IRCloggy #git 2021-09-28

Logs Search ←Prev date Next date→ Channels Documentation

Provider of IRC logs since 2005.
WARNING: As Freenode became unjoinable and lost all warnings in topics, we cannot log channels on Freenode anymore.

2021-09-28

→ pinage404 joined00:01
← hornhack left00:01
→ tuxayo joined00:02
→ Helmholtz joined00:05
→ christall joined00:05
→ Vipulk joined00:06
→ ormaaj joined00:08
← alzgh left00:10
→ alzgh joined00:11
→ nabaiste^ joined00:12
← R2robot left00:19
→ R2robot joined00:23
→ cliluw joined00:25
← christall left00:30
← ZacSharp left00:31
→ hbautista_ joined00:31
← hbautista left00:34
← Samian left00:39
→ SpeakerToMeat joined00:42
SpeakerToMeat Hi00:42
Quick stupid question, is there any form of add which will add all the changed files to the commit, but none of the untracked ones?00:42
← subopt left00:45
→ Samian joined00:47
← cliluw left00:48
← nyah left00:49
← hbautista_ left00:50
→ drew` joined00:50
bremner isn't that one of the documented options (maybe -A?)00:52
→ hbautista joined00:52
→ ttree joined00:54
→ Null_A joined00:56
→ hqdruxn08_ joined00:57
→ hqdruxn08__ joined00:59
→ getSomethingGood joined00:59
← hqdruxn08 left01:01
← hqdruxn08_ left01:03
BtbN SpeakerToMeat, commit -a will do that, but it's not exactly the same obviously01:03
SpeakerToMeat I found a hint somewhere.01:05
git add -u01:05
← Null_A left01:07
← bambanxx left01:08
← alzgh left01:11
← lxsameer left01:13
→ Null_A joined01:14
← emf left01:18
← Null_A left01:26
→ Null_A joined01:26
← EvilDMP left01:27
→ jcbjoe2021 joined01:38
← _xor left01:40
→ christall joined01:41
← kilobyte_ch left01:42
→ kilobyte_ch joined01:44
← ChmEarl left01:45
← christall left01:45
← retroid_ left01:49
← Samian left01:51
← adlternative left01:53
← gast0n left01:54
→ stats4647 joined02:01
→ gast0n joined02:03
← Null_A left02:06
← stats4647 left02:06
→ ferdna joined02:08
← kilobyte_ch left02:14
→ Null_A joined02:24
→ kilobyte_ch joined02:28
← Null_A left02:28
← ttree left02:30
→ _xor joined02:32
→ mooz joined02:37
← FinnElija left02:39
→ FinnElija joined02:40
→ jcbjoe2021_ joined02:41
← terrorjack left02:49
→ gouldsam joined02:52
→ sudoforge joined02:52
← jcbjoe2021 left03:07
jcbjoe2021_ → jcbjoe202103:07
→ vysn joined03:12
→ Treleure joined03:21
→ node1 joined03:29
← AbleBacon left03:30
← mooz left03:32
→ ttree joined03:34
→ justsomeguy joined03:37
→ Tetractys joined03:40
← justsomeguy left03:41
→ christall joined03:42
→ node1_ joined03:44
← node1_ left03:44
→ node1_ joined03:44
← node1 left03:48
→ Samian joined03:50
← node1_ left03:51
← navidr left03:58
← lgc left04:01
→ bytehackr_ joined04:06
← ttree left04:10
← drogas left04:14
→ srinidhi joined04:16
← christall left04:18
← fjmorazan left04:20
→ fjmorazan joined04:20
→ mooz joined04:25
← gast0n left04:27
← stkrdknmibalz left04:34
→ darkstardevx joined04:37
→ zumba_addict joined04:45
→ ThomasD13 joined04:48
→ aniruddha joined04:51
← nabaiste^ left04:56
← gouldsam left05:00
← hbautista left05:02
← Treleure left05:07
defjam → clouding05:07
← YuGiOhJCJ left05:10
clouding → defjam05:14
→ dimi1947 joined05:19
→ stats4647 joined05:20
← Murr left05:21
→ Murr joined05:22
← stats4647 left05:24
← getSomethingGood left05:27
Hash → JudyRuliani05:35
JudyRuliani → JudyRiuliani05:35
← crabbedhaloablut left05:36
→ crabbedhaloablut joined05:37
→ rfuentess joined05:46
← jkordani left05:49
→ retroid_ joined05:49
← mic4ael left05:55
→ mic4ael joined05:55
→ nightstrike joined05:55
← ferdna left05:56
→ twb joined05:58
twb This annoys me constantly: error: object a6727941433ee1c91a20ede6cb381af1d18c566d: missingSpaceBeforeDate: invalid author/committer line - missing space before date05:59
← sudoforge left05:59
twb How can I have transfer.fsckobjects=true by default, *BUT* turn it off for a single "git clone" command?05:59
bytehackr_ → bytehackr05:59
twb i.e. I want git clone -o transfer.fsckobjects=false https://github.com/coreutils/coreutils05:59
Oh damn. it's just git -c not git -o06:00
The trick was looking in "man git" not "git --help" or "git config --help"06:00
gitinfo the git manpage is available at https://gitirc.eu/git.html06:00
→ af joined06:08
→ christall joined06:15
→ kenanmarasli joined06:20
← thiago left06:22
drew` → drew06:24
← dimi1947 left06:32
→ otisolsen70 joined06:35
← Furai left06:46
hendry i have a strange issue where i'm modifing a file, commit it, and then on my git show it shows that i've deleted it. wth.06:48
→ Furai joined06:50
→ sQVe joined06:51
← christall left06:52
twb hendry: "git show" by default shows the most recent commit06:56
hendry: perhaps you also did "git mv" or "git rm" and forgot?06:57
→ theoceaniscool joined06:59
→ alkino joined07:00
→ EvilDMP joined07:04
← EvilDMP left07:04
← unluckyshrubbery left07:04
← skapata left07:05
→ EvilDMP joined07:05
← sQVe left07:06
← Betal left07:06
→ mithrin joined07:08
mithrin Hey, how to add custom CA to git certificate store?07:08
twb What git certificate store07:09
mithrin: http.sslCAPath ?07:09
→ alzgh joined07:11
twb mithrin: on Debian it looks like git (mostly) uses libcurl-gnutls so put your certificate in /usr/local/share/ca-certificates/fuck.crt and then run "sudo update-ca-certificates"07:12
mithrin git config --global http.sslBackend schannel helped07:12
twb mithrin: if you mean an SSH certificate, that goes in ~/.ssh/authorized_keys on the SSH server07:12
mithrin thank you all07:12
twb mithrin: interesting. Out of curiosity, what OS/distro are you using?07:12
mithrin Windows 10 Enterprise07:13
twb Ah OK07:13
← CalimeroTeknik left07:13
→ CalimeroTeknik joined07:13
twb Then yeah I guess schannel causes it to use the Windows system keyring?07:14
osse one of them07:15
sounds relevant only for https07:15
← molt left07:17
twb Does git:// support TLS? I never really thought about it07:18
← reset left07:22
← ackyshake left07:22
→ skapata joined07:23
← skapata left07:23
→ skapata joined07:23
osse no clue07:23
twb I mean presumably the use case is something like "git clone https://github.com/…/euthanasia-howto"07:24
→ sQVe joined07:25
← jazzy left07:25
→ ackyshake joined07:25
