IRCloggy #git 2022-11-17

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.

2022-11-17

← bindu left00:00
→ arescorpio joined00:00
← odoood left00:00
→ odoood joined00:03
← arescorpio left00:03
← odoood left00:03
→ bindu joined00:04
→ gast0n joined00:05
← nyah left00:05
→ nyah joined00:06
← dvanloon left00:07
→ softirq joined00:14
→ dvanloon joined00:15
← oprypin left00:15
→ hiroot joined00:15
← zebrag left00:16
← Fusl left00:16
→ madewokherd joined00:17
← blahboybaz left00:18
→ oprypin joined00:20
← Murr left00:21
← NorrinRadd left00:22
→ Murr joined00:22
← Echoz left00:22
→ Guest40 joined00:22
Guest40 hello00:22
a doubt I have about Git00:23
I do commit A at 12:0000:23
then I do commit B --amend at 20:0000:24
I was expecting the history entry of the commit to have the 20:00 timestamp00:24
→ NorrinRadd joined00:24
Guest40 but i has 12:00 timestamp00:24
→ Techcable joined00:24
Guest40 is it the supposed behavior?00:25
→ Fusl joined00:28
Guest40 ping?00:29
→ cdown joined00:30
→ trafficjam joined00:32
→ Xenguy joined00:36
← softirq left00:42
→ softirq joined00:43
ajak that sounds right00:50
← Guest40 left00:50
← geosmile left00:51
→ TheCatBe1 joined00:52
← TheCatBehemoth left00:53
← NorrinRadd left00:53
→ bdeferme2 joined00:54
← squirrel left00:56
← duxsco left00:58
→ NorrinRadd joined00:58
← delay left00:59
← TheCatBe1 left01:00
→ audio joined01:00
← zeenk left01:08
→ blahboybaz joined01:09
← perrierjouet left01:11
← softirq left01:16
→ perrierjouet joined01:17
→ bywaterloo joined01:18
→ Echoz joined01:25
← bywaterloo left01:37
← chexum left01:40
← mei left01:41
→ chexum joined01:41
→ mei joined01:41
← ferdna left01:42
← derpydoo left01:43
→ TheCatBe1 joined01:43
→ srinidhi joined01:45
→ softirq joined01:47
→ lucasta joined01:47
→ wallabra joined01:51
← lucasta left01:54
← rustyshackleford left01:54
← bodiccea left01:55
← jacobk left02:00
→ atrigent joined02:02
→ bywaterloo joined02:05
→ infernix joined02:07
← bket left02:07
← gxt left02:09
← umbramalison left02:10
→ umbramalison joined02:11
→ kexec joined02:11
→ jacobk joined02:12
→ gxt joined02:13
→ bket joined02:13
← TheCatBe1 left02:22
← blahboybaz left02:22
→ xx joined02:23
← mgedmin left02:30
← jacobk left02:33
→ mgedmin joined02:33
← nyah left02:37
← softirq left02:40
→ lucasta joined02:42
→ eXtr3m0_ joined02:43
← eXtr3m0_ left02:44
→ eXtr3m0_ joined02:44
← eXtr3m0 left02:44
eXtr3m0_ → eXtr3m002:45
→ dulange__ joined02:45
← dulange_ left02:48
← mandeep left02:48
→ YuGiOhJCJ joined02:52
← YuGiOhJCJ left02:56
→ YuGiOhJCJ joined02:56
→ softirq joined03:00
← lucasta left03:00
← Stilgar left03:06
← SuperLag left03:08
→ SuperLag joined03:08
← gxt left03:13
→ gxt joined03:14
← xx left03:14
→ xx joined03:14
← softirq left03:19
→ lucasta joined03:20
→ softirq joined03:21
→ SirRippovMaple50 joined03:24
← rgrinberg left03:24
← lucasta left03:34
← cdown left03:35
← bywaterloo left03:35
→ Stilgar joined03:35
← Cleverness left03:38
← softirq left03:40
← perrierjouet left03:41
← kexec left03:41
→ perrierjouet joined03:42
→ bodiccea joined03:44
→ softirq joined03:46
→ finn_elija joined03:47
← FinnElija left03:47
finn_elija → FinnElija03:47
→ dsrt^ joined04:01
→ mven joined04:04
→ lucasta joined04:04
← Stilgar left04:13
→ Stilgar joined04:14
← sudoforge left04:15
→ jacobk joined04:20
← lucasta left04:24
← gsi left04:24
→ gsi joined04:26
← softirq left04:26
→ softirq joined04:26
← softirq left04:31
← lechner left04:31
→ lechner joined04:32
← lechner left04:36
← gast0n left04:37
→ zebrag joined04:38
→ TheCatBehemoth joined04:39
← zebrag left04:39
→ Jong joined04:47
→ blahboybaz joined04:50
→ lolok joined04:53
→ lucasta joined04:54
→ lechner joined04:56
← srinidhi left05:00
→ __ELFrederich joined05:20
← _ELFrederich left05:20
← lechner left05:21
← skapata left05:21
→ lechner joined05:30
← Lord_of_Life left05:31
→ Lord_of_Life joined05:31
← atrigent left05:32
→ atrigent joined05:33
← small left05:38
← lucasta left05:43
← cation left05:47
→ nate4 joined05:49
← Shentino_ left05:50
→ small joined05:54
← nate4 left05:55
← makara1 left05:56
← makara left05:56
→ makara joined05:58
→ makara1 joined05:59
→ Shentino joined05:59
→ maret joined06:01
← Jong left06:02
→ cation joined06:04
→ han-solo joined06:04
← audio left06:07
→ FH_thecat joined06:10
→ blaklistd joined06:14
← TheCatBehemoth left06:19
→ madewokherd` joined06:25
← madewokherd left06:29
→ AyLaa23 joined06:32
← fstd left06:34
→ fstd joined06:34
← shailangsa left06:35
→ ajfriesen joined06:35
← AyLaa23 left06:35
← MaverickFischer left06:35
← specing left06:36
← majors left06:36
→ MaverickFischer joined06:36
→ majors joined06:36
← NightMonkey left06:36
← jacobk left06:36
→ sco1_ joined06:36
← sco1 left06:36
← arahael left06:37
→ NightMonkey joined06:37
→ jacobk joined06:38
→ specing joined06:38
← elastic_dog left06:42
← moldorcoder7 left06:42
→ elastic_dog joined06:43
→ csm310599 joined06:44
← thiago left06:44
← bgs left06:46
← kyshoc left06:47
→ kyshoc joined06:47
← kyshoc left06:47
→ kyshoc joined06:47
