikke	You clone it, and on that clone, you can work with submodules	04:37
thiago	.gitmodules only contains URLs	04:37
eschwartz	unfortunately not in a truly parseable format, since you cannot just split on "{sha1} {path}" due to the end containing git describe bits	04:37
	← subopt left	04:37
eschwartz	but only sometimes!	04:37
thiago	yeah, don't use spaces in paths	04:37
eschwartz	I personally would not, but people can be crazy and when you try to make a build system for general consumption, someone will probably do it...	04:38
	fortunately .gitmodules inherently refuses those clowns who skip spaces and jump straight to newline chars	04:38
	→ aniruddha joined	04:39
thiago	it could get a -0 option	04:39
twb	Note that you also need to know which commit each submodule is pointing at, which is in the main repo's tree	04:40
	← hbautista_ left	04:41
eschwartz	the `submodule status` does list that	04:41
twb	ah k	04:41
eschwartz	$ git submodule	04:42
	fb2bd257f6089ed9890264bc4ef5ef96eaa2ef98 submod (heads/master)	04:42
	$ git submodule deinit submod && git submodule	04:43
	-fb2bd257f6089ed9890264bc4ef5ef96eaa2ef98 submod	04:43
	→ rnkn joined	04:44
	← The_Blode_ left	04:45
thiago	I had noticed it before because all of mine are prefixed with +	04:49
	← arunkumar413 left	04:50
	→ MrMagic joined	04:50
twb	This is the point where I go "can git just get a --format=json for literally every command please"	04:50
ikke	https://lore.kernel.org/git/CACPiFCLzsiUjx-vm-dcd=0E8HezMWkErPyS==OQ7OhaXqR6CUA@mail.gmail.com/	04:52
	← skapata left	04:53
	→ The_Blode joined	04:53
rnkn	anyone know a good way I can keep a work tree on a separate device without a repo?	04:55
ikke	did you look at git worktree?	04:56
rnkn	I have a writing device that does not handle lots of files well (e.g. a git repo)	04:56
	← MrMagic left	04:56
ikke	man git worktree	04:56
gitinfo	the git-worktree manpage is available at https://gitirc.eu/git-worktree.html	04:56
rnkn	ikke: as in git --work-tree=PATH ?	04:56
	→ MrMagic joined	04:57
rnkn	ah yes, I did, so I'd like that workflow, but with the caveat that sometimes the device worktree and the local worktree could contain conflicting changes	04:58
thiago	you want a checkout without a .git dir?	04:58
	that's a git archive	04:58
rnkn	I thought git archive was for creating tarball releases?	04:58
thiago	"device that does not handle lots of files well" does not sound like a sensible choice for a repository	04:59
	git archive \| tar -x	04:59
	I am assuming you're not going to use git in the extracted sources	04:59
	if you are, I really, really advise you to reconsider your choice of devices and filesystems	05:00
rnkn	thiago: correct, I want to avoid having a repo on the device	05:01
	I want a work tree on the device, while keeping the repo on my local machine	05:01
thiago	then use git archive and immediately extract onto the device	05:01
ikke	does the local repo require a worktree?	05:02
rnkn	yeah that's not a good way	05:02
thiago	no, git archive works on bare repositories	05:02
rnkn	ikke: preferably	05:02
ikke	then git-worktree is your best option	05:03
thiago	I think it's not	05:03
	how are you accessing the device's storage from your work machine?	05:03
rnkn	ikke: okay cool	05:03
	I've been trying to build got (gameoftrees) on macOS as it's designed for separate work-tree and repo, but alas no luck	05:05
	uses a .got dir to track the work-tree's commit	05:05
	← filePeter left	05:08
eschwartz	what does "keeping the repo on my local machine" mean?	05:10
rnkn	eschwartz: on my laptop	05:10
eschwartz	given the reference to gameoftrees supporting separate worktree and "repo", I'm assuming you actually do mean "worktree and --git-dir"	05:11
	so... what is wrong with man git-worktree	05:11
gitinfo	the git-worktree manpage is available at https://gitirc.eu/git-worktree.html	05:11
eschwartz	it does exactly what you say this gameoftrees thing does	05:11
rnkn	eschwartz: not quite, it handles conflicts between multipe worktrees	05:11
eschwartz	also for clarity, please feel free to start saying "git-dir" instead of "repo"	05:12
	→ mei joined	05:13
eschwartz	what does "conflicts between multiple worktrees" mean, and what does it have to do with your initial question which was "I have a writing device that does not handle lots of files well" and "anyone know a good way I can keep a work tree on a separate device"	05:13
rnkn	e.g. I check out a worktree on the dumb device, alter file foo.txt, then also happen to edit foo.txt on my laptop's worktree	05:13
	← mei left	05:13
eschwartz	then the dumb device and the laptop's live worktree are on different branches, so there is no conflict whatsoever	05:13
	OR	05:14
thiago	git worktree can do that, so long as they have checked out different branches	05:14
	however, mind what I said about git not being forgiving of dumb filesystems	05:14
eschwartz	you could have a bare clone on your laptop, and a worktree on the dumb device	05:14
thiago	I advised you to rethink	05:14
eschwartz	only one worktree	05:14
rnkn	eschwartz: yeah I was wondering that, I like the idea of this separate bare repo workflow	05:15
	but say I have that, how best do I deal with the foo.txt scenario?	05:16
	or did you mean to not have a worktree on the laptop?	05:17
	← homura left	05:17
	→ homura joined	05:17
	← gast0n left	05:17
mackerman	One .git dir per device. Zero or one worktrees per device.	05:18
	As they are different repos, remote workflows apply. commit, push, fetch, merge.	05:18
rnkn	whoa wait, I have been confusing git-worktree with git --work-tree=PATH	05:19
	→ mei joined	05:20
eschwartz	incidentally, referring back to my earlier question about git archive + submodules	05:20
	my current WIP rewrite https://paste.xinu.at/HcbgJs/ :D	05:21
twb	rnkn: in what way does the device suck? Is it eMMC, or JFFS2, or what	05:21
eschwartz	so much less terrible already :D	05:21
rnkn	twb: seeing you outside of #emacs makes me uncomfortable	05:22
	twb puts pants on	05:22
rnkn	twb: it's this: https://www.kingjim.co.jp/pomera/dm30/	05:23
	it gets noticably slower with git repos on board	05:23
twb	OK so yeah almost certainly eMMC	05:24
	i.e. basically an SD card solderde onto the mainboard	05:24
	rnkn: can you control what filesystem it uses?	05:25
rnkn	likely, but also I think the OS is only built to handle small text files of limited number	05:25
	← durham left	05:25
rnkn	nope, FAT32	05:25
twb	If so, try JFF- OK	05:25
	Yeah the problem there is simply that eMMC sucks and FAT32 doesn't work around that	05:25
	← vysn left	05:25
	→ durham joined	05:26
rnkn	it's okay, I am pretty sure git-worktree is the panacea I want	05:26
	thank you ikke and sorry for not fully investigating git-worktree (i.e. not --work-tree=)	05:29
twb	rnkn: that is a pretty cool unit	05:30
rnkn	twb: it does seem to be a cafe conversation starter	05:30
twb	I remember homeless people asking about my 7" EeePC	05:31
	That was back bfeore smartphones existed	05:31
thiago	I'm rewatching Leverage right now (2008, 2009) and it's funny to see flip phones and phones with physical keyboards	05:35
	→ YuGiOhJCJ joined	05:39
	→ tristan__ joined	05:45
	← rnkn left	05:45
	← Murr left	05:49
	→ Murr joined	05:49
	→ Samian joined	05:51
	→ madewokherd` joined	05:52
Samian	Hello magnificent minds of #git. How can I search a repo to find the first commit to contain a specific string in a file?	05:53
	← madewokherd left	05:54
Samian	like the first commit that has in file x.py the string. "torch.ops.fb.split_embedding_codegen"	05:54
twb	Samian: git log --reverse -1 -G'Underpants.*Gnomes'	05:54
	→ rnkn joined	05:54
twb	Samian: not exactly what you asked for, but close	05:55
	→ LACampbell joined	05:55
Samian	underpants?	05:56
bookworm	South Park?	05:56
LACampbell	I'm looking at a branch called "foo" on gitlab. gitlab tells me to do 'git checkout -b "foo" "origin/foo". git tells me "origin/foo" is not a commit	05:56