ikke no, not natively07:28
You could use some tls tunneling tool, but you'd have to use it on both sides of the connection07:29
twb ikke: thanks07:29
← Samian left07:31
← aniruddha left07:31
→ christall joined07:35
← durham left07:35
← vysn left07:37
← nrg left07:39
→ gouldsam joined07:39
← zumba_addict left07:39
← christall left07:40
→ nrg joined07:40
→ stats4647 joined07:41
→ durham joined07:41
← bn_work left07:45
→ gas51627 joined07:47
← ackyshake left07:47
← stats4647 left07:49
→ ackyshake joined07:50
← weyhmueller left07:55
→ weyhmueller joined07:55
→ aniruddha joined07:58
→ Guest1 joined07:59
← Guest1 left07:59
→ hornhack joined08:01
→ node1 joined08:01
node1 Hi08:02
osse my name is08:02
node1 Is there anyway to do silent installation of "git config with username/ secret tocken and remote repository"08:03
secret token == password*08:03
← daoudr_ left08:03
node1 silent here means unattended configuration.08:03
← hackinghorn left08:04
twb node1: are you just saving credentials into .git/config?08:04
← Tetractys left08:04
node1 twb, yes if the action is unattended or silent.08:05
→ rsx joined08:05
node1 twb, do you know anyway?08:06
twb node1: I don't know a way; that seems like something you should avoid08:07
node1 why i should avoid?08:07
twb Because storing secrets outside of a TCB is shit08:07
node1: when you say secret password, are you specifically referring to an SSH password?08:08
node1: if so, upgrading to SSH keys is usually very easy. SSH certificates are even better, but are more work.08:09
node1 github token which has least privilege access.08:09
twb ah OK I'm not familiar with that08:09
node1 ok08:10
twb Because I am a lazy hipocrite, I just told github to trust my existing ~/.ssh/ed2551908:10
* ~/.ssh/id_ed2551908:10
node1 will promote ssh in my next version, till yet things should be simple.08:10
twb (Every time I go to clean up, I get bogged down trying to find a trustworthy handset, and ragequit)08:11
← mat001 left08:11
→ lxsameer joined08:12
← Oddly left08:14
→ Oddly joined08:14
node1 !man gitcredentials08:16
gitinfo The git man pages are available online at https://gitirc.eu/git.html. Or were you looking for the "man git-foo" syntax (without the !)?08:16
the gitcredentials manpage is available at https://gitirc.eu/gitcredentials.html08:16
→ pretty_dumm_guy joined08:17
twb node1: I'm just guessing, but have you tried git clone https://alice:swordfish@github.com/pornhub/memcachedb08:18
← hnOsmium0001 left08:19
node1 twb, Your intention does not looks good to me. :(08:19
node1 need to read and develop lot of things08:20
← node1 left08:20
← bytehackr left08:23
← rahl left08:29
→ furrymcgee joined08:30
→ rahl joined08:31
→ terrorjack joined08:32
← EvilDMP left08:38
→ EvilDMP joined08:40
← gouldsam left08:43
→ vdamewood joined08:50
← Lunatrius left08:52
← igemnace left08:52
→ igemnace joined08:54
→ drogas joined08:57
← elastic_dog left08:58
← ano left08:59
→ elastic_dog joined08:59
→ unluckyshrubbery joined09:09
→ hackinghorn joined09:12
← hornhack left09:14
← cmc left09:14
→ cmc joined09:15
← cmc left09:15
→ cmc joined09:16
← pulse left09:21
← Murr left09:21
→ Murr joined09:21
→ cousteau joined09:25
← bodiccea left09:38
→ natrys joined09:41
→ vladoski joined09:45
→ masber joined09:47
→ meator joined09:52
← durham left09:57
→ durham joined09:57
→ AnapodoPsalidi joined10:00
← AnapodoPsalidaki left10:00
→ ConstantDesc joined10:07
← furrymcgee left10:14
→ furrymcgee joined10:14
→ vlado joined10:16
→ christall joined10:16
← ConstantDesc left10:18
← vladoski left10:18
→ saroy joined10:32
→ rewrit3 joined10:33
← lxsameer left10:33
← af left10:36
→ mthall joined10:37
← skapata left10:39
→ af joined10:41
→ elkalamar joined10:41
← igemnace left10:42
← meator left10:46
→ Xaldafax joined10:55
← christall left10:56
→ amosbird_ joined10:56
← amosbird left10:56
→ ConstantDesc joined10:59
→ betelgeuse joined10:59
→ meator joined11:01
→ PCatinean joined11:05
→ tuskkk____ joined11:10
→ enoq joined11:14
gxt_ → gxt11:18
← Vonter left11:26
→ Vonter joined11:27
→ SpeakerToMeat__ joined11:32
← SpeakerToMeat left11:33
← elkalamar left11:36
← ConstantDesc left11:38
→ lxsameer joined11:39
→ speckz joined11:41
← momomo left11:43
→ momomo joined11:47
← durham left11:47
→ durham joined11:48
→ ConstantDesc joined11:51
← speckz left11:51
→ bodiccea joined11:53
← srinidhi left11:54
→ sandipan joined11:58
← saroy left11:58
→ AbleBacon joined12:02
← durham left12:02
→ stats4647 joined12:03
→ spaceone joined12:05
spaceone is there some command to do git reset $file && git checkout $file in one step?12:05
→ durham joined12:06
→ tolot joined12:06
osse git checkout HEAD $file12:06
spaceone git reset @ "$file"12:06
ah, thanks12:06
osse I think @ will work too12:07
jast but not with reset :)12:07
@ is a shorthand for HEAD12:07
← stats4647 left12:08
jast quick reference:12:08
!checkout12:08
gitinfo checkout does two different things! (1) no file/path argument: 'git checkout <ref>' = switches to a branch or commit. Uncommitted changes are carried over; if they don't apply cleanly the whole operation rolls back. (2) 'git checkout [<commit>] -- <path>' = overwrites the given files/paths with different versions, from <commit> or (if not given) from the index.12:08
jast !reset12:08
gitinfo reset does two things! (1) without file/path argument: 'git reset [flags] [<commit>]' = make the current branch point to <commit> (default: HEAD). --soft = don't do anything else. --mixed (default) = overwrite the index to match. --hard = overwrite the working files to match. (2) 'git reset [<commit>] -- <path>' = overwrite the index entries for <path> with the content from <commit> (default: HEAD)12:08
cousteau Didn't know @ was shorthand for HEAD12:10
jast it didn't always exist12:10
← ghost43_ left12:10
jast not sure when it was added12:10
→ ghost43 joined12:11
cousteau Also, what is reset+checkout supposed to do? Forget all uncommitted changes (indexed or not)? Isn't that what reset --hard does?12:12
osse pretty sure it's been around for a fairly long time. But only recently did it start working for e.g. push12:12