→ arahael joined06:49
→ sa0 joined06:50
→ agarr joined06:53
→ otisolsen70 joined06:54
→ igemnace joined06:56
← otisolsen70 left06:57
→ otisolsen70 joined06:58
→ shailangsa joined07:02
→ KnoP joined07:10
← FH_thecat left07:10
→ FH_thecat joined07:11
→ srinidhi joined07:11
→ mihael joined07:13
mihael Is it possible to get a repo size without cloning the repo?07:13
← NorrinRadd left07:18
→ agarr1 joined07:20
← bn_work left07:21
← agarr left07:23
agarr1 → agarr07:23
grawity not really07:28
the closest you could get is directly running `du -hs /path/to/repo` on the git server07:28
but even that will likely be different from a fresh clone07:28
← otisolsen70 left07:29
← LuxuryMode left07:35
→ carl- joined07:35
→ carl-- joined07:37
→ hamster joined07:39
← ham left07:40
← carl- left07:40
← agarr left07:41
→ peirik__ joined07:41
← xx left07:43
→ carl--- joined07:43
→ rfuentess joined07:43
← dulange__ left07:44
← carl-- left07:46
← trafficjam left07:51
← blahboybaz left07:51
→ Jong joined07:54
→ zeenk joined08:01
→ agarr joined08:03
← ali1234 left08:04
→ ali1234 joined08:05
← FH_thecat left08:07
→ theoceaniscool joined08:07
→ FH_thecat joined08:08
← Jong left08:08
→ aidalgol joined08:10
→ oxymoron93 joined08:14
→ hqdruxn08_ joined08:14
→ feriman joined08:17
← hqdruxn08 left08:17
← Murr left08:21
→ Murr joined08:21
← jazzy left08:22
← dermoth left08:23
→ xx joined08:25
→ mncheck joined08:25
→ otisolsen70 joined08:25
← agarr left08:26
→ zmt01 joined08:29
← bodiccea left08:30
→ bodiccea joined08:30
← zmt00 left08:32
← GrapeSlush left08:33
← peirik__ left08:35
→ dermoth joined08:35
← jfsimon1981 left08:35
→ GrapeSlush joined08:37
→ jfsimon1981 joined08:39
→ TheCatBehemoth joined08:41
→ MajorBiscuit joined08:42
→ belsirk joined08:44
← TheCatBehemoth left08:45
→ peirik__ joined08:46
← rfuentess left08:47
← BlessJah left08:50
→ BlessJah joined08:50
→ agarr joined08:52
→ cami||e_ joined08:54
→ coot joined08:56
→ EvilDMP joined08:59
→ dulange__ joined09:03
← maret left09:03
→ maret joined09:07
→ Guest40 joined09:10
← troyt left09:16
→ troyt joined09:18
→ sedzcat joined09:23
→ rfuentess__ joined09:27
← belsirk left09:30
→ dionysus69 joined09:32
← agarr left09:35
→ Nintendo joined09:35
Nintendo I have a branch called “docker” that hasn't been accepted by upstream yet, but I need it to run a Docker container. Then, I have a branch named “feature” that I need to push upstream. I want to test the feature branch (which is based off upstream's master), but I need my Docker branch's changes in order to do so.09:38
I thought the correct way of doing this was: git rebase --onto docker feature09:39
But I ran that, and all the changes from the feature branch disappeared. I don't know how to get them back, but I can see the last (and only) commit in .git/COMMIT_EDITMSG.09:40
grawity if you have an older "good" commit id, you can e.g. `git reset --keep` to it09:41
→ gareppa joined09:41
grawity if you don't have one, try to find it in `git reflog feature`09:41
commits might disappear after a successful rebase if that particular change was *already* in the "base"09:41
Nintendo Thank you. I do have a good commit, but I don't see it in git log, so I don't know the ID.09:42
It is in git reflog, though.09:43
So I am saved... I suppose the question now is: What command should I have run instead of rebase --onto?09:44
← gareppa left09:50
→ tirnanog joined09:50
→ donofrio joined09:50
→ nate4 joined09:52
← donofrio__ left09:53
← peirik__ left09:53
→ duxsco joined09:55
→ peirik__ joined09:55
→ strk joined09:56
← nate4 left09:57
osse Nintendo: if you wanted to rebase feature onto docker then you should presumably either have run: git rebase docker, or git rebase --onto docker master when stood on the feature branch09:58
replace master with whatever branch the feature branch is branched off from09:58
→ fandre1986 joined10:01
gsi another approach would be: keep the two branches as they are, start another one from master, and temporarily merge what you need into there, test, then drop10:02
branches are cheap and useful, but nobody said they must be kept after they served their purpose :)10:02
the first merge probably will be fast forward, the next may or may not conflict, depends on the types of changes in the branches, and maybe their order, but that's to be expected I guess10:04
← Aminda left10:04
gsi you kind of "do in advance" during testing what later merges between test and regular use happens, it's useful to see those conflicts, too10:05
Nintendo Yes, thank you both. I ended up doing just that, gsi, in order to test.10:05
git rebase -i --onto testing master docker # with “testing” being my cheap and useful temp branch off “feature”10:06
git rebase -i --onto docker master testing # Also, for whatever reason, the reverse failed.10:07
← perrierjouet left10:07
→ Aminda joined10:08
→ thblt joined10:10
Nintendo Back to actual testing... nothing breaks. After pushing, I will be a free man once again. 😃10:10
→ perrierjouet joined10:11