twb	bookworm: normally my example variables are fuck/duck/cluck/suck but since I'm not a regular here I was trying to be clean	05:57
LACampbell	I know git is just graph theory but we didn't cover "checkouts" at all in undergrad math	05:58
	← gabin left	06:00
bookworm	you are in a clone from the gitlab repo?	06:00
twb	LACampbell: are you completely new to git?	06:00
bookworm	or in a newly initiated repo that isn't actually linked to the one in gitlab?	06:01
LACampbell	twb: been using it for 6 years or so. does it ever get any easier?	06:01
bookworm	yes	06:01
twb	LACampbell: OK, sorry, it sounded like a super noob question	06:01
LACampbell	I rm -rf'd the whole directory, cloned it again, and everything works	06:02
twb	LACampbell: ah OK never mind then	06:02
LACampbell	twb: yeah I suck at git hard. I don't know why. I just can't wrap my head around it	06:02
twb	LACampbell: I would have checked "git remote -v", and "git fetch --all"	06:02
	LACampbell: are you a haskell person?	06:02
	LACampbell: if so check out hstorage which is a straight port of git's content-oriented filesystem	06:02
LACampbell	I'm not bad at FP, haskell makes more sense than git	06:03
twb	LACampbell: that's how I first "got" git, coming from a background of RCS and darcs	06:03
LACampbell	I guess I don't understand 'checkout'	06:03
	it does so many different things	06:03
bookworm	https://imgs.xkcd.com/comics/git.png	06:03
	← euouae left	06:03
twb	checkout does stuff to the working tree. checkout -b also tells "git branch" to do stuff.	06:04
LACampbell	is it normal to just delete your directory from time to time? is it kind of like old school windows where you need to defrag	06:04
	→ gabin joined	06:04
bookworm	you can use switch instead if that is any easier for you	06:04
	absolutely not	06:04
	→ envex joined	06:04
twb	LACampbell: no, defrag is callde "git gc"; it runs automatically as needed	06:05
bookworm	that means you have no idea how to use git... you never need a fresh clone, you have the history right there	06:05
twb	I only "delete and start over" when I've done something deeply broken.	06:05
	I did it more often when I was using rcs, darcs, git, hg, and monotone at the same time	06:05
	Now I use only git it's easier to keep it in my head	06:05
eschwartz	https://xkcd.com/1597/	06:06
bookworm	reflog, and if you've messed that up what on earth did you do?	06:06
twb	Oh the other time I would "delete and start over" is when I didn't have enough RAM to "git gc"	06:06
bookworm	eschwartz: to late ;P scroll up	06:06
twb	So e.g. linux.git on a 1GB RAM system	06:06
bookworm	too*	06:06
eschwartz	aw, man...	06:06
LACampbell	why would the same command tell me "origin/foo" is not a commit, then after I fresh clone, do something else? Is it a commit or isn't it? and i wanted a branch not a commit. why is it telling me about a commit with the -b flag?	06:06
eschwartz	I was only looking at comments after the question	06:07
	( you could delete it periodically, but then you're the noob from the comic who has no clue how git works )	06:07
	LACampbell: perhaps the remote is not called "origin"?	06:08
LACampbell	I used the exact same command, it had the exact same name both times	06:08
bookworm	LACampbell: you nuked it so now we'll never know. Alas checkout does many things so depending how you name stuff you might need to be more specific	06:08
	and tell git which of the two things that are applicable you actually mean	06:09
eschwartz	because you literally get that exact message when "origin/foo" is not a valid man gitrevisions	06:09
gitinfo	the gitrevisions manpage is available at https://gitirc.eu/gitrevisions.html	06:09
LACampbell	so what's the thing to do in that scenario then? "i have a super weird error". git status and go from there?	06:09
	if you never need to nuke it	06:09
eschwartz	if "origin/foo" is a valid remote_name/branch_name then it resolves to a commit, and git checkout works	06:10
	if not, then it doesn't	06:10
	"when doesn't it resolve" -> generally, when you accidentally have the remote named something else	06:10
bookworm	status first, then depending on what you are doing look at it. git remote -v / git branch -r stuff like that	06:10
LACampbell	so in got the concept of a branch is really just a single commit. there's no higher abstraction	06:11
eschwartz	if you git checkout -b "foo" <TAB> then tab completion should list all the valid possibilities and you can eyeball those for sanity	06:11
bookworm	^	06:11
	a branch is not a commit, it points to one	06:12
Samian	if I search on google: "oss install" github	06:12
eschwartz	to be more accurate, a branch is not a commit, it's just a moving label pointing at a commit	06:12
Samian	a lot of stuff comes up	06:12
	what does OSS mean?	06:12
eschwartz	the commit it points to is "the most recent commit on that branch"	06:12
bookworm	open source software?	06:12
LACampbell	so when you commit to a branch, you are moving that branches label to the thing you just comitted	06:12
eschwartz	right	06:12
bookworm	strange way to google though	06:12
LACampbell	so that's why checkout is used for branches and commits. because git reduces them down to being the same thing	06:13
	ok, that kinda helps. I'll try tab completion next time	06:13
bookworm	and files	06:13
LACampbell	at least briefly, before I nuke it	06:13
thiago	well, kinda. They're also different enough that we now have git switch	06:13
eschwartz	any/all git revisions are either: 1) a commit, 2) a pointer to a commit, 3) an annotated tag blob	06:13
Samian	bookworm aaahhh! open source software!	06:13
	thanks!	06:13
bookworm	and other things :D again checkout is heavily overloaded	06:13
Samian	next q. What's the diff between HEAD and trunk?	06:13
eschwartz	and 3 doubles as pointing to a commit	06:13
Samian	are they synonyms?	06:13
eschwartz	HEAD is a special reference to the current branch	06:14
bookworm	Samian: there is no trunk in git vocabulary as such	06:14
eschwartz	trunk means nothing, though in some repos it might be a branch name	06:14
thiago	anyway, I good analogy for a branch is that it's a place where you stick post-it notes with the SHA-1 of the commit	06:14
bookworm	there's trunk based development, if that's what you mean	06:14
thiago	to commit is to write a new post-it note with the SHA-1	06:14
bookworm	thiago: wrong order, the branch is the sticky note and you attach it to wherever	06:15
thiago	to checkout a new branch is to move the pen that is pointing to the currently checked out one	06:15
eschwartz	trunk is the traditional name for other VCS software, i.e. subversion, to name the default branch	06:15
thiago	no, the commit is the sticky note because of the reflog	06:15
eschwartz	(not that subversion can really be said to have true branches)	06:15
thiago	you can inspect the notes below it	06:15
bookworm	you do you	06:15
eschwartz	thiago: I got you with "you post sticky notes to the branch", I got lost when you suggested the commit is, itself, the sticky note	06:17
thiago	not the commit itself	06:17
	the branch is the note	06:17
eschwartz	> no, the commit is the sticky note	06:17
thiago	$ cat .git/refs/heads/main	06:17
	b72cc01f793fc9ba474b38780e116de6011f1016	06:17
	the analogy is not very good, of course	06:18
eschwartz	the reflog contains old sticky notes which used to be the topmost sticky note on the branch, fair enough	06:18
thiago	if you're sticking notes on your whiteboard, the branch is the section and you probably wrote the name on top	06:18
bookworm	doesn't work for me	06:18
thiago	the notes are entries in the reflog, with the topmost being the current	06:18
bookworm	in my world commits and branches are an actual tree, there's no white board	06:19
thiago	there's no tree in git	06:19
	in git commits, I mean	06:19
bookworm	graph / tree 🤷‍♂️	06:20
thiago	thre is of course the tree object	06:20
	sure, but that's a different discussion	06:20
	yes, the commits form a DAG	06:20
bookworm	it's really not	06:20