cousteau: --hard doesn't work for a single file12:12
jast it's easy to get the two different forms of reset mixed up12:12
the syntax for checkout and reset is one of the biggest design flaws in Git's CLI IMO12:13
cousteau osse: oh, oops12:13
Yeah I just saw that in the manpage12:13
jast: now there's git switch to partly fix the weirdness of checkout12:14
→ average joined12:14
cousteau But I don't know to which extent it fixes it12:14
jast git switch and git restore each cover a little less than half of what "git checkout" does12:15
it's not a good improvement IMO12:15
cousteau Also I don't see why you wouldn't be able to reset --soft/--mixed/--hard a single file. I mean it should be possible and would make sense, right?12:16
osse those flags only make sense if you change the value of HEAD12:16
cousteau At least reset to the HEAD commit12:16
Yeah good point12:16
osse I suppose git could accept them if the commit argument is absent or equal to HEAD...12:17
cousteau soft changes the HEAD, mixed changes HEAD and index, and hard changes HEAD, index, and tree; right?12:18
osse yep12:18
← furrymcgee left12:21
cousteau Perfect12:24
anddam howdy, I am using log --color --graph --abbrev-commit --format='…' with a lengthy format in a repo with lot of commmits, the output takes some time to pop up and I figure this is because output has to be collected as a whole before parsing it12:29
→ pulse joined12:29
anddam is there a way to speed up the output? namely I am looking for changes in a single directory of a packages repo, that is need to only list a handful of entries out of many thousands12:30
for comparison git log --oneline is pretty fast to produce output, I'd like to add date to that12:31
osse I'd gradually reduce the format until yuo find the culprit12:33
but I am fairly sure that git does not read the whole log before starting to produce output12:33
My guess is that the main culprit is --graph12:33
→ gh34 joined12:37
bremner --reverse would be another suspect option12:39
→ mat001 joined12:39
→ christall joined12:48
← nightstrike left12:50
← ConstantDesc left12:50
← christall left12:50
→ gouldsam joined12:50
→ christall joined12:51
→ Swahili joined12:51
← bambams left12:58
→ bambams joined13:00
→ zer0bitz joined13:00
jast btw guys I made an overview of checkout vs reset, comments welcome13:01
https://gitirc.eu/help/checkout-reset.html13:01
← bambams left13:03
→ nyah joined13:03
xD- nice13:03
jast msg gitinfo .weblogin13:04
whoops13:04
typing is hard13:04
let's go shopping13:04
← ThorMojito left13:04
osse Did you write all of that just now?13:05
jast yes13:05
cbreak jast: for git checkout [--] <path> you write that it doesn't update the index13:05
but ... does it matter?13:05
it could update the index with itself13:05
jast mainly for understanding what's going on13:05
basically changing the index isn't what you're after when you use this13:06
cbreak it'd look nicer if the whole column was green13:06
and no, I don't just say that because I played too much tetris13:06
but it's almost a 4 stack13:06
jast but then why have the column in the first place13:06
maybe I should change the labels to be more imperative13:07
cbreak good question13:07
jast I changed the column headers13:07
cbreak I would make a column about "current branch updated"13:07
jast it's a bit subtle13:07
that's "checkout without path"13:08
section below13:08
→ igemnace joined13:08
jast but maybe it doesn't hurt being clear about it multiple times13:08
osse as long as it's fewer than 813:09
cbreak you write that git reset --soft doesn't change the branch pointer13:09
jast how about this13:09
← igemnace left13:09
cbreak ah, right, same issue as before13:09
it changes it to HEAD in my mind13:09
I find it easier to remember that git reset (without path) always changes the branch pointer13:10
jast again I changed the column headers13:10
cbreak and with path always changes the index13:10
jast I added footnotes :)13:12
← cmbengue left13:12
→ cmbengue joined13:12
→ sudoforge joined13:15
← cmbengue left13:15
cbreak :)13:16
I like it13:17
→ speckz joined13:17
jast .trigger_edit checkout-reset Having trouble keeping all of the various forms of "git checkout" and "git reset" apart? Here's an overview: https://gitirc.eu/help/checkout-reset.html13:18
gitinfo jast: Okay.13:18
→ ThorMojito joined13:18
→ bambanxx joined13:19
→ bn_work joined13:23
→ ApostleInTriumph joined13:26
← ApostleInTriumph left13:27
→ ApostleInTriumph joined13:27
ApostleInTriumph hello. I'm trying to push to a private repo which exists. but I get the following error from terminal, fatal: repository 'https://github.com/user/Project.git/' not found13:28
→ furrymcgee joined13:28
jast ApostleInTriumph: just to clarify, this is also the error message you get if your Git client doesn't try to authenticate (which can happen because GitHub pretends that repos don't exist if you're unauthenticated)13:30
what can help is including your username in the URL or in your local repo's config13:31
git config credential.user your-github-username13:31
sorry, it's credential.username, not credential.user13:31
ApostleInTriumph jast thanks! I tried it, still the error persists13:32
→ Guest0 joined13:32
jast does it ask for a password?13:33
Guest0 → jkordani13:33
← furrymcgee left13:34
← jkordani left13:34
← meator left13:35
→ jkordani joined13:35
→ mtor joined13:35
ApostleInTriumph yes13:35
← mtor left13:35
→ meator joined13:35
ApostleInTriumph jast when I enter wrong password, it says remote: Support for password authentication was removed on August 13, 2021. Please use a personal access token instead.13:35
when I enter the right one (the personal access token), it says remote: Repository not found.13:36
jast do you have a personal access token set up? with full "repo" access enabled?13:36
→ otisolsen70_ joined13:36
jast that's the bold "repo" checkbox at the top of the list when setting up a token13:36
← rfuentess left13:37
ApostleInTriumph okay my bad that was the issue13:37
jast oh, good. I like issues that are explainable. :)13:37
