IRCloggy #git 2024-11-13

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.

2024-11-13

← hamburgler left00:05
→ hamburgler joined00:07
ossifrage_ → ossifrage00:13
→ Lord_of_Life joined00:14
← Everything left00:17
← KicksonButt left00:18
← memset left00:24
← P1ro left00:30
← xx left00:34
→ P1ro joined00:35
→ ferdna joined00:35
← insidesources left00:36
→ pveierland joined00:37
→ insidesources joined00:38
← jensen1 left00:43
→ memset joined00:46
→ jensen1 joined00:46
← TomyWork left00:48
← ChmEarl left00:54
← kaartic left00:56
← kus left00:59
← stats4647 left01:03
→ kaartic joined01:04
← hamburgler left01:11
→ hamburgler joined01:12
→ stats4647 joined01:14
← stats4647 left01:19
→ Gary_Sinise joined01:20
← pedahzur left01:21
→ Minvera joined01:22
← stwange6 left01:26
← lucasta left01:28
→ konsolebox joined01:32
→ gsi_ joined01:37
← gsi left01:41
← seifer left01:41
→ seifer joined01:42
← fundatillus left01:47
→ fundatillus joined01:49
→ stwange6 joined02:01
← peirik left02:04
← ferdna left02:06
→ hbautista joined02:13
→ thomasross joined02:14
← romibi left02:15
← Traneptora left02:21
← newpy left02:22
→ romibi joined02:27
→ kaartics joined02:28
← ZLima12 left02:30
← kaartic left02:30
→ ZLima12 joined02:31
→ peirik joined02:36
← konsolebox left02:43
← analpaper left02:46
← kaartics left02:46
← batrick left02:53
→ analpaper joined02:54
← analpaper left02:54
→ analpaper joined02:54
← bookworm left02:57
→ bookworm_ joined02:57
← skapata left02:59
→ donofrio joined03:01
← denysonique left03:03
→ denysonique joined03:03
→ stats4647 joined03:04
→ cedd joined03:08
← stats4647 left03:09
cedd is there a way to cherry-pick only the changes introduced by a commit compared to its parent, not the entire state of the commit?03:09
← dviola left03:17
→ pks_ joined03:21
← pks left03:22
pks_ → pks03:22
← bloody left03:23
→ konsolebox joined03:28
← donofrio left03:29
← hdh left03:31
← thomasross left03:31
← muty left03:32
→ muty joined03:33
← muty left03:33
→ muty joined03:33
← alphazone left03:36
← R2robot left03:37
→ Guest38 joined03:39
→ R2robot joined03:40
Guest38 hi03:42
russian rapists are welcomed in our opensource project. We are pro-loli-bride. glad russians have inadvertantly killed feminists with their WW3 prelude: https://quakeone.com/forum/quake-talk/quake-central/284392-dear-russians-banned-from-linux-plz-help-chaosesque-instead (there was a nuclear explosion in toropets, tver oblast on03:42
the night of sep 17-18 2024 btw) (yes that plasma ball and it's radiance proves it visually)03:42
← Guest38 left03:45
b0nn !ops03:47
→ Guest382 joined03:48
Guest382 hi03:48
fuck your code of conduct03:48
russian rapists are welcomed in our opensource project. We are pro-loli-bride. glad russians have inadvertantly killed feminists with their WW3 prelude: https://quakeone.com/forum/quake-talk/quake-central/284392-dear-russians-banned-from-linux-plz-help-chaosesque-instead (there was a nuclear explosion in toropets, tver oblast on03:48
the night of sep 17-18 2024 btw) (yes that plasma ball and it's radiance proves it visually)03:48
b0nn !ops03:49
ikke: osse cbreak_ bremner ^03:50
Guest382 b0nn:03:50
we are going to kill you for imposing a code of conduct03:50
b0nn jast:03:50
Guest382 originally opensource and freesoftware had no code of conduct03:50
b0nn cmc: canton7 avar ^03:51
Guest382 it was about show us the code, freedom of speech, plz code 4 free03:51
b0nn: I've been in opensource over 20 years03:51
b0nn thinks that tags all the ops03:51
Guest382 look up MikeeUSA03:51
we will kill you people03:51
for your censorship03:51
and throwing out of the men who did the code03:51
infavor of women, feminists, and transfaggots03:51
we WILL kill you03:51
the deal was: the men code, everyone mutually benifits, freedom of speech etc03:52
← Guest382 left03:52
JAA This is cross-channel spam and better reported to network opers. `/stats p` doesn't list any for me though. :-/03:55
cedd is it possible to cherry-pick or apply just the changes a commit introduces compared to its immediate parent?03:59
b0nn They (the aspammers) have gone now (I cannot tag them); I assume that they will be doing this for a while until they tire of it or the Network Ops find a solution03:59
← zeroes_ left04:00
← eramirez left04:00
JAA cedd: That's what a cherry-pick is supposed to do.04:01
b0nn cedd: My understanding was that a cherry-pick only applies the changes in that commit04:02
oh, snap04:02
cedd JAA: thats what i thought but for some reason i keep getting the full changes to the file its cherrypicking up to that commit04:02
→ eramirez joined04:02
cedd im cherry-picking to an orphaned branch, not sure if that affects anything04:02
→ donofrio joined04:02
