IRCloggy #git 2020-08-22

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.

2020-08-22

→ shush joined00:00
→ navidr joined00:00
→ downunder joined00:01
← tiin57 left00:04
← shush left00:05
→ tiin57 joined00:08
← Hello71 left00:10
← mindCrime left00:12
→ mindCrime joined00:12
→ Hello71 joined00:13
← kupi left00:14
→ mns joined00:16
→ orbyt_ joined00:17
→ AlexMax joined00:31
→ Case_Of joined00:31
AlexMax Is there a way for me to directly pull a specific branch from an un-named remote directly into a brank-spanking-new branch?00:32
I basically want to pull a branch from somebody elses' repository into a new local branch without creating a remote and without having to merge it into an existing branch00:33
← tiin57 left00:33
rafasc AlexMax: git fetch url theirbranch:yourbranch00:36
AlexMax You're a peach, thank you.00:36
→ tiin57 joined00:37
→ shush joined00:38
← mns left00:38
→ arcatech joined00:42
← shush left00:43
← blurkis left00:45
→ blurkis joined00:46
→ sudoforge joined00:46
→ shush joined00:55
← shush left01:03
← tiin57 left01:03
← cdown left01:07
→ tiin57 joined01:07
← vicfred left01:07
← rafasc left01:07
← meropyri left01:08
→ meropyri joined01:08
→ cdown joined01:08
→ TheSecondNik joined01:12
← lukey left01:14
→ lukey joined01:16
← TheSecondNik left01:18
← jaziz left01:25
→ jaziz joined01:27
← darkbit left01:28
← m0viefreak left01:31
← tiin57 left01:34
→ disi joined01:36
← cdown left01:36
disi is there a way to clone into a detached HEAD state (at the tip of the repo's HEAD)?01:36
→ tiin57 joined01:37
→ leeN joined01:39
← F0rTex left01:40
→ F0rTex joined01:42
← dfee left01:42
← raven-au left01:53
→ raven-au joined01:54
→ dfee joined01:54
← dfee left01:55
← Brainium left01:56
← raven-au left01:59
← tiin57 left02:04
← gpanders left02:06
→ gpanders joined02:07
→ tiin57 joined02:07
→ blowry joined02:07
← walidvb left02:12
← bambanx left02:21
→ bambanx joined02:22
← bambanx left02:23
→ vicfred joined02:24
→ bambanx joined02:24
→ raven-au joined02:27
→ mustache_ride joined02:27
← leeN left02:28
→ uenotan joined02:29
← Cabanossi left02:30
→ Cabanossi joined02:32
← tiin57 left02:33
→ itok joined02:35
→ mithro joined02:36
→ snowkidimd joined02:36
→ bitonic joined02:36
→ tiin57 joined02:37
← johnny56 left02:38
← itok left02:38
← uenotan left02:39
→ itok joined02:39
← bambanx left02:40
→ noahmg123 joined02:41
→ envex joined02:44
→ Newami joined02:48
← Newami left02:48
← gpanders left02:48
→ gpanders_ joined02:48
← Maxattax left02:51
→ johnny56 joined02:53
← lagothrix left02:54
→ lagothrix joined02:54
← dunk left03:03
← tiin57 left03:03
← zalun left03:04
→ zalun joined03:05
→ kwilczynski joined03:06
→ dunk joined03:07
← prit left03:07
← SrPx left03:07
→ tiin57 joined03:07
← navidr left03:07
← dunk left03:08
→ SrPx joined03:08
→ Goplat joined03:09
→ prit joined03:09
→ dunk joined03:09
→ navidr joined03:10
→ omnidapps joined03:11
→ disruptivenl joined03:11
→ heyitsrama joined03:12
← johnny56 left03:14
← rumbler31_ left03:17
← FFY00 left03:18
← sindns left03:18
→ FFY00 joined03:18
→ bolovanos joined03:18
→ Muimi joined03:19
→ bambanx joined03:20
→ bjs joined03:21
→ libertyprime joined03:21
← envex left03:23
← FFY00 left03:25
→ FFY00 joined03:25
→ johnny56 joined03:28
→ envex joined03:30
← tiin57 left03:33
← johnny56 left03:34
→ tiin57 joined03:37
→ z8z joined03:41
→ AnAverageHuman joined03:44
← monr0e left03:46
→ johnny56 joined03:47
← g00s left03:49
← AnAverageHuman left03:52
← lucasb left03:54
← johnny56 left03:55
← bolovanos left03:58
→ johnny56 joined03:58
→ bket joined04:00
← orbyt_ left04:00
← navidr left04:02
← tiin57 left04:03
← johnny56 left04:04
← thomasross left04:05
→ tiin57 joined04:07
← g4570n left04:11
← FFY00 left04:16
→ FFY00 joined04:16
→ johnny56 joined04:18
← zalun left04:22
→ berken_ joined04:23
→ zalun joined04:23
← Hello71 left04:23
berken_ → berken04:23
← johnny56 left04:24
→ Hello71 joined04:31
← tiin57 left04:33
→ tiin57 joined04:37
→ johnny56 joined04:38
← johnny56 left04:44
← FFY00 left04:46
→ FFY00 joined04:47
← Hello71 left04:52
← FFY00 left04:57
← Muimi left04:58
→ FFY00 joined04:58
→ shush joined05:01
← tiin57 left05:03
← astronavt left05:04
→ astronavt joined05:04
← shush left05:06
→ orbyt_ joined05:06
← snowkidimd left05:07
→ tiin57 joined05:07
→ johnny56 joined05:08
← arcatech left05:14
← johnny56 left05:20
→ Guest20 joined05:26
→ johnny56 joined05:28
→ ferdna joined05:32
← causasui left05:33
← tiin57 left05:33
← johnny56 left05:34
→ puppy joined05:36
→ rustyshackleford joined05:36
→ tiin57 joined05:38
← libertyprime left05:44
→ libertyprime joined05:46
→ johnny56 joined05:48
→ Hello71 joined05:52
→ fatalhalt joined05:53
← dedowsdi left05:56
→ dedowsdi joined06:00
← tiin57 left06:03
← Guest20 left06:03
← johnny56 left06:04
← Hello71 left06:05
→ tiin57 joined06:07
→ Phylock joined06:08
← F0rTex left06:10
→ johnny56 joined06:17
← thiago_ left06:22
← johnny56 left06:24
← geosmile left06:26
← mindCrime left06:30
→ calcul0n joined06:32
← Phylock left06:32
← libertyprime left06:33
← tiin57 left06:34
→ libertyprime joined06:34
→ tiin57 joined06:37
→ johnny56 joined06:37
→ fstd_ joined06:37