ApostleInTriumph why was password deprecated at the first place to replace with this painfully long random string that's hard to store and impossible to remember?13:38
jast you'll have to ask the GitHub folks that13:38
though the password was never usable if you enabled two-factor auth13:38
maybe they just didn't want to support two things in their Git auth13:38
← otisolsen70 left13:39
jast maybe they just realized that many people use git's credential store to save their git HTTP passwords and think that the credential store getting stolen is an important attack vector13:39
→ jwillikers joined13:40
osse more like attack unordered map, amirite13:40
jast deallocates osse13:40
osse forks+execs in his destructor13:41
jast anyway, the security theatre in all these web services annoys the hell out of me, too13:41
if you want to avoid all of that, use SSH access to repos13:41
→ igemnace joined13:41
← ThomasD13 left13:41
jast osse: what's a destructor13:41
osse the thing that runs when a thing goes out of scope13:42
or is deallocated13:42
cbreak passwords are bad, because people are too lazy to use proper ones, which are long and impossible to remember13:42
jast is that some kind of C++ madness13:42
cbreak ssh keys are much nicer to work with.13:42
← mthall left13:42
← otisolsen70_ left13:42
jast agreed13:42
osse jast: that too. lots of languages have them13:42
jast though I have both passwords and SSH keys in my password manager13:42
→ mthall joined13:42
jast destructors are one of the things that I find a little too magical13:43
cbreak I have passwords in my password manager, and ssh keys on my physical keychain via yubikeys13:43
jast for instance, what happens if a destructor raises an exception?13:43
cbreak jast: the object fails to be destroyed13:43
osse sub DESTROY { } is perhaps more up your alley13:43
cbreak if this happens while an exception is already being processed, your program will be terminated13:44
and by default, if you let an exception escape your destructor, your program will also be terminated13:44
jast it was a rhetorical question btw13:44
anyway I prefer RAII style code13:45
→ ConstantDesc joined13:45
osse huh13:46
isn't that the point of destructors13:46
cbreak raii without destructors is like pizza without the dough.13:46
osse or are you yanking our chains13:46
12speed dura ace chains13:47
jast I don't really know13:47
I think as long as exceptions aren't a thing it's fine13:47
osse If a destructor raises an exception I'd wager that the code you would have to write if destructors weren't a thing would raise an exception too13:48
I bet ... ONE THOUSAND DOLLARS13:49
cbreak a destructor that raises an exception should also raise an eyebrow or two during code review.13:49
jast the problem is that if you do semi-complex things in a destructor, often it's very hard to see any exceptions it may raise13:49
that's also the main downside of exceptions13:50
cbreak it's ok.13:50
by default, exceptions aren't allowed to leave destructors.13:50
jast at least with a concept like Rust's Result types you'll always be forced to explicitly panic on errors13:50
cbreak so if you screw up handling them inside, it's termination.13:50
jast that's pretty terrible though13:50
imagine getting a termination in 1 out of 10000 runs13:50
cbreak why?13:50
if you fail to handle errors, that's what you get.13:51
jast I'm saying exceptions can make it hard to even be aware of where you might need to handle errors13:51
cbreak -> go handle your errors13:51
osse Speaking of failing, I fail to see how the concept of destructors makes the real life problem any worse than it would be without them13:51
jast well you did start out mentioning a destructor that forks-and-execs :)13:52
I don't really want to live in a world where that's a thing13:52
cbreak fork and exec are more a legacy C thing13:53
it has many problems with non-prehistoric code13:53
avoid it if possible13:53
jast it's what pretty much everything does under the hood13:53
cbreak there are plenty of alternatives13:53
← igemnace left13:54
osse jast: Yeah but that was a meaningless joke :P It was intended to be about as realistic as the concept of deallocating a human13:55
jast AFAIK on Linux there is nothing else except fork() that can create new processes13:55
cbreak vfork with exec13:55
posix_spawn13:55
osse bro do u even clone()13:55
jast AFAIK posix_spawn wraps fork and exec13:55
cousteau cbreak: mind you, good passwords are long AND easy to remember13:56
cbreak cousteau: why easy to remember?13:56
cousteau But those are kinda hard to come up with13:56
jast yeah well, in the end clone() does the same thing as fork(), except with more options and slightly different semantics13:56
cbreak You can use jrbaCMEook898+DGO09Xh, for free13:56
cousteau Because if they're hard to remember you'll end up writing them down or having to reset the password every time13:57
osse "cousteau and cbreak are fighting over passwords" is long and easy to remember \o/13:57
jast use a password manager, done13:57
osse: damn you, stop leaking my master password on IRC13:57
cbreak cousteau: the idea is to write them down, indeed13:57
here, have a new one: 4JVamhqxFBL9nDdDzegB5GYgV. Put it into a password manager, just to be safe.13:58
cousteau I'd say that's a potential security risk13:58
cbreak how so?13:58
there are various types of password managers, for varying degrees of paranoia13:58
jast getting your remembered passwords confused in your head is a potential security risk, too, just of a different kind13:58
cousteau Writing them down, I mean. Password managers may or may not be secure enough. I haven't made up my mind yet.13:58
jast the thing is13:58
cbreak and the stand-alone local-only ones should be more secure than anything you can practically keep into your head reliably13:59
jast I have 361 distinct passwords13:59
and that's just for personal accounts13:59
there's a limit to how many passwords you can remember and never get mixed up13:59
cbreak cousteau: think about the attacker model first, that should allow you to determine which kind of password manager is acceptable13:59