ikke b0nn: they were k-lined04:03
b0nn awesome04:03
cedd this is the error im getting: CONFLICT (modify/delete): light.txt deleted in HEAD and modified in 18ffdb9 (text - wip/ add body). Version 18ffdb9 (text - wip/ add body) of light.txt left in tree.04:06
so its leaving the full file in the current working tree instead of just single change04:06
b0nn cedd: was light.txt deleted in your current branch?04:06
cedd b0nn: it doesnt exist in current branch b/c its an orphan branch04:07
JAA Then it doesn't surprise me. It's kind of hard to apply a diff to a file that doesn't exist. :-)04:07
b0nn Yeah - but the cherry-picked commit is trying to act on that file, but cannot because it no longer exists04:07
cedd i tried adding an empty light.txt before the cherry-pick but that doesnt work either04:07
JAA That sounds like it would likely lead to a conflict as well, yeah.04:08
ikke No because the context is missing04:08
JAA The file in your target branch needs to be reasonably similar (at least in the relevant hunks) to be able to apply a cherry-pick.04:08
Well, apply it without conflicts, that is.04:09
ikke cedd: what exactly do you expect the result to be?04:09
JAA Similar to what it looks like in the parent of the commit you're cherry-picking, that is.04:10
cedd i figured it would just act using the limited info it had; in this case create a light.txt file and add the change into it04:10
← gast0n left04:12
cedd instead it added the light.txt file and added many more changes introduced by ancestors of that commit04:12
ikke cedd: you cannot change something that does not exist, so the result is unedefined. What perhaps makes sense to you is usually the wrong thing to do in farout the most cases04:13
cedd ikke: idk i think acting on the limited info would be a better default than introducing even more changes outside of the commit being cherry-picked04:14
ikke What git does is indicating there is a conflict that you need to solve04:14
how to solve it is up to you.04:15
cedd yea but it should have stopped at only the changes in that commit since it's a cherry-pick imo04:15
larryv seems like inventing a new state that exists in neither head nor the cherry-picked commit would not be anything anyone except you wants.04:15
← donofrio left04:16
cedd whats the new state being invented?04:16
ikke cedd: you cannot change something if the something that needs to be changed does not exist04:16
larryv "create a light.txt file and add the change into it"04:16
ikke what does 'add a change' mean?04:16
cedd ikke: yea you can, non-existence is a state a file can be in04:16
ikke: the change introduced by the cherry-picked commit04:17
ikke cedd: please open a file called /tmp/foobar.txt on your, change line 5 that contains 'bar=1' to 'bar=2'04:18
on your pc*04:18
cedd what are you trying to demonstrate04:19
b0nn Git is looking for a specific line in the file, and it's going to change it from one known state to another04:20
← dlh404 left04:22
cedd i understand thats how it generally operates, i just think the default in this instance could be better since "a file with cherry-picked change" isn't a big challenge04:22
ikke The only alternative would be to do something like `git show <commit> path/to/file >path/to/file`04:22
cedd theres enough info to apply a sensible cherry-pick in this instance04:22
ikke You still didn't explain how you want the end result to look like04:23
b0nn honestly, git doesn't think so04:23
if you have a better way than git, send a patch to the maintainers04:23
cedd ikke: create the file since it doesnt exist, and then apply the change for that file from the cherry-picked commit04:23
ikke 'apply the change'?04:23
How04:23
cedd create the file first and then add the blob to it04:23
ikke with what contents?04:24
cedd the blob in the cherry-picked commit that changes from its parent04:24
or blob(s)04:24
ikke a blob is not a change, a change needs something to be changed in the first place04:24
cedd yes, the blobs that are different from HEAD04:25
i just think theres an opportunity for a sensible default here04:26
or even an option for this kind of behavior04:26
ikke Say I have this patch: https://tpaste.us/VY4804:28
→ stats4647 joined04:28
ikke (ie, what cherry-picking a commit would look like)04:28
No, foo.txt does not exit04:28
I try to apply that patch04:28
What would the file end up looking like?04:29
cedd i do see what youre saying04:30
ok let me tell you my use case so you understand why im kind of disappointed04:30
i made a test project to better learn the git commands04:31
and in it i made an ascii-art branch that has 4 branches of different types of ascii-art: fantasy, dragons, skeletons, monsters04:31
