IRCloggy #git 2019-01-18

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.

2019-01-18

← surfist left00:00
← sanscoeur left00:00
→ surfist joined00:00
← sanscoeu_ left00:01
→ Kaisyu joined00:02
→ peacememories joined00:02
← finalbeta left00:04
← dpyro left00:05
→ dpyro joined00:08
← andietnad left00:10
← thatpythonguy left00:15
← hbautista left00:19
→ BenderRodriguez joined00:20
← spacesuitdiver left00:24
← DolphinDream left00:24
→ beatzz joined00:25
← n3wborn left00:26
← thiago left00:28
← elsheepo left00:29
← Noodlewitt left00:32
→ Noodlewitt joined00:32
→ andyhuzhill joined00:35
→ xrexeon joined00:35
← jstein left00:35
← topdownjimmy left00:36
→ topdownjimmy joined00:36
← cdown left00:38
← matsaman left00:39
→ Raed joined00:41
← irdr left00:43
→ irdr joined00:43
← elichai2 left00:45
Disconsented_ → Disconsented00:45
← venmx left00:47
→ DolphinDream joined00:50
← peacememories left00:50
← manuelschneid3r left00:50
→ horribleprogram joined00:50
← yyy left00:52
Zexaron hello00:55
before I push I would want to compare diff between two branches00:55
that's using git diff or is there any "prepush preview" feature that could be handy?00:56
like git push --checkdiff00:56
I'll see git push00:57
← akushner left00:58
→ Tibian joined00:59
← dfee left00:59
← jrnieder left01:00
← Tibian left01:02
bremner Zexaron: git diff origin/branchname, for example01:03
or git log origin/branchname..01:03
→ ferdna joined01:04
← DolphinDream left01:05
Zexaron bremner: oh sorry, diff betewen commits of these two branches, not the whole branch, but it worked, it's just a LOT because i'm comparing 4 months of master changes01:06
rewt `git diff <ref> <ref>` will diff the 2 refs01:09
bremner if the second command gives you the commits in HEAD but not in the remote branch01:09
← causasui left01:09
← \void left01:10
→ thatpythonguy joined01:10
Zexaron github shows different hash than git local, that normal ?01:13
toothe it is if you have different commits01:14
otherwise, no01:14
Zexaron I'm looking at the log at the same date it is on github01:14
toothe is the commit message the same?01:15
→ DolphinDream joined01:15
Zexaron yes01:16
well, I made a new branch based off the remote branch because the original I deleted locally01:17
I also rebased01:18
← Puffball left01:18
← mat001 left01:19
Zexaron i used log before rebase, i can see the commit being the same hash, rebasing makes new hash on totally unchanged commit01:19
well, it may have been changed during something rebase did i guess, I wasn't editing any conflicts manually01:20
← Sasazuka left01:20
← m0viefreak left01:21
Zexaron rewt: i did git diff hash1 hash2 and it didn't work, still compares full branch01:23
I need to refresh what <ref> is lol01:24
→ Wetmelon joined01:25
← ngui left01:25
→ ngui joined01:26
← DolphinDream left01:28
← ngui left01:30
→ ngui joined01:30
rewt hashes are refs01:30
what do you mean by compares the full branch?01:31
a branch is just a pointer to a commit01:31
if you mean compares all files, then yes, that's expected, because each git commit is basically a snapshot of the working tree01:31
if you want to only diff certain files, you can do `git diff <ref> <ref> -- <file>`01:32
Zexaron okay it does differently now that I put 3 dots in middle01:32
hash...hash01:32
rewt the triple-dot means compare the 2nd ref to the last common ancestor of the 2 refs... basically what changed in ref2 since the branches diverged01:33
Zexaron but it always diffs the whole branch, what I mean is, it starts from begining such as .cmake or .md ... like totally off what i need01:34
→ akushner joined01:34
Zexaron one of the refs is on remote branch01:34
but it exists locally inside the log ofcourse, before I rebased01:34
range-diff was outputting commit names ... so that doesn't count01:35
rewt yes, so if you `git diff master featurebranch`, and there were merges into master since featurebranch was created, then those would also show in that diff... but if you `git diff master...featurebranch`, those additional merges into master would not show up01:36
→ llinguini joined01:36
rewt i think that may have been what you saw01:36
← Peetz0r left01:38
→ jstimm joined01:39
→ igemnace joined01:39
Zexaron there's little difference in any case I made so far01:43
with one hash, two hasehes, 3 dots, 2 dots01:44
← beatzz left01:45
← ngui left01:45
→ beatzz joined01:45
Zexaron rewt: infact, the output is the same no matter what I enter01:45
doesn't even need to be a valid hash01:46
← CodeSlingerPaul left01:46
Zexaron i entered 3eu29u293u2929u and got the same output, gibberish01:46
← boombatower left01:46
Zexaron okay sorry, I did it only now01:46
that didn't work, looks like it does work if the part of the has is missing, it still recognizes it as valid01:47
so I guess nevermind01:47
rewt yes, hashes can be shortened as long as they stay unique in the repo01:47
Zexaron Well none of the suggestions, do you have any ideas what else can I do ?01:48
rewt even in huge repos, you can go as short as about 5 characters01:48
sometimes01:48
Kira canton7, osse: a much safer method seems to be using xhost to test the existence of an x session.01:48
Zexaron none of the suggestions worked*01:48
rewt are you looking at a public repo?01:48
Zexaron I'm not dealing with master right now01:48
I have a local branch based on the old remote branch, remote branch is still 4 months old01:49
Kira canton7, osse: that's what I ended up doing.01:49
Zexaron I made a copy, rebased, updated01:49
Yes it is public on github01:49
but I'm comparing it to my fork on remote, still I have it ofcourse fetched so it's downstream/my_branch01:49
I'm not dealing directly with the public repo on upstream01:50
rewt so there was an existing branch, you rebased it, and added more commits to it?01:50
Zexaron no I did not add more commits, there's only one commit01:50
I rebased the branch, and amended the same commit01:50
→ drbean_ joined01:51
→ jottr_ joined01:51
rewt when you rebase, you can't really diff the old to the new... there is no direct relationship between them01:51
→ Elon_Satoshi joined01:51
← akushner left01:52
→ mar77i_ joined01:52
rewt a rebase is the same as creating a new branch, and copying all the commits from the old branch to the new branch, and then deleting the old branch01:52
Zexaron I diffed the changes prior to amending, that's what im looking for, but that's a partial diff, only the changes, not the whole01:52
rewt so anything that was added to the source branch between creating the old and new branches are present in the new branch and will show up in any diff you try01:52
→ toothe_ joined01:53
→ SuperKaramba joined01:53
→ Werring` joined01:53
→ Zohnuqu3pah5iej- joined01:53
Zexaron the rebase most likely did not change anything in my commit, but I was curious01:53
→ jargan joined01:53
rewt the rebase very likely recreated every single commit since your branch was created01:54
Zexaron the rebase is only for getting rid of the conflicts, right, those commits that were added by rebase are totally irrelevant01:54
rewt there is something you can try though:01:54
Zexaron to me*01:54
← alyptik left01:54
rewt diff <( git diff source...oldbranch ) <( git diff source...newbranch )01:54
that will diff the diffs of each branch01:54
Zexaron what is "source" ?01:54
→ boshhead_ joined01:54
rewt source is the branch that it was branched off of... what was active at the time the branch was created01:55
← boshhead left01:55
boshhead_ → boshhead01:55
rewt likely master01:55
Zexaron I usually switch to master before creating anything new yes01:55
→ rafalcpp joined01:55
→ mikap joined01:56
Zexaron but the source would be different for the old branch right, that's way way long ago, any way I could figure out it's hash ?01:56
→ wolfshappen_ joined01:56
← jottr_ left01:56
→ waveclaw_ joined01:56
→ Kronuz_ joined01:57
rewt that could've been master too01:57
→ imMute^ joined01:57
Zexaron i don't think the log goes back 4 months, and I think I did some log cleaning and fsck and object and repacking01:57
→ irdr_ joined01:58
→ thealphanerd joined01:59
→ kjartan_ joined01:59
← ferdna left02:00
← irdr left02:00
← xrexeon left02:00
← BenderRodriguez left02:00
← kjartan left02:00
← jottr left02:00
← kfoonamalik left02:00
← InfinityIO left02:00
← catkiller left02:00
← xnox left02:00
← rts-sander left02:00
← waveclaw left02:00
← CommanderViral left02:00
← Kronuz left02:00
← Pistos left02:00
← Trel left02:00
← chckyn left02:00
← ahf left02:00
← Werring left02:00
← ghormoon left02:00