cousteau Are they the same password with two added letters at the end (each letter having 19 possible variations)?14:00
jast no14:00
all of them are randomly generated14:00
cousteau Oh14:00
jast the next best thing instead of a password manager is to generate all of your passwords out of a master password and a "domain" value14:01
but then you have to remember which domain value is used where14:01
cousteau So the fact that 19*19=361 was just a coincidence? Dammit :(14:01
← aniruddha left14:01
jast yeah, just a coincidence14:01
I have an additional 291 credentials for work14:02
some of those are for testing, but still14:02
cousteau How secure is that "master password + site-specific value" method, out of curiosity?14:02
jast depends mostly on your master password and the algorithm being used14:02
if you use something like HMAC-SHA256 you should be good on the second point14:03
cousteau Is strcat an acceptable algorithm?14:03
jast or, better yet, a good key derivation function14:03
→ thiago joined14:03
jast a big downside of this is that many services have weird password restrictions14:04
some will force you to use a special character from a specific narrow range, some will reject special characters from the same range, etc.14:04
some even have a maximum number of characters in a password14:04
cousteau Yeah and some have the fantastic idea of limiting the MAX length to some ridiculously small value14:05
jast the only sane way to deal with all of that is to use a password manager and having your password DB encrypted with a strong master password14:05
cousteau I remember the "password must have a min length of 6 and a max length of 8" requirement at uni14:05
cbreak cousteau: remember: Getting cleartext access to some of your passwords isn't all that hard14:08
would access to passwords compromise others?14:08
with a stupid derivation scheme it would14:08
with purely random individual keys it would not14:08
with a good derivation scheme it probably wouldn't14:08
→ Trailblazer51 joined14:08
cousteau Maybe you are right, the thing is that password managers imply (1) me trusting my passwords to software someone else made (kinda easy actually, I'm very little paranoid), (2) lower portability (what if I want to use my password from another PC?), and (3) the fear that they'll somehow access my laptop and figure out my password and then I'm screwed14:09
None of these would be a real concern if I knew more about security, probably14:09
cbreak 1: You have to trust every web page and application you log in to anyway14:09
and you have to trust your OS, your browser, ... this is one more software to trust14:09
2. Depending on the password manager you chose, portability should be OK.14:10
I use 1password, which has applications for MacOS, Linux, Windows, Android, ... and a web interface14:10
(but the web interface is a pain to use...)14:10
3. Yes.14:11
encrypt your file system with FDE. And use a good password manager password :)14:11
cousteau The thing with knowing little about security is that I get paranoid about all the wrong things14:12
← kenanmarasli left14:13
← Trailblazer51 left14:13
samir Hi Everyone. I've been trying to get a program called stagit (https://codemadness.org/stagit.html) working on a VPS which hosts a couple of my remote git repositories.14:14
In the stagit documentation, it mentions having a `url` file, `owner` file and `description` file. This enables you to specify a clone URL, show the repo's owner and a description of the repo, respectively.14:14
I'm familiar with the `description` file as I see one created every time I run `git init --bare` in a remote repo. But I have never seen the `url` file or `owner` file.14:14
Are those maintained just in the local repository under `repo/.git`? Or are they managed in the remote repo with all of the other files such as `config`, `description`, `hooks`, etc?14:15
← hackinghorn left14:17
bremner samir: you should ask the stagit authors your stagit questions14:17
osse my guess is you can simply create the url and owner files next to the description file14:18
samir bremner: I have but they have mentioned that I need to maintain these files. I'm not familiar with them so was wondering if they are normal git files14:18
osse: will try that. Would they be in the local repo or remote repo?14:18
osse they are not14:18
samir ok14:19
→ hackinghorn joined14:19
osse surely it must be in the repo that stagit generates from14:20
samir osse: yeah, I think you are right. I have stagit running in the VPS (remote repo). It looks like it can pick up the git clone url from the `url` file from there. Will try the same thing for the owner.14:22
Thanks!14:22
→ skapata joined14:24
← skapata left14:24
→ skapata joined14:24
cousteau Yes, if it's not something specific to git I wouldn't expect it to be in the .git directory, but on a regular file within the project14:25
Otherwise git won't bother cloning that14:25
vdamewood → vinleod14:27
samir cousteau: Makes sense. Thanks!14:28
← af left14:29
cousteau For example, GitHub searches for GitHub-specific config files in the .github directory14:30
→ af joined14:31
cousteau So if you clone a project from GitHub that has those config files, they will be cloned as well, and if you upload the project to another GitHub repo it'll preserve those files. From git's point of view, they're just regular files in the project.14:32
→ rfuentess joined14:38
← speckz left14:39
→ LuxuryMode joined14:43
→ subopt joined14:45
samir got it. Thanks for that explanation14:47
→ hnOsmium0001 joined14:49
← bambanxx left14:49
→ node1 joined14:51
← twb left14:53
jast samir: I think they did it this way because stagit also supports bare repositories. there is no officially defined way to put custom metadata in a bare repository, so the options you have is adding it to the config file or just creating another plain file next to it (like the "description" file). git itself ignores these.14:53
← average left14:54