thiago	but you don't need branches to do that. You could write the output of git commit-tree in post-it notes	06:20
bookworm	the nice thing with analogies is that they only need to make sense intuitively, how it actually works under the hood isn't that important as long as the behavior matches	06:21
thiago	right	06:21
	I find my analogy is helpful when trying to explain the reflog	06:21
bookworm	and what works for you doesn't necessarily work for anyone else	06:21
thiago	which wasn't the topic here, so it may be unnecessarily complex	06:21
bookworm	I never use sticky notes the way you are suggesting	06:21
	a reflog is just a list of things you've done, a list	06:22
	there's no sticky notes involved	06:22
	conceptually at least	06:22
	← sh4 left	06:22
bookworm	it's easily explained as simple "history"	06:22
	everyone can understand that concept	06:23
eschwartz	the real problem here is that physical sticky notes would be way too slow to implement an actually performant VCS	06:23
thiago	the reflog file is actually a queue, not a stack like the post it note analogy	06:23
	entries are always added on top and are expired from the bottom	06:23
	eschwartz: well, but would it be faster than git push over Avian Carriers?	06:24
	I discovered today a colleague had never heard of April Fools RFCs...	06:25
eschwartz	birds can fly faster than people can find the right post-it note in a messy pile	06:25
bookworm	(or read my handwriting, future me included)	06:25
eschwartz	so it depends on how organized you are... :)	06:25
	bookworm: oh dear, the latency of correcting your handwriting when avian roundtripping is required to send corrections	06:26
thiago	in my experience, people who use post-it notes aren't very organised	06:26
eschwartz	thiago: that was my leading assumption, yeah	06:26
thiago	they use them to keep themselves somewhat organised, with varying levels of success	06:26
	← thiago left	06:33
	→ Guest3026 joined	06:36
	← Guest3026 left	06:36
	← FFY00 left	06:41
	tristan__ → gtristan	06:41
	→ christall joined	06:49
	→ lesihctej joined	06:50
	← jetchisel left	06:50
	→ theoceaniscool joined	06:52
	→ Hax0r joined	06:53
	← christall left	06:54
	→ pyeverything joined	06:56
	→ rtur joined	07:00
	← pyeverything left	07:00
	→ furrymcgee joined	07:01
	← ttree left	07:02
	→ Gurkenglas joined	07:02
	← unluckyshrubbery left	07:05
	→ srinidhi joined	07:12
	← furrymcgee left	07:13
	→ furrymcgee joined	07:13
	← vicfred left	07:16
	← sudoforge left	07:28
	→ sudoforge joined	07:32
	→ roadie joined	07:37
	→ christall joined	07:39
	→ ishutin_ joined	07:41
	← christall left	07:48
	← LiENUS left	07:50
	← AbleBacon left	08:05
	→ EvilDMP joined	08:07
	← durham left	08:09
	→ feep[work] joined	08:11
feep[work]	hi, I have a weird situation.	08:11
	→ poebel joined	08:11
feep[work]	I have a repo that I'm trying to merge with another repo	08:11
	I've copied all the files from the other repo into the current repo, made a commit, then did a merge. this is the wrong order and I'm trying to rebase them into one commit	08:11
	is there any way to squash a merge while keeping it a merge? Ie. I'm trying to turn the previous commit into the "real merge"	08:12
	← rnkn left	08:15
	← ecraven left	08:16
feep[work]	nm got it using lots of copying files elsewhere and copying them back	08:21
	"easy merge"	08:21
	← feep[work] left	08:23
	→ oxymoron93 joined	08:31
	← EvilDMP left	08:33
	← hnOsmium0001 left	08:34
	→ Spydar007 joined	08:38
	→ LuKaRo joined	08:41
	→ EvilDMP joined	08:45
	← Samian left	08:52
	→ Dev0n joined	08:53
	→ Samian joined	08:53
	← Samian left	08:58
	← Betal left	09:00
	← YuGiOhJCJ left	09:01
	→ YuGiOhJCJ joined	09:01
	→ xx joined	09:02
	→ cousteau joined	09:04
	→ keypusher joined	09:05
cousteau	There was a way to have multiple branches "active" at the same time, in different folders, right?	09:06
	← gtristan left	09:07
	→ gtristan joined	09:07
cousteau	I'm thinking of possible ways to perform this whole "use git to track a remote software project that isn't git" idea	09:08
	("isn't git" used to be "is a thing called Perforce", but now I'm encountering a project that doesn't use any SCM)	09:09
	→ unluckyshrubbery joined	09:09
cousteau	My current situation is: there's a project that I needed to copy to a folder and modify a couple of files for it to work (mostly due to the use of absolute paths instead of relative). Now it was modified, and I don't want to go over again modifying all the files, so I thought that git might be able to help me as a merging tool	09:12
	→ xa_ joined	09:13
	← FH_thecat left	09:17
	← roadie left	09:20
	← xa_ left	09:21
	→ daoudr joined	09:27
	← momomo left	09:28
	→ momomo joined	09:29
	← The_Blode left	09:29
	← austin987 left	09:29
	← YuGiOhJCJ left	09:31
	→ pretty_dumm_guy joined	09:32
	← oxymoron93 left	09:33
	← momomo left	09:35
avar	Is there no way to merge two branches and resolve & record a conflict with rerere that not only conflicts in file A, but requires an addition in file B only present in one of the two branches?	09:37
	→ YuGiOhJCJ joined	09:37
	→ momomo joined	09:38
avar	I don't think there is, have been trying to make it do it. I've got a merge where branch X has changed file A, but branch Y has moved the relevant function from file A to B. When I get a conflict I can just resolve/record the resolution to A, but changing the relevant thing in file B & doing "git merge --continue" with rerere.enabled=true & rerere.autoUpdate=true will just record the resolution to A, not B.	09:39
	→ Guest77 joined	09:40
	← Guest77 left	09:40
cousteau	...ah, "git worktree" is what I was looking for	09:41
	→ natrys joined	09:42
	→ austin987 joined	09:43
	→ christall joined	09:45
	→ oxymoron93 joined	09:46
	← gtristan left	09:48
	← AsenMx left	09:57
	→ dsrt^ joined	10:00
	→ vysn joined	10:04
	← audiocat left	10:05
	→ Masklin_Gurder joined	10:05
	→ lgc joined	10:06
	← bin101 left	10:08
	→ enoq joined	10:08
	← lesihctej left	10:10
	→ bin101 joined	10:10
	→ lgc_ joined	10:11
	← daoudr left	10:12
	← lgc left	10:12
	← christall left	10:14
	→ xa_ joined	10:19
	→ jetchisel joined	10:23
	→ reset joined	10:27
	→ pyeverything joined	10:33
	→ AsenMx joined	10:37
	← pyeverything left	10:38
	→ Ecophobia joined	10:42
	← Hax0r left	10:43
	→ Hax0r joined	10:50
	← Hax0r left	10:50
	→ FFY00 joined	10:52
	→ spuz joined	10:54
spuz	I would like to push a local git repo to a remote server via SSH. The repository does not yet exist on the remote server. How can I do this?	10:54
ikke	spuz: git does not support pushing to no-existing repositories, it needs to be created first	10:55
	ssh <remote> 'git init repo.gi'	10:55
spuz	ok thanks	10:55
	→ rfuentess joined	10:59
	← momomo left	11:01
	← Lunatrius left	11:06
cousteau	Curious; will that implicitly create a bare repo (since a dir name ending in .git was specified)? or not unless you explicitly add --bare?	11:07
ikke	oh, no it won't	11:08
	forgot --bare	11:08
cousteau	(I would guess if you do this you probably want the remote repo to be bare)	11:08
	→ goldfish joined	11:08
ikke	you are right	11:08
cousteau	I remember there were some restrictions when pushing to non bare repos but I don't remember them	11:08
spuz	I actually don't want a bare repo but I think I figured out the right configuration. I just ran "git init my-project" which created "myproject/.git" then "cd my-project; git config receive.denyCurrentBranch ignore". Then from my local machine, "git push --all my-remote". Then finally on the remote server "git reset --hard master"	11:11
	← AnapodoPsalidaki left	11:12
	→ Lunatrius joined	11:16
cousteau	It'd be easier with a bare repo, though. But if that works for you...	11:17
	← pretty_dumm_guy left	11:17
cousteau	That'll only be useful if you want the the file tree to be available in the remote computer. Usually the remote computer just keeps a bare repo which other computers clone.	11:18
	← xa_ left	11:19
	← lasr[m] left	11:21