← drbean left02:00
← brandonkal left02:00
← Copenhagen_Bram left02:00
← jast left02:00
← moei left02:00
← _xor left02:00
← dhollinger left02:00
← stutter left02:00
← powerbit left02:00
← errr left02:00
← gde33 left02:00
← leah2 left02:00
← strugee left02:00
← mika left02:00
← Zialus left02:00
← wolfshappen left02:00
← jrzz_ left02:00
← paraxial left02:00
← mar77i left02:00
← comptroller left02:00
← rafalcpp_ left02:00
← APic left02:00
← swalladge left02:00
← toothe left02:00
← Zohnuqu3pah5iejo left02:00
← anddam left02:00
← imMute left02:00
← MylesBorins left02:00
← sweatsuit left02:00
← satifant left02:00
Kronuz_ → Kronuz02:00
jargan → jast02:00
Werring` → Werring02:00
mikap → mika02:00
thealphanerd → MylesBorins02:00
Kira netsplit! yay02:00
rewt you could use gitk or similar to see what the repo actually looks like; it would give you all relationships between commits and hashes for all commits02:00
→ strugee joined02:00
→ chckyn joined02:00
→ sweatsuit joined02:01
Zexaron rewt: well I just reminded that I don't need hashes, i can do branches ofcourse, so i just put master in both sources, and branch names accordingly specifiying downstream/oldbranchname and the current branch I'm on which is local02:02
→ paraxial joined02:02
Zexaron rewt: and I think it worked!!!02:02
→ ahf joined02:02
← jstimm left02:03
Zexaron So i got the full diff, not just the latest changes diff02:03
→ cdown joined02:03
← angelo_ts left02:04
→ jubal joined02:05
→ xrexeon joined02:06
→ beatzz_ joined02:06
→ Peetz0r joined02:06
→ ferdna joined02:07
→ errr joined02:07
beatzz_ → elsheepo02:07
→ Trel joined02:07
→ Pistos joined02:07
→ swalladge joined02:07
→ rts-sander joined02:07
→ satifant joined02:07
→ gde33 joined02:07
→ anddam joined02:07
→ leah2 joined02:07
← ferdna left02:07
→ dhollinger joined02:08
→ ferdna joined02:08
← jubal_ left02:08
→ angelo_ts joined02:09
→ jstimm joined02:09
← beatzz left02:10
→ finalbeta joined02:11
← horribleprogram left02:12
→ DanDare joined02:13
← F0rTex left02:13
→ roygbiv joined02:13
→ F0rTex joined02:14
← troyt left02:15
→ troyt joined02:15
DanDare Hi. I'm new to git. In a simple project lets say I want to test a bug fix or improvement, what's preferable? Testing everything before commit or mark that in git history?02:21
Well, not exactly a git question but what's common? Just commit everything else youre doing (or trying) to do?02:22
b0nn DanDare: Unit tests?02:23
→ thiago joined02:23
b0nn I would run unit tests relevant to that branch until I was ready to merge, then I'd run the full suite02:23
DanDare In case i decide to register/commit that try but then I see its not the case and want previous state, its 'git revert' im looking for?02:24
b0nn your CI/CD setup should cover this02:24
DanDare b0nn, well not unit test02:24
b0nn, I mean, im confused to even answer you. Im new to git and version control all together. Thanks anyway.02:25
b0nn ok, git's only reason for living is to track changes in your code02:26
people often pair it with a continuous integration/continuous deployment system02:27
which really means, they run certain types of unit tests when you commit, and when you merge02:27
→ davidfetter joined02:27
DanDare Nice I understand this. Thanks02:28
← cdown left02:28
→ alyptik joined02:29
← lss8 left02:31
← thatpythonguy left02:31
→ lss8 joined02:32
← steven left02:33
→ steven joined02:35
← orbyt_ left02:38
→ sgn_ joined02:39
← Wetmelon left02:40
← zulutango left02:41
→ orbyt_ joined02:42
→ fattredd_ joined02:43
→ cdown joined02:44
← Peetz0r left02:44
→ hbautista joined02:48
← Essadon left02:48
→ zulutango joined02:48
← korzq left02:53
← duderonomy left02:53
← zulutango left02:53
→ sanscoeur joined02:54
← Zexaron left02:55
→ ngui joined02:56
→ jottr_ joined02:56
→ zulutango joined02:56
← skered left03:00
← ngui left03:00
← ketas left03:00
→ ngui joined03:00
toothe_ → toothe03:00
roygbiv DanDare: probably the most commonly used method is to create a branch for your experiments and do your work there03:00
← cdown left03:00
roygbiv but it's flexible so you can't really say there is the One True Way™03:01
← JanC left03:02
→ cdown joined03:02
→ JanC joined03:03
→ appleguru joined03:03
DanDare roygbiv, ok thanks. And yeah that's how I said my question wasn't exactly about git :p I think it makes sense, create branch, try what you have in mind, then forget that forever if its not good03:03
→ watabou joined03:03
→ DolphinDream joined03:04
roygbiv DanDare: yep. the idea is to keep your master branch clean and use it as your "release" branch. if you like your experiments, then you can merge them to master and make a new release03:04
some people don't like using master as a release branch though. it can be really complicated but for a simple project with just you or a couple of other devs, using master should be fine.03:05
appleguru git newb here… I need to do a rebase, but I don’t think I am doing it quite right03:06
roygbiv 😲03:06
→ prakashdanish joined03:06
appleguru I have a local branch I want to open a PR against a remote branch with… but right now it has a ton of unrelated commits in its history, even though the actual diff is very small03:07
how do I get rid of them all?03:07
← elsheepo left03:07
roygbiv appleguru: you mean you want to combine a bunch of commits into a single one?03:07
← daegontaven left03:07
appleguru maybe?03:07
roygbiv heh03:07
DanDare roygbiv, ok I got it thanks. Just using master is perfectly fine for me03:07
appleguru I started with the remote branch03:08
← Geezus42 left03:08
appleguru and cherry picked commits from a different local branch that I needed03:08
roygbiv DanDare: there's something called "git-flow" if you want to read up on it. it'll make your brain hurt for a small project, but it might be worth look at for edification03:08
← jstimm left03:08
roygbiv appleguru: ok i don't mean to sound like a git god here but what i usually do is rebase my local branch, then force push it up to the remote.03:09
← jottr_ left03:09
→ dfee joined03:09
appleguru roygbiv: thats exactly what I was trying to do03:09
← DolphinDream left03:09
appleguru but, I got a whole bunch of merge conflicts when I tried03:09
DanDare roygbiv, sure! I take note. My brain hurts already so I think it doesnt matter :p03:09
appleguru so.. maybe I was doing it wrong?03:09
roygbiv appleguru: hmmm, weird03:09
appleguru: did you do something like 'git rebase -i HEAD~5' or some-such?03:10
appleguru I tried `git rebase -i HEAD~1`03:11
here, I’ll show you specifically what I am trying to do… may help :D03:11
https://github.com/BogGyver/openpilot/pull/2303:11
roygbiv see that's only going to show you one commit. if you want to combine several into one then you'll need to do like HEAD~5 or however many previous commits you need to see03:11
appleguru diff is 6 files, pretty small03:11
(but is showing 176 commits)03:12
← zulutango left03:12
roygbiv well you can show a whole bunch. you can say HEAD~176. but that doesn't mean you are going to do anything necessarily with all 176 of them03:12
→ esrse joined03:12
roygbiv that's just how many you want to see. once git drops you into an editor, you then give it instructions on what to do with various commits. if you don't do anything to a commit, it's left alone03:13
appleguru I guess I don’t really understand how all of thsoe commits got into the history to begin with?03:15
since I in theory started at the latest head and then just cherry picked a few things on top03:15
→ DolphinDream joined03:15
appleguru (or at least tried to)03:15
roygbiv and the 'f' option is probably the one you want to use. it'll combine a commit with the one above it and remove its log message03:15
hmmm ok that question is probably out my pay range. i'm pretty modestly skilled about git when it comes to anything outside the typical operations03:16
Redrambles may be totally off here - but would something like 'git reset --soft <earliest relevant commit>' and then an add, commit and force push work?03:16
appleguru I mean, I can pull out the 6 files that changed, revert to the latest and make a new commit with them… that would work fine of course, but isn’t very git like (and loses the history, which doesn’t resally matter too much here)03:18
← Dirak left03:19
roygbiv if it were me i would just use the 'f' option in an interactive rebase to roll a bunch of commits into a single one. then force push03:19
is that the best or smartest way? i have no idea heh but it's the one i know03:20
← ngui left03:20
→ ngui joined03:20
→ zulutango joined03:21