thblt Some days ago I asked about issues with gpg signatures from git with multiple subkeys on multiple smartcards. I finally came up with a working solution to refresh git config (and also sshcontrol) with the currently available keys, here it is: https://raw.githubusercontent.com/thblt/dotfiles/master/.local/bin/refresh-gpg-cards10:11
(There is no question, this is a follow-up) (also "crossposting", sorry to the people at the intersection of #gnupg and #git)10:12
← mihael left10:15
→ drogas joined10:21
← Murr left10:21
→ Murr joined10:21
← maret left10:26
← hamburgler left10:26
← hiroot left10:26
→ maret joined10:36
→ roadie joined10:36
→ delay joined10:44
→ agarr joined10:45
← delay left10:51
→ mobidrop joined10:53
← feriman left10:54
→ mei_ joined10:57
→ biapy joined10:59
← mei left10:59
gsi am not affected myself, but: thank you for following up with the solution! :-D11:01
← farzat left11:02
← Stilgar left11:03
→ Stilgar joined11:03
→ mei joined11:04
← mei left11:04
← YuGiOhJCJ left11:05
← duxsco left11:05
→ duxsco joined11:06
→ farzat joined11:07
→ jackneill joined11:07
← mei_ left11:07
← peirik__ left11:08
← dvanloon left11:10
← wavesong left11:13
→ peirik__ joined11:14
← agarr left11:15
← Guest40 left11:15
→ TheCatBehemoth joined11:16
← TheCatBehemoth left11:21
→ Guest96 joined11:22
→ iomari891 joined11:23
→ mei joined11:23
Guest96 during initial clone, the remote says something like 'remote: Compressing objects: 100% (20249/20249), done.', and spends an considerable amount of time in that step. Is there a way to disable this step? In a LAN the compression may not gain much. Depending on what the remote actually compresses in that step...11:24
→ agowa338 joined11:24
ikke It's delta compression, which is baked in the protocol11:25
Guest96 so I would need to throw more MHz into the issue, which I do not have. Guess I need to live with it.11:26
gsi "compression" may not matter because the LAN is fast, but still _packing_ is useful because even LANs have latency11:27
→ atrigent_ joined11:27
← phylaz left11:27
grawity I guess you could tune git upload-pack on the server to do less work (no idea what it does *after* 100%, but it's probably something to do with git-pack-objects), or even clone through a file-based protocol (like dumb-http or nfs) instead of the smart protocols11:27
Guest96 does core.compression affect this compression?11:28
I may set it to 111:28
→ swamp_ joined11:28
grawity it influences pack.compression so probably, but for packfiles there's a lot more going on than just zlib compression11:28
ikke Guest96: it's the compression used for individual objects, so I would not expect it to affect that part11:28
grawity I wouldn't be surprised if that's already the cheapest part...11:29
← atrigent left11:29
Guest96 pack.compression defaults to -n1, according to the docs, so zlib decides.11:29
grawity see the other pack.* options11:29
→ dvanloon joined11:30
← zmt01 left11:31
→ Major_Biscuit joined11:31
← oxymoron93 left11:31
← Guest96 left11:31
→ vitali64 joined11:33
→ oxymoron93 joined11:34
← small left11:35
← MajorBiscuit left11:35
→ szkl joined11:39
→ wavesong joined11:40
← LuKaRo_ left11:41
← cami||e_ left11:43
→ LuKaRo joined11:44
← xx left11:46
→ lucasta joined11:48
→ Sulky joined11:54
← lucasta left11:55
← lolok left11:56
← Sulky left11:58
→ delay joined11:58
← Aminda left12:00
← Celelibi left12:03
← derjanni_ left12:04
→ derjanni joined12:06
→ Sulky joined12:06
→ Aminda joined12:06
→ wootehfoot joined12:07
→ Celelibi joined12:08
→ cami||e_ joined12:08
← peirik__ left12:08
→ peirik__ joined12:10
← Sulky left12:11
→ Sulky joined12:12
← Millicent left12:16
→ Millicent joined12:17
← roadie left12:18
← Murr left12:21
→ Murr joined12:21
← vitali64 left12:25
← oxymoron93 left12:36
→ oxymoron93 joined12:37
→ small joined12:37
← duxsco left12:41
→ duxsco joined12:41
← fandre1986 left12:54
→ ccapndave joined12:56
← delay left13:01
→ gas51627 joined13:04
→ blackchief joined13:05
→ softirq joined13:05
→ smeeagle joined13:05
← Xenguy left13:07
→ Xenguy joined13:08
← llh left13:09
← dvanloon left13:12
→ dvanloon joined13:16
← t3nj1n left13:17
→ t3nj1n joined13:17
← peirik__ left13:18
← __ad left13:19
→ ad__ joined13:19
→ phylaz joined13:20
→ nyah joined13:20
← ccapndave left13:21
→ peirik__ joined13:21
← iomari891 left13:22
→ ccapndave joined13:23
← KnoP left13:24
← rfuentess__ left13:24
← ccapndave left13:25
→ ccapndave joined13:25
← peirik__ left13:27
→ sd5 joined13:27
→ llh joined13:28
← ccapndave left13:29
← smeeagle left13:30
← maret left13:32
→ delay joined13:34
→ smeeagle joined13:35
→ thuna` joined13:39
→ bywaterloo joined13:39
→ iomari891 joined13:42
← phylaz left13:43
← mei left13:45
→ mei joined13:45
→ Earnestly joined13:53
→ KnoP joined13:53
→ nate4 joined13:53
→ peirik__ joined13:54
Earnestly What is an appropriate solution to using local mirrors (which become local clones) for submodules? With 2.38.1 this is now considered unsafe: https://github.blog/2022-10-18-git-security-vulnerabilities-announced/#cve-2022-3925313:54