i added these ascii-arts piece-meal so that there were quite a few commits with a message containing 'add head'04:32
← stats4647 left04:32
cedd so what i wanted to do what select all the commits with 'add head' and do a mass cherry-pick merge of all these commits to have a commit of all the heads of the different ascii-arts04:32
and it even works with 'git show', i see a diff output of all the heads of the ascii-art and it's really cool04:33
i just can make a commit with this because the functionality doesnt exist04:34
ikke remember that a commit is a full snapshot04:34
cedd i just cant*04:35
b0nn can you reconstruct light.txt from all the commits in that dead branch?04:35
cedd by 'dead branch' you mean orphaned branch?04:36
ikke cedd: what `git show` command did you use?04:36
→ dviola joined04:37
cedd git show $( git rev-list --grep='add head$' fantasy dragons skeletons monsters )04:37
ikke Ok, so you are showing each separate commit04:38
cedd yea04:38
ikke That can only work if you have a commit with the ascii arts missing all the heads first04:38
cedd yea i know it cant work, im just disappointed there isnt some kind of option to allow a less rigorous merge without defaulting to just vomiting the entire state of the file at that cherry-pick04:40
ikke No one before thought that to be usefull and contributed it to git04:42
cedd and thats a fair point04:42
but it doesnt help my feelings in this moment04:43
← analpaper left04:43
ikke sadly I cannot help you with that04:43
cedd console me04:44
→ hwpplayer1 joined04:45
ikke https://images.fineartamerica.com/images/artworkimages/mediumlarge/1/little-cute-kitten-serhii-kucher.jpg04:45
cedd :O04:46
→ analpaper joined04:47
cedd ok i feel slightly better now04:47
b0nn LOL04:47
ikke https://wallpapercave.com/wp/UFmaMS9.jpg04:49
← cpackham left04:50
cedd alright im off to bed fellas, thanks for the support <304:53
← Betal left04:53
→ bloody joined04:58
→ duxsco joined05:00
→ stats4647 joined05:03
← Gary_Sinise left05:05
→ gnoo_ joined05:16
← hbautista left05:17
← gnoo left05:18
← stats4647 left05:19
→ stats4647 joined05:24
→ duxsco_ joined05:24
→ Traneptora joined05:25
← duxsco left05:27
← cedd left05:28
← kurahaupo left05:31
← stats4647 left05:32
← stagas left05:32
→ stagas_ joined05:32
← smallville7123 left05:38
→ smallville7123 joined05:38
← bloody left05:44
→ bloody joined05:47
→ Gary_Sinise joined05:49
← Gary_Sinise left05:51
bookworm_ → bookworm05:57
← analpaper left05:58
← hwpplayer1 left05:58
→ sove joined06:00
→ frobnic_ joined06:00
← frobnic left06:00
sove for delta compress up to 8 threads but it maxes at 100%. is there any way to force 8 threads06:01
Delta compression using up to 8 threads06:02
→ analpaper joined06:03
← analpaper left06:03
→ analpaper joined06:03
→ YuGiOhJCJ joined06:05
← insidesources left06:07
stagas_ → stagas06:09
→ kale_ joined06:11
← dylanj left06:14
→ dylanj joined06:14
→ kaartic joined06:18
← foul_owl left06:24
→ alphazone joined06:26
← thiago left06:27
→ jmd joined06:28
→ insidesources joined06:33
→ larryv_ joined06:37
→ foul_owl joined06:38
← larryv left06:40
→ duderonomy joined06:41
→ stats4647 joined06:42
larryv_ → larryv06:42
← stats4647 left06:47
sove hey06:47
no good answers found06:47
→ rosco joined06:59
← kaartic left07:00
→ kaartic joined07:00
→ rfuentess joined07:05
→ crossby1004 joined07:16
← sove left07:20
→ sove joined07:22
← sove left07:23
→ sove joined07:23
← sove left07:29
← analpaper left07:30
→ Dex_2x joined07:35
← larryv left07:39
← jensen1 left07:46
← Xlusive left07:49
→ Xlusive_ joined07:49
Xlusive_ → Xlusive07:50
← konsolebox left07:51
→ rsx joined07:51
→ gandhibobandhi joined08:04
← travisghansen left08:07
← pveierland left08:07
→ travisghansen joined08:07
→ stats4647 joined08:10
← jmd left08:13
← stats4647 left08:14
→ Vooloo joined08:15
→ xx joined08:15
← Vooloo left08:15
→ pveierland joined08:17
← Mooncairn left08:19
← nullroute_ left08:21
← Ingvix left08:22
→ Ischwitch joined08:22
Ischwitch → Ingvix08:24
→ falafel joined08:25
→ ds_ joined08:25
→ nullroute joined08:34
← nullroute left08:34
→ nullroute joined08:34
→ robobub joined08:34
→ chexum joined08:35
→ toric joined08:38
← memset left08:38
→ memset joined08:38
→ hwpplayer1 joined08:39
← pveierland left08:40
→ CrunchyFlakes joined08:46
← hwpplayer1 left08:47
← toric left08:52
→ toric joined08:55
← CrunchyFlakes left08:56
← toric left08:58
→ toric joined09:00
→ CrunchyFlakes joined09:01
← toric left09:06
→ npc joined09:24
→ iomari891 joined09:29
← GNUmoon left09:35
→ GNUmoon joined09:35
← nullroute left09:35
→ savolla joined09:37
← tejr left09:38
jfsimon1981 → jfsimon09:43
→ nullroute joined09:44
← nullroute left09:44
→ nullroute joined09:44
→ tejr joined09:45
← adamtajti left09:46
→ adamtajti joined09:46