→ skered joined03:22
appleguru roygbiv: how do I do that?03:24
roygbiv something like "git rebase -i HEAD~10" or however many commits you want to see, then just start marking them 'f' for the ones you want to combine into the commit above it03:25
→ andyhuzhill_ joined03:25
← fattredd_ left03:25
appleguru but there are like ~170 completely unrelated commits?!03:25
→ korzq joined03:25
appleguru (that I essentially want to just toss)03:25
I just want to tell git “please take what I have, and put it on top of this remote branch with one new commit”03:26
how can I do that?03:26
roygbiv ok you want to delete the commits completely, including any changes associated with them?03:26
appleguru yeah03:26
roygbiv ok well instead of marking them 'f' you can mark them 'd' but hmm, 170 of them. i have no idea what's going to happen heh03:27
appleguru really needs to learn git03:27
roygbiv i've deleted one or two commits from a repo but 170 that is really something. i thought you were wanting to keep the changes, but just combine the commits03:27
← andyhuzhill left03:28
andyhuzhill_ → andyhuzhill03:28
→ Goplat joined03:30
← ferdna left03:31
appleguru I just did it the old fashioned way *shrug*03:33
looks much cleaner now :P03:33
roygbiv all's well that ends well 😎03:34
→ andyhuzhill_ joined03:34
→ mobile_c joined03:34
→ Dirak joined03:35
→ breakingsad joined03:36
→ ketas joined03:36
← dfee left03:38
← andyhuzhill left03:38
andyhuzhill_ → andyhuzhill03:38
← andyhuzhill left03:39
→ andyhuzhill joined03:40
← q-chymera left03:40
→ q-chymera joined03:41
← q-chymera left03:42
→ q-chymera joined03:42
← andyhuzhill left03:44
→ andyhuzhill joined03:44
← ngui left03:45
→ ngui joined03:45
← sanscoeur left03:45
→ Peetz0r joined03:46
→ dfee joined03:47
← zulutango left03:47
→ zulutango joined03:48
← appleguru left03:48
← roygbiv left03:49
← andyhuzhill left03:49
→ andyhuzhill joined03:49
← lagothrix left03:50
→ lagothrix joined03:50
← ngui left03:50
← Khisanth left03:50
→ ngui joined03:51
← Peetz0r left03:51
→ Peetz0r joined03:52
← alyptik left03:52
← Dirak left03:52
→ thebope joined03:53
← andyhuzhill left03:54
← skered left03:54
→ andyhuzhill joined03:55
← andyhuzhill left03:59
→ andyhuzhill joined03:59
→ ferdna joined04:01
→ alyptik joined04:03
← andyhuzhill left04:04
→ andyhuzhill joined04:04
← blackmesa1 left04:05
← DolphinDream left04:05
← ferdna left04:05
SuperKaramba → BenderRodriguez04:06
→ ir7466 joined04:07
ir7466 hi there04:07
i do a git fetch, and then git remote -a04:07
but i am definitely not seeing the same list of remote branches i see on github04:07
could it be someone else introduced some sort of fetch "rule" that is blocking some branches?04:07
i do notice that the only ones it's fetching have particular names04:07
but i don't know where to look for such a "filter"04:08
→ Khisanth joined04:08
→ stitched_unicorn joined04:08
← ngui left04:10
→ ngui joined04:10
← stitched_unicorn left04:12
← Cabanoss- left04:14
→ Cabanossi joined04:14
← mooasaurus left04:15
← korzq left04:15
← orbyt_ left04:15
→ stitched_unicorn joined04:15
→ sanscoeur joined04:16
→ ferdna joined04:18
→ buzzlewoo joined04:20
← Trieste left04:22
→ Trieste joined04:25
← buzzlewoo left04:29
← prakashdanish left04:32
→ kapil____ joined04:35
→ brandonkal joined04:38
← Noodlewitt left04:38
← brandonkal left04:41
→ brandonkal joined04:41
← stitched_unicorn left04:44
→ sauvin joined04:45
← ferdna left04:45
→ ferdna joined04:46
← dfee left04:46
← xrexeon left04:47
→ rsrx joined04:49
→ Noodlewitt joined04:55
→ \void joined04:58
→ dfee joined05:00
→ jottr_ joined05:05
→ benharri joined05:06
← jottr_ left05:10
← bket left05:15
waveclaw_ → waveclaw05:17
→ bket joined05:18
→ netj joined05:21
→ duderonomy joined05:22
→ inkbottle joined05:24
← libertyprime left05:26
→ figurelisp joined05:28
toothe anyone understand the idx file format?05:29
thiago what are you trying to do?05:32
← ferdna left05:36
toothe thiago: sorry, I should be clear05:36
I am trying to understand how the idx format works05:36
i gethat...there is an incrementing count based on the first byte of the sha1 digest05:37
then the digests05:37
what comes after that?05:37
← netj left05:37
toothe I'm reading Documentation/technical/pack-format.txt05:37
→ netj joined05:38
← inkbottle left05:38
← scrptktty left05:38
→ akushner joined05:39
← cdown left05:43
→ cdown joined05:43
→ xmate joined05:47
→ inkbottle joined05:47
← bn_work left05:49
→ tsdh joined05:51
← akushner left05:51
→ Puffball joined05:52
← snowgoggles left05:53
→ akushner joined05:58
← flying_sausages left05:59
← Elon_Satoshi left06:00
← ngui left06:00
→ Elon_Satoshi joined06:02
→ libertyprime joined06:03
← cdown left06:05
→ ngui joined06:05
← kjartan_ left06:05
thiago toothe: sure, but just for curiosity or are you trying to do something with those files?06:07
toothe yes - I am writing an implementation of git06:08
at least parts of it06:08
personal project06:08
← benharri left06:10
← pR0Ps left06:10
→ kjartan joined06:10
→ Dirak joined06:14
toothe i think its in write_idx_file06:21
in the file pack-write.c06:21
tsdh I have the task to merge fixes from the latest release branch into master every now and then. Is there any reason to favor "git merge release_x_y_z" over "git merge origin/release_x_y_z"? I'd rather use the latter because then I can just fetch and go and have no need to checkout release_x_y_z before.06:22
← figurelisp left06:24
→ pR0Ps joined06:27
→ venmx joined06:29
→ dre_ joined06:29
→ n3wborn joined06:29
← victorqueiroz left06:31
kadoban tsdh: No, git doesn't care06:32
tsdh kadoban: Ok, good.06:33
← rsrx left06:34
kadoban Do make sure you fetch though, if that's your intention. Or at least know exactly what you're merging. Git will happily merge a week/month/year old origin/release if you just forget to fetch06:34
← libertyprime left06:35
← kadoban left06:37
← hbautista left06:37
tsdh kadoban: Could that make any problems? It would merge whatever the last merge-base is. In the worst case, I could have merged some more commits now which I'll then merge the next time, no?06:38
Oh, he's gone...06:38
→ mat001 joined06:40
← ngui left06:40
→ ngui joined06:41
← andyhuzhill left06:44
← Goplat left06:45
→ andyhuzhill joined06:45
← ngui left06:45
→ ngui joined06:45
→ exmate joined06:47
← Inline left06:47
← xmate left06:48
← Dirak left06:50
→ andyhuzhill_ joined06:50
→ Dirak joined06:50
← ngui left06:50
→ ngui joined06:51
← \void left06:52
← andyhuzhill left06:53
andyhuzhill_ → andyhuzhill06:53
→ powerbit joined06:55
← ngui left06:55
→ ngui joined06:55
← thiago left06:57
← Dirak left06:57
→ keepLearning512 joined06:58
← andyhuzhill left06:58
← dre_ left07:01
← yuriii left07:03
→ rick8024 joined07:04
→ Dirak joined07:04
← KnoP left07:05
rick8024 → KnoP07:05
← venmx left07:05
→ jottr_ joined07:06
→ flying_sausages joined07:09
→ andyhuzhill joined07:10
← jottr_ left07:10
← thebope left07:11
→ thebope joined07:12
← Dirak left07:13
← igemnace left07:14
→ chele joined07:15
← Noodlewitt left07:15
→ Dirak joined07:16
→ ansraliant joined07:16
ansraliant morning07:17
← thebope left07:18
← andyhuzhill left07:18
← Dirak left07:21
→ dreiss joined07:26
← Fusl left07:28
→ Fusl joined07:30
→ andrzejv joined07:30
← nowhere_man left07:34
→ rsrx joined07:34
→ T_UNIX joined07:35
← mat001 left07:37
→ sQVe joined07:39
← ir7466 left07:41
→ cluelessperson_ joined07:44
cluelessperson_ how do you mirror a repo without saving it in a *.git directory?07:44
→ andyhuzhill joined07:45
→ Phylock joined07:45
→ Dirak joined07:45
m1dnight1 → m1dnight_07:46
→ acidjnk joined07:47
→ horribleprogram joined07:48
← rsrx left07:48
← breakingsad left07:48
← andyhuzhill left07:49
→ thebope joined07:49
→ andyhuzhill joined07:49
→ pks joined07:52
← ngui left07:55
→ ngui joined07:56
→ oxymoron93 joined07:59
← dreiss left07:59
← ngui left08:00
→ ngui joined08:00
osse cluelessperson_: what do you mean by mirror in that case?08:04
→ dege joined08:05
osse cluelessperson_: maybe you want man git-bundle08:09
gitinfo cluelessperson_: the git-bundle manpage is available at https://gitirc.eu/git-bundle.html08:09
← horribleprogram left08:11
→ pks_ joined08:12
← pks left08:14
pks_ → pks08:14
← andyhuzhill left08:14
→ bsaboia joined08:15
→ andyhuzhill joined08:15
← andyhuzhill left08:19
→ andyhuzhill joined08:20
← ngui left08:20
anddam howdy08:20
→ ngui joined08:21
← GreenJello left08:22
← dpyro left08:23
← andyhuzhill left08:24
→ andyhuzhill joined08:24
← ngui left08:25
→ ngui joined08:25
anddam I have a local master that 'is ahead of 'origin/master' by 7 commits"08:26