← fstd left06:41
fstd_ → fstd06:41
← jaziz left06:44
→ Phylock joined06:46
← libertyprime left06:50
← orbyt_ left06:52
← heyitsrama left06:53
← johnny56 left06:54
→ mindCrime joined06:54
→ mirrorbird joined06:54
→ feriman joined06:59
→ g00s joined06:59
← unluckyshrubbery left07:01
→ libertyprime joined07:02
→ shush joined07:02
← tiin57 left07:04
→ F0rTex joined07:06
→ tiin57 joined07:07
→ johnny56 joined07:09
← shush left07:11
→ Kulrak joined07:12
← johnny56 left07:14
← Goplat left07:14
← kjartan left07:17
→ kjartan joined07:20
→ johnny56 joined07:24
← fatalhalt left07:29
← johnny56 left07:30
← tiin57 left07:33
← deAtog left07:36
→ deAtog joined07:37
← pmercado left07:37
→ tiin57 joined07:37
← dedowsdi left07:37
→ dedowsdi joined07:39
j416 disi: can you clarify?07:42
→ johnny56 joined07:43
← g00s left07:49
← johnny56 left07:50
← thansen left07:50
→ thansen joined07:50
← akem left07:53
→ akem joined07:55
← ferdna left07:56
← envex left07:59
← Phylock left08:01
← mindCrime left08:02
→ johnny56 joined08:03
← tiin57 left08:03
→ tiin57 joined08:07
← johnny56 left08:09
← Betal left08:09
← cd left08:13
Jookia is there a way to fetch remote branches and have them auto update without checking out and git pull08:14
→ johnny56 joined08:23
→ qswz joined08:26
qswz hmm, I've entered `git am` mistakenly, then after git rebase --abort I lost my currect changes08:27
j416 Jookia: the common way to work is to update your local branches as you work on them08:28
← johnny56 left08:29
bookworm qswz: define changes, stuff from before the rebase?08:29
you can get it back with the !reflog08:29
gitinfo The git reflog (`git log -g`) temporarily (90 days by default) snapshots your branch states at each operation that changes the branch, making it easy to undo e.g. merges and rebases. The usual warnings about !rewriting/undoing history apply. See https://sukima.github.io/GitFixUm/ for full details.08:29
qswz bookworm: yes, pending changes, not commited08:29
bookworm ah, then that's bad yes08:29
qswz ah ok, reflog, I remember, thanks08:29
Jookia j416: ah. so i guess rebase on to origin stuff instead?08:30
→ sozuba joined08:30
j416 Jookia: depends on what you want to do, really08:30
qswz can I visualize the reflog on gitk?08:31
j416 Jookia: 'git fetch' will update your remote tracking branches (which are your local branches that reflect the remote, the output of git branch -r)08:31
← sozuba left08:31
j416 Jookia: then you'll have to choose what to do with that; you can merge, rebase, reset, etc.08:31
← bambanx left08:31
→ sozuba joined08:32
Jookia hmm08:33
that doesn't seem to work on my machine. is that new behaviour?08:33
→ unluckyshrubbery joined08:33
← tiin57 left08:33
→ tiin57 joined08:37
j416 Jookia: !doesntwork08:39
gitinfo Jookia: Sorry to hear it doesn't work, but how about some details? At this point we don't even know whether it did nothing, spit out an error message, or abducted your pet unicorn in a flying saucer. That makes it hard to troubleshoot – unicorn abduction requires a completely different solution than an error message, for instance. So, please be specific...08:39
→ johnny56 joined08:40
Jookia j416: http://ix.io/2uSz08:45
j416 Jookia: so it worked08:46
Jookia: fetch does not affect your local branches (output of 'git branch08:46
Jookia: fetch does not affect your local branches (output of 'git branch'), only the remote tracking branches (output of 'git branch -r')08:46
Jookia ah08:46
j416 so you need to merge, rebase, etc.08:46
Jookia ugh08:47
j416 pull means fetch + merge08:47
Jookia can i just directly check out remote tracking branches?08:47
j416 until recently when it changed to fetch + nag to configure whether it should be merge or rebase or something else08:47
Jookia: yes08:47
Jookia: you'll end up with detached HEAD08:47
but that's fine if you just want to view them08:47
Jookia ok i'll do that instead08:48
j416 !detached08:48
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 !reattach08:48
← sudoforge left08:49
Jookia yeah that's fine08:49
← johnny56 left08:50
j416 o/08:50
Jookia: or otherwise, just 'git merge' and you're done.08:51
your branches haven't diverged so it'll be a clean fast-forward (no merge commit)08:51
Jookia j416: it's annoying08:53
j416 I see.08:53
Jookia i don't want to check out and merge every branch i might use08:53
j416 so don't?08:53
just the one you're currently working on is sufficient08:54
Jookia i'm not working on these branches08:54
j416 even easier then08:54
Jookia i'm working on a WIP branch and i rebase on stable-dev-board which isn't updated after git fetch08:54
and that does nothing because it's not updated08:54
j416 then rebase on origin/stable-dev-board after 'git fetch'08:54
Jookia yeah, that's what i'll do08:55
← duderonomy left08:55
j416 no need to keep a local copy of it if you're not working on it08:55
just keep the remote tracking branch08:56
→ duderonomy joined08:56
Jookia longer to type08:56
j416 tab completion08:58
→ jaziz joined09:00
Jookia ugh09:00
← mirrorbird left09:00
← duderonomy left09:00
Jookia it also means i have to remember which branches i interact with09:00