jast in fact no part of git (except gitweb if you want to count that) ever looks at the "description" file, either14:54
← drogas left14:56
jast typically if you run something like this on the server side, it will be dealing with bare repos (that's what you normally want for hosting purposes)14:56
← jwillikers left14:58
→ Null_A joined14:59
→ Lunatrius joined15:03
← sQVe left15:04
→ hornhack joined15:05
← hackinghorn left15:06
hornhack → hackinghorn15:06
samir jast: Thanks! You all have been very helpful! I think I need to crack open https://git-scm.com/book15:07
jast you wouldn't get any specific information about these files in the book, though, because they're not a standard thing in git :)15:07
samir good point!15:07
jast that said, reading up a bit probably can't hurt15:08
but we're also happy to explain things you find confusing, or point you to more material15:08
samir I'm not a developer or programmer but feel that it would be good to put some of my files and documents under version control15:08
and given the ubiquity of git, that would be a good way to go15:08
jast sure, that can definitely be useful15:08
samir Thanks, jast!15:08
jast a git repository tends to work best if all of the files in a repo are related15:08
but it's not an absolute requirement if you don't plan on doing things like, say, branching15:09
samir branching is probably a feature I don't need at the moment. My biggest "need" is being able to revert back to a prior version of a file or document15:10
jast sure, that's always an option with git15:10
samir say I updated a config file or "dotfile" and things went bad15:10
→ elkalamar joined15:10
jast it's more geared towards reverting everything back, but both things are possible15:10
samir right now, I have files like `program.conf.1`, `program.conf.bak`, `program.conf.crap`, etc.15:11
jast yeah, that's no way to live :-)15:11
samir sad, but true15:11
jast preprint2 of final Thesis draft2 - Copy.docx15:12
samir LOL!15:12
imMute needs more Final Final15:13
jast more - Copy - Copy, too15:13
ikke and Really Final15:14
jast also, remember to store a subtly different and slightly more up-to-date copy in the temp/ subdir15:14
anyway15:14
the most important thing in versioning things with git is that if you ever want to look back, you'll appreciate having written good commit messages that briefly describe the changes and your motivation for making them15:15
← rfuentess left15:15
→ yates joined15:16
yates is there a way to find a tag given the hash?15:16
jast of the commit?15:17
git tag --points-at15:17
yates jast: thanks, but i mean the reverse15:17
jast so you have the hash of the tag object?15:18
yates yes15:18
← ApostleInTriumph left15:18
→ shored joined15:18
yates specifically i'm tryign to see if one of the linux kernel tags matches this hash: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tag/?h=v5.10.415:18
→ kenanmarasli joined15:18
jast I'm not sure if there's an Optimal(tm) way, but there's always "git show-ref | grep hash"15:18
yates there is no 5.10.4 tag15:19
→ rfuentess joined15:19
→ vysn joined15:20
← masber left15:20
jast the tag name is v5.10.4 and at least in that hosted repo it exists15:20
vinleod → vdamewood15:20
← Murr left15:21
→ Murr joined15:21
yates is this not the "master" linux repo? https://github.com/torvalds/linux.git15:22
→ ChmEarl joined15:23
→ pyeverything joined15:23
jast I don't know exactly how all of the linux repos are connected15:23
← node1 left15:23
jast historically the official repo has been on git.kernel.org15:23
and I *think* linus isn't the maintainer of the stable kernels15:23
bookworm GKH is15:24
jast yeah, that's what I thought15:24
bookworm and the other guy, can't remember the name15:24
→ ttree joined15:24
bookworm the problem with the kernels is that development is at least in part based on sending patches around, meaning for stable that some commits actually differ in their hashes as they are slightly rewritten and cherry picked15:25
→ tirnanog joined15:26
bookworm iow, comparing the stable tree with torvald's isn't exactly trivial15:26
← gouldsam left15:28
→ bambanxx joined15:31
→ junktext joined15:36
← Null_A left15:39
→ Gaboradon joined15:42
→ ash_worksi joined15:46
→ cmbengue joined15:48
← tuskkk____ left15:49
ash_worksi so, I had to leave yesterday, but I still need help with difftool15:52
the last suggestion, which was to use the -t option, yielded the same results as before (no sign the application has or will launch; exit status 0)15:53
→ nate_ joined15:55
← pyeverything left15:56
slaughtr → slaughter15:56
→ ysavd^ joined15:56
← justache left15:57
ash_worksi oh wow15:58
→ justache joined15:58
ash_worksi I removed all local difftool configuration and removed the guitool and it worked15:59
yay15:59
→ Maxattax joined16:03
→ beshr joined16:08
← enoq left16:14
← vlado left16:19
→ stats4647 joined16:19
← ttree left16:29
← sandipan left16:32
→ sandipan joined16:32
→ Brumaire joined16:34
← ConstantDesc left16:34
→ bambams joined16:35
← rfuentess left16:38
→ jazzy joined16:39
→ Narrat joined16:40
← jkordani left16:42
→ zebrag joined16:44
→ pcatinean_ joined16:47
← PCatinean left16:50
← ysavd^ left16:50
← alzgh left16:51
→ alzgh joined16:51
→ ysavd^ joined16:51
← Narrat left16:57
→ Narrat joined16:57
← sandipan left16:58
→ saroy joined16:58
→ sandipan joined17:00
← cousteau left17:01
→ ano joined17:02
→ node1 joined17:02
← saroy left17:03
← elkalamar left17:03
→ emf_ joined17:03
→ koolazer joined17:09
← stats4647 left17:09
← alzgh left17:09
→ alzgh joined17:10
→ stats4647 joined17:10
→ ConstantDesc joined17:10
← christall left17:11
← ConstantDesc left17:11
yates how do i show the hash of a tag in git?17:12
→ pyeverything joined17:13
another| annotated tag or simple tag?17:14
yates the tags such as v5.10.4 in https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git17:16
if i've already cloned it and checked the tag out?17:16
← pyeverything left17:18
another| those are likely annotated tags17:18
→ christall joined17:19
← ash_worksi left17:19
nedbat yates: just curious: why do you want the hash?17:19