how do I reset this to origin/master?08:26
also why do I sometime specify "repo refspec" and other times "repo/refspec"?08:27
oxymoron93 origin/master is remote tracking branch, it is the state of remote repo as of the last time you fetched from origin anddam08:28
← keepLearning512 left08:28
← andyhuzhill left08:29
→ andyhuzhill joined08:30
← Puffball left08:30
anddam but I just fetched origin08:31
osse anddam: "repo/refspec" is actually just "refspec". Some refs have the remote repo name in their name08:31
anddam: if you want to nuke all your own crap on master you can do git reset (--hard) origin/master08:32
oxymoron93 sorry for bailing out, had to quickly satisfy some other customers anddam :D so you use that ^08:32
anddam osse: is that a different namespace though? I mean is "origin/fubar' just part of the ref name?08:33
osse anddam: yes. the full name is "refs/remotes/origin/fubar"08:33
anddam oxymoron93: no problem at all, I would never pretend help, even the less in real time08:33
→ mat001 joined08:34
← andyhuzhill left08:34
anddam osse: I think I'm missing a bit then, that full name is a "local" concept?08:34
osse anddam: not sure what you mean by "local concept". But it is indeed "local"08:34
anddam: refs/remotes/origin/fubar is like a "mirror" of the branch named "fubar" that exists on the remote named "origin"08:35
anddam I thought it was split in two domains, repository and ref, so (origin, master) would reference the ref named "master" on remote "origin"08:35
← sgn_ left08:35
anddam osse: oh ok, that mirror part implies it is "local" for what I perceive as local08:35
→ keepLearning512 joined08:35
osse anddam: in "git push origin master" you decide you want to push to origin, and what you want to push is master08:37
anddam osse: let me show an example of what makes me confused08:38
← mat001 left08:38
→ manuelschneid3r joined08:38
anddam git branch -a has master remotes/origin/HEAD -> origin/master remotes/origin/master08:38
I want to check all "my own crap" before resetting master to be sure there is nothing important so I run git diff origin/master..master08:39
osse anddam: the first thing is your own master branch. the last thing is that mirror i talked about.08:39
anddam ok, but here I have two issues08:39
osse the stuff in the middle is pretty uninteresting to be honest. at list for this discussion08:39
anddam 1) the diff I ran says warning: refname 'master' is ambiguous.08:39
how's that ambiguous if it is an exact match with a branch name?08:40
osse anddam: do: git for-each-ref and show me the output08:40
anddam 2) the part you defined uninteresting puzzles me, it "links" origin/master but there's no such line in branches08:40
← ali1234 left08:42
← exmate left08:43
anddam osse https://gist.github.com/adab4d/f904478a1d820e1fa2853f47e966da3708:43
→ andyhuzhill joined08:45
← ngui left08:45
→ ngui joined08:46
→ Silmarilion joined08:46
oxymoron93 anddam: issue is that you have tag named `master`08:48
what was the purpose for such naming? playing around? :D08:48
← andyhuzhill left08:49
anddam no, I remember that was the suggested name for stable in gitflow08:49
→ andyhuzhill joined08:49
← ngui left08:50
→ ngui joined08:50
anddam oh the tag, not the ref08:50
osse anddam: you have both a branch and a tag named master08:51
that's why "master" is ambiguous08:51
oxymoron93 well tag is also ref as you can see : refs/tags/<tag>08:51
→ rsrx joined08:52
anddam yes I figured after oxymoron93's line and deleted it08:52
→ exmate joined08:53
← ansraliant left08:57
← freeman42x left09:00
← ngui left09:00
← wrouesnel left09:01
→ floppydh joined09:03
→ jottr_ joined09:07
← shentino left09:07
→ shentino_ joined09:08
amosbird Hello09:10
when doing rebase09:10
https://la.wentropy.com/A6mC09:10
← troyt left09:10
amosbird which one is my local commit?09:10
→ ngui joined09:10
→ stitched_unicorn joined09:11
← jottr_ left09:11
→ jrzz joined09:11
→ troyt joined09:12
→ mikecmpbll joined09:13
← rsrx left09:13
→ cswl joined09:14
→ fairuz joined09:15
← fairuz left09:15
→ rsrx joined09:16
→ leprechau joined09:18
→ mpodien joined09:18
← andyhuzhill left09:19
→ andyhuzhill joined09:20
→ rnmhdn joined09:22
→ Makaveli7 joined09:22
oxymoron93 amosbird: usually, your side is shown before `=======`, in this case plain nothing09:24
amosbird oxymoron93: hmm, my local commit has the message "."09:25
which is after ======09:25
I don't quite understand if that really means my local changes09:25
osse anddam: as for the uninteresting puzzle. HEAD is a symbolic ref, like a pointer to a ref. That puzzling bit is a copy of the origin repo's HEAD and what it points to09:25
amosbird that "onto" and "join" keywords are also confusing09:26
→ GreenJello joined09:27
← mpodien left09:27
→ jottr_ joined09:30
← ngui left09:30
← pks left09:31
amosbird it's indeed my local changes09:32
← andyhuzhill left09:34
→ nowhere_man joined09:35
→ andyhuzhill joined09:35
← andyhuzhill left09:35
→ ansyeb joined09:35
ansyeb hello. I am doing: git ls-tree -r --name-only master | grep Project/Crm/Classes/System/Cron/Jobs/UpdateDriverDroveAtReport. it show me some path, starting at application/src/.. and so on. where can I find this directory on a system?09:36
anddam osse: the symbolic ref is pointing to 'origin/master' whose full name is 'remotes/origin/master' is that right?09:38
→ interrobangd joined09:39
← mikecmpbll left09:42
→ mikecmpbll joined09:43
→ pks joined09:49
anddam different topic, I'm trying to learn a good habit09:49
I have a master ref tracking origin/master, origin receives a new commit, I fetch origin09:49
so far so good09:49
now I want to update my master in order to receive that commit in my working directory09:50
shouldn't a 'git checkout master' do that?09:50
I get Your branch is behind 'origin/master' by 1 commit, and can be fast-forwarded. (use "git pull" to update your local branch)09:50
but why do I need to pull if I already fetched?09:50
→ igemnace joined09:51
anddam from what I know pull is fetch+merge or fetch+rebase if so selected by options09:51
Seveas_ → Seveas09:51
← rnmhdn left09:52
oxymoron93 well merge is indeed needed if you want your local master to have those changes09:52
anddam so I should not do a checkout but a merge09:52
oxymoron93 as described, git fetch <remote> only updates refs/remotes/<remote>/<branches> which are remote tracking branches, simple local master stays the same09:52
anddam this is counterintuitive for me, I want to move my working directory to a different ref09:53
and I already fetched that ref09:53
in fact I'm here just after the fetch, git log -1 shows * e4c19cd - (HEAD -> master) Update tasks09:54
oxymoron93 in that case sure, `git checkout origin/master` would suffice anddam09:54
but not git checkout master, as master isn't updated09:54
osse anddam: yes09:54
well, refs/...09:54
anddam while the last commit is * 633cd8d - (origin/master, origin/HEAD) Whitespace change (5 minutes ago)09:54
I did not expect log output to be contextual, I thought it was an absolute concept in a repo09:55
i.e. log has tho show "the one history to rule them all"09:55
oxymoron93: checking out origin/master put me in a detached state09:56
that was not desired09:56
oxymoron93 yes it does09:56
← xcm left09:56
oxymoron93 in that case go back to master and run `git merge origin/master`09:56
anddam but not desired09:56
I ended up doing pull, that is the merge,09:56
but I really don't get the hang of it09:56
sorry if I seem rambling, but it's the frustration09:57
→ xcm joined09:58
osse anddam: Git is distributed, thus there is no log to rule them all.09:58
→ blackmesa1 joined10:00
→ venmx joined10:01
anddam osse: well I can clearly see it at "the set of all commits this repository knows"10:01
ansyeb how to show content of a commited file in git?10:01
anddam in fact git log -1 origin/master shows the new commit10:02
osse: it's not the distributed parts the bothers me, it's something about the relationships between repo, branches and tracking branches10:03
osse: I have no issues at all with mercurial, since a few years actually. But everybody and their brother is using github/gitlab nowadays10:03
and frankly I can see why, bitbucket's UX is horrid in comparison10:04
thanks for the info guys10:05
→ irqq joined10:05
→ ngui joined10:05
osse anddam: !remote_tr10:06