Currently clone from a mirror and then use git submodule to change the path to the local repo13:55
← duxsco left13:56
Earnestly Which now produces the error: > fatal: transport 'file' not allowed13:56
← delay left13:56
Earnestly Is there perhaps a more sensible alternative to just using: git -c protocol.file.allow=always?13:56
→ ccapndave joined13:57
← softirq left13:58
← nate4 left13:58
← smeeagle left14:00
sco1_ → sco114:03
← Aminda left14:04
→ Aminda joined14:04
→ YellowSuB joined14:05
← c0dehu1k left14:06
→ cdown joined14:06
→ cdown_ joined14:08
→ irrgit_ joined14:09
← csm310599 left14:09
← Dotz0cat left14:10
← cdown left14:10
cdown_ → cdown14:10
← thuna` left14:11
← ccapndave left14:12
→ thuna` joined14:12
→ Dotz0cat joined14:13
→ phil170 joined14:14
→ ccapndave joined14:15
← oxymoron93 left14:17
→ csm310599 joined14:17
← Xenguy left14:19
→ rama joined14:19
→ oxymoron93 joined14:20
← Murr left14:21
→ Murr joined14:21
→ eroux joined14:23
→ tga joined14:26
→ gast0n joined14:27
tga greetings. I'm building a kind of audit data lake, recording (probably in csv/parquet files) a bunch of API interactions -- I was just considering using git for versioning, to make sure nothing is ever lost14:27
← jfsimon1981 left14:27
← peirik__ left14:27
→ phil170_ joined14:27
tga my question: what's a repository size that would already be too much for git?14:27
→ jfsimon1981 joined14:27
← ccapndave left14:28
→ ccapndave joined14:28
← phil170 left14:29
→ phylaz joined14:30
← ccapndave left14:33
← RiFo left14:35
→ phil170 joined14:36
← phil170_ left14:36
→ gh34 joined14:37
← ProperNoun left14:38
→ ProperNoun joined14:39
→ peirik__ joined14:40
stutz morning, is there a setting/config to make git use -4 by default?14:45
← rama left14:45
← peirik__ left14:45
→ TheCatBehemoth joined14:46
→ rama joined14:47
← rama left14:48
selckin GIT_SSH_COMMAND="ssh -4" and ~/.curlrc for http14:48
→ rama joined14:48
← rama left14:50
→ rama joined14:51
tga would you consider git for versioning say ~100GB of text files?14:52
so a size large enough that it doesn't fit in memory14:53
ikke tga: I personally would not14:54
tga I want something like S3 versioning but locally14:54
I guess the alternative is to run an actual S3 implementation14:54
ikke tga: maybe something like lfs14:54
tga oh hmm I always thought of LFS as something for single large files, but I guess it would also work for many many small files14:55
stutz selckin thanks14:55
← Sulky left14:56
ikke tga: it would add overhead though14:56
tga I am mostly protecting against wrong deletes and I also want to be able to "prove" that the archive wasn't trivially modified14:56
so nobody can go into a log file and change a few numbers14:56
→ rgrinberg joined14:57
tga lfs can still deal with a file being changed/replaced, right?14:57
← strk left14:59
→ geri joined15:00
← edgars left15:00
← wootehfoot left15:02
← sedzcat left15:02
← mobidrop left15:02
→ thiago joined15:04
← cdown left15:06
→ cdown joined15:06
← atrigent_ left15:06
← rama left15:07
← thiago left15:09
→ rama joined15:12
→ atrigent joined15:12
← phil170 left15:13
→ delay joined15:16
← umbramalison left15:18
→ smeeagle joined15:18
→ umbramalison joined15:19
→ duxsco joined15:19
← smeeagle left15:23
← dsrt^ left15:25
→ kostkon__ joined15:26
← kostkon left15:26
← thuna` left15:27
→ sudoforge joined15:27
← atrigent left15:29
→ peirik__ joined15:29
→ NorrinRadd joined15:29
__ELFrederich → ELFrederich15:32
ELFrederich Hey guys... I have a new Windows 11 laptop. I'm more of a Linux guy but I need to keep Windows on it because I also use the laptop for music. What's the best way to do development on it?... just do native everything? I've seen people doing cool things between VSCode and a WSL Linux distro. Anyone mess around with that?15:34
sorry, thought I was in #golang. Please disregard15:35
leah2 i'd dual boot if i can :p15:35
← delay left15:35
→ softirq joined15:36
→ phil170 joined15:36
→ humanface joined15:37
ELFrederich leah2, unfortunately it's only 512G so I'd rather not. It does have a micro-sd card though. I wonder if some of my larger VSTs can install onto it. Though I'd probably be better off getting a 1G NVMe drive than adding a 512G sd card.15:38
→ atrigent joined15:42
→ jfsimon1981_b joined15:46
→ delay joined15:47
← delay left15:47
← jfsimon1981_b left15:50
← softirq left15:50
← phil170 left15:50
ELFrederich actually, they're helping me out in #go-nuts right now, but I guess now I do have a Git question for anyone running VSCode with WSL2... do you run Git on Windows or withing the WSL2 Linux VM?15:50
ikke ELFrederich: I can answer the same here :P15:51
ELFrederich On my work computer where I don't have WSL let alone WSL2, I use the Windows Git and then end up using Git Bash quite a lot for ssh'ing around. But now on my personal laptop with Windows 11 and WSL2 should I even bother with the Windows version of Git?15:51
ikke, ;-)15:51