	→ nyah joined	11:25
	→ jwillikers joined	11:26
	→ xa_ joined	11:27
	→ keir joined	11:27
	← Ecophobia left	11:29
	← wdouglass left	11:32
eschwartz	ikke, spuz: some servers are running software that auto-creates the repo for you on demand, though	11:33
spuz	cousteau: the remote server needs a local file tree. It's not a build server but something like that	11:38
	← xa_ left	11:40
	← AsenMx left	11:42
	→ AsenMx joined	11:42
	→ jrofd joined	11:44
	→ Hax0r joined	11:45
	→ xa_ joined	11:45
	→ dimi1947 joined	11:46
	← dimi1947 left	11:46
	→ AnapodoPsalidaki joined	11:49
	→ rewrit3 joined	11:50
	← Hax0r left	11:50
	← jrofd left	11:58
cousteau	I see	11:59
	Then an "elegant" (actually quite ugly) approach would be to have a bare repo on the server AND a clone of that repo also on the server, so that you push to the server and then pull to the clone. But that doesn't seem to provide any benefit over the solution you proposed (push to the server and then reset master)	12:01
	OK I'm gonna go crazy with this whole "track an external project so that it looks like it was made with git" idea	12:03
	→ roadie joined	12:05
	→ tristan__ joined	12:06
	← goldfish left	12:11
	→ christall joined	12:11
	← AsenMx left	12:12
	→ AsenMx joined	12:13
	→ skapata joined	12:14
	← skapata left	12:14
	→ skapata joined	12:14
	← AsenMx left	12:14
	→ AsenMx joined	12:14
	→ stpa joined	12:15
	→ otisolsen70 joined	12:16
	← AsenMx left	12:19
	→ AsenMx joined	12:19
	→ The_Blode joined	12:20
	← _Mocramis left	12:21
	→ jrofd joined	12:21
	→ ppang joined	12:24
	→ avar joined	12:28
	→ Bayes joined	12:29
	← Bayes left	12:29
	→ Bayes joined	12:29
	← EvilDMP left	12:31
	← tchan left	12:32
	→ gh34 joined	12:32
	→ EvilDMP joined	12:32
	← lgc_ left	12:33
	→ lgc joined	12:34
	→ AnAverageHuman joined	12:34
	← christall left	12:37
	← LordLion left	12:41
	→ LordLion joined	12:41
	← EvilDMP left	12:47
	→ pretty_dumm_guy joined	12:48
	← roadie left	12:51
	→ EvilDMP joined	12:51
	→ momomo joined	12:53
	→ christall joined	12:53
	→ hackinghorn joined	12:55
	→ jazzy2 joined	12:56
	→ hwrd joined	13:00
	← Xenguy left	13:01
	← stpa left	13:02
	→ stpa joined	13:03
	→ stefanct joined	13:03
	stpa → spampel	13:08
stefanct	afaict git diff replaces tabs with spaces (unlike format-patch for example). i can see why this might be done but is there a way around this?	13:09
	← spampel left	13:11
	→ onizu joined	13:11
	→ gjnoonan_ joined	13:13
	← keir left	13:14
	← CalimeroTeknik left	13:15
	→ CalimeroTeknik joined	13:16
	→ humanface joined	13:17
	→ keir joined	13:18
	← CalimeroTeknik left	13:20
	→ CalimeroTeknik joined	13:20
	← frobnic left	13:20
	→ frobnic joined	13:22
	← jazzy2 left	13:22
	→ tchan joined	13:24
	← shokohsc8 left	13:25
	→ shokohsc8 joined	13:26
	← EvilDMP left	13:27
	→ computeiro joined	13:31
	← keir left	13:31
	→ keir joined	13:32
	← pretty_dumm_guy left	13:39
	→ hsiktas joined	13:40
	→ alfredb joined	13:42
	← YuGiOhJCJ left	13:47
	→ roadie joined	13:50
	← The_Blode left	13:51
cousteau	stefanct: you sure that's not your terminal emulator doing funny things?	13:52
	→ The_Blode joined	13:52
	← keir left	13:52
	→ keir joined	13:53
ikke	must be, if you do git diff \| base64, the tabs are preserved	13:53
	← oxymoron93 left	13:53
	→ EvilDMP joined	13:54
	← humanface left	13:55
	← lgc left	13:56
cousteau	I was thinking git diff \| hd, or git diff >patch.diff	13:56
	→ lgc joined	13:56
cousteau	OK I can confirm that git diff >git.diff uses TABS, and doesn't convert them to spaces	13:57
	← roadie left	13:57
ikke	The patches would be corrupt otherwise	13:58
cousteau	BUT, if I copy-paste the output of git diff from the git bash (Windows) to a text editor, I do get spaces. But that's the terminal emulator, not git	13:58
ikke	correct	13:58
cousteau	Well, git does some magic with CRLF, maybe there's an option to do it with tabs as well and they accidentally toggled it on	13:59
	But yeah it shouldn't happen	13:59
	← ppang left	14:00
	→ gnit joined	14:02
	→ humanface joined	14:06
stefanct	it's the pager apparently	14:06
	(less in my case)	14:06
cousteau	Oooh	14:06
ikke	aaaaah	14:06
	so git --no-pager diff should have tabs then	14:07
	or git diff \| cat	14:07
cousteau	Hmmm nope, I tried `git diff \| cat` and still got spaces when I copied from the terminal. Or it might be both.	14:08
stefanct	cat is fine too for me	14:09
	← keir left	14:09
	→ keir joined	14:10
cousteau	So the terminal emulator I'm using ALSO replaces tabs with spaces, apparently	14:10
	← keir left	14:10
	→ wender joined	14:11
	→ Torr joined	14:11
	→ keir joined	14:11
cousteau	printf 'a\tb\n' # also prints spaces	14:11
	→ ajfriesen0 joined	14:12
	← ajfriesen left	14:12
	ajfriesen0 → ajfriesen	14:12
cousteau	But `printf 'a\tb\n' \| wc -c` says 4 so it should be fine.	14:12
	Moral of the story: don't copy from the terminal.	14:13
	← computeiro left	14:13
	→ hnOsmium0001 joined	14:14
cousteau	OK, I think I've figured out the proper way to "fake a proper use of git merge"	14:15
	Ideal situation: someone goes into the git repo I'm using, pushes their changes from the "official external repo" on top of master (which contains an older version of the external repo), then merges my devel branch into master, and copies the result to the external repo	14:17
	← keir left	14:17
	→ keir joined	14:17
cousteau	Actual situation: someone goes into the git repo I'm using, copies a bunch of files I modified into the external repo, add some changes of their own, and then expect my git to continue making sense.	14:19
mehwork	trying to pull changes made on a remote branch with `git pull origin feature/ticket-431` but getting: `:fatal: Not possible to fast-forward, aborting`. Any idea why?	14:20
cousteau	Possible solution: if I want my git to look as if "ideal situation" happened, I need to (1) merge the commits they should have merged into master, and (2) replace my repo with the content of the external one.	14:21
mehwork	do i have to just add --no-ff	14:21
stefanct	my moral is quite different... git --no-pager diff ftw	14:22
cousteau	mehwork: maybe someone pushed to that branch and you need to merge manually	14:24
	← mteo left	14:24
mehwork	cousteau: git pull --rebase ?	14:25
	First, rewinding head to replay your work on top of it..	14:27
cousteau	I'd do a fetch and a merge, just in case	14:27
mehwork	isn't that what git pull does	14:27
	← gnit left	14:28
cousteau	That's what I thought, but I would have sworn that git fetch will freely move remote branches without complaining	14:28
	Apparently you get that error if someone rewinds the branch on the server	14:29
	← Lunatrius left	14:33
cousteau	Hmm nope, I tried to do that with two local repos and it did the merge right away	14:33
	→ Lunatrius joined	14:34
	→ sh4rm4^bnc joined	14:35
	→ Tobbi_ joined	14:35
	sh4rm4^bnc → sh4	14:36
	← sh4 left	14:37
	→ sh4 joined	14:37
cousteau	mehwork: which git version do you have? Maybe this is something that happened in older versions	14:38
	(if someone has a better idea of what is happening feel free to suggest)	14:39
homura	gitgud.io	14:40
	→ subopt joined	14:41
	→ thiago joined	14:41
	← homura left	14:47
	← sh4 left	14:50
	→ sh4 joined	14:50
	← sh4 left	14:51
	→ sh4 joined	14:51
	→ mteo joined	14:51
	← poebel left	14:53
	→ Tobbi_1 joined	14:54
	← Tobbi_ left	14:55
	Tobbi_1 → Tobbi_	14:55
	← keir left	14:59
	→ Pent joined	15:01
	→ junktext joined	15:02
	→ WishBoy joined	15:02
	← WishBoy left	15:02
	→ lantech19446 joined	15:03
lantech19446	hey everyone, I just signed up for github and created my first repo and now I want to push all my working files to my repo, can I use git add . ?	15:03
	→ WishBoy joined	15:04
	← WishBoy left	15:04
	← stefanct left	15:04
	→ roadie joined	15:06
	← tristan__ left	15:08
	→ cyberpear joined	15:11
	← tchan left	15:13
	→ tchan joined	15:13
	→ AbleBacon joined	15:14
	→ Guest25 joined	15:14
	← vysn left	15:14
	Guest25 → samuelrajan747	15:17
mehwork	cousteau: 2.24.3	15:17