gitinfo anddam: [!remote_tracking_branch] Remote-tracking branches (branches which start with e.g. 'origin/', listed by 'git branch -r') are read-only mirrors of the branches in another repository. They're updated by 'git fetch'. You can't edit them directly (trying to check them out results in a !detached HEAD), but you can create a local branch based on a remote-tracking branch using e.g. 'git checkout -b <branch> <remote>/<branch>10:06
osse Maybe it's this part that's either missing or making it worse? :P10:06
→ pfleck joined10:07
← blackmesa1 left10:08
oxymoron93 ansyeb: is this what you want? `git show <commit>:<path_to_the_file>`10:10
← n3wborn left10:10
ansyeb I dont yet understand where files are stored on a server yet even ;/10:11
or theyre hidden under hash-names and I can never find them with expected nam ein FS10:11
← stitched_unicorn left10:13
canton7 ansyeb, they're stored in .git/objects, under hash-names (along with commits, tags, records of directories, etc)10:13
ansyeb, !bottomup has a good description (as do other books)10:13
gitinfo ansyeb: 'Git from the bottom up' starts with explaining the building blocks of git and proceeds to tell you how they fit together. https://jwiegley.github.io/git-from-the-bottom-up/10:13
ansyeb git show <commit>:<path_to_the_file> was indeed what I am loking for, ty10:16
and for gitlab its repo/objects?10:17
← catsup left10:18
← kjartan left10:18
← exmate left10:18
→ catsup joined10:18
canton7 gitlab uses git10:18
← catsup left10:19
→ catsup joined10:20
→ styler2go joined10:20
← Kaisyu left10:21
→ kjartan joined10:23
← ngui left10:25
→ ngui joined10:26
→ goodafternoon joined10:26
→ TJ- joined10:26
← cd left10:27
← keepLearning512 left10:27
← pfleck left10:27
→ keepLearning512 joined10:30
→ vavkamil joined10:31
← hussam left10:31
→ xmate joined10:34
← pR0Ps left10:36
→ pR0Ps joined10:36
← pks left10:40
→ pks_ joined10:40
→ cur8or joined10:41
pks_ → pks10:41
ansyeb anyways, got the whole issue solved. thanks to you. finding the source revealed the query, and executing it manually made the issue clear10:41
→ pfleck joined10:41
← keepLearning512 left10:41
← Dirak left10:42
mar77i_ → mar77i10:44
→ blackmesa1 joined10:49
← dege left10:50
← hofmann3900 left10:52
← goodafternoon left11:03
← esrse left11:04
← xmate left11:05
→ HZun joined11:08
shentino_ → shentino11:08
← ngui left11:10
→ ngui joined11:10
← pfleck left11:12
→ peacememories joined11:18
← peacememories left11:22
→ aw1 joined11:25
→ courrier joined11:27
← TJ- left11:28
← xcm left11:29
← bsaboia left11:29
→ xcm joined11:29
← ngui left11:30
→ pfleck joined11:31
→ keepLearning512 joined11:33
← blackmesa1 left11:33
← cur8or left11:35
→ mooasaurus joined11:37
← igemnace left11:39
→ ngui joined11:40
← pfleck left11:42
→ jungsubk joined11:42
← ngui left11:45
← pks left11:49
→ ngui joined11:50
→ blackntan joined11:55
← ngui left12:00
→ ngui joined12:05
← ngui left12:10
→ ngui joined12:11
→ pks joined12:14
← ngui left12:15
→ ngui joined12:15
← jungsubk left12:17
← alyptik left12:20
← ngui left12:20
→ ngui joined12:21
Luyin if I have two files with the exact same contents, and both are in different git repositories, should I be able to merge something in in repo A, make "git diff HEAD~ > foo.patch" and git apply this patch in repo B with "git apply foo.patch"?12:22
→ sgn_ joined12:22
→ Munt joined12:23
→ apoos_maximus joined12:24
← ngui left12:25
→ ngui joined12:25
→ blackmesa1 joined12:26
← z8z left12:30
← jottr_ left12:33
← ngui left12:40
→ ngui joined12:41
← blackmesa1 left12:42
→ pfleck joined12:42
← flying_sausages left12:43
← ngui left12:45
→ ngui joined12:45
← brandonkal left12:46
→ flying_sausages joined12:48
← ngui left12:50
→ ngui joined12:51
→ gxt joined12:52
← courrier left12:54
← ngui left12:55
→ ngui joined12:55
← rsrx left13:05
← topdownjimmy left13:05
→ topdownjimmy joined13:06
→ DolphinDream joined13:10
← pks left13:11
← ngui left13:15
← blackntan left13:16
← Makaveli7 left13:17
← DolphinDream left13:20
→ ngui joined13:20
→ jottr_ joined13:20
→ wildermind joined13:22
→ xrexeon joined13:25
← xrexeon left13:25
← ngui left13:25
→ xrexeon joined13:25
→ fission6 joined13:26
→ ngui joined13:26
→ hussam joined13:29
← ngui left13:30
→ ngui joined13:30
← jottr_ left13:33
→ alyptik joined13:33
→ courrier joined13:34
← ngui left13:35
→ ngui joined13:36
← wildlander left13:36
← yates_home left13:36
→ dege joined13:37
← ngui left13:40
→ ngui joined13:40
← ngui left13:45
→ xmate joined13:48
Luyin hi there, I have a txt file in UTF-16LE encoding (on windows). I want to place it under git version control, so I tried to put "working-tree-encoding=UTF-16LE" into .gitattributes. now when I try to "git add", I get the following error: https://paste.xinu.at/YOLf/ anyone got any idea how to resolve this? I can't change the encoding to UTF-16 (BE), because I don't know what the implications might be for my13:49
project. this txt file needs to remain UTF-16LE13:49
→ ngui joined13:50
→ wildlander joined13:52
→ Essadon joined13:56
→ pks joined13:56
← Essadon left13:57
canton7 Luyin, the man page says that UTF-16 files with BOM must use "UTF-16", not "UTF-16BE" or "UTF-16LE". Are you sure that "UTF-16" implies UTF-16LE, even if the file has a BOM which indicates it's UTF-16BE?13:57
→ Essadon joined13:57
canton7 I guess if git's storing it as utf-8 internally, it'll lose the BOM, so it might check it back out again as UTF-16BE?13:58
Luyin I'm not sure anything is implied... I figured out the encoding information I needed with "file MAIN.txt" before I tried anything in .gitattributes13:58
if that's even answering your question :-/13:58
← ngui left14:00
→ bsanford joined14:01
← bsanford left14:02
← apoos_maximus left14:02
→ apoos_maximus joined14:03
canton7 From some cursory reading, it looks like iconv likes to write 'utf-16' in the machine's endianness, and there might not be a way to convince it otherwise14:04
← quackgyver left14:05
canton7 you're probably only using little-endian systems? So 'utf-16' is probably fine, and it'll get checked out as utf-16le with BOM14:05
otherwise, is it possible to remove the BOM?14:05
← mensvaga left14:06
← aw1 left14:08
← courrier left14:09
→ benharri joined14:10
→ lacrymology joined14:10
← fission6 left14:11
sgn_ IIRC, it's not possible to remove the BOM14:15
git requires the BOM for UTF-1614:15
mns_ → mns14:15
→ igemnace joined14:16
sgn_ the Unicode specification requires UTF-16BE for UTF-16-w/o-BOM but a lot of Windows software out there doesn't respect it14:16
← Essadon left14:16
sgn_ Linux with musl libc chose different extreme path, to not emit the BOM on UTF-16 (w or w/o BE notation, I forget the details), and there is a test case in git failed14:17
xcm → Guest2704014:18
← Guest27040 left14:18
→ xcm joined14:18
→ Essadon joined14:18
sgn_ → sgn14:18
→ Inline joined14:19
← KnoP left14:19
← rkta left14:21
→ KnoP joined14:21
→ rkta joined14:21
← pfleck left14:21
→ Puffball joined14:23
← nowhere_man left14:26
→ d^sh_ joined14:26
← d^sh left14:29
← andrzejv left14:30
→ hbautista joined14:31
← kjartan left14:31
→ th3m1s joined14:35
→ ngui joined14:35
→ kjartan joined14:36
← totte left14:38
→ czart joined14:38
← yn left14:40
→ blackmesa1 joined14:43
→ stitched_unicorn joined14:43
→ totte joined14:45
→ yn joined14:45
→ jottr_ joined14:47
← floppydh left14:48
→ scr267 joined14:50
→ jstimm joined14:50
← YuGiOhJCJ left14:50
← rafalcpp left14:52
← queip left14:52
→ fission6 joined14:54
← ngui left14:55
→ ngui joined14:56
→ codebam joined14:56
→ dpyro joined14:57
→ boombatower joined14:57
→ agowa338 joined14:58
→ yoh joined14:59
← dpyro left14:59
← sgn left14:59
→ orbyt_ joined14:59
→ dpyro joined14:59
← dpyro left15:00
→ dpyro joined15:01
→ rafalcpp joined15:02
← codebam left15:03
→ codebam joined15:04
→ queip joined15:05
← ngui left15:10
← igemnace left15:10
→ ngui joined15:10
xcm → Guest9661015:11
← Guest96610 left15:11
→ xcm joined15:11
→ cdown joined15:12
← Case_Of left15:18
→ skered joined15:18
→ pfleck joined15:19
→ Case_Of joined15:20
← ngui left15:20
→ ngui joined15:21
→ greggerz joined15:25
← ngui left15:25
→ ngui joined15:25
← tsdh left15:25
← agowa338 left15:27
→ Siecje joined15:29
← ngui left15:30
Siecje I have two commits. I want to update the first commit. (I accidentally included some debugging code). I have updated the first commit (git commit --amend).15:31
How can I set the branch to this new commit?15:31
→ Envil joined15:32
→ cdown_ joined15:33
← pfleck left15:35
Siecje This is what I have done. https://dpaste.de/2drd15:35
→ ngui joined15:35
← stitched_unicorn left15:36
← cdown left15:36
cdown_ → cdown15:36
→ jnewt joined15:38
← interrobangd left15:38
→ fattredd joined15:40
← inkbottle left15:40