← vicfred left09:00
→ vicfred joined09:02
Jookia and detached HEADs won't update if i fetch, so i'd need a branch and to do git pull sometimes09:03
← tiin57 left09:03
→ johnny56 joined09:04
→ tiin57 joined09:07
← johnny56 left09:07
Jookia looks like hub does this: https://hub.github.com/hub-sync.1.html09:07
← theoceaniscool left09:12
← qswz left09:15
← nioncode left09:16
→ johnny56 joined09:18
→ nioncode joined09:19
→ i7c joined09:19
→ mirrorbird joined09:20
← johnny56 left09:22
← dedowsdi left09:29
← tiin57 left09:34
← Dara left09:34
→ YuGiOhJCJ joined09:35
→ Dara joined09:36
→ tiin57 joined09:36
→ johnny56 joined09:38
← FH_thecat left09:40
→ darkbit joined09:42
→ gpanders joined09:42
← gpanders_ left09:42
← johnny56 left09:44
→ emmanuelux joined09:45
← xelxebar left09:45
→ xelxebar joined09:46
→ paul424 joined09:47
paul424 What does the git propose : " git push --set-upstream origin new-editor" as in warning : fatal: The current branch new-editor has no upstream branch.09:48
what that cmd would do ?09:48
???09:53
Jookia paul424: it tells git that you push/pull from origin for that branch09:56
→ benfelin joined09:57
→ dedowsdi joined09:57
→ johnny56 joined09:57
→ TheSecondNik joined09:58
← tiin57 left10:04
← johnny56 left10:04
→ tiin57 joined10:06
→ johnny56 joined10:08
← sozuba left10:11
← johnny56 left10:15
→ powerhouse joined10:16
→ johnny56 joined10:19
← TheSecondNik left10:22
← iNs_ left10:23
→ iNs joined10:23
← johnny56 left10:25
→ snowkidi_ joined10:26
← snowkidi_ left10:26
→ johnny56 joined10:29
← tiin57 left10:33
← deltam left10:36
→ tiin57 joined10:37
← johnny56 left10:41
→ FH_thecat joined10:42
→ deltam joined10:43
← FH_thecat left10:43
→ snowkidimd joined10:46
← theorangeone_ left10:52
→ millido joined10:54
→ johnny56 joined10:54
→ theorangeone_ joined10:56
← theorangeone_ left10:56
← johnny56 left10:59
→ FH_thecat joined11:01
→ sudoforge joined11:03
← z8z left11:03
← tiin57 left11:03
→ johnny56 joined11:04
→ cdown joined11:05
→ tiin57 joined11:07
→ shush joined11:08
← johnny56 left11:09
→ TheSecondNik joined11:10
← shush left11:13
→ johnny56 joined11:22
← Aquazi left11:27
← sudoforge left11:28
← johnny56 left11:30
→ wrobinson joined11:31
wrobinson hi11:31
gitinfo wrobinson: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.11:31
→ calcul0n_ joined11:32
wrobinson how might I go about doing a pull request for a branch of a forked project?11:32
← tiin57 left11:33
→ johnny56 joined11:33
wrobinson i forked a repo; cloned locally; checked out the branch i'm interested in; did `git branch patch` and `checkout patch`; made changes and commited; pushed to my public repo; ....11:34
Not sure how now to do the request-pull properly (or if I've made any mistakes in my process already)11:35
← calcul0n left11:35
← Thorn left11:36
→ tiin57 joined11:37
→ Thorn joined11:38
← peeps[zen] left11:38
← johnny56 left11:39
→ h3ck joined11:42
→ Brainium joined11:43
→ wrobinso1 joined11:44
→ Nickeeh joined11:46
← TheSecondNik left11:47
← wrobinson left11:47
→ johnny56 joined11:53
← FFY00 left11:54
→ FFY00 joined11:55
→ TheSecondNik joined11:55
wrobinso1 → wrobinson11:57
→ tarkus joined11:58
← johnny56 left11:59
mackerman wrobinson: Git has no concept of pull request.12:01
TheSecondNik → TheNik12:01
mackerman If you are using one of several git based systems that provde things like forks and PRs, ask their support. #github and #gitlab have channels.12:01
← tiin57 left12:03
mackerman Also, projects or authors may have put their own procdures in place for how "merge requests" work.12:04
→ tiin57 joined12:06
← Brainium left12:07
wrobinson mackerman: thanks, but what's this then: https://www.git-scm.com/docs/git-request-pull ?12:08
Jookia wrobinson: unrelated to web-based workflows12:09
mackerman An email generator.12:09
wrobinson ah i see. appreciated12:10
I guess that would be explained somewhere, if I read the book :s12:11
→ Hello71 joined12:12
mackerman There is more than one !book but Pro Git does have a chapter on GitHub12:14
gitinfo There are several good books available about git; 'Pro Git' is probably the best: http://git-scm.com/book but also look at !bottomup !cs !gcs !designers !gitt !vcbe and !parable12:14
→ sgen joined12:14
← YuGiOhJCJ left12:15
→ johnny56 joined12:19
→ relipse joined12:24
relipse Good morning12:24
why isn't anything showing up in my diff: diff master..12:25
.3b05302279a112:25
BtbN Probably because nothing is different?12:28
→ m0viefreak joined12:28
relipse no i just added a new file myclass.php in master12:29
→ liori joined12:29
← HumanG33k left12:30
← tiin57 left12:33
← Hello71 left12:34
→ onizu joined12:35
← snowkidimd left12:36
→ Hello71 joined12:37
→ tiin57 joined12:37
→ HumanG33k joined12:37
→ snowkidimd joined12:38
← snowkidimd left12:38
← HumanG33k left12:39
→ HumanG33k joined12:39
→ cuppajoeman joined12:41
→ mjh4386 joined12:48
→ computer2000 joined12:50
← computer2000 left12:51
← millido left12:51
← jaziz left12:54
← igemnace left12:57
← tiin57 left13:03
→ tiin57 joined13:06
→ Newami joined13:08
→ Jackneilll joined13:09
← Jackneill left13:09
→ Lucas_Gray joined13:10
← sgen left13:11
← TheNik left13:17