	→ ChmEarl joined	15:18
	← hackinghorn left	15:21
	→ durham joined	15:23
	← alfredb left	15:26
	→ neltraus_ joined	15:31
	← jinsun left	15:32
	→ jinsun joined	15:32
	← ajfriesen left	15:34
	→ push-f joined	15:34
	→ ajfriesen joined	15:35
j416	lantech19446: that's a good start indeed.	15:35
	lantech19446: commit would be next, and then github will have instructions on how to add a remote and push your commit.	15:35
lantech19446	I just got it done, I found an article online that walked me through it	15:35
j416	nice.	15:36
	→ starfarer joined	15:36
lantech19446	the hardest part was the damn PAT they're forcing you to use now	15:36
j416	what's that	15:36
lantech19446	personal access token	15:36
j416	use ssh, make your life simpler™	15:36
lantech19446	I'm ssh into my server but I don't even know how to connect to my repo via ssh	15:36
j416	ssh to github.	15:37
lantech19446	ok i'll have to look up info for it because i just cloned the repo to my computer so I can work in my IDE and then push it so it wouldn't be very hard to do now	15:37
j416	you'll find instructions on github on how to set it up. tl;dr create a keypair if you don't have one already, give github your public key, then use the ssh URL foryour repo instead of https.	15:38
lantech19446	I thought they said I had to do it over https now	15:38
j416	if they did, they are lying	15:38
lantech19446	ok	15:38
j416	unless of course "they" is your company that has a firewall or such that blocks ssh.	15:39
lantech19446	I have to say this is some really amazing technology I mean seconds to rebuild my entire site and it works	15:39
	→ pupskuchen joined	15:39
lantech19446	j416: I'm the one who would be configuring the firewall anyway lol	15:39
j416	:)	15:39
lantech19446	I'm actually a sys admin learning html /css/js to help out with our website since the guy who was helping left	15:40
j416	sounds like great fun	15:40
lantech19446	I'm loving it I think I actually want to transition to webdev next	15:40
j416	haha. good for you.	15:40
lantech19446	would you like to see the site I've built?	15:40
j416	nope	15:40
	(sorry)	15:40
lantech19446	it's ok you aren't obligated too	15:41
	→ ajfriesen9 joined	15:41
lantech19446	one more question though, once i update a file on my computer and push it to the repo how do I update the files on the server, I don't have to reclone every time i assume	15:42
	→ g2anj joined	15:42
j416	!deploy	15:43
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	15:43
	← ajfriesen left	15:43
	ajfriesen9 → ajfriesen	15:43
lantech19446	there must be some functionality to keep the files in sync though so that when i start working from another location i'm working on the newest files	15:44
j416	man git-push	15:44
gitinfo	the git-push manpage is available at https://gitirc.eu/git-push.html	15:44
j416	man git-fetch	15:44
gitinfo	the git-fetch manpage is available at https://gitirc.eu/git-fetch.html	15:44
	← humanface left	15:45
j416	you would typically push to a bare repo (such as one on github or your own host). You can fetch from both bare and non-bare (ones that have work dirs) repos.	15:45
	a bare repo is just the .git directory.	15:45
lantech19446	yea it looks like fetch is what I want I'm gonna test it in a moment	15:47
j416	!book explains in more detail	15:47
gitinfo	There are several good books available about git; 'Pro Git' is probably the best: https://git-scm.com/book but also look at !bottomup !cs !gcs !designers !gitt !vcbe and !parable	15:47
lantech19446	ok cool	15:48
	huh I just ran into something you probably know the answer too I added a directory to my repo on my computer and did git add test and then git status and it tells me there's nothing to commit, does it automatically take the directory with the files or something?	15:49
imMute	lantech19446: is test/ empty?	15:50
lantech19446	yes	15:50
imMute	lantech19446: git tracks files not directories. so you can have empty directories stored in git. a common work around is putting a .gitkeep file in the folder.	15:50
lantech19446	ok	15:50
imMute	note that ".gitkeep" is not special - any name will work - it's just a common name used for this purpose	15:50
lantech19446	ok so that worked and I did a commit but it's not showing up in my repo so I'm assuming I have to do git push after commit?	15:53
	← rfuentess left	15:53
imMute	yes, commit is local. you need to push to share it elsewhere	15:53
lantech19446	ok cool, i'm going to remember this it's going to become very useful to me	15:54
	I wonder how it'll deal with our bigger site, the guy who built it used a ton of symlinks and it's always been a bitch to back up or transfer	15:55
	ok it seems like I actually wanted git pull it might be replacing all the files i'm not sure but it updates the repo which is all i really need	15:59
cousteau	mehwork: that should be new enough	16:00
	→ Samian joined	16:09
	← roadie left	16:11
	→ roadie joined	16:13
	← Samian left	16:14
	← saroy left	16:19
	← Dotz0cat left	16:24
	← spuz left	16:28
	→ bdeferme1 joined	16:31
	→ rpkilby joined	16:33
	← christall left	16:33
	→ christall joined	16:34
	→ guest8765 joined	16:34
	← rpkilby left	16:36
guest8765	came here to ask for public bug tracker to search for some issues.. found it in the topic desc! (:	16:36
	← christall left	16:36
	→ christall joined	16:38
	← christall left	16:43
	← guest8765 left	16:45
lantech19446	ok I successfully pushed my 2nd site to a new repo everything succeeded but when I tried to clone it to my desktop i get clone succeeded but checkout failed and when i run git status it has deleted every single file even though it only objected to 1 or 2 so I just downloaded the zip and i'm extracting it to my repo but will this get me around my issue?	16:49
	→ humanface joined	16:56
	→ roadie` joined	16:57
	→ wgrant joined	16:58
cousteau	lantech19446: shot in the dark: are you on Windows? Maybe the files have forbidden names	16:58
mackerman	lantech19446: Sure, "git archive" or similar extraction of one version of the tree is useful. Especially for deployment or release workflows.	16:59
cousteau	(Also, do you have enough space on your drive? You wouldn't believe how often that's caused trouble for me...)	16:59
	Does it mention WHY checkout failed?	17:00
lantech19446	cousteau: I am on windows not a forbidden name it's chucky.jpg and chucky.png and I have 3Tib free	17:00
	← roadie left	17:00
	→ christall joined	17:00
lantech19446	it's trying to say invalid path for chucky.jpg	17:00
	← enoq left	17:00
cousteau	OK, assuming neither of those files are more than 3 TiB in size...	17:01
lantech19446	they're like Kib	17:01
mackerman	Did you not clone into a new directory? clone is for grabbing the first copy of a remote.	17:01
lantech19446	I did	17:01
	cousteau once had trouble with files named "aux.whatever"	17:01
lantech19446	yea i saw that when i was googling around for the error	17:01
	→ goldfish joined	17:02
cousteau	Apparently aux.c is not a valid file name on Windows.	17:02
	→ meator joined	17:03
cousteau	What's the path for this chucky.jpg? Is it on the root of the repo or on some subfolder?	17:03
lantech19446	images\chucky.jpg	17:04
	→ Samian joined	17:05
cousteau	What git are you using? The one that comes with a "git bash" console? Or some fancy program with git built into it?	17:05
lantech19446	great and now that i downloaded the repo it wants to add all of them instead of just the new file. I've gotta get this working	17:05
	I'm just using git on powershell	17:06
	I could try it on git bash i'm familiar with bash	17:06
cousteau	I recommend using git bash	17:06
lantech19446	any chance of it getting me past this issue?	17:06
cousteau	Although I would expect it to work in powershell if git supports that	17:07
lantech19446	it was one of their recommended ways	17:07
	← goldfish left	17:08
cousteau	Honestly I have no idea, it's just that that's how I use it on Windows, and I've noticed that drifting away from bash tends to cause trouble	17:08