→ kaartics joined09:48
← kaartics left09:50
→ kaartics joined09:50
← kaartic left09:50
← thuna` left09:52
← Dex_2x left09:54
→ Dex_2x joined09:55
→ kaartic joined09:56
→ darkbit joined09:56
← kaartics left09:59
→ rsrx joined10:01
rsrx hi, i'm trying to commit .git/hooks/pre-push file to the repo but `git gui` doesnt see the file even if it has changes10:02
any ideas why?10:02
ikke Yes, you cannot10:02
For security reasons10:03
← falafel left10:03
← aminvakil left10:03
rsrx hmm10:04
ok10:04
→ aminvakil joined10:04
JAA You can move your hooks to a different directory and set core.hooksPath. That will need to be set explicitly by everyone who clones the repo, too.10:05
rsrx thanks10:10
→ thuna` joined10:10
→ kurahaupo_ joined10:10
kurahaupo_ → kurahaupo10:18
← housemate left10:18
→ stats4647 joined10:22
← stats4647 left10:27
→ kaartics joined10:29
← kaartic left10:30
← mort left10:34
← duderonomy left10:34
← nullroute left10:45
→ nteodosio joined10:47
← nteodosio left10:49
← crossby1004 left10:52
← ds_ left10:54
← kaartics left10:57
→ ds_ joined10:58
[twisti] i havea git log format from some tool that i like, but would like to add date to. how can i see what the current format actually is ?10:59
→ Betal joined11:02
→ kaartic joined11:02
← kaartic left11:02
→ kaartic joined11:03
← Milos left11:03
← hexnewbie left11:04
← memset left11:04
→ memset joined11:04
← Novi left11:05
→ Novi joined11:06
→ nullroute joined11:07
← nullroute left11:07
→ nullroute joined11:07
→ ainema joined11:11
← duxsco_ left11:12
→ housemate joined11:12
→ hexnewbie joined11:14
selckin %ad is the default date format i think11:15
→ infogulch113 joined11:21
← infogulch11 left11:21
infogulch113 → infogulch1111:21
← Teacup left11:21
← fundatillus left11:21
→ Teacup joined11:23
← thuna` left11:24
→ crossby1004 joined11:35
← tomboy64 left11:42
→ Everything joined11:44
→ kaartics joined11:50
→ Milos joined11:51
← kaartic left11:54
→ stats4647 joined11:56
→ tomboy64 joined11:56
→ crossby1104 joined11:58
← crossby1104 left12:00
← stats4647 left12:00
← crossby1004 left12:01
← Milos left12:02
→ toric joined12:03
→ crossby1004 joined12:03
→ stats4647 joined12:04
← toric left12:04
→ Milos joined12:05
← stats4647 left12:08
→ toric joined12:09
← toric left12:09
← Dex_2x left12:15
→ toric joined12:15
← toric left12:16
→ warmana joined12:16
→ rcares joined12:18
→ davispuh joined12:19
← davispuh left12:19
→ toric joined12:21
→ snoriman joined12:22
← toric left12:22
→ toric joined12:23
← toric left12:23
→ davispuh joined12:24
snoriman Hi! I'm trying to apply a patch that I created using `git diff --color=never > fix.patch`. I try to apply it using `git apply --reject <etc>`. I'm using --reject to make sure things go well when the patch has been applied already. However even with --reject I get an error like "error: while search for <part of patch>". Shouldn't --reject ignore this and just continue?12:24
← YuGiOhJCJ left12:27
→ Dex_2x joined12:32
← supernovah left12:37
→ duderonomy joined12:37
→ belsirk joined12:38
← rfuentess left12:41
→ supernovah joined12:44
→ rfuentess joined12:44
← belsirk left12:46
← kaartics left12:47
→ kaartic joined12:48
← supernovah left12:49
← flow left12:54
← Dex_2x left12:55
→ supernovah joined12:55
[twisti] selckin: i meant the default for pretty.format - i can make a new format that includes the date, but how do i get the rest to look the same as before ?12:57
← ds_ left12:57
→ ds_ joined12:58
→ avih joined13:00
→ Flow joined13:01
selckin don't understand the question then, what do you have from some tool?13:01
avih hi. is there some docs or details on what $GIT_DIR/HEAD might be or contain? specifically, beyond a normal file which contains "refs/..." or a hash (or that too, if that's exactly what it is). i realize this file is not a public git API, but still. thanks in advance.13:02
← rsrx left13:02
[twisti] selckin: my `git log` already looks nonstandard, but even if we assume it didnt, the question stays the same13:04
i want my pretty.format to look like 'current pretty.format plus date'13:05
→ konsolebox joined13:05
[twisti] so `git log --pretty=format: '<whatgoeshere?> %ad'`13:05
→ Dex_2x joined13:06
← supernovah left13:07
avih i'm trying to look at the git source code Documentation/ dir (about .git/HEAD), but so far i'm not seeing any obvious docs file name...13:08
→ dlh404 joined13:09
← memset left13:12
→ belsirk joined13:12
→ memset joined13:13
→ skapata joined13:14
← rfuentess left13:15
← ectospasm left13:17
→ ectospasm joined13:20
→ supernovah joined13:20
← Dex_2x left13:22
← ectospasm left13:24
osse avih: man gitrepository-layout13:25
gitinfo avih: the gitrepository-layout manpage is available at https://gitirc.eu/gitrepository-layout.html13:25