Windows Git gives me anxiety every time I install it... too many questions, checkboxes. I get nervous about checking the wrong one.15:52
ikke Why not use it in wsl2 then?15:52
jast unfortunately WSL2 forces the windows hypervisor on you15:55
ikke yes15:55
ELFrederich ikke, was just looking for confirmation that it's a good place to use it from. Not sure if I'll run into issues, for instance... will VSCode running on Windows be able to highlight added, modified and deleted lines within a file?15:55
→ atrigent_ joined15:55
jast virtualbox gets very, very slow when not being able to use its own hypervisor15:55
← Techcable left15:56
← atrigent left15:56
← csm310599 left15:56
jast personally I kind of doubt a Windows VS Code can use a WSL Git properly\15:56
but I haven't tried it myself, I don't really use VS Code15:57
ikke vscode has wsl integration15:57
So you can run the back-end inside wsl15:57
jast true that15:57
ikke That's how I use it, and it works well15:57
You just work entirely in your wsl environment and ignore windows15:57
jast of course that means your project files better be inside WSL, too15:58
ikke yes15:58
That's what I do and prefer15:58
No windows fs shennanigans15:58
→ vladoski joined15:58
jast personally I have everything in a traditional VM... and if I absolutely have to use Windows, I have my editor and tools inside msys2, too15:58
e.g. if I want to do fancy 3D or audio stuff15:59
← NorrinRadd left16:00
jast as for the VST thing, I have a USB SSD for my larger instruments16:00
I'd recommend against using an SD card, SSDs tend to be faster and more long-lived16:01
← rama left16:01
→ jazzy joined16:01
← peirik__ left16:01
→ rama joined16:01
jast since you don't need high write speeds, lower-end SSDs will do the trick16:01
e.g. Crucial X1, WD Blue SSDs, Sandisk Portable series16:03
→ zenobius0 joined16:03
← kostkon__ left16:04
→ kostkon joined16:04
→ softirq joined16:04
jast that said, the faster ones will make your instruments load faster still :) (and you'll want to make sure you're using the drive on a USB "super speed" port)16:06
→ cdown_ joined16:06
← cdown left16:07
cdown_ → cdown16:07
absence if a file full of conflict markers has somehow been marked as resolved, can i undo that? without losing all my changes, which git checkout --conflict=merge seems to do16:09
← Earnestly left16:09
absence i assume it's just a flag or something, but web search is failing me today16:10
jast "git checkout --conflict..." is for when switching branches16:10
← carl--- left16:10
absence jast: ah, i don't know what it's for, it just showed up when i searched for this scenario16:11
jast if you haven't committed yet, you can just keep editing your file and then "git add" it again once it looks clean. of course that means you won't be able to use a merge tool16:11
→ NorrinRadd joined16:11
jast i.e. that's really only an option if the conflict markers are easy to read :)16:11
absence i've already worked more than an hour on this conflict, and i wasn't half finished, so i'd very much like to use a merge tool :D16:12
jast makes sense :>16:13
the problem is that marking the file as resolved removes the pointers to the different versions16:13
it's still possible to manually dig up the correct versions and massage the index accordingly but it's not that easy16:13
oh wait, I think there's a different option16:14
1. make a backup copy of your half-resolved file; 2. "git checkout -m <file>"; 3. move your copy back in place16:15
→ peirik__ joined16:15
→ thiago joined16:15
jast absence: ^ :)16:16
absence i was afraid you'd say that... oh well, it's a LOT of files, but better than doing it over! thanks for confirming16:16
jast wait, you resolved *all* files?16:16
← SuperLag left16:17
absence jast: yes. visual studio code has this button that says "unstage all", which i assumed meant "unstage all the staged files", not "resolve all the conflicted files". you learn something every day...16:18
jast well16:19
it most likely did what it says, i.e. reset the index back to the state in HEAD16:19
under normal circumstances that just unstages everything... but since conflict states are in the index, too, they disappear as a side effect16:20
honestly that button probably shouldn't be active, or should be re-labeled, during merge resolution :)16:20
however here's a trick to do all of them at once16:21
← Murr left16:21
← cdown left16:21
→ Murr joined16:21
jast cd <top of repo>; git ls-files -m -z | xargs -0 tar c -zvf mychanges.tgz; git checkout -m .; tar x -f mychanges.tgz16:22
→ delay joined16:23
→ SuperLag joined16:23
jast btw this only works properly if everything is unstaged because "git ls-files -m" ignores files which have all of their changes staged16:24
→ durham joined16:24
jast you can replace that bit with "git diff --diff-filter=M --name-only -z HEAD | xargs ...", then it will pick up everything that differs from HEAD16:27
→ vitali64 joined16:27
← epicout left16:28
tirnanog that could end up creating a tarball that only contains a subset of the given files. one way to protect against that, if using gnu tar, would be: tar --null -T -16:28