	→ jstein joined	17:09
lantech19446	same error and git restore fails too	17:09
cousteau	Huh	17:09
	→ Betal joined	17:10
	← EvilDMP left	17:10
	← ninjin left	17:10
	→ ninjin joined	17:11
cousteau	Honestly I don't think I can help here	17:11
lantech19446	error: invalid path 'images/chucky.jpg?_subject_uid=7888084&w=AAAfmveVu2oZofBbzRKou625rHOAD9ztRtM1Jyrs4Xh3Bw'	17:11
	fatal: unable to checkout working tree	17:11
cousteau	Oh	17:11
	Yeah that file isn't called "chucky.jpg" but "chucky.jpg?_subject_uid=blahblah"	17:12
lantech19446	I wonder if it's even used on the site	17:12
cousteau	File names can't contain "?" on Windows	17:12
	That looks like someone downloaded an image from a site using some script and the downloaded file name includes the '?xxx=yyy&zzz=ttt' part of the URL	17:14
	You won't be able to have such a file in your PC	17:14
lantech19446	they don't even seem to be legitimate files on the server i'm not sure what the hell they're used for	17:14
cousteau	Your best bet would be to rename those files from Linux or similar, maybe amend the commit if possible. Try to figure out why those files are there, maybe you can even delete them	17:16
lantech19446	official word from my boss we don't need them anymore	17:16
	and if for some reason he decides we do i'll still have the first version on github because the push succeeded from the server	17:16
cousteau	If they were accidentally added you might even want to take them out of git's history by amending/rebasing commits	17:17
	Or just delete them (using an OS/platform that supports those files) and commit	17:18
lantech19446	that's what i'm doing right now	17:18
cousteau	Or, instead of deleting them, rename them	17:18
	Take out the ?... part	17:18
	That way you can easily recover them, even on Windows	17:19
lantech19446	they were from back in 2019 we did a fundraiser with chucky cheese, if I don't delete them now i'm gonna delete them in 2mos when i'm getting rid of all the shit no longer used on the site	17:20
	← Strom left	17:20
lantech19446	goddamnit it found another one of those files	17:21
	→ Strom joined	17:22
	← cousteau left	17:24
lantech19446	those images aren't even valid on linux, I have no fucking idea how they ever got displayed on our site	17:24
	← Samian left	17:27
	← ninjin left	17:30
	← shokohsc8 left	17:30
	→ ninjin- joined	17:30
	← roadie` left	17:33
	→ elf_fortrez joined	17:33
	→ roadie` joined	17:33
lantech19446	hazzah I have accomplished the task	17:34
	← meator left	17:37
	← samuelrajan747 left	17:38
	→ pretty_dumm_guy joined	17:40
	→ shokohsc8 joined	17:42
	→ hqdruxn08 joined	17:47
	← elf_fortrez left	17:49
	← ajfriesen left	17:49
	← gnoo left	17:54
	→ meator joined	17:55
	→ koolazer joined	17:57
	← meator left	17:58
	→ meator joined	17:59
lantech19446	I got the error error: src refspec master does not match any I added the files committed the files and got that error when I tried to push the files now it says my working tree is clean and won't let me add them and the changes didn't push	17:59
	→ Brumaire joined	18:00
	← meator left	18:02
	→ meator joined	18:02
lantech19446	never mind when I used git push # instead of -u origin master it worked	18:05
	← pretty_dumm_guy left	18:07
	→ stkrdknmibalz joined	18:09
	← Vooch left	18:10
	→ sniperwolf joined	18:12
	→ FabioP joined	18:15
	→ hbautista_ joined	18:22
	→ txtsd joined	18:23
	→ EvilDMP joined	18:24
lantech19446	so apparently some of my changes still didn't propagate	18:28
	I don't get what the issue is	18:28
	→ pretty_dumm_guy joined	18:31
	← roadie` left	18:33
	→ Samian joined	18:33
	→ roadie` joined	18:39
	→ zumba_addict joined	18:40
	← EvilDMP left	18:40
	j416 serves lantech19446 some tea. there there.	18:40
j416	triggered by profanity	18:41
	← tmz left	18:41
	→ tmz joined	18:42
lantech19446	thank you j416 I just managed to get all the issues resolved now if my blood pressure would come down a little bit I'll be just fine	18:42
	→ absc joined	18:43
	← meator left	18:47
	→ meator joined	18:48
	← APic left	18:49
	→ alfredb joined	18:50
	→ Hax0r joined	18:51
	← pmcnabb left	18:57
	← alfredb left	19:05
	← meator left	19:07
	→ meator joined	19:08
	→ odoood joined	19:08
Torr	Where in the <.git> folder are commit messages stored?	19:11
imMute	Torr: .git/objects/ all objects are in there (commits, tags, trees, blobs, etc)	19:12
	← kyshoc left	19:13
	→ kyshoc joined	19:13
	← kyshoc left	19:13
	→ kyshoc joined	19:13
Torr	In which binary format are they stored?	19:14
	← rahl left	19:15
imMute	Torr: https://git-scm.com/book/en/v2/Git-Internals-Git-Objects	19:16
Torr	Thank you	19:16
	→ bradyt joined	19:21
bradyt	i don't really feel like mentioning some other project's "master" in my code. is there another reference like "bleeding edge", "HEAD", etc?	19:21
	er, HEAD isn't right either	19:22
imMute	"main branch"?	19:22
ikke	"The branch which shall not be named"	19:22
bradyt	yeah basically	19:23
DoofusCanadensis	upstream	19:23
	→ Dotz0cat joined	19:23
DoofusCanadensis	but that usually refers to the entire repository no any specific branch	19:23
another\|	you-know-which-branch	19:23
bradyt	i feel like i accidentally used the term "upstream" when it's actually a completely different repo, only same ecosystem	19:23
	these don't have any commits in common, if that makes sense	19:24
	"bleeding edge of your client breaks my client"	19:24
	i like that #git is not above humor, i thought lighter moods were more rare on irc	19:24
	*humour	19:25
eschwartz	bradyt: "the branch used by those racist people"	19:26
bradyt	i mean, it's now amusing that a project /hasn't/ switched to main	19:26
eschwartz	why?	19:27
bradyt	like, you're not that good of a programmer?	19:27
	→ Dotz0cat_ joined	19:27
bradyt	not the "you" "you"	19:27
eschwartz	assuming one wants to switch, who in their right mind would ever use "main"?	19:27
	← Dotz0cat left	19:27
eschwartz	consider e.g. "trunk"	19:27
	← Torr left	19:27
DoofusCanadensis	it's what github settled on	19:27
bradyt	i noticed "trunk" somewhere in docs	19:27
	i think github offers "main", "trunk" and something else	19:28
	Dotz0cat_ → Dotz0cat	19:28
bradyt	as ideas	19:28
eschwartz	$ cd git-git/; git log origin/main<TAB> -> origin/maint	19:28
bradyt	origin/roller-coaster	19:28
twb	eschwartz: I thought they were D/s people	19:28
bradyt	hmm, can't find that list of alternatives to "master"	19:29
eschwartz	so github, trying to distract people from talking about ICE, got people to talk about master (without slaves) instead, and they... convinced people to rename to main, because "ma<TAB>", even though that's even worse, now you need 4 characters to disambiguate main vs. maint, instead of 3 to disambiguate master and maint	19:30
	so much muscle memory :/	19:30
bradyt	huh, hadn't heard about their involvment with ICE, just AWS	19:30
	*involvement	19:30
eschwartz	everyone bought their distraction hook line and sinker, but the real indignity is people actually respect their ability to come up with a good naming scheme... what was wrong with trunk, might I ask?	19:31
bradyt	oh, probably "development"	19:31
eschwartz	or "develop", as suggested by git-flow	19:32
bradyt	SVN trunk	19:32
DoofusCanadensis	trunk belonged to SVN	19:32
bookworm	or you know... master	19:32
DoofusCanadensis	don't reuse it!!oneone	19:32
eschwartz	in fact git-flow has a master and a develop branch, and the master branch has no earthly reason to exist since it acts as a duplicate of the third git-flow branch, "release"	19:32
bradyt	this might be where i noticed the term "trunk": https://github.com/cli/cli/issues/929	19:33
	i guess they never suggested in renaming docs	19:33
	"bleeding edge" seems too snarky for comments in tests	19:34
DoofusCanadensis	eschwartz: no, you can merge hotfixes into master as well with git-flow	19:34
eschwartz	yes, but then it still only has tagged releases which are additionally present in release-${version}	19:35