← ngui left15:40
→ ngui joined15:41
fattredd Hey _ikke_, I wanted to thank you for helping me out yesterday. I had been struggling for hours and I really appretiate your assistance15:41
→ olabaz joined15:41
← dege left15:42
→ cdown_ joined15:42
olabaz Hi, I am new to working with git. I have a local directory, one on a server I ssh into and, a git I formed in said server. I normally write files locally and send them to server where I modify to a final version and then I push those changes to git. I now want to make my local folder have the files in git. What is the best way to do this?15:44
also, my git is hosted on github*15:44
← cdown left15:44
cdown_ → cdown15:44
olabaz I hope this question makes sense.15:45
← thebope left15:45
← ngui left15:45
← akushner left15:45
→ thebope joined15:45
→ ngui joined15:45
fattredd To clarify, you're saying that you edit your code on your computer as well as a personal server. Then you want to have github host your code. Is that right?15:47
← HZun left15:48
olabaz fattredd: I edit my code on my computer and then I modify it on server to get things working better. That directory on the server is already linked to my github account15:49
now I just want the final changes to be reflected on my local computer15:49
I can just download the files I changed but I would like to have that handled through git if possible15:49
fattredd Okay. You can install git on your computer too. Then you just need to clone the repository to your computer15:50
← cdown left15:51
olabaz ok, is there a way to ignore a file15:51
fattredd This is exactly what git is good at. Are you using command line, or a visual interface for git?15:51
olabaz commandline15:51
Luyin olabaz: man gitignore :)15:51
gitinfo olabaz: the gitignore manpage is available at https://gitirc.eu/gitignore.html15:51
Luyin olabaz: perhaps reading Pro Git (at least the first few chapters) will help you: https://git-scm.com/book/en/v2/Getting-Started-About-Version-Control15:52
olabaz Luyin: ok I'll take a look. thanks15:52
Luyin np. it's a really good read15:53
and it explains a lot of things a lot better than the manpages definitely :D15:53
→ inkbottle joined15:55
← apoos_maximus left15:55
fattredd For sure. Highly recomended.15:55
← sQVe left15:56
→ pfleck joined15:58
→ thiago joined15:59
← oxymoron93 left16:00
→ TJ- joined16:00
← ngui left16:00
← tureba left16:02
→ ngui joined16:05
→ aw1 joined16:07
← aw1 left16:07
← treehug88 left16:08
→ aw1 joined16:08
→ al-damiri joined16:09
→ Toadisattva joined16:15
← sanscoeur left16:15
← ngui left16:15
Siecje How do you update the second last commit?16:16
→ tureba joined16:16
benharri as in edit it and rewrite the history?16:17
man git-rebase16:17
gitinfo the git-rebase manpage is available at https://gitirc.eu/git-rebase.html16:17
benharri probably something like git rebase -i HEAD~216:17
← chele left16:20
→ durham joined16:20
← keepLearning512 left16:21
← Xiti left16:22
← blackmesa1 left16:23
← durham left16:23
→ thebetrayer joined16:23
Siecje I created a new branch. I moved the branch back to the commit I wanted to update. I updated it.16:24
Now I have three commits in the new branch. How do I rebase the most recent two onto the newly ammended commit?16:24
→ aw1_ joined16:25
→ ngui joined16:25
← codebam left16:26
← aw1 left16:26
→ Xiti joined16:27
← ericboehs left16:27
Siecje This is what my repo looks like https://dpaste.de/L68J16:27
← wadadli left16:28
← pfleck left16:29
→ wadadli joined16:30
→ codebam joined16:30
→ ericboehs joined16:30
← thebetrayer left16:30
← ngui left16:30
→ pfleck joined16:30
← rts-sander left16:31
→ thebetrayer joined16:31
→ Zexaron joined16:31
→ apoos_maximus joined16:32
apoos_maximus is it possible to create a new repository on github from the command line ?16:34
← thiago left16:34
benharri Siecje: look into man git-cherry-pick if you just want to grab two commits16:34
gitinfo Siecje: the git-cherry-pick manpage is available at https://gitirc.eu/git-cherry-pick.html16:34
benharri apoos_maximus: not without an external tool16:35
apoos_maximus like ?16:35
→ ngui joined16:35
← thebetrayer left16:35
Siecje benharri: I did git rebase --onto feature HEAD~216:36
benharri apoos_maximus: i'm not sure; i haven't used any that do that, but i think github offers a tool called hub16:36
apoos_maximus alright ill look into it16:37
→ thebetrayer joined16:37
apoos_maximus so as of now all we could do to publish our local reps is by creating one manually on github.com and pulling from it then pushing to it16:38
_ikke_ You don't need to pull from that repo if it's a new empty repo16:39
→ mat001 joined16:39
apoos_maximus yeah imean just in case we put a readme16:39
_ikke_ You could do it, but note that you would have 2 root commits16:40
not that anything is wrong with that, could just be unexpected16:40
Luyin I have that in one or two of my repos /o\16:41
well, I guess you learn with time and error16:41
apoos_maximus yeah ! |_o_|16:42
thanks _ikke_ Luyin16:43
Luyin didn't do anything :D16:43
← mat001 left16:44
apoos_maximus well it felt good to not be the only one to have been thinking that ... !16:44
Luyin ha, ok :) glad to "help" with that :)16:45
apoos_maximus :)16:45
← ngui left16:45
olabaz Hey, if git is installed on a school server that I ssh into. Can I still do git --global user.name for my repositories in /home/me?16:47
← pfleck left16:47
Luyin --global will use the ~/.gitconfig file to read/write16:47
olabaz ok good thanks16:48
← codebam left16:49
← jstimm left16:49
_ikke_ olabaz: there is also --system, that uses /etc/gitconfig fyi16:50
← xrexeon left16:50
→ ngui joined16:50
olabaz _ikke_: ah ok don't want that one. Good to know thanks16:51
← Siecje left16:52
→ AtumT joined16:53
→ jamiejackson joined16:55
jamiejackson happy friday! we're planning an svn-to-git migration, and i need tips on what to do about converting our comprehensive svn config file over to .gitattributes: svn config: https://gist.github.com/jamiejackson/8c76439019aa177a0a271696afa99ff516:57
← thebope left16:58
→ thebope joined16:59
→ JeffH joined16:59
jamiejackson does it make sense to just convert every svn auto prop line to the .gitattribute equivalent, or are there things in there that are extraneous/problematic? i'd be grateful for any advice from git veterans.17:00
→ Admish joined17:02
→ mat001 joined17:02
← thebope left17:03
← inkbottle left17:06
← alyptik left17:08
canton7 jamiejackson, well, .gitignore for the ignores, .gitattributes for the line endings stuff17:08
← JeffH left17:08
canton7 the mime type stuff doesn't have an equivalent in git as far as I know -- not even sure what it's used for in svn17:09
→ akushner joined17:09
canton7 the eol stuff has equivalents in .giattributes17:09
← mikecmpbll left17:10
← olabaz left17:10
→ causasui joined17:12
→ blackpharaoh joined17:13
→ TheRuckus joined17:13
← Arcaelyx left17:14
← mat001 left17:15
← ngui left17:15
jamiejackson canton7, okay, focusing on the eol end executable settings: is there anything in there that doesn't make sense to just convert to the .gitattributes equivalent? should i just convert the syntax of those over to .gitattributes, line by line, or are there changes that would make more sense for git?17:17
→ thiago joined17:19
→ Arcaelyx joined17:19
← fission6 left17:19
→ \void joined17:20
→ ngui joined17:20
→ mat001 joined17:20
← thebetrayer left17:20
→ thebetrayer joined17:21
→ prakashdanish joined17:21
← prakashdanish left17:22
→ mat001_ joined17:23
← thebetrayer left17:24
→ prakashdanish joined17:24
← prakashdanish left17:24
→ prakashdanish joined17:25
← acidjnk left17:25
← ngui left17:25
← mat001 left17:25
→ ngui joined17:26
→ thebetrayer joined17:26
← qqx left17:26
→ qqx joined17:26
→ max12345 joined17:27
max12345 hello, I feel like I haven't really understood submodules and how to use them. Specifically how to update them properly. I have found that particular stackoverflow answer that explains it, but I'm not sure it's working or if I maybe misunderstand submodules.17:28
← tombusby left17:29
← ngui left17:30
→ ngui joined17:30
← blackpharaoh left17:32
← T_UNIX left17:34
→ m0viefreak joined17:35
← akushner left17:36
→ tombusby joined17:36
→ akushner joined17:37
← max12345 left17:37
_ikke_ moritz: Can you explain what you are confused about?17:40
← akushner left17:41
→ Regon joined17:41
jamiejackson hmm, let me ask this a different way, now that i just read about `.gitattributes`'s `* text=auto`. do you suppose that will cover what i had manually specified in my svn config? https://gist.github.com/jamiejackson/8c76439019aa177a0a271696afa99ff517:43