→ causasui joined13:23
→ Brainium joined13:26
← tiin57 left13:34
→ wootehfoot joined13:34
→ tiin57 joined13:36
→ z8z joined13:36
← z8z left13:37
← tarkus left13:39
← feriman left13:40
→ leonardus joined13:41
leonardus hello, is there any way to configure Git to use Tor for all of its operations over the internet?13:42
BtbN git does not have explicit tor support, like pretty much most other software.13:45
you will have to route traffic through tor yourself. Or use a http proxy, if you only use http for git.13:46
→ Muimi joined13:47
inkbottle Is there an option to reproduce `git log|grep foo`?13:51
← m0viefreak left13:52
→ monr0e joined13:53
ikke git log --grep13:53
→ royal_screwup21 joined13:54
inkbottle of course ;)13:54
ikke depends on what you are trying to grep13:54
→ CodeSlingerPaul joined13:56
→ Phylock joined13:56
→ AnAverageHuman joined14:01
→ rafasc joined14:01
← dviola left14:02
← Phylock left14:03
← tiin57 left14:03
← royal_screwup21 left14:05
→ royal_screwup21 joined14:05
← mjh4386 left14:07
→ wunderbar joined14:07
→ tiin57 joined14:07
← diogenese left14:07
→ diogenese joined14:07
→ dmlloyd joined14:08
→ sgen joined14:09
← royal_screwup21 left14:10
inkbottle Is git language aware? I'm very surprised the way `git grep --function-context` is really showing the whole function.14:10
ikke in a very limited way14:11
inkbottle quite effective though14:11
← Dara left14:11
← vdamewood left14:12
→ aw1 joined14:12
← Lucas_Gray left14:12
→ Dara joined14:12
rafasc inkbottle: you should also look into git log -L14:12
→ xace joined14:13
rafasc the :<funcname> part in particular.14:13
inkbottle I'm on it14:13
← dedowsdi left14:14
→ heyitsrama joined14:15
← astronavt left14:17
inkbottle It is impressive14:17
← heyitsrama left14:20
aw1 does the gh cli (https://cli.github.com/manual/) wrap around the git binary like hub (https://github.com/github/hub) does?14:25
→ roadie joined14:26
→ tarkus joined14:27
aw1 I did not quite get "...wraps git in order to extend it.." part from the readme of hub14:27
ikke it means it forwards subcommands that it does not handle to git14:28
I don't think gh does the same from reading that page14:28
→ jstein joined14:30
aw1 thanks ikke14:31
← tiin57 left14:33
→ tiin57 joined14:36
inkbottle I want to create a new remote repo like so: `git push "no-history" --set-upstream [email@hidden.address] master`14:40
A web search say there can be some complications...14:41
→ mitigate joined14:41
inkbottle says*14:41
→ arcatech joined14:41
inkbottle I do not want to squash or rebase before pushing14:42
→ thiago joined14:42
→ JanC_ joined14:43
JanC_ → JanC14:43
→ royal_screwup21 joined14:43
inkbottle That, might be what I'm looking for: https://stackoverflow.com/questions/12543055/how-to-push-new-branch-without-history/3495485214:44
create a new branch with --orphan14:44
← royal_screwup21 left14:49
rafasc If you explain why you want this, we may be able to explain the complications.14:50
inkbottle (just as before; I think --orphan is what I'm looking for; there is no depth option with `git push`; I think I have the elements I need to do that) I want to create a remote repo with a "view" of my repo, not the complete repo itself14:55
I post the solution when it's done ;)14:57
→ davispuh joined15:01
inkbottle (--orphan) According to [1] that should be the right case scenario [1](https://stackoverflow.com/a/50605543)15:03
← tiin57 left15:03
→ tiin57 joined15:07
← crose left15:08
→ lucasb joined15:08
← aw1 left15:08
→ aw1 joined15:10
→ royal_screwup21 joined15:16
→ shush joined15:19
→ caskd joined15:19
← greggerz left15:22
← roadie left15:24
← royal_screwup21 left15:31
→ enoq joined15:31
→ anakimluke joined15:31
→ royal_screwup21 joined15:31
→ independent joined15:32
← tiin57 left15:33
← relipse left15:34
→ greggerz joined15:34
→ tiin57 joined15:37
anakimluke Hi! I have a picky small question. Consider the following output https://paste.ubuntu.com/p/zzHyq2mtgP/ . On line 13 the git advice message uses `--` before `<file>`; whereas on line 6, there isn't. Why is that?15:38
→ dfee joined15:38
anakimluke I have already read the manual sections about the `--` on both `reset` and `checkout`, they have the same meaning.15:39
ikke anakimluke: You are suplying 2 arguments to git reset, which means the 2nd argument must be filenames15:40
so there is no ambiguity15:40
for git checkout, you only supply one argument, so git has to guess whether it's a branchane or a filename15:41
by adding '--', you tell git that what follows must be filenames15:41
anakimluke ikke, cool!! I hadn't thought of that!15:42
← dfee left15:51
→ dfee joined15:53
anakimluke thanks :)15:53
← lukey left15:59
→ lukey joined16:01
← tiin57 left16:03
→ heyitsrama joined16:05
← tarkus left16:07
→ tiin57 joined16:07
← heyitsrama left16:09
← Gustavo6046 left16:11
→ Gustavo6046 joined16:11
← caskd left16:13
← royal_screwup21 left16:14
→ royal_screwup21 joined16:15
rafasc it also helps to disambiguate certain scenarios: E.g. if you have a branch named config and a file named config.16:16
git checkout config -- ; would checkout the branch while git checkout -- config; would checkout the file.16:16