	I have never heard a good explanation for why git-flow has a "master" branch which all release branches are merged into, other than "yay, even more bureaucratic merges"	19:36
	← meator left	19:36
eschwartz	it is not even conceptually a "master" copy of the code in the sense that people typically name a branch master, trunk, or develop. Because it never has the latest code, only releases...	19:37
	→ braxas joined	19:38
DoofusCanadensis	because git-flow was created when master was considered the default branch	19:39
	← roadie` left	19:39
DoofusCanadensis	ideally, you'd rename master to production and flow from develop to release to production	19:39
	it was just a workflow that got designed and then popular	19:41
	nobody says you have to use that	19:42
	← Gustavo6046 left	19:50
	→ qbit joined	19:51
	→ Gustavo6046 joined	19:51
qbit	hi!	19:51
	can one tell a specific repo to always signoff ? i don't see a config options specifically for it	19:52
	format.signOff ahh!	19:52
	→ nvmd joined	19:55
qbit	thanks :D	19:55
	← qbit left	19:55
	← nvmd left	19:55
	→ nvmd joined	19:55
eschwartz	DoofusCanadensis: I mean, I don't actually like git-flow on any level, so I don't use it. I just also have nosy opinions about things making sense...	19:57
DoofusCanadensis	heh	19:58
	I used it briefly at the beginning and I like the concepts, but it didn't fit with our current workflow	19:59
	→ roadie` joined	20:00
	→ john_johnk joined	20:01
	← roadie` left	20:05
	← absc left	20:06
	← mei left	20:09
	→ dflick_ joined	20:10
	← dflick left	20:12
	→ absc joined	20:12
	← absc left	20:15
	→ absc joined	20:17
	→ roadie` joined	20:19
	← crabbedhaloablut left	20:24
	← roadie` left	20:24
	→ crabbedhaloablut joined	20:24
	→ cmc joined	20:24
	← durham left	20:24
	→ durham joined	20:24
	← john_johnk left	20:27
	← absc left	20:30
	→ john_johnk joined	20:31
	→ gadget joined	20:32
	→ clime joined	20:34
	→ durham_ joined	20:34
	← otisolsen70 left	20:35
	← durham left	20:36
	→ Guest71 joined	20:49
	← onizu left	20:51
	→ Lord_of_Life_ joined	20:51
	← Lord_of_Life left	20:52
	Lord_of_Life_ → Lord_of_Life	20:54
	→ vysn joined	20:56
	← clime left	20:58
	← durham_ left	21:01
	← gh34 left	21:02
	→ durham joined	21:02
	← gadget left	21:04
	→ gadget joined	21:04
	← FFY00 left	21:05
	← furrymcgee left	21:05
mehwork	is there a way to make it so i can just do 'git co branchFoo ; git pull' to pull the latest changes on a branch that someone else committed, instead of 'git pull origin branchFoo'?	21:06
	→ roadie` joined	21:06
	→ mago joined	21:07
mago	Hello need help	21:07
	Im trying to push to a repository, the web interface wont let me upload the folder cuz its more than 25 mb	21:07
	Im stuck at this command: git remote add origin https://github.com/yourUsername/yourRepository.git	21:07
	Dond understant what should be in there.. I have a folder but not a git:	21:07
	← john_johnk left	21:08
mago	https://github.com/Formula-HP-Computer-Vision-on-Wheels/Group2	21:08
	→ cweiss joined	21:09
mehwork	do i have to do: git branch -u origin/branchFoo ?	21:11
	if i just type `git fetch` it says the branch fetched stuff as (forced update)	21:13
	← gabin left	21:13
	← roadie` left	21:13
	→ ackyshake joined	21:17
	← gadget left	21:19
	→ gadget joined	21:20
	← mago left	21:20
	← austin987 left	21:21
	→ FFY00 joined	21:22
mehwork	oh, origin/foo means local branch called origin/foo, whereas 'origin foo' (with a space) means a remot ebranch named 'foo'?	21:22
	→ zebrag joined	21:24
	→ roadie` joined	21:24
mehwork	All i really want to know is when i create a new branch like 'git co -b feature/foo' is am i supposed to push it like `git push origin feature/foo' and pull it like `git pull origin feature/foo` even if someone else commits to it?	21:24
	→ m0viefreak joined	21:26
	← aniruddha left	21:29
	← roadie` left	21:29
	← alkino left	21:30
	← gxt left	21:30
	← Bayes left	21:31
	→ gxt joined	21:32
	→ ajfriesen joined	21:32
	→ anddam joined	21:34
	→ austin987 joined	21:34
anddam	I have "error: The following untracked working tree files would be overwritten by checkout:", how do I get a list of those files on stdout?	21:34
	← christall left	21:34
osse	anddam: A bit difficult since they're not tracked. What git does under the hood is take the list of files in the commit you want to check out, subtract the list in the commit you have checked out now, then filter the result on which files exist on disk.	21:37
	Not too difficult to script, but therw exists no direct command for it	21:37
	This is a start:	21:37
anddam	osse: ok, no worries then	21:38
osse	git diff --name-only --diff-filter=A HEAD othercommit	21:38
anddam	I thought there was something like status --short	21:38
osse	The files among those that already exist on disk are the problem	21:39
	← AsenMx left	21:39
anddam	in my case I'll just grab stderr for lines starting with a tab	21:39
	a bit of sed, a pinch of xargs and presto	21:40
osse	You must have quite a few such files for the effort to be worth it. In which case why the hell do you have so many files on disk that just happen to match (by name, anyway) the files in another commit? :P	21:42
anddam	long story	21:43
	to make it short I am trying to create an impossible workflow	21:44
	seems I am succeeding at that	21:44
	osse: what effort?	21:44
osse	The effort og scripting it	21:44
anddam	osse: git checkout whatever-is-giving-conflicts 2>&1 \| ag '^\t' \| tr -d \\\t \| xargs rm	21:44
	→ EvilDMP joined	21:45
osse	If that works well for you then I'm all for it	21:47
	← theoceaniscool left	21:48
m0viefreak	simply deleting those problematic files sounds like you might as well just `git clean -fdx` before doing the checkout?	21:50
osse	But what about their precious build products	21:51
m0viefreak	:)	21:51
	a build product that's suddenly checked in in another branch sounds strange	21:52
	← odoood left	21:52
	→ Quietus joined	21:52
osse	No I mean that such a clean command would also remove build products	21:53
	← Guest71 left	21:53
m0viefreak	ah, true	21:53
anddam	osse: it still doesn't	21:54
m0viefreak	what about simply `git checkout -f` then?	21:54
	it should simply overwrite everything without question	21:55
	← zumba_addict left	21:55
anddam	basically I have different hosts where I have a codebase, and I am trying to edit those on a single machine, mine, that is somehow configured how I like	21:55
	osse: going through git would mean continuously committing on one side and pulling on the other	21:55
osse	m0viefreak: ooh, perfect	21:55
anddam	osse: I am trying to sync those using unison (data synchronizer) but I am using a three way setup, since only one of my system is accessible from all the others	21:56
	turns out the directory I had were not on the same starting point, so I synced files that were not under revision control (yet)	21:57
	s/directory/repository/	21:57
	→ vdamewood joined	22:03
	← Brumaire left	22:04
	→ christall joined	22:05
	← EvilDMP left	22:06
	← jwillikers left	22:06
	→ EvilDMP joined	22:08
	→ mago joined	22:10
mago	Im having trouble pushing..	22:10
	HELP SOS	22:10
	fatal: the remote end hung up unexpectedly	22:10
	← Lord_of_Life left	22:10
mago	First I got that, then i increased the buffer..	22:10
	git config http.postBuffer 999288000	22:11
	Tried some random values..	22:11
	Then I get that:	22:11
	fatal: protocol error: bad line length 81920.72 MiB/s	22:11
	→ Lord_of_Life joined	22:11
mago	I need to push large folders, the web interface wont do..	22:11
	Is github trying to make it as hard of posiblo, lol..	22:11
	← molt left	22:12
lantech19446	hopefully the people who helped me this morning are still here, I was kind of rude and didn't thank you guys for all your help, it really appreciated I couldn't have gotten up and running without you and I apologize for not saying it sooner my head was just spinning this morning	22:14
	→ molt joined	22:14
mago	If you are up for it you could help me..	22:14