→ epicout joined16:28
tirnanog without xargs.16:28
jast good point16:28
that's what I wanted to do initially but I was too lazy to look up the correct options in tar :)16:29
← zeenk left16:31
tirnanog in this case, they spare us from --files0-from or the like. thankfully. I wonder why they don't have -0 as an alias of --null though.16:32
→ c0dehu1k joined16:32
← c0dehu1k left16:33
→ c0dehu1k joined16:33
absence thanks, i'll tinker with it a bit16:33
→ madewokherd joined16:33
absence wish git could just remember what operation it was doing when it encountered the conflicts to begin with, and do all of this automatically16:34
a pony would be nice too16:34
← madewokherd` left16:36
← durham left16:36
← nyah left16:37
→ durham joined16:37
jast separation of concerns :)16:39
conflicts and conflict resolution is a separate thing from merging and rebasing (and the other things that do file level merges that might cause conflicts)16:39
it would be possible to wrap more UI around this, I suppose16:40
but essentially "git checkout -m <paths>" re-creates the merge in those paths, and merging does its work in the working tree, so I don't think there'd be a straightforward file to restore conflicts without overwriting the file16:41
though it might be possible to do trickery with tree-level merges16:41
let me try and come up with something magical16:41
← coot left16:41
← tga left16:42
→ tga joined16:42
← peirik__ left16:42
← TheCatBehemoth left16:42
jast huh, I just found a machine on which I didn't have the git repo checked out yet... weird :)16:42
→ funhouse joined16:42
← NorrinRadd left16:44
← oxymoron93 left16:46
→ TheCatBehemoth joined16:47
→ bgs joined16:47
→ wootehfoot joined16:49
← delay left16:52
jast so, right now the code path that re-creates the conflict state always writes out the file16:55
but looking at the code I think it should be possible to skip that, it's just not supported right now16:55
→ delay joined16:58
osse git update-index --index-info can do it16:58
without actually touching any files16:59
jast yeah, but then you have to look up the correct blobs yourself16:59
right now I'm trying to find out where "checkout" even gets this info from when stages 1-3 have been removed16:59
← geri left17:00
→ skapata joined17:00
→ blahboybaz joined17:02
jast oh, I see, the index has a "resolve undo" array17:04
hah, found it17:04
absence: "git update-index --unresolve" might prove helpful17:05
... if this happens again, at least17:05
← dionysus69 left17:05
jast I *think* you can just pass it the top-level tree and it will unresolve all files it remembers being conflicted before...17:06
in any case it shouldn't overwrite any files17:06
but if you want to try it, backup your repo, just to be on the safe side17:06
→ feriman joined17:06
← softirq left17:06
→ cdown joined17:07
jast I've always wondered how unmerging worked, so this has been educational17:07
→ derpydoo joined17:07
← cami||e_ left17:08
→ JordiGH joined17:08
JordiGH I can't gpg sign a commit that already exists, right? Signatures are part of the commit, hence new commit and new hash ?17:09
→ cdown_ joined17:09
jast JordiGH: correct17:09
it's technically possible to sign commits after the fact in two ways:17:09
1) create a signed tag17:09
2) put a signature in a "git notes" ref17:10
← derpydoo left17:10
jast AFAIK there is no standardized way to do the latter, but the basic idea goes like this:17:11
→ Betal joined17:11
← umbramalison left17:11
← cdown left17:12
cdown_ → cdown17:12
← rama left17:12
jast git cat-file commit <commit ID> | gpg --sign ... | git notes add -F - --ref notes/signature <commit ID again>17:12
→ NorrinRadd joined17:13
→ rama joined17:13
jast FWIW notes don't scale well to, say, hundreds of thousands of commits17:13
JordiGH I wonder if that will appease github. https://github.com/newrelic/newrelic-lambda-layers/pull/11717:13
← tga left17:13
jast no :)17:14
you'll have to sign all of the commits, e.g. using interactive rebase, and force push17:14
→ softirq joined17:15
JordiGH This is weird, they don't even care what the signature is, do they?17:15
← cweiss076532124 left17:15
JordiGH I can just generate a new signature that isn't verified by anyone.17:15
jast probably, yeah17:15
JordiGH What even is the point of having a signature that isn't trusted by anyone.17:15
osse jast: Success! http://sprunge.us/WZc9Us17:15
I guess17:15
jast possibly they'll reject questionable signatures manually17:15
osse: hah, you missed me finding a built in option for it17:15
osse yeah I saw it17:16
→ dionysus69 joined17:16
jast should work even for rebase and such, provided that the resolve_undo structure is up-to-date17:16
← mven left17:18
→ mven joined17:18
← jacobk left17:21
→ umbramalison joined17:21
← TheCatBehemoth left17:21
→ cweiss076532124 joined17:22
← rgrinberg left17:22
← iomari891 left17:22
→ Murr- joined17:23
← Murr left17:25
Murr- → Murr17:25
← coldtom left17:26
→ tga joined17:27