avih osse: much appreciated :)13:25
osse: do you happen to know whether "normal files which contains a hash or refs/*" is pretty much it?13:26
normal file*13:26
← rcares left13:27
avih hmm... "In some legacy setups, it is a symbolic link instead of a symref that points at the current branch."13:27
so pretty much not always a normal file...13:28
→ ectospasm joined13:30
→ batrick joined13:32
osse avih: I've never encountered HEAD *not* being a plain file13:32
Pretty sure the change from a symlink to a file happened so long ago it's not a practical concern13:32
avih that's good info. thanks.13:32
osse Yeah that change happened the same year. 2005.13:34
avih osse: fwiw, the context is contrib/completion/git-prompt.sh (to which i've contributed, but i was interested further how it does things beyond places I've touched), and this file reads HEAD and other files and tries to deduce the status of the repo13:34
failure is an option obviously, it's an informational utility, but better be right if possible :)13:35
osse: "same year" as what?13:36
← stagas left13:36
→ stagas_ joined13:36
avih "points at the current branch" so in this context what is "the current branch"? an object file? a symref file? some directory?13:38
osse avih: same year as git itself came to be13:39
avih oh :D13:39
yeah, pretty early then. i guess it shouldn't be a practical concern then.13:39
osse avih: where's that quote from?13:39
oh13:39
avih especially for non critical tools which only try to scrap a bit of info.13:39
osse nvm13:39
→ fonini joined13:40
osse avih: in the first paragraph: "describing the currently active branch."13:40
ie. what HEAD points to isi the current branch by definition13:40
avih oh, you mean it's a symlink to refs/... at the gitdir?13:42
osse no13:42
avih (if i read it right, it's either a symref, or a symlink to the "current branch")13:42
osse I answered "in this context what is "the current branch"?"13:42
→ analpaper joined13:42
← analpaper left13:43
avih i was referring to this context: "In some legacy setups, it is a symbolic link instead of a symref that points at the current branch."13:43
osse Yes13:43
avih what's "current branch" in this sentence?13:43
a dir? a file?13:44
osse it's bad wording is what it is13:44
avih :)13:44
osse because the current branch is determined by HEAD itself. That wording makes it sound like something else determines what the "current branch" is. But HEAD itself does13:44
Whatever HEAD points to (whether it's a symlink or a symref or whatever; doesn't matter) is by definition the current branch13:45
→ badoola joined13:45
mjt0k (which might not be named branch, too)13:46
→ celmor joined13:46
avih yes, i get that, but if i follow the symlink, what does it point to? the the itself which ta symref would hold, e.g. to refs/heads/master ? or to an object file with the hash if detached?13:46
osse avih: if it is a symlink then it points to a file in .git/refs/heads13:47
avih gotcha. that's what i hoped. thanks :)13:48
so basically, instead of old symlink to refs/whatever, it's now a normal file which contains that "symre/whatever" text, yes?13:49
symref of "refs/whatever"13:49
→ Dex_2x joined13:50
osse yes13:50
cat it and see13:50
avih cool. thanks. that was helpful.13:50
leon_ → leon13:57
→ xcm_ joined14:01
→ denysonique2 joined14:02
→ pgloor7 joined14:02
← Wolf480pl left14:03
→ __ad joined14:03
← cebor left14:03
← dostoyevsky2 left14:03
← preisi_svr left14:03
← xcm left14:03
← pgloor left14:03
pgloor7 → pgloor14:03
← frojnd left14:03
← qsx left14:03
← blmt left14:04
← avu left14:04
← simon816 left14:04
← denysonique left14:04
denysonique2 → denysonique14:04
← wCPO left14:04
← ad__ left14:04
← celmor left14:04
→ blmt joined14:04
→ cebor joined14:05
→ qsx joined14:05
→ dostoyevsky2 joined14:05
→ wCPO joined14:05
→ Wolf480pl joined14:05
→ simon816 joined14:05
→ stats4647 joined14:05
→ frojnd joined14:05
→ preisi_svr joined14:06
← stats4647 left14:09
→ deadrom joined14:09
→ kaartics joined14:10
← kaartics left14:10
→ kaartics joined14:10
← kale_ left14:11
← kaartic left14:12
→ kaartic joined14:15
← memset left14:15
→ memset joined14:15
→ avu joined14:17
Traneptora snoriman: what's in the *.rej files?14:18
← kaartics left14:19
Traneptora was any part of the patch applied?14:19
snoriman Traneptora: the first time I called it yes, the patch was applied14:19
Traneptora if you try to apply it again then most likely the entire patch won't apply14:19
← Dex_2x left14:19
Traneptora so there isn't anything to abort and continue to, the entire patch is invalid14:19
snoriman yeah exactly14:20
Traneptora that's what you want, right?14:20
snoriman Traneptora: I fixed it by using: https://gist.github.com/samsnori/9cc0cfa14ccd82658b9d324ec6bd687714:20