	← cmc left	22:15
	→ cmc joined	22:15
lantech19446	I can try, what do you need help with?	22:16
	→ odoood joined	22:16
lantech19446	mago: are you able to use ssh?	22:19
mago	I need to push a folder to github but its not working..	22:19
	← Wulf left	22:19
mago	lantech19446 dont know..	22:19
lantech19446	what OS are you on?	22:19
mago	lantech19446 i thinks im not using it right now.. I want to upload here:	22:19
	https://github.com/Formula-HP-Computer-Vision-on-Wheels/Group2	22:19
	Im following those steps:	22:21
	https://github.community/t/how-to-upload-an-entire-folder/1056/7	22:21
lantech19446	I don't get why you'd push it twice	22:21
	is it succeeding the first time?	22:21
	when you do git push -u origin master?	22:22
mago	It fails..	22:23
lantech19446	ok, are you on windows?	22:23
mago	Ubuntu 20.04..	22:23
	First I get: fatal: the remote end hung up unexpectedly	22:23
	Then i increase the buffer: git config http.postBuffer 999288000	22:23
	Then I get: fatal: protocol error: bad line length 81920.72 MiB/s	22:23
lantech19446	this might be completely off the mark but make sure that openssh is installed	22:23
	I'm really so new to this i'm grasping at straws just like you are but it seems like we're the only ones here so it's worth a shot	22:24
	← Bombe left	22:25
	→ Bombe joined	22:25
eschwartz	http post buffers do sound more like http than like ssh...	22:26
mago	Ok, so i should be doing this through ssh.. How?	22:26
	← christall left	22:26
lantech19446	eschwartz: that was my next attempt I just found someone who says they solved it with git config --unset http.postBuffer but they were on windows	22:27
	mago: I wasn't saying to do it through ssh it's possible to but the way I understand it is that git requires it maybe eschwartz can shed some light on that like I said i'm really new like today new	22:28
	I found a solution online that agrees with eschwartz about it being http buffer if you run the command I just told him about it may help	22:28
mago	Ok, ill try to unset the buffer..	22:28
lantech19446	if it makes you feel any better I had a ton of troubles getting git up and running this morning a couple people here were helping me all morning	22:30
kyshoc	mago: are you behind a proxy of any sort?	22:30
lantech19446	eschwartz: I thought I knew you from one of the channels lol we're both in the arch channels	22:30
mago	kyshoc no	22:31
eschwartz	that is true	22:31
mago	git clone is easy but git push OMG	22:31
lantech19446	I knew i recognized your nick I just couldn't place it right away I haven't been very active in there recently	22:31
eschwartz	to be fair, neither have I	22:31
kyshoc	mago: okay, you mentioned this is a "large folder", what do you mean exactly — are there very large individual files in this folder, or a lot of small ones? what's the total size of the folder, and the largest single file?	22:32
lantech19446	sometimes i look and they're having really heady conversations and i just want to relax	22:32
	→ roadie` joined	22:33
	→ forgotmynick joined	22:33
mago	It wont let me look now while i try to pus but its a few hundreds Mb folder below 800 Mb if im correct	22:33
lantech19446	oh that's not that big man I did close to a gig earlier and it worked	22:34
mago	The biggest file is like 60 Mb I think	22:34
	The folder is 540 MB kyshoc	22:35
kyshoc	okay, that should be fine — just fyi though, GitHub itself (not git) has a hard limit of 100MB per file	22:36
	cool, that should be fine too	22:36
lantech19446	huh I'm glad you said that kyshoc cuz i had no idea	22:36
kyshoc	mago: what's your git version (run `git --version` to check)?	22:37
mago	git version 2.25.1	22:37
	← roadie` left	22:37
kyshoc	lantech19446: I learned it the hard way :-) docs here: https://docs.github.com/en/github/managing-large-files/working-with-large-files/conditions-for-large-files	22:38
mago	After unsetting the buffer I get: fatal: the remote end hung up unexpectedly	22:38
	Could it be a permissions issue?	22:38
lantech19446	thanks kyshoc	22:39
	→ lechner joined	22:40
mago	It says I have direct acces and that i am admin	22:40
	I guess ssh is the way.	22:42
kyshoc	mago: try running `GIT_TRACE=1 GIT_CURL_VERBOSE=1 git push origin master`, it won't fix anything but might give us a better idea of what's going on	22:42
mago	Ok	22:43
	← keypusher left	22:43
	← palasso left	22:45
	→ kaisyu joined	22:48
	→ johnny joined	22:48
	→ Adoi joined	22:49
	← EvilDMP left	22:49
johnny	howdy folks. there seems to be some tooling out there to rebase incrementally. is there one anyone can recommend?	22:50
	i just wanna redo the rebase on top of every commit since. it's a lot of minor changes, but it's easy to miss the differences	22:51
	← sniperwolf left	22:53
	← Samian left	22:55
	← lgc left	22:55
	→ keypusher joined	22:55
	← Gurkenglas left	22:56
	← wender left	22:56
mago	kyshoc thats what i get:	22:56
	https://pastebin.com/raw/yctitubT	22:56
	← odoood left	22:57
mago	The 1.49GB seems off, i only got <600Mb, but i did run a command thrice i think..	22:57
	← jstein left	22:59
	← molt left	23:00
	→ molt joined	23:01
mago	Did someone get to have a look at the paste?	23:02
	← pizdets_ left	23:02
	→ pizdets joined	23:04
mago	How do i see what i already added to git through git add? Will running the command multiple times with the same folder add it multiple times?	23:04
	→ keypushe- joined	23:05
	← keypusher left	23:05
	← Adoi left	23:06
	→ fstd joined	23:06
	keypushe- → keypusher	23:08
	→ howdoi joined	23:08
kyshoc	mago: to that last line: `git status` will tell you what files you've "staged" to git but haven't committed yet, and `git diff --staged` will show you their contents. running `git add` multiple times will not add multiple copies of your files to Git unless you also commit and change those files in between runs of `git add`.	23:08
	→ christall joined	23:09
mago	kyshoc is just that when I do: git push origin master	23:09
	The number of objects and the size doesnt seem to match with what im actually uploading..	23:09
	← christall left	23:10
	→ christall joined	23:11
	← vdamewood left	23:11
kyshoc	mago: yeah, 1.6GB worth of objects for ~500MB of files does seem strange. what do you get when you run `git log`? this will show you your commit history	23:13
mago	Ok	23:13
	← christall left	23:14
mago	OK, it shows like I have alread comitted two times, but they were fails and i cant see anythign where its supposed to be..	23:14
	→ christall joined	23:14
mago	https://pastebin.com/raw/5wyszTDm	23:15
kyshoc	okay, yeah having two commits might be contributing	23:15
	also, I noticed in that pastebin from earlier that git is using HTTP/2, if you think your ISP might be doing something silly you could try forcing git to use HTTP/1.1 with `git config --global http.version HTTP/1.1`	23:16
	^ I'd suggest trying that just to rule it out as a cause	23:16
mago	Ok, that and then rerun the command i was using? Something silly like what?	23:17
	← natrys left	23:17
kyshoc	correct, run that git config I just shared, then try `git push origin master` again	23:18
mago	kyshoc usually higher version number is better isnt it?	23:18
	← vysn left	23:19
kyshoc	usually, but not always :-)	23:20
	← beshr left	23:21
	← humanface left	23:23
mago	kyshoc i s the command: git push -u origin master	23:25
	the same?	23:25
kyshoc	yes, that'll work!	23:25
mago	i got error again	23:25
	fatal: the remote end hung up unexpectedly	23:25
	Everything up-to-date	23:25
	→ Guest23 joined	23:30
	→ Xenguy joined	23:30
	← mago left	23:33
	← FFY00 left	23:34
	→ bambanxx joined	23:34
	→ fabio__ joined	23:34
	→ platta joined	23:35
	← FabioP left	23:39
	← Guest23 left	23:39
	→ beshr joined	23:44
	→ FFY00 joined	23:45
	← FFY00 left	23:47
	→ FFY00 joined	23:48
	← FFY00 left	23:49
	← cliluw left	23:54
	→ durham_ joined	23:57
	← durham left	23:57
	← durham_ left	23:59
	→ durham joined	23:59

Logs Search ←Prev date Next date→

Channels Documentation