→ TheNik joined16:16
ikke that's the ambiguity I was refering to :)16:16
inkbottle `git branch` after `git checkout --orphan="newbranch"` shows only one branch: "master"16:17
ikke inkbottle: the branch is only actually created after you commit16:18
if you run git init, git branch does not list master either16:18
← shush left16:18
→ clime joined16:18
← lukey left16:19
inkbottle okay16:19
← h3ck left16:21
→ h3ck joined16:21
→ tomty89 joined16:23
tomty89 what should i do when git bisect(ing the kernel) points to a commit that is "based" on an older tag16:24
e.g. i'm trying to bisect between v5.6 and v5.7-rc116:24
but at some point it goes to a commit that is "described" as v5.6-rc2 blah blah16:25
→ DogManJr joined16:25
← enoq left16:25
tomty89 and apparently if i reference with that commit the whole tree is indeed "old"16:25
(i.e. excludes certain/many commits between the range)16:26
inkbottle Apparently I should have commited `master` before doing `git checkout --orphan newbranch`; I say that because `git rm --cached -r` did fail. Can I do `git checkout master`; `git commit -a`; then `git checkout --orphan newbranch` again?16:28
cbreak sure16:30
inkbottle okay16:30
cbreak but I'd recommend not doing commit -a16:30
→ Phylock joined16:30
cbreak I'd recommend to always git add -p, and/or git diff --cached before git commit16:30
→ TheresaJayne joined16:31
inkbottle Seems good advice16:32
← tiin57 left16:33
← dfee left16:34
→ tiin57 joined16:37
tomty89 https://paste.linux.community/view/d105df01 anyone?16:39
← Hello71 left16:42
→ sudoforge joined16:42
demize `git commit -v` is also useful, often helps for writing better commit messages as well.16:42
mitigate given a commit id can i get the refs that refer to it (if any)16:44
ikke git branch --contains <hash>16:45
→ Hello71 joined16:45
→ tarkus joined16:45
cbreak mitigate: directly?16:45
git rev-parse --symbolic-full-name can do it sometimes16:46
→ roadie joined16:48
← DogManJr left16:49
mitigate that doesnt seem to cut it. i'm trying to get it so i can stick it into a commit message of a new commit with a new history but with the same commit info and tree16:49
← sudoforge left16:49
mitigate i'm not sure yet if git-branch --contains | head -1 would always work16:51
→ sudoforge joined16:52
ikke I would not rely on it16:53
→ pmcnabb joined16:54
→ sindns joined16:54
← paul424 left16:55
ikke tomty89: using git bisect with complex history can sometimes be hard to follow16:58
tomty89: does git rev-list v5.6..v5.7-rc1 | grep <commit> return it?16:58
← Phylock left16:59
→ heyitsrama joined17:02
tomty89 ikke: yes17:02
but, to make the problem more obvious: https://paste.linux.community/view/af72974417:03
← tiin57 left17:03
← heyitsrama left17:05
← Dara left17:05
→ tiin57 joined17:07
→ dviola joined17:12
→ Phylock joined17:13
→ astronavt joined17:14
← clime left17:18
← mitigate left17:20
← Hello71 left17:20
← blurkis left17:20
→ blurkis joined17:21
→ Hello71 joined17:23
← wootehfoot left17:23
← gpanders left17:26
→ gpanders joined17:28
← pmcnabb left17:29
← Muimi left17:30
→ sozuba joined17:30
→ Guest20 joined17:30
→ cosimone joined17:32
→ pmcnabb joined17:32
→ tgunr joined17:33
← tiin57 left17:33
← k3yp left17:34
← royal_screwup21 left17:35
→ royal_screwup21 joined17:35
← TheNik left17:35
→ k3yp joined17:36
→ tiin57 joined17:37
rafasc tomty89: I don't immediately see a problem here.17:37
← cosimone left17:37
← Guest20 left17:37
rafasc what are you trying to show with: git rev-list --count HEAD ?17:38
tomty89 rafasc: to show that the commit references a tree that is much older than the last "good" coomit17:40
as in it lacks many commits after it but before the current testing commit17:41
→ wootehfoot joined17:42
tomty89 rafasc: just to make this more obvious: https://paste.linux.community/view/d105df0117:42
(note that v5.6-rc2 is older than v5.6)17:42
→ ferdna joined17:45
→ averell joined17:46
← sozuba left17:47
← Hello71 left17:47
← pmcnabb left17:48
← ferdna left17:49
← royal_screwup21 left17:50
→ royal_screwup21 joined17:50
→ Hello71 joined17:51
← royal_screwup21 left17:55
→ sozuba joined17:57
rafasc you can have more than one one good commit, and thus, multiple paths that lead good commits into the bad one.17:58
The search is anchored on the bad commit, not the good ones.17:59
was v5.6 the only good version started bisect?18:00
tomty89 rafasc: i started with v5.6 and then the first test was good18:01
as you can see in my paste; if that's what you are asking18:01
rafasc your paste doesn't show the bisect start.18:02
← causasui left18:03
← tiin57 left18:03
← astronavt left18:04
tomty89 rafasc: ah right. i started with git bisect bad v5.7-rc1 and git bisect good v5.618:04
→ Goplat joined18:04
tomty89 rafasc: the thing is it doesn't really have to do with bisect, apparently; but like a commit in a revision list can be referencing a tree much older than the list range18:06
→ Forkk joined18:06
→ tiin57 joined18:06
tomty89 rafasc: i end up trying to remedy it by merging v5.6 to the problematic commit; and then there's conflict, and then i look for some "merge v5.6 into blah blah" commit and merge that instead18:07