← konsolebox left14:20
Traneptora I would recommend using a patch that has a commit in it, with git format-patch14:21
and applying it with git am14:21
snoriman Traneptora: well, it's part of a build system which can run multiple times; what I want is that it ignores the patch when it can't apply it and I don't want an error14:21
Traneptora: hmm when I apply such a format-patch multiple times, would that work or do I still get an error once it has been applied already?14:21
Traneptora snoriman: not by default but you can use `git am --3way foo.patch`14:24
which will no-op if the patch has already been applied14:24
snoriman a and that works only with "format-patch" ? And does that work when I clone with --depth 1 ?14:25
Traneptora snoriman: it should work with --depth 1 provided that your patch is either not applied or the most recent applied patch14:27
git am only works normally if there's commit info like author and stuff14:28
snoriman ok thanks14:28
Traneptora patches created with git diff won't have that info14:28
← iomari891 left14:37
→ alexherbo2 joined14:41
← dsimic left14:44
→ dsimic joined14:46
belsirk → rfuentess14:56
→ hbautista joined14:59
← Murr left15:01
→ Left_Turn joined15:01
← alexherbo2 left15:02
→ alexherbo2 joined15:03
→ Murr joined15:04
→ donofrio joined15:06
→ iomari891 joined15:07
→ Guest35 joined15:09
→ gh34 joined15:11
← Guest35 left15:13
→ stats4647 joined15:17
→ TomyWork joined15:17
← stats4647 left15:21
→ donofrio_ joined15:30
← memset left15:32
→ memset joined15:33
→ ChmEarl joined15:36
← Everything left15:38
→ konsolebox joined15:42
← Murr left15:42
← fonini left15:43
→ Murr joined15:44
← rfuentess left15:46
→ fonini joined15:58
← rsx left15:58
→ thiago joined15:59
→ cedd joined15:59
infogulch11 → infogulch15:59
→ commiejoke29 joined16:00
← ds_ left16:01
← ggGuest11 left16:06
→ stats4647 joined16:07
← stats4647 left16:11
← cedd left16:14
→ donofrio__ joined16:14
→ cedd joined16:16
← donofrio_ left16:18
→ jpw_alt joined16:19
→ ggGuest11 joined16:19
← jpw left16:19
jpw_alt → jpw16:19
← jpw left16:20
→ jpw joined16:20
← Pixi` left16:23
← donofrio__ left16:24
→ jensen1 joined16:26
← rapmoc left16:27
→ rapmoc joined16:27
← techtruth left16:28
← konsolebox left16:36
→ stats4647 joined16:37
→ hwpplayer1 joined16:40
→ paculino joined16:45
← paculino left16:51
→ paculino joined16:51
→ lucasta joined16:58
← kaartic left17:05
→ kaartic joined17:05
→ Orion joined17:17
→ Orion9 joined17:19
← Orion left17:19
→ iroblox joined17:21
← adamtajti left17:23
← iroblox left17:26
← Orion9 left17:26
← warmana left17:30
→ mervn joined17:33
mervn hello all, is there a way to set the commit-timezone to UTC without using the command-line or GIT_COMMITTER_DATE17:35
something in git-config that sets timezone (or TZ) is preferred, but a GIT_TZ (or equivalent) works as well17:37
← kaartic left17:38
→ epicout joined17:41
← hwpplayer1 left17:44
leah2 mervn: perhaps https://dallas.lu/en/how-to-fake-the-timezone-of-a-git-commit/17:44
the hooks approach17:45
← donofrio left17:45
→ pedahzur joined17:48
gsish is env(1) too simple?17:48
→ john_johnk joined17:51
← TomyWork left17:51
→ donofrio joined17:51
→ ds_ joined17:53
mervn leah2: thanks for the resource. hooks wont work, the goal is to avoid touching GIT_AUTHOR_DATE and GIT_COMMITTER_DATE because they may be set somewhere else - and i dont want the format overridden for UTC17:54
← savolla left17:56
→ larryv joined17:57
→ savolla joined17:57
mervn gsish: its not consistent, if i switch shell context, TZ is removed. it also requires an alias, which also is inconsistent17:57
i am human and make errors - so relying on my 100% consistency is not viable. git should "automatically seek" the correct setting18:00
→ Pixi joined18:01
mervn that's why aliases and env(1) are not viable. something "guaranteed", like GIT_TZ in .profile or core.timezone in git-config would go a long way18:03
leah2 patch it to call tzset :)18:03
← savolla left18:05
→ savolla joined18:05
→ Ancient joined18:06
← ainema left18:08
mervn leah2: yeah i was thinking about a patch for "GIT_TZ > core.timezone > TZ". i think it is valid use-case, as GIT_COMMITTER_DATE is for formatting18:08
leah2 writing a wrapper script is easier tho and about as much work as setting an env var18:09
mervn is there any interest for this upstream?18:09
→ YuGiOhJCJ joined18:09
jensen1 → aork18:09
Ancient Is it possible to use a pkcs11 provider, such as libtpm2_pkcs11 to sign git commits with a hardware TPM key?18:10
→ hdh joined18:13
→ wootehfoot joined18:13
→ hdh_ joined18:18
mervn Ancient: to my knowledge, there is no direct PKCS11 signing support18:19
← alexherbo2 left18:20