another| do you need the hash of the tag itself or the commit it's referencing?17:19
← christall left17:23
← vysn left17:25
← gas51627 left17:26
→ shabius joined17:28
yates what is the long number in parentheses next to the "tag name" here? https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tag/?h=v5.10.417:28
perhaps i'm abusing the syntax: is it called a "hash" or a "blob"?17:29
nedbat: i want to verify i've got the right repo and tag version.17:30
another|: the commit17:30
i've already drug through the dirt from using https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git instead of https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git17:31
just want to be sure i've got the right thang (as they say in Lousianna)17:32
another| `git show v5.10.4` ?17:32
yates another|: yes, that's what i converged to after RTFM. thanks17:33
← _xor left17:34
← sandipan left17:34
→ Guest10 joined17:35
← Guest10 left17:36
→ christall joined17:36
→ _xor joined17:39
→ ScottMichaud joined17:41
→ elkalamar joined17:42
→ jstein joined17:42
← pcatinean_ left17:43
← Narrat left17:44
→ Narrat joined17:44
→ ash_worksi joined17:44
← rsx left17:45
← node1 left17:46
← ScottMichaud left17:47
→ YoungChiefBTW joined17:48
→ hbautista joined17:49
→ orbyt joined17:51
← LuxuryMode left17:53
← ash_worksi left17:55
← emf_ left17:55
→ aniruddha joined17:56
← zebrag left17:59
→ phoenixh joined18:03
← elkalamar left18:05
→ lgc joined18:09
nvmd ok so how can one move around in a repo with submodules without getting "failed to unpack tree object ..." and "Submodule X could not be updated"?18:10
← phoenixh left18:11
nvmd at this point I've had so many issues with submodules I wonder why people ever bothered adding it to git18:13
← christall left18:13
→ christall joined18:13
← christall left18:18
→ hbautista_ joined18:20
→ christall joined18:21
← darkstardevx left18:21
← hbautista left18:21
→ reset joined18:23
→ liefer joined18:23
→ pyeverything joined18:24
→ impermanence joined18:24
← christall left18:25
← meator left18:32
→ darkstardevx joined18:38
← EvilDMP left18:39
← hackinghorn left18:42
→ hackinghorn joined18:43
→ christall joined18:46
→ emf joined18:54
→ fling joined18:55
← rtjure left18:59
← SpeakerToMeat__ left19:02
← pyeverything left19:02
→ pyeverything joined19:03
← christall left19:04
← beshr left19:05
→ average joined19:07
← emf left19:11
→ mei joined19:12
→ emf joined19:13
→ SpeakerToMeat__ joined19:20
→ Adoi joined19:21
← zer0bitz left19:23
→ Betal joined19:23
→ rwp joined19:23
← kenanmarasli left19:26
← emf left19:28
→ christall joined19:31
→ ZacSharp joined19:32
← defjam left19:33
← Adoi left19:38
← geirha left19:38
→ EvilDMP joined19:40
→ emf joined19:42
drew I want to cherry-pick the two patches at the tip of one branch to the current branch. Would that be `git cherry-pick my_branch:HEAD~2..HEAD` ?19:47
→ Samian joined19:49
← christall left19:58
osse drew: git cherry-pick onebranch~2..onebranch20:01
drew thanks osse20:02
osse drew: cherry-pick (along with most commands) do their changes on the current branch, so all you need is a way to specify the commits you want to cherry-pick20:05
my_branch:HEAD is invented syntax :P I guess it means the HEAD of my_branch. but "my_branch" alone is the head (lowercase) of my_branch, and HEAD (uppercase) always means the current commit absolutely20:06
← ZacSharp left20:08
← fling left20:12
→ fling joined20:13
→ geirha joined20:16
← Samian left20:17
← pyeverything left20:19
← alkino left20:20
→ Samian joined20:21
drew gotcha lol20:21
→ christall joined20:24
→ mukul joined20:28
← tolot left20:30
← yates left20:31
→ pyeveryt_ joined20:31
→ ash_worksi joined20:32
← mukul left20:35
→ yates joined20:39
← terrorjack left20:42
← vdamewood left20:43
← christall left20:48
→ christall joined20:48
← cdown left20:52
→ zebrag joined20:53
→ cdown joined20:57
← ysavd^ left21:02
→ vysn joined21:07
← ano left21:07
→ Lord_of_Life_ joined21:12
← Lord_of_Life left21:13
→ rafasc joined21:14
→ Null_A joined21:14
Lord_of_Life_ → Lord_of_Life21:15
← rafasc left21:16
→ rafasc joined21:16
← Murr left21:21
→ Murr joined21:21
← elastic_dog left21:23
→ squirrel joined21:24
→ onizu joined21:25
yates if i have a valid commit, how do i determine what branch that commit belongs to?21:25
→ elkalamar joined21:26
rafasc yates: git branch --contains <commit>21:26
yates is it via "git show <commit>"?21:26
← Brumaire left21:26
→ rtjure joined21:27
yates rafasc: thanks: http://paste.debian.net/1213638/21:27
but now i'm still confused: i thought "HEAD" is simply the tip of whatever branch you're in21:28
rafasc and it is, in a sense.21:28
the problem here is that you are in no branch.21:29
you have a commit checked out directly.21:29
yates isn't each and every commit a part of some branch?21:29
rafasc try: git branch -a --contains <commit>21:29
yates: not necessarily.21:30
→ elastic_dog joined21:30
rafasc e.g. you can tag a commit, push it, and delete the branch.21:30
then you have a commit, that exists, can be checked out, and does not belong to any kind of branch.21:31
!detached21:31
gitinfo A detached HEAD (aka "no branch") occurs when your HEAD does not point at a branch. New commits will NOT be added to any branch, and can easily be !lost. This can happen if you a) check out a tag, remote tracking branch, or SHA; or b) if you are in a submodule; or you are in the middle of a c) am or d) rebase that is stuck/conflicted. See !reattach21:31
← gh34 left21:31
yates https://www.youtube.com/watch?v=CDeG4S-mJts21:32
nedbat yates: branches are references to commits. a commit doesn't have to be referenced by a branch, or be in the ancestors of any branch.21:34
yates rafasc: so that extended -a command gives this: http://paste.debian.net/1213640/21:34
→ phylaz joined21:34
nedbat yates: does that tell you what you need?21:35
rafasc that means that commit is reachable from all those branches.21:35
← phylaz left21:35
yates i guess so21:35
thanks rafasc, nedbat21:35
rafasc !next then?21:35
gitinfo Another satisfied customer. NEXT!21:35
yates i'm going back to svn...21:36
rafasc NO!21:36