← Forkk left18:07
rafasc tomty89: imagine that I branch before v5.6 and merge it to v5.7. the fact that v5.6 is good does not tell anything about this side branch, (the one that branches before 5.6 and merges to 5.7)18:07
tomty89 i think i sort of know "why" it happens, but what i really care is how to deal with it (when i bisect)18:09
→ Maxattax joined18:09
→ bolovanos joined18:09
→ Forkk joined18:10
rafasc when you mark a commit as good, you only say: ignore all descendent of this commit, ancestors of it are still uncertain.18:10
how to deal with it in what way?18:11
tomty89 rafasc: that have some reference to checkout with so that i get the real v5.6+something18:11
instead got sling shot back to where the testing commit was first introduced or so18:12
(when it was in another repo)18:12
rafasc run git bisect skip ..v5.6 ?18:13
← Hello71 left18:14
tomty89 rafasc: doesn't seem to do anything?18:15
git describe is still v5.6-rc2-876-ga2ae604da74d18:16
rafasc try v1.0..v5.618:16
→ feriman joined18:16
→ Hello71 joined18:17
tomty89 rafasc: fatal: ambiguous argument 'v1.0..v5.6': unknown revision or path not in the working tree.18:17
rafasc (v1.0 needs to exist)18:17
The ideia is that you pass a range of commits that you're not interested.18:17
I don't think there's a shortcut for "everything up to v5.6"18:18
tomty89 i'm trying v5.518:18
hmm maybe i should try v5.6-rc218:18
seem to have stalled forever18:22
rafasc git bisect skip v5.5..v5.6, but you may make it worse, as when it decides it needs something in that range, it will probably checkout 5.4.18:22
← darkbit left18:22
rafasc It has to write the skip list to a file, maybe it is doing that.18:22
another thing you may try is use: git replace --graft; to fake the history and make bisect thing there's no ancestors of v5.618:24
I believe bisect honors grafts, but I'm not certain.18:24
← Phylock left18:28
tomty89 rafasc: what i really need to deal with is not exactly bisect, i think, because what i want/need is a tree (for building) with the testing commit merged onto v5.618:30
rafasc so you're looking for a merge in 5.6 that merged a certain commit you know?18:32
tomty89 rafasc: i think so18:32
rafasc https://github.com/mhagger/git-when-merged18:32
← greggerz left18:32
rafasc or looking into git log --graph --oneline --ancestry-path <yourcommit>..v5.718:33
tomty89 rafasc: ugh no, that's not what i mean18:33
i mean like, to "re-merge" it so that i get such tree18:33
← tiin57 left18:33
tomty89 because now if i checkout the testing commit it doesn't seem to give me a tree that is based on v5.618:33
(while the commit was "ultimately" merged before the v5.7-rc1 tag)18:34
→ cosimone joined18:34
→ tiin57 joined18:36
tomty89 hmm, seem like i can actually do that if i checkout to a branch18:37
← roadie left18:37
→ robogoat joined18:38
← sudoforge left18:39
rafasc bisect is designed to find the good->bad transition, marking as good/bad are merely hints to bisect. It still chooses which ones should be tested. If you need to skip revisions, (and you seem to want to skip everything up until 5.6), you have to bisect skip. good/bad purpose is completely different.18:40
← tgunr left18:40
→ sudoforge joined18:40
rafasc Or, I didn't understand your problem, which would not surprise me to be honest.18:40
tomty89 it's odd to me that no one seem to understand me lol18:41
or maybe i don't understand git18:41
← sudoforge left18:41
rafasc I think you may be assuming git bisect works in a way it doesn't.18:42
→ sudoforge joined18:42
tomty89 or git overall18:42
rafasc is your problem that bisect is giving you commits you cannot test, because a certain commit is missing?18:43
have you tried cherry-picking that commit before testing?18:43
tomty89 it seem to me the problem is due to the fact that, it is okay to git that commits between two (consecutive) tags are out of order and references trees that are older than "the older tag"18:43
rafasc: more like many commits are missing, i think18:44
→ roadie joined18:44
← roadie left18:44
→ greggerz joined18:44
tomty89 at least in torvalds linux master, commits seem to be far from being linearly ordered18:45
→ pmcnabb joined18:46
rafasc what you said is indeed true. And can be even common, for example for a bug that affects 3.0, people sometimes branch from 3.0, then merge the result into 4.0,5.0,6.0, usually retagging the result as 4.1 5.1 6.1.18:47
→ orbyt_ joined18:47
tomty89 rafasc: i think i can fix my building tree in every round with the checkout v5.6 first then merge/fast forward approach; just that i'm not sure if git bisect needs any fix18:48
rafasc It is a relatively common workflow, especially for "backfixes". But in this case, they're more like "forward fixes", because the fix is done on the older commit that is relevant, and "forwarded" to the newer releases.18:48
→ mindCrime joined18:48
rafasc Maybe, maybe it needs a better way to know "skip everything up to 5.6".18:49
← mindCrime left18:50
rafasc I'm not sure, but I think when you skip a range, it just writes every hash into a "skip" file.18:50