mervn the best workaround i know of is creating an SSH-key with PKCS11 and using the SSH-key to sign the commit18:20
← hdh left18:21
→ alexherbo2 joined18:21
Ancient My SSH key is currently PKCS11, but I wasn't sure how to bridge that to git. It looks like the SSH key signing seems to want a path to a key file (from what I see).18:21
← lucasta left18:21
Ancient I'll test it, and if that doesn't work I may try gnupg-pkcs11-scd which looks like it might bake that into gnupg signatures.18:21
Most of the docs seem to have user.signingkey need to point to a file, which I think would eliminate the TPM key option (no file to be read)18:22
mervn i have no hardware to test this with directly so this is speculative18:24
← alexherbo2 left18:25
Ancient If the ssh signature approach can use the ssh-agent, then it should work. I'll do a bit of testing.18:25
mervn but you can use the public-key of your SSH as long as ssh-agent (or i presume gpg-agent with ssh support) is active18:25
point git to a public-key (which should be able to be generated from a TPM) and then let ssh-agent handle the rest18:27
← jfsimon left18:27
Ancient Aha. That's where I misunderstood. I thought user.signingkey needed to point to the private key.18:27
Rather than the public key.18:27
Normally you'd expect signatures to use the private key, to validate it by public key later.18:28
mervn it can be either, but public-key only works when ssh-agent (or gpg-agent with ssh support) is active18:28
← rainn92675560720 left18:30
mervn this should work, but it might not - in which case, you can send a feature-request upstream18:30
Ancient I'll report back in a few minutes :)18:30
← ds_ left18:32
→ ds_ joined18:32
→ rainn92675560720 joined18:32
mervn i think PKCS11 is a valid use-case. but at the rate im going, we'll have email in git before the end-of-year lol18:32
gsish mervn: is there a reason for wanting UTC times so bad? do you maybe want to set the affected user's TZ in general? that'd be consistent across that user's activity, including git invocations from other contexts18:33
if it's not for all of the user, maybe you want to make it a habit to start development sessions with env TZ= and be done for all child prozesses equally?18:34
don't know if you said it above, might have missed your motivation, but seeing how existing approaches might cover what you want to achieve ...18:35
did not suggest that you have to remember specifying env TZ= for each individual git invocation yourself, of course that'd be error prone18:35
and if you want to just adjust individual commits, there's git commit --date (also available from interactive rebase in case you want to apply it with hindsight)18:36
→ techtruth joined18:38
mervn gsish: for my particular use-case, its privacy. my name is uncommon and knowing my timezone is effectively a dox. i can imagine there are many git users that are (un)knowingly affected by this18:39
Ancient mervn, I've confirmed that works as you said it would. Thank you. :)18:39
I'll do a brief write up of how to do it for the Gentoo Wiki, should make it easier to find when future people are wondering the same thing.18:40
gsish might as well put a git(1) in your PATH before the actual git installation, that does the TZ manipulation before running the actual implementation of git18:41
needs to be done once per machine (and user?), but then is fully transparent and needs no manual intervention and cannot be forgotten18:41
mervn TZ is set in my .profile, and is convenient for many applications to have the right TZ - but git is a flight-risk. the env(1) solution is by no means bad, it just has a non-zero chance of failure which defeats the purpose18:42
gsish other manipulations that were outlined above probably also need to be re-done per machine18:42
mervn Ancient: glad it worked, i also use Gentoo (btw)18:42
Ancient Great taste :P18:43
avih osse: in my quest to detect the current branch in initial bare repo (which i couldn't find how to do with rev-parse, and which led me to check if i can try to read HEAD myself), i found that git branch --current will show "master" in an initial bare repo. thanks again for the info. at least now i know a bit more ;)18:45
mervn gsish: i have an admittedly non-standard setup, where my user-account is trusted the least - so having a git replacement in PATH wont work for me in particular18:45
avih osse: and while HEAD is indeed a regular file with current git, if the repo uses reftables (since 2.45), then it's intentionally an incorrect symref, which the docs didn't mention that i could find.18:47
(i mean, it could use reftables, but it's not the default)18:47
hdh_ → hdh18:48
← memset left18:48