nedbat yates: i have confidence in you. you can do this.21:36
yates nedbat: lol.21:36
nedbat: hand me my cane, please21:36
← zebrag left21:37
rafasc hits yates in the head with his own cane.21:38
rafasc s/his/their21:38
you meme video saying someone charging 200$/h for git advice.. everyone here doing it for free.21:40
sometimes I wonder if that meme is funny for people that understand german.21:42
→ ZacSharp joined21:42
yates rafasc: it's a reflection on the (over-)complexity of git, not on the kindness of people here.21:43
i wonder if that meme is funny for people who ARE german...21:44
nedbat yates: git is more complex than svn. but it's not over-complex. it could use better ux sometimes, but the facilities are all there for good reason.21:45
yates hits nedbat with his cane21:45
nedbat yates: is the cane a reference to your age? :)21:46
→ Gustavo6046_ joined21:46
← elkalamar left21:46
yates yes. don't tell me you're older than me... 195721:46
← Gustavo6046 left21:47
yates old, yeah, but i grew up with the Beatles!21:48
nedbat yates: ok, you got me, but only by five years.21:48
yates i remember rotary-dial phones and 300-baud acoustic modems...21:48
nedbat yup21:49
osse FIIIIIVE YEEEAEAARS21:49
Gustavo6046_ → Gustavo604621:49
nedbat osse: ? :)21:49
osse bowie21:49
yates osse: +121:49
nedbat osse: i don't know that song21:50
yates David Bowie21:50
← nvmd left21:50
nedbat yes, i got that :)21:53
← Narrat left21:53
← stats4647 left21:53
osse https://www.youtube.com/watch?v=2ObjtVdsV3I21:54
→ stats4647 joined21:54
→ terrorjack joined21:57
yates yeah, i don't remember that song either21:58
← Samian left21:58
yates osse: you a bowie fan? (probably a stupid question..)21:59
osse not in particular22:00
i just like a lot of the songs that everyone likes22:00
→ ConstantDesc joined22:01
← christall left22:03
→ christall joined22:04
yates nedbat: didn't we interact on #python some months ago?22:04
nedbat yates: probably. i'm more active there22:05
yates countenance falls at his apparent lack of popularity22:05
← christall left22:08
yates in git-speak, what is the nomenclature of the hash representing a commit? hash? blob? what?22:09
commit b1313fe517ca3703119dcc99ef3bbf75ab42bcfb (HEAD, tag: v5.10.4)22:10
what is b1313fe517ca3703119dcc99ef3bbf75ab42bcfb called?22:10
during this century?22:11
← ConstantDesc left22:11
tang^ IRC is not instant messaging22:12
and it's a hash22:12
squirrel how is irc not instant messaging?22:13
yates because tang^ doesn't want it to be instant messaging, whatever instant messaging means to tang^... ?22:14
tang^ it's not guaranteed to be instant22:14
yates close enough for me.22:14
what's the delay, two seconds?22:14
osse concidering you gave us just over 78 years to reply I'd say you got an answer pretty quickly22:14
yates tang^: thanks22:15
osse: where did 78 come from?22:16
tang^ number of whole years left in this decade22:16
osse i was jsut referring to the "during this century" thing. but maybe you meant something else22:16
squirrel heh22:17
tang^ err, that's what I meant, this century22:17
nedbat yates: sometimes people refer to the hash as the sha also22:17
rafasc commit id also an acceptable term.22:18
tang^ SHA-1 hash if you wanted to be detailed22:18
rafasc eventually we'll change to another sha.22:19
tang^ true enough22:19
rafasc git help glossary also mentions object name22:19
object name22:20
The unique identifier of an object. The object name is usually represented by a 40 character hexadecimal string. Also colloquially called SHA-1.22:20
→ nehsou^ joined22:20
→ cagmz joined22:21
yates i didn't know there was a "git help glossary" command. that's pretty cool/useful22:21
← aniruddha left22:21
rafasc try git help -g22:23
← vysn left22:24
← travisghansen left22:24
rafasc man git config can be very informative.22:24
gitinfo the git-config manpage is available at https://gitirc.eu/git-config.html22:24
yates rafasc: very cool too!22:25
rafasc but that's more useful to look things up.22:26
osse also man gitrevisions22:26
gitinfo the gitrevisions manpage is available at https://gitirc.eu/gitrevisions.html22:26
yates query: does everyone here know what the slang term "hosed" means? "Jennifer really hosed my stick shift when I loaned her the car yesterday."22:26
→ travisghansen joined22:26
yates someone actually asked me about it on another channel recently... made me feel OLD22:27
osse screwed up? broke?22:27
yates yes22:27
ruined22:28
osse hah22:28
and english isn't even my mother's thong!22:28
→ jstein_ joined22:28
bremner thanks for that image22:28
rafasc osse: is she called Jennifer?22:28
yates i hear you. my wife is a Tamilian from Bangalore and speaks better English than 95 percent of Americans.22:29
rafasc bremner: I didn't have the image until you mentioned it.22:29
bremner rafasc: you're welcome22:29
yates i'm not sure i even want to ask..22:29
JudyRiuliani → Hash22:29
→ ConstantDesc joined22:30
rafasc zzz time. G'night all o/22:30
yates rafasc: good night old person22:30
← rafasc left22:31
yates oh sorry, that's nedbat22:31
← ConstantDesc left22:31
→ jess joined22:37
→ christall joined22:37
← SpeakerToMeat__ left22:38
→ SpeakerToMeat__ joined22:39
← pyeveryt_ left22:42
→ sdasd joined22:43
← emf left22:44
← squirrel left22:45
sdasd → squirrel22:45
← rtjure left22:45
← jstein_ left22:48
← jstein left22:48
→ vicfred joined22:51
← theoceaniscool left22:52
→ pyeveryt_ joined22:53
← orbyt left22:55
← pyeveryt_ left22:58
← christall left22:58
→ atrigent joined23:02
← pizdets left23:03
→ pizdets joined23:04
→ igemnace joined23:07
← squirrel left23:09
← R2robot left23:17
→ ttree joined23:22
→ R2robot joined23:22
→ christall joined23:25
← JanC left23:40
→ JanC joined23:40
← ZacSharp left23:41
← bambanxx left23:47
← mei left23:49
← natrys left23:51
← christall left23:54
→ Duckle9 joined23:54
→ stkrdknmibalz joined23:55
← rtur left23:55
← mat001 left23:56
← Duckle left23:56
Duckle9 → Duckle23:56
→ zebrag joined23:56
→ rtur joined23:57
← cagmz left23:58
→ mat001 joined23:58

Logs Search ←Prev date Next date→ Channels Documentation