→ saveNexit joined18:50
rafasc tomty89: also, git rev-list has --bisect.18:50
You could try write your own version of bisect around that.18:51
← sudoforge left18:51
rafasc tomty89: I don't think the merge ff will work. Because once the ff finishes, you'll end up marking v5.6 as good over and over again.18:53
→ gioyik joined18:53
← jbeaudoin left18:55
tomty89 rafasc: well, i mean i have another tree for building18:56
so i just mark the commit in the bisecting tree with the result of the building tree18:57
→ zillolo joined18:57
rafasc but that would be lying to bisect. Ater that you cannot trust their decisions, because they're based on false assumptions.18:58
a 'bad' commit can be fastforwarded to a 'good' commit. That doesn't make it good.18:59
← johnny56 left18:59
rafasc you seem to be confusing the purpose of marking good/bad with skipping.18:59
→ nvmd joined18:59
tomty89 rafasc: the thing is i don't know what to do other than lying lol19:01
it doesn't even seem like i can always ff either19:01
literally a mess19:01
rafasc you say: git bisect skip19:01
tomty89 rafasc: i tried, but either it doesn't seem to do anything or it just skip to something even older19:02
feels like i'm thrown to a broken time machine geez19:02
← tiin57 left19:03
tomty89 it seems like the only way is to manually hunt for some commit that happens to be v5.6+something19:03
(where manually is like randomly)19:04
rafasc maybe automate it with bisect run. If the bug was introduced in a branch that was based before 5.6 and got merged into 5.6.x, not tranversing the older commits is a bisect fail.19:04
there's a feature in progress for bisect, that would probably help you find that. Some patches were sent, and I think another person is picking it up again.19:05
It's about teaching git bisect "--first-parent".19:06
← shored left19:06
tomty89 ugh i don't know how much i understand that; probably gonna give up19:06
← Hello71 left19:06
tomty89 let me reboot and test my last miracle first19:06
← tomty89 left19:06
→ tiin57 joined19:07
← saveNexit left19:08
→ saveNexit joined19:08
→ Hello71 joined19:08
→ shored joined19:08
← Hello71 left19:13
→ johnny56 joined19:13
→ ferdna joined19:14
→ Betal joined19:16
← gioyik left19:16
← mirrorbird left19:17
→ gioyik joined19:20
→ mindCrime joined19:23
← gioyik left19:25
→ royal_screwup21 joined19:26
← saveNexit left19:30
→ saveNexit joined19:30
← tiin57 left19:33
→ gioyik joined19:34
→ tiin57 joined19:37
← tarkus left19:38
→ clime joined19:38
← royal_screwup21 left19:40
→ royal_screwup21 joined19:41
← h3ck left19:41
→ cd joined19:42
← bolovanos left19:44
← royal_screwup21 left19:46
→ Phylock joined19:46
← vicfred left19:47
→ royal_screwup21 joined19:52
← saveNexit left19:53
→ saveNexit joined19:53
← johnny56 left19:55
→ Hello71 joined19:56
← puppy left19:56
→ h3ck joined19:57
← royal_screwup21 left19:58
→ royal_screwup21 joined19:59
← royal_screwup21 left20:01
← andreas303 left20:01
→ royal_screwup21 joined20:02
← Hello71 left20:02
← tiin57 left20:03
→ andreas303 joined20:04
→ Hello71 joined20:05
← to1ne left20:05
← royal_screwup21 left20:07
→ tiin57 joined20:07
→ syr joined20:08
← CodeSlingerPaul left20:08
→ johnny56 joined20:09
← zillolo left20:10
← clime left20:10
→ to1ne joined20:14
← saveNexit left20:15
→ saveNexit joined20:15
← feriman left20:16
→ duderonomy joined20:17
← cosimone left20:20
← johnny56 left20:20
→ tarkus joined20:23
→ royal_screwup21 joined20:29
inkbottle I've amended a commit that I previously pushed on gitlab, and as a consequence, gitlab does not allow me to push it again20:29
rewt push -f will force push, but that will !rewrite !history20:30
gitinfo Rewriting public history is not recommended. Everyone who has pulled the old history will have to do work (and you'll have to tell them to), so it's infinitely better to just move on. If you must, you can use `git push --force-with-lease <remote> <branch>` to force (and the remote may reject that, anyway). See http://goo.gl/waqum20:30
[!clean_history] For a pointer on how to handle history in git see this e-mail from Linus: http://thread.gmane.org/gmane.comp.video.dri.devel/3474420:30
← tiin57 left20:33
← ferdna left20:34
→ johnny56 joined20:35
→ tiin57 joined20:37
inkbottle rewt: Thanks, it was quite messy.20:37
← saveNexit left20:38
→ saveNexit joined20:38
→ Lucas_Gray joined20:39
cbreak force pushing is fine, if you know what you're doing :)20:41
← royal_screwup21 left20:44
→ royal_screwup21 joined20:44
← royal_screwup21 left20:49
inkbottle Actually, gitlab did not accept the --force option; so, since it was mostly experimenting with git workflow, I deleted the project on gitlab and pushed it again. Not grand, but it thought me that I shouldn't have done it in the first place.20:49
← Hello71 left20:51
→ zetneteork joined20:51
→ Hello71 joined20:53
→ royal_screwup21 joined20:54
ikke inkbottle: Did you try to force push to a protected branch?20:55
← johnny56 left20:55
← royal_screwup21 left20:57
→ royal_screwup21 joined20:58
← saveNexit left21:00
→ saveNexit joined21:00
← tiin57 left21:03
← jstein left21:04
→ cosimone joined21:05
→ tiin57 joined21:06