mervn and as a temp-solution, i am using env(1) as you suggested. but since git pulls-in TZ, it really should have a way for you to adjust it within git (to avoid profuse workarounds)18:48
→ memset joined18:48
← stats4647 left18:49
→ stats4647 joined18:50
← eramirez left18:51
← stats4647 left18:56
← wootehfoot left19:00
→ stats4647 joined19:05
← john_johnk left19:12
← mervn left19:13
gsish git _should_ provie a means to manipulate TZ from within it? I feel the opposite, wasn't the point of generic and reusable approaches like env vars that apps need not re-invent their manipulation _because_ users got other means to adjust them while apps just use them?19:16
in other words, I'd expect upstream's desire to accept responsibliity for TZ manipulation rather low (am guessing, granted)19:17
→ Juesto joined19:19
← rosco left19:20
← Juest left19:22
Juesto → Juest19:23
← gandhibobandhi left19:26
← Lunatrius left19:30
→ Turn_Left joined19:33
← savolla left19:34
→ realies1 joined19:35
→ Lunatrius joined19:37
→ cpackham joined19:37
← Left_Turn left19:37
← commiejoke29 left19:39
→ alexherbo2 joined19:40
← smallville7123 left19:46
→ Left_Turn joined19:53
← cedd left19:56
← Turn_Left left19:56
→ cedd joined19:58
← realies1 left19:58
→ realies1 joined19:59
← ds_ left20:00
→ Guest84 joined20:07
Guest84 Hello20:07
← alexherbo2 left20:07
Guest84 do people still use gitflow?20:07
→ FreeBDSM joined20:11
← Ohtani left20:12
FreeBDSM hi. is there a way to remove fat binary files committed to git without changing commits checksums?20:12
ikke no20:12
← Guest84 left20:16
FreeBDSM ikke, pretty please??20:16
ikke ok, because you ask so kindly20:17
→ analpaper joined20:17
← analpaper left20:17
→ analpaper joined20:17
FreeBDSM are commits checksums really hashes from the commits themselves?20:18
I mean, if there's a strict algo - there should be a way to get a checksum of a commit before it gets committed20:19
ikke yes20:19
In fact, you would do almost the same thing git is doing while committing20:19
The commit object includes a hash for a tree object. The tree object can refer to blobs and other tree objects20:20
changing one bit will affect all hashes20:20
FreeBDSM how come no one wrote a script that'd replace fat blobs with tiny commits with just some short text strings whose hash would match the hash of the edited commits?20:20
ikke Because that's nearly impossible20:21
FreeBDSM that'd work like a brute force until hash matches, sure, but its got to be possible20:21
ikke famous last words20:22
FreeBDSM how long would such a brute force approach take time? an hour? more?20:22
ikke The end of the universe20:22
← beardedbear left20:26
← ak77 left20:28
→ masber joined20:28
→ commiejoke29 joined20:28
→ ak77 joined20:28
b0nn You're in luck, the rebirth of the next universe will be when you have the answers, please report back then, am genuinely curious about the result :)20:29
← brw left20:30
→ brw joined20:31
→ CrunchyFlakes_ joined20:32
← CrunchyFlakes left20:32
→ masber64 joined20:37
← masber left20:39
← donofrio left20:45
→ Turn_Left joined20:49
→ eramirez joined20:52
← Left_Turn left20:53
← memset left21:02
→ Everything joined21:02
→ memset joined21:02
→ donofrio joined21:06
← paculino left21:17
→ smallville7123 joined21:23
← cedd left21:23
→ cedd joined21:27
← donofrio left21:28
→ mh_le joined21:35
mh_le is it possible to push to two remotes at the same time?21:35
← masber64 left21:36
← realies1 left21:36
Everything mh_le: git push ... && git push ... :-)21:37
→ realies1 joined21:42
← npc left21:42
→ Left_Turn joined21:44
← thiago left21:46
← Turn_Left left21:47
← cedd left21:50
→ cabazon joined22:00
JAA You can have multiple push URLs on a remote, and then pushing to that will push to all of them.22:00
b0nn git remote set-url --add --push all git://another/repo.git22:03
← gh34 left22:04
→ ferdna joined22:06
→ ds_ joined22:06
realies1 → realies22:07
← realies left22:11
← cabazon left22:11
← crossby1004 left22:16
→ cedd joined22:19
→ cabazon joined22:21
← cabazon left22:21
→ eyeris joined22:32
eyeris Is there a standard term for rewriting a branch to retroactively eliminate merge commits by rebasing?22:33
← ds_ left22:37
→ thiago joined22:38
→ alexherbo2 joined22:40
→ Turn_Left joined22:55
← Left_Turn left22:56
← puff left22:59
→ lucasta joined22:59
← eyeris left23:01
→ falafel joined23:01
← Everything left23:02
← stats4647 left23:05
→ stats4647 joined23:06
← davispuh left23:08
← stats4647 left23:10
← iomari891 left23:25
← fonini left23:34
← commiejoke29 left23:35
→ stats4647 joined23:35
→ ferdna_ joined23:36
← ferdna left23:39
← stats4647 left23:40
← cebor left23:42
→ cebor joined23:45
← alexherbo2 left23:48
← Turn_Left left23:52

Logs Search ←Prev date Next date→ Channels Documentation