→ coldtom joined17:27
← zenobius0 left17:27
← rama left17:27
→ rama joined17:28
→ iomari891 joined17:28
← JordiGH left17:30
← Major_Biscuit left17:30
← softirq left17:33
← tga left17:34
← rama left17:35
← cweiss076532124 left17:35
→ tga joined17:36
→ cweiss076532124 joined17:36
← han-solo left17:37
→ irrgit__ joined17:38
← irrgit_ left17:41
→ jacobk joined17:42
→ wagle joined17:43
← tga left17:43
← vladoski left17:45
→ softirq joined17:46
→ thuna` joined17:47
→ rama joined17:49
→ tga joined17:49
← tga left17:49
← dulange__ left17:49
← rama left17:53
→ nate4 joined17:55
→ derpydoo joined17:58
← sa0 left17:58
→ sa0 joined17:58
→ lucasta joined17:59
← nate4 left18:00
← dionysus69 left18:02
← srinidhi left18:03
← lucasta left18:04
← dvanloon left18:04
→ phil170 joined18:04
→ Praxiteles joined18:07
→ dvanloon joined18:07
→ phil170_ joined18:08
→ rama joined18:08
← rama left18:10
→ rama joined18:10
→ DibbTubberOne joined18:11
← phil170 left18:11
← Praxiteles left18:13
→ Praxiteles joined18:14
→ kostkon__ joined18:14
← kostkon left18:15
→ jundran joined18:16
→ coot joined18:17
← iomari891 left18:18
← phil170_ left18:18
← rama left18:20
→ rama joined18:20
← irrgit__ left18:21
← kostkon__ left18:22
→ kostkon joined18:22
→ srinidhi joined18:26
→ audio joined18:26
← kostkon left18:26
→ TheCatBehemoth joined18:27
→ kostkon joined18:27
← rama left18:27
→ irrgit_ joined18:29
→ rama joined18:33
← rama left18:34
→ rama joined18:34
→ lucasta joined18:34
← rama left18:36
→ rama joined18:37
← cdown left18:38
→ rsx joined18:38
→ dulange__ joined18:41
← softirq left18:43
← jacobk left18:45
← delay left18:46
→ softirq joined18:47
← Praxiteles left18:49
→ hamburgler joined18:50
ikke gpg signatures cannot be part of the thing it signs18:51
← Stilgar left18:51
← wpcarro left18:52
→ wpcarro joined18:53
ikke oh, ignore me18:54
gpg signs the object, and that's hashed18:54
← DibbTubberOne left18:55
→ cdown joined18:56
→ cdown_ joined18:58
← rama left18:59
→ rama joined18:59
← gast0n left19:00
← cdown left19:01
cdown_ → cdown19:01
← dvanloon left19:05
→ jacobk joined19:05
→ karabaja4 joined19:06
← rsx left19:07
→ dvanloon joined19:07
→ causasui joined19:08
→ thebombzen_ joined19:10
→ odoood joined19:12
← igemnace left19:13
← rama left19:13
← cdown left19:14
→ rama joined19:14
← softirq left19:16
← truepassion left19:17
→ Stilgar joined19:18
← NorrinRadd left19:20
← vitali64 left19:20
→ NorrinRadd joined19:21
← user____1 left19:29
→ int3r3st3d joined19:36
← int3r3st3d left19:36
→ user____ joined19:37
→ int3r3st3d joined19:37
→ phil170 joined19:37
→ softirq joined19:39
← rama left19:40
← subopt left19:40
→ rama joined19:40
← rama left19:44
← lucasta left19:45
← mei left19:45
← phil170 left19:46
→ phil170 joined19:46
ztrawhcse what an odd thought for Github to require signed commits in a PR19:48
→ donofrio_ joined19:48
ztrawhcse considering that github itself doesn't provide a *method* to merge the PR via fast forward19:48
← phil170 left19:50
← donofrio left19:51
→ JordiGH joined19:56
→ rama joined19:59
→ mei joined20:00
→ KZ-Spectra joined20:01
→ lenkovitsa joined20:03
→ zumba_addict joined20:05
→ remedan_ joined20:07
← remedan left20:08
lenkovitsa → neilim20:09
← odoood left20:09
← omnigoat left20:11
← fling left20:15
← durham left20:15
→ lucasta joined20:16
→ durham joined20:16
→ fling joined20:16
← durham left20:18
← atrigent_ left20:20
→ atrigent joined20:20
← Murr left20:20
→ Murr joined20:21
→ omnigoat joined20:22
← rama left20:24
→ rama joined20:25
← duxsco left20:25
→ lgc joined20:25
← softirq left20:26
← lgc left20:26
← thebombzen_ left20:26
← lucasta left20:30
← agowa338 left20:32
← neilim left20:33
← sentriz left20:34
← rama left20:34
→ sentriz joined20:35
→ rama joined20:35
→ oliveoli joined20:36
← blackchief left20:39
→ denl^ joined20:40
← TheCatBehemoth left20:42
→ m0viefreak joined20:46
→ TheCatBehemoth joined20:48
← zumba_addict left20:51
→ rgrinberg joined20:51
← rgrinberg left20:52
← TheCatBehemoth left20:53
← int3r3st3d left20:58
← visualshock left20:59
→ lucasta joined20:59
→ Praxiteles joined21:00
← rama left21:00
→ rama joined21:01
← bloody_ left21:04
→ polo_ joined21:05
← dvanloon left21:05
→ TheCatBehemoth joined21:06
← coot left21:06
→ bloody joined21:07
→ dvanloon joined21:07
← fling left21:09
→ vladoski joined21:10
→ stef204 joined21:12
→ fling joined21:12
← karabaja4 left21:14
← polo_ left21:15
→ polo_ joined21:15
→ rgrinberg joined21:16
→ ferdna joined21:17
← rama left21:18
← wootehfoot left21:19
→ rama joined21:19
← polo_ left21:20
→ polo_ joined21:20
polo_ → money21:20
→ Techcable joined21:21
← bloody left21:21
← money left21:23
← fling left21:23
→ fling joined21:24
← theoceaniscool left21:25
← lucasta left21:26
→ softirq joined21:27
← oliveoli left21:27
→ polo joined21:28
← rama left21:28
→ rama joined21:29
→ fweht joined21:30
→ blackchief joined21:31
→ imp joined21:31
→ lucasta joined21:32
← fstd left21:36
← softirq left21:38
← otisolsen70 left21:38
← KZ-Spectra left21:38
→ softirq joined21:39