cbreak inkbottle: you should fix the gitlab settings then21:07
→ jaziz joined21:08
→ sudoforge joined21:08
→ johnny56 joined21:10
← Hello71 left21:10
→ sunova joined21:11
→ Hello71 joined21:13
← johnny56 left21:15
← royal_screwup21 left21:19
→ royal_screwup21 joined21:19
← rafasc left21:19
← zetneteork left21:23
← saveNexit left21:26
→ wnbeser joined21:27
← lagothrix left21:27
→ johnny56 joined21:28
→ lagothrix joined21:28
← blurkis left21:29
→ blurkis joined21:31
← arcatech left21:31
→ Human_G33k joined21:33
← tiin57 left21:33
← HumanG33k left21:35
→ tiin57 joined21:37
→ causasui joined21:38
← libertyprime left21:38
inkbottle Yes, I should. I'll try that some other time.21:39
← duderonomy left21:43
→ libertyprime joined21:49
→ kantlivelong joined21:50
← tarkus left21:52
← johnny56 left22:01
→ g00s joined22:01
tinga How does "git submodule update" behave if the current repo is asking for a sha1 that is not in the submodule unless "git remote update" or "git fetch" is run for the submodule? I can't see it explained in the man page (or don't understand) and am a tad too lazy to set up a test.22:02
← tiin57 left22:03
→ astroid joined22:05
→ saveNexit joined22:06
→ tiin57 joined22:07
→ saveNexist joined22:08
→ johnny56 joined22:10
← Hello71 left22:10
← saveNexit left22:11
→ arcatech joined22:11
inkbottle How do you easily switch between branch when some files are tracked in one branch but not in the other?22:12
tinga (re myself: It does have a --no-fetch flag so it appears that by default, it *does* fetch. A colleague couldn't get it to work and because "git submodule update" is so fast for me (who already has all the commits locally) I was suspecting it does not fetch.)22:12
→ Hello71 joined22:13
← johnny56 left22:15
tinga inkbottle, heh, dunno; but where is the non-tracked file coming from? A history that relies on a non-tracked file doesn't seem to make much sense unless it's generated, and if it is, just generate it?22:16
You could also use `git cat-file -p $sha1 > file` to get a copy from the other branch.22:17
inkbottle I think `git checkout -f master` would be the right solution (initially I've created the branch I'm in now with `git checkout --orphan` for the reason "you want to publish the tree from a commit without exposing its full history". So I've done some `git rm -r --cached`)22:18
The files that are going to be overwritten are identical22:19
tinga But if you check out the other branch again it will delete that file.22:19
inkbottle Only I didn't want to push them, so I `git rm --cached` them22:19
Not delete IMO, but overwrite with an identical file22:20
Oh, "again"...22:20
tinga What I mean is if master tracks foo, but otherbranch does not, you can go from otherbranch to master via "git checkout -f master" and it won't complain;22:21
inkbottle BAck from master to unrelated-orphan-branch22:21
tinga but if you want to go from master to otherbranch, it will unlink foo, which may not be what you want.22:21
But you can get a copy of the file via git cat-file as I showed.22:21
inkbottle I agree with waht you said22:22
Not sure the file will be unlinked... However it wouldn't really matter in my case, plus, I can retrieve it if needed with the method you suggested22:23
← cbreak left22:25
inkbottle The `-f` didn't cause any issue; actually it worked very well22:26
→ cbreak joined22:26
← royal_screwup21 left22:29
→ royal_screwup21 joined22:29
← saveNexist left22:30
→ blahboybaz joined22:30
→ saveNexist joined22:30
saveNexist → saveNexit22:31
→ leeN joined22:31
← orbyt_ left22:32
→ Anticom joined22:32
← tiin57 left22:33
← royal_screwup21 left22:35
← jaziz left22:35
← Hello71 left22:35
← sunova left22:37
→ tiin57 joined22:37
→ Hello71 joined22:38
← cosimone left22:44
→ johnny56 joined22:45
→ zebrag joined22:47
← causasui left22:47
← inkbottle left22:48
← johnny56 left22:51
← saveNexit left22:53
→ saveNexit joined22:53
← Anticom left22:53
← calcul0n_ left22:57
← powerhouse left22:57
→ sindns_ joined22:59
← kjartan left22:59
→ orbyt_ joined23:02
← sindns left23:02
→ kjartan joined23:02
← tiin57 left23:03
→ johnny56 joined23:04
→ dqx_ joined23:06
→ tiin57 joined23:07
→ royal_screwup21 joined23:09
← johnny56 left23:11
← leeN left23:11
← aw1 left23:13
← royal_screwup21 left23:14
→ aw1 joined23:14
← Lucas_Gray left23:14
← emmanuelux left23:15
← saveNexit left23:15
→ saveNexit joined23:15
← dviola left23:16
→ dviola joined23:18
→ dedowsdi joined23:19
→ johnny56 joined23:25
← dedowsdi left23:27
→ theoceaniscool joined23:29
← johnny56 left23:30
← saveNexit left23:30
→ dedowsdi joined23:31
← sozuba left23:31
← mindCrime left23:31
← AnAverageHuman left23:31
→ AnAverageHuman joined23:31
← tiin57 left23:33
→ sozuba joined23:33
← Phylock left23:33
→ tiin57 joined23:37
← wnbeser left23:37
← kjartan left23:40
→ kjartan joined23:40
→ byanka joined23:43
→ johnny56 joined23:43
→ powerhouse joined23:44
← theoceaniscool left23:44
→ leeN joined23:45
← leeN left23:45
→ leeN joined23:45
← nvmd left23:46
← leeN left23:46
→ vicfred joined23:47
→ g4570n joined23:47
← johnny56 left23:51

Logs Search ←Prev date Next date→ Channels Documentation