_ikke_ jamiejackson: what do you expect to happen? (I'm not that familiar with svn)17:44
→ Brilpikk3wyn joined17:46
grawity looks like automatic CRLF conversion on checkout, mostly17:46
Brilpikk3wyn → Pikk3wyn17:46
← prakashdanish left17:48
← Toadisattva left17:48
→ bn_work joined17:49
jamiejackson _ikke_, svn treats everything as binary, and merges/diffs when there are windows/linux/osx working copies can be problematic, so it's best to tell svn when a file is text and svn converts to the native EOL for any given checkout17:50
← Jackneilll left17:50
← ngui left17:50
→ ngui joined17:51
_ikke_ I would not do * text=auto17:51
well, auto might not be that bad17:51
→ mat001 joined17:51
_ikke_ should suffice17:51
→ Jackneill joined17:52
→ mikecmpbll joined17:52
→ bmbouter joined17:52
← Arcaelyx left17:52
bmbouter anyone have an idea of how a project could make all of its developers have a git command that adds some syntax to the commit message?17:52
← jamiejackson left17:54
← mat001_ left17:54
_ikke_ bmbouter: there is no way to get that automatically17:55
→ crissae joined17:55
_ikke_ there is always user interaction necessary17:55
← ngui left17:55
→ cdown joined17:55
gitinfo set mode: +v17:55
bmbouter that is ok this just let's us know we can accept that solution as our best option17:55
→ ngui joined17:55
→ sanscoeur joined17:56
_ikke_ this command could be an alias17:56
→ akushner joined17:57
→ fission6 joined17:57
→ mat001_ joined17:58
→ rsrx joined17:59
→ chingao joined17:59
→ p0s1x joined17:59
→ jamiejackson joined18:00
← ngui left18:00
← p0s1x left18:00
jamiejackson got disconnected _ikke_, but i looked at channel logs, and i think your conclusion was that `* text=auto` is fine?18:01
← mat001 left18:01
→ treehug88 joined18:02
← thiago left18:02
← apoos_maximus left18:03
→ apoos_maximus joined18:03
_ikke_ yes18:05
← cdown left18:05
→ ngui joined18:05
_ikke_ * text would be bad, but text=auto lets git decide itself whether it's binary or not.18:05
You could of course make sure known extensions are always treated as binary18:05
→ n3wborn joined18:06
→ cdown joined18:06
up|ime bmbouter: you might be looking for .git/hooks/prepare-commit-msg18:07
← chingao left18:07
← apoos_maximus left18:08
up|ime bmbouter: what I did once is with our particular branching model you could extract the ticket ID and put it in the commit msg18:08
→ apoos_maximus joined18:09
→ keepLearning512 joined18:09
→ inkbottle joined18:10
← keepLearning512 left18:14
← ngui left18:15
← fission6 left18:16
→ ngui joined18:16
→ Arcaelyx joined18:16
→ moei joined18:19
← ngui left18:20
→ ngui joined18:20
← tombusby left18:20
bmbouter up|ime: ty I think we will use prepare-commit-msg18:21
the branching idea makes sense too18:21
ty18:21
→ tombusby joined18:21
← irqq left18:21
→ jstimm joined18:23
→ fission6 joined18:24
← sauvin left18:25
← ngui left18:25
→ ngui joined18:26
← ngui left18:30
→ ngui joined18:30
→ alyptik joined18:31
← jstimm left18:31
← gxt left18:33
← ngui left18:35
→ ngui joined18:36
← jelmer left18:37
→ jelmer joined18:37
→ aw1__ joined18:38
← ngui left18:40
← aw1_ left18:40
→ ngui joined18:40
← aw1__ left18:41
← fission6 left18:42
→ thiago joined18:44
← kapil____ left18:44
← kjartan left18:44
← iam730 left18:46
→ iam730 joined18:47
→ Geezus42 joined18:47
← iam730 left18:49
→ kjartan joined18:49
→ kenlee joined18:50
→ MonoGreyMatter joined18:50
← MonoGreyMatter left18:50
→ MonoGreyMatter joined18:51
→ ali1234 joined18:52
← rsrx left18:53
→ folkrav joined18:53
→ xrexeon joined18:58
← xrexeon left18:59
→ xrexeon joined18:59
← xrexeon left18:59
→ xrexeon joined19:00
← ngui left19:00
→ {HD} joined19:00
→ sanscoeu_ joined19:01
← sanscoeur left19:04
→ fission6 joined19:04
← jamiejackson left19:07
→ freeman42x joined19:07
← Zexaron left19:08
→ irqq joined19:10
← irqq left19:11
→ Sasazuka joined19:11
→ irqq joined19:11
← wildermind left19:12
← jwest left19:12
← venmx left19:13
← mat001_ left19:13
← inkbottle left19:13
nikivi how to list all submodules in a current git repo?19:13
that is list all dirs that have .git in them19:13
there is no .gitmodules file19:14
→ Dirak joined19:14
→ plexigras joined19:15
→ jwest joined19:15
_ikke_ then they are not officially submodules19:15
git ls-tree HEAD | grep '16000 commit'19:16
you need to add one 019:16
But that only works if they have been added to the parent repo19:17
← akushner left19:17
_ikke_ If they haven't you have to use something like find .git19:17
find . -name .git19:17
qqx That find command would also list "submodules" of "submodules"19:18
→ cur8or joined19:18
_ikke_ ie, works recursively19:19
→ ngui joined19:20
→ akushner joined19:22
← freeman42x left19:29
→ freeman42x joined19:29
→ pfleck joined19:31
→ codebam joined19:31
← alyptik left19:32
← ngui left19:35
→ ngui joined19:36
→ dege joined19:36
← cswl left19:38
→ snowgoggles joined19:38
anddam osse: it kinda helps19:38
← anddam left19:40
← bebbet left19:41
→ bebbet joined19:42
→ mat001 joined19:46
→ Darcidride_ joined19:47
← ngui left19:50
→ ngui joined19:50
← Envil left19:51
← ngui left19:55
→ thiago_ joined19:59
← thiago left19:59
→ ngui joined20:00
← dpyro left20:00
→ dpyro joined20:01
→ cd joined20:04
→ gxt joined20:06
→ Makaveli7 joined20:08
← sanscoeu_ left20:08
→ acebrianjuan joined20:09
→ friendofafriend joined20:13
friendofafriend Howdy, I'd like to find just the differences between a repo and its fork. Could someone refer me to an example of this?20:15
← ngui left20:15
← Dirak left20:18
→ alyptik joined20:18
← hbautista left20:20
→ ngui joined20:20
← codebam left20:21
parsnip friendofafriend: add both repos as remotes and browse history locally.20:22
→ iam730 joined20:23
← mobile_c left20:24
friendofafriend parsnip: Is there some way to see what codebase a fork actually used?20:24
I've got this repo with patches to the Linux kernel, and I'm just trying to generate a patchset I could (try to) apply to a newer kernel.20:25
← th3m1s left20:27
→ horribleprogram joined20:27
parsnip do the parent hashes not exist in the public linux kernel?20:28
→ GoGi joined20:29
← Pikk3wyn left20:29
GoGi How can I see in the history of which refs (if any) a certain commit is contained?20:30
→ codebam joined20:30
← troyt left20:30
→ troyt joined20:30
← horribleprogram left20:30
acebrianjuan Hi all20:31
friendofafriend parsnip: I really don't know anything about parent hashes or otherwise. Are changes from upstream not flagged in a fork?20:31
→ horribleprogram joined20:31
acebrianjuan I've got a question about sharing patches and applying them20:31
friendofafriend I'd actually have to just go hash checking, or something?20:32
acebrianjuan Please, have a look at this issue: https://github.com/gnss-sdr/gnss-sdr/issues/22920:32
You'll see that a contributor shares his patch in .txt format20:33
I'm familiar with the pull request mechanism20:33
But sharing patches is new for me20:33
I'd like to know how does working with patches work20:34
friendofafriend acebrianjuan: This is a pretty good reference. https://www.thegeekstuff.com/2014/12/patch-command-examples/20:35
acebrianjuan friendofafriend: thank you!20:36
Do patches retain authorship when they are applied/merged?20:37
friendofafriend Retain authorship? I'm not sure what you mean.20:37
acebrianjuan friendofafriend: meaning that whoever wrote the patch is kept as the author of the code20:38
in the git logs20:38
friendofafriend The patch has no interaction with git.20:39
← czart left20:39
friendofafriend So it will be as though you've made the modification yourself, like any other modification.20:39
← codebam left20:40
acebrianjuan If you check the shared .txt patch, there's a header with some metadata20:41
friendofafriend The "From:" and "Date" stuff? That's not applied to the file.20:41
acebrianjuan friendofafriend: ok, that's what I was referring20:42
← xmate left20:42
→ mobile_c joined20:42
acebrianjuan friendofafriend: but check this out: https://github.com/gnss-sdr/gnss-sdr/commit/63e90f862ffde4b65e6d12185cff433de10e5f4320:42
you'll see that GitHub states: "sergey-nik authored and carlesfernandez committed on Dec 13, 2018"20:43
friendofafriend acebrianjuan: That was done externally somehow.20:45
acebrianjuan so in some way or the other, the author is credited for his patch20:45
friendofafriend That was not accomplished by the patch.20:45
acebrianjuan ok20:45
friendofafriend: https://stackoverflow.com/questions/25327743/what-flow-causes-github-commits-that-are-authored-by-one-user-but-committed20:47