← jundran left21:40
← softirq left21:41
← ELFrederich left21:41
→ softirq joined21:42
← srinidhi left21:42
→ fstd joined21:44
→ jundran joined21:45
← jundran left21:46
← polo left21:46
→ jundran joined21:50
← softirq left21:50
→ polo joined21:50
→ nate4 joined21:56
← sd5 left21:59
→ Erisa joined22:00
← nate4 left22:01
← stef204 left22:03
← chorduser2 left22:07
→ chorduser2 joined22:07
← polo left22:10
← KimK left22:12
← jundran left22:12
→ polo joined22:12
← piwu left22:12
→ piwu joined22:12
→ mathsboy joined22:13
mathsboy Hi all, I'm wondering how to take some commits from one branch and insert them in between two commits on another branch. Not sure if merge -squash or rebase or what is needed. I have a base branch which I checked out so base branch has commits b1 -> b2 and I checked it out at b2. I then made commits on my feature branch: b1 -> b2 -> f1 -> f2. The base branch has been updated and I'd like to take its commits into my feature branch so it22:16
looks like: b1 -> b2 -> b3 -> b4 -> f1 -> f222:16
Any help/advice greatly appreciated :)22:16
ajak interactive rebase, 'edit' where you want to insert the commits, then cherry pick then i guess?22:17
s/pick then/pick them/22:17
mathsboy so I want to be on feature branch, then rebase "onto" base branch?22:17
ajak no you want to do an interactive rebase, 'edit' the commit where you want to insert the commits, then when you get to where you can 'edit' that commit, cherry pick the commits from elsewhere, the continue the interactive rebase22:18
mathsboy ok cool... trying now :)22:19
→ KZ-Spectra joined22:21
← rama left22:22
→ rama joined22:23
← rama left22:23
→ rama joined22:23
→ thebombzen_ joined22:24
ztrawhcse ajak: you can also just add "pick <sha1>" in the interactive rebase list22:27
you're not restricted to only the commits that were originally in between the base and HEAD22:27
ajak oh cool, i figured that might be the case but wanted to avoid the slight complexity of mangling the text22:28
when you could simply do `cherry-pick start..finish` in one go22:28
mathsboy okay,, I think I got it... couple of unntidy things, but it seems to make sense now. Thanks a lot ajak, this is my first ever rebase (not as terrifying as expected)22:29
← m0viefreak left22:30
ajak yay22:30
→ gast0n joined22:30
ajak yeah, easy concept, but not all that intuitive if you've never done it before :)22:30
← polo left22:32
ztrawhcse yeah, doing cherry-pick is easier because you don't have to trust yourself to get the todo-list formatted correctly22:37
→ manjaro-mate joined22:38
← gh34 left22:38
→ arescorpio joined22:38
← arescorpio left22:39
→ beencubed joined22:39
← lucasta left22:41
→ arescorpio joined22:41
→ giu- joined22:44
→ cdown joined22:45
← bgs left22:46
← manjaro-mate left22:46
← feriman left22:46
← bywaterloo left22:49
→ cdown_ joined22:49
← Stilgar left22:50
→ Stilgar joined22:50
← mathsboy left22:51
← cdown left22:52
cdown_ → cdown22:52
→ Yruama joined22:53
→ coot joined22:55
← rama left22:55
← Aminda left22:55
→ rama joined22:56
→ Aminda joined22:56
← fling left22:57
→ fling joined22:59
← rama left23:01
→ rama joined23:01
alexia <ztrawhcse> considering that github itself doesn't provide a *method* to merge the PR via fast forward <-- i thought "rebase and merge" would do that but apparently github rewrites the committer23:02
they solve that by signing github-made commits with their own key23:02
i think only either gpg keys added in the author/committer's profile or github-made signatures will show as verified though23:03
← imp left23:03
← NorrinRadd left23:04
→ imp joined23:04
ztrawhcse alexia: yes, that's a "rebase" for you after all ;)23:05
using their own key does, of course, break the entire point of having PGP in the first place23:05
alexia i mean yeah, but clearly they don't care about web of trust and only about authenticating that "this commit was indeed made by github user X"23:06
ztrawhcse or someone that possesses access to github user X's email address for account reset purposes :p23:06
... or found a github vulnerability23:07
← vladoski left23:08
→ epony joined23:08
← thiago left23:09
→ Guest47 joined23:14
← Guest47 left23:15
← derpydoo left23:15
← luke-jr left23:16
→ mathsboy joined23:16
→ lukedashjr joined23:17
lukedashjr → luke-jr23:18
→ Grace joined23:19
← mncheck left23:21
← vicecea left23:21
→ edgars joined23:25
← BlessJah left23:26
← luke-jr left23:27
→ zeenk joined23:28
→ BlessJah joined23:30
← edgars left23:34
→ Xenguy joined23:34
← causasui left23:36
← rama left23:36
→ rama joined23:36
→ dionysus69 joined23:37
← humanface left23:38
← EvilDMP left23:38
→ EvilDMP joined23:40
→ softirq joined23:41
→ giu-- joined23:41
← dionysus69 left23:41
← thebombzen_ left23:43
← giu- left23:44
← giu-- left23:45
→ giu- joined23:45
← EvilDMP left23:49
→ luke-jr joined23:50
← mathsboy left23:53
← coot left23:53
→ daoudr joined23:54
← TheCatBehemoth left23:57
← clime left23:57
← sa0 left23:59

Logs Search ←Prev date Next date→ Channels Documentation