"Causing the two to point to different people can happen on rebasing, editing a commit (e.g. amending), doing a commit on behalf of someone else (e.g. by specifying --author), applying patches (git am), squashing (e.g., on merge or rebase), or cherry-picking."20:48
friendofafriend Sure, but that's not being caused by your patch.20:48
Someone is fooling around with git, after the file is patched, to cause attribution to an author.20:49
acebrianjuan friendofafriend: what about applying patches with git am?20:49
just asking20:50
e yes, git am will preserve the authorship info20:51
→ th3m1s joined20:52
acebrianjuan e: indeed, just checked that with: git help am | grep -i "author"20:52
"The commit author name is taken from the "From: " line of the message"20:53
e the signed-off-by line makes me think that's what happened with the example you linked20:54
in general, on github specifically, rebases are probably the more likely cause of author != committer20:54
acebrianjuan ok, thank you both :)20:56
← horribleprogram left20:56
friendofafriend Good luck with your project, acebrianjuan.20:56
← dilfridge left20:58
friendofafriend Is there some way to generate an archive of only files that have been changed between one repo and its fork?20:58
Or better yet, a patchset?20:58
→ dilfridge joined20:59
→ Dirak joined21:00
← Darcidride_ left21:03
← th3m1s left21:04
→ sanscoeur joined21:04
→ th3m1s joined21:04
→ codebam joined21:05
← apoos_maximus left21:05
→ Sasazuka__ joined21:08
← th3m1s left21:10
← mobile_c left21:11
← pfleck left21:11
← fission6 left21:11
← Sasazuka left21:11
→ fission6 joined21:16
→ th3m1s joined21:17
← th3m1s left21:22
→ th3m1s joined21:22
← th3m1s left21:22
→ z1haze joined21:24
z1haze Hello! I have a question about --force destructive pushes.. i understand the it can rewrite history, etc.. but the commits are never really 'gone' right? they are just orphaned.. so if someone rebased onto a branch then force pushes,, and then a week later they want to reset back to the last original commit to that branch before the rebase.. they could in theory couldnt they?21:25
koala_man will such orphaned commits disappear in a gc?21:26
z1haze i mean, i guess you could also just reset the branch, and just cherry pick the actual branch commits right21:26
just tryin to go about this disaster in the 'best' way21:27
_ikke_ orphaned objects will be deleted after about 3 months (the default reflog expire time)21:27
→ th3m1s joined21:31
← fattredd left21:31
← n3wborn left21:33
→ neilthereildeil joined21:36
→ ferdna joined21:36
→ pfleck joined21:36
← Trieste left21:37
neilthereildeil hi i just checked out an old commit and git committed something on top of it. and then i created a new branch with "git branch BranchName HEAD". I saw the branch was created. now when i commit something on this same branch, why does the commit go ahead of the branch? the branch should be updated to point to the tip, right?21:38
← cdown left21:38
← Dirak left21:38
→ Trieste joined21:39
→ ozcanesen joined21:40
_ikke_ neilthereildeil: you need to check out the branch first21:41
just creating it is not enough21:41
← Phylock left21:44
← ngui left21:45
→ Phylock joined21:46
← rullie left21:47
→ Dirak joined21:48
→ ngui joined21:50
→ hofmann3900 joined21:50
← th3m1s left21:53
edman007 Hey, so I'm trying to find the commit that deleted a particular variable, so when I do git blame --reverse v3.4, when I know the line existed in v3.4, it shows a commit 'adc1ef1e37358' as the one I want, but doing git diff adc1ef1e37358~ adc1ef1e37358 doesn't appear to show that line changing...22:00
→ hbautista joined22:00
edman007 so when a reverse git blame shows a commit, how do I view that commit?22:00
← Sasazuka__ left22:01
← dpyro left22:01
→ dpyro joined22:02
neilthereildeil _ikke_: yea i did check out that branch22:02
→ Sasazuka joined22:02
neilthereildeil _ikke_: ok, i shouldnt git checkout the commit or head? i have to say the branch name specifcaly?22:03
hmm i see that i did that...22:03
im looking in bash histoy22:03
← plexigras left22:04
neilthereildeil git checkout refs/heads/branchName; make changes; git add changes; git commit22:04
what am i doing wrong?22:05
← dege left22:10
← codebam left22:10
← fission6 left22:12
← neilthereildeil left22:15
→ qassim joined22:16
→ libertyprime joined22:17
← xrexeon left22:18
→ dre_ joined22:19
→ fission6 joined22:20
→ blackmesa joined22:21
→ dege joined22:23
← ngui left22:25
→ ngui joined22:26
← SpeakerToMeat left22:28
→ Toadisattva joined22:29
← friendofafriend left22:29
← ngui left22:30
→ ngui joined22:30
← blackmesa left22:32
→ klotzhauer joined22:33
← ngui left22:35
← orbyt_ left22:35
→ ngui joined22:36
← ngui left22:40
→ ngui joined22:40
← scr267 left22:41
→ brandonkal joined22:45
← rkta left22:45
→ rkta joined22:46
← MonoGreyMatter left22:50
← fission6 left22:50
→ codebam joined22:51
← pfleck left22:54
← skered left22:56
← foul_owl left22:56
← kjartan left22:57
→ fission6 joined22:58
← mat001 left22:58
← greggerz left23:00
→ mat001 joined23:02
→ kjartan joined23:02
z1haze i have 2 branches that have the same parent commit.. what determines if i can merge 1 branch into another?23:04
← tombusby left23:04
z1haze its not giving me an option to merge the direction I want to merge, and I want to understand why that is23:04
→ tombusby joined23:04
← thiago_ left23:07
→ sQVe joined23:09
→ orbyt_ joined23:11
→ foul_owl joined23:11
← ferdna left23:13
→ skered joined23:17
← dre_ left23:17
mud z1haze: What isn't giving you the option? What happens when you try?23:19
→ Celmor joined23:20
← fission6 left23:21
Celmor I work with a git repo on 2 computers, one of them has gotten behind and I can't pull as it keeps reporting 'error: Your local changes to the following files would be overwritten by merge:' etc., how can I just "force" the pull and overwrite the old state of my local repo on that computer? git pull --force won't do it23:22
→ thejunkjon joined23:22
thejunkjon has anyone seen an issue before where git-remote-http just hangs and uses 100% cpu?23:22
mud Celmor: you just want the local state and history of that branch to exactly equal the remote one?23:23
Celmor yes23:23
thejunkjon I am using git version: git version 2.20.123:23
Celmor someone the local docs got stuff like '<<<<<<< HEAD' or '>>>>>>> 5d30c9436419b383afdc0dd2b8ba0c06a24b810d' inserted so it's different from any tracked commit23:23
← ozcanesen left23:24
mud That's not a merge (or pull). That's git fetch && git reset --hard @{u} i believe. Take a backup in case i'm wrong23:24
Celmor I have the current repo on my other computer anyway23:24
mud Those look like conflict markers. Are you in the middle of a merge?23:24
Celmor well, I tried editing the a file and commit, then deleted that commit as I noticed those docs were way behind23:25
mud You might have to git merge --abort if you are, i dunno23:25
← tombusby left23:25
Celmor fatal: There is no merge to abort (MERGE_HEAD missing).23:25
→ tombusby joined23:25
Celmor thanks 'git reset --hard @{u}' did it23:25
mud Might be a merge where someone accidentally commited the conflict markers then.23:26
Cool23:26
Celmor what does @{u} refer to?23:26
mud It's the 'upstream' branch. The one that would be pulled from23:26
Man gitrevisions should have it23:27
Hm, bot don't like me maybe23:27
Locally 'git help revisions' should show it23:28
← AtumT left23:28
Celmor No manual entry for gitrevision23:28
oh, plural23:29
→ horribleprogram joined23:29
Celmor got it, thanks23:30
mud Anytime23:30
z1haze what is the easist way to undo a rebase? create a new branch and cherry pick the original commits?23:30
mud z1haze: That won't result in exactly the original commits. The reflog would be the way to go23:31
← horribleprogram left23:32
z1haze i dont know how to do that :\23:32
← xcm left23:32
→ xcm joined23:32
mud Look for the original ref in git reflog thebranchname and use that how you like, create a new branch from it, reset the old, etc23:32
← jakogut left23:33
← causasui left23:33
← dilfridge left23:33
← orbyt_ left23:34
← irqq left23:35
→ causasui joined23:37
← libertyprime left23:47
← akushner left23:48
→ orbyt_ joined23:49
← sQVe left23:49
→ sgn joined23:50
→ victorqueiroz joined23:50
← Revan- left23:51
← Makaveli7 left23:52
← manuelschneid3r left23:54
→ sanscoeu_ joined23:56
→ Bobdude joined23:58
← mat001 left23:59
→ leeN joined23:59
← sanscoeur left23:59

Logs Search ←Prev date Next date→ Channels Documentation