Perl 6 - the future is here, just unevenly distributed

IRC log for #git, 2016-07-08

| Channels | #git index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:02 SwiftMatt joined #git
00:04 jstimm joined #git
00:08 rkazak joined #git
00:09 ochorocho__ joined #git
00:15 jstimm_ joined #git
00:17 cqi joined #git
00:18 jeffreylevesque joined #git
00:22 cyan__ joined #git
00:22 ejb joined #git
00:23 jaguarmagenta joined #git
00:23 fuchstronaut joined #git
00:24 fstd_ joined #git
00:24 mischat joined #git
00:26 allcentury joined #git
00:26 shgysk8zer0 joined #git
00:28 dsrtrck joined #git
00:28 dsrtrck hi
00:28 a3Dman joined #git
00:32 dsrtrck how can i merge these commits into one? https://ibin.co/2nIDe45TBBTM.gif I don't have the option in TortoiseGit I think because of the merge commit.
00:34 dsrtrck I don't have that option in SourceTree either, but I haven't used SourceTree to merge commits into one before.
00:38 Faylite You need to rebase and squash the commits into one
00:38 Faylite Be careful
00:41 pur3eval joined #git
00:42 xelxebar joined #git
00:43 pur3eval joined #git
00:43 romerocesar joined #git
00:44 dsrtrck thank you... should i right-click on the first commit (the last item in the list) and select to "rebase 'search-properties' onto this ..." (in TortoiseGit)?
00:46 tuxlovesyou joined #git
00:46 lugzer joined #git
00:46 hazelux joined #git
00:47 EvilPenguin joined #git
00:48 Ldip joined #git
00:49 Faylite dsrtrck, sorry, I haven't used SourceTree
00:52 Faylite Read the documentation well, re-writing history is dangerous ;)
00:55 dsrtrck ok, thank you! :)
00:56 tterrag joined #git
00:58 daniel_rose joined #git
00:59 tesuji_ joined #git
01:01 shinnya joined #git
01:02 hahuang61 joined #git
01:04 romerocesar joined #git
01:04 vigilvindex joined #git
01:08 Tim-SEM joined #git
01:09 daniel_rose joined #git
01:09 dsrtrck great! that worked! :D
01:10 Tim-SEM left #git
01:12 liquidjamm joined #git
01:14 romerocesar joined #git
01:15 Literphor joined #git
01:18 shgysk8zer0 joined #git
01:21 Lyqyd joined #git
01:23 mellernoia joined #git
01:24 romerocesar joined #git
01:25 pks joined #git
01:26 mischat joined #git
01:26 durham_ joined #git
01:31 regfop joined #git
01:36 soerenb joined #git
01:37 v0n joined #git
01:40 fracting joined #git
01:40 runlevl4 joined #git
01:46 SwiftMatt joined #git
01:46 mehola joined #git
01:46 nitrix joined #git
01:46 romerocesar joined #git
01:46 BombStrike joined #git
01:46 Literphor joined #git
01:49 HoierM joined #git
01:51 ekinmur joined #git
01:51 watersoul joined #git
01:53 stantonk joined #git
01:53 Lunatrius joined #git
01:54 johngilbrough joined #git
01:54 romerocesar joined #git
01:55 phenx joined #git
01:56 watersoul joined #git
01:57 zacts joined #git
01:58 EvilPenguin joined #git
01:58 ascarter joined #git
02:02 titanium17 joined #git
02:02 ejb joined #git
02:04 fahadash joined #git
02:04 githogori joined #git
02:05 b1tchcak_ joined #git
02:09 ochorocho__ joined #git
02:15 nettoweb joined #git
02:19 harish joined #git
02:20 kriskropd joined #git
02:23 hexagoxel joined #git
02:23 jstimm_ joined #git
02:24 jaguarmagenta joined #git
02:24 fuchstronaut joined #git
02:24 romerocesar joined #git
02:25 a216 joined #git
02:27 johngilbrough joined #git
02:27 mischat joined #git
02:31 johngilbrough Are there any git flowcharts that describe the basic interactions of two individuals sharing a single directory?
02:32 Goplat joined #git
02:33 jaguarmagenta joined #git
02:34 dreiss joined #git
02:34 romerocesar joined #git
02:35 lugzer joined #git
02:41 ojacobson It's a big empty box labelled "don't."
02:41 ojacobson If you need to share space to collaborate, do that through git's remote operations (fetch, push), not by sharing a single repository or work tree.
02:43 jstimm joined #git
02:44 johngilbrough ojacobson, That's starting to make sense.  But again, do you know of a flowchart that describes this?
02:46 chachasmooth joined #git
02:48 pokergod joined #git
02:48 Literphor joined #git
02:48 pokergod i have a merge conflict.  How do i git checkout -- file and have it discard my merge conflict
02:49 shgysk8zer0 joined #git
02:52 grassass joined #git
02:54 romerocesar joined #git
02:55 stantonk joined #git
03:02 dmto joined #git
03:03 hahuang61 joined #git
03:06 dreiss joined #git
03:11 RustyShackleford joined #git
03:12 RustyShackleford So this project is migrating to git. Because of some dumb company policies, the easiest way to manage our dependencies IMO is to add them to the project's git repo in a local maven repository
03:12 ojacobson okay..?
03:12 mehola joined #git
03:13 RustyShackleford binaries in git can slow down the repo, right?
03:13 ojacobson Not particularly.
03:13 RustyShackleford is there anything I can do to mitigate that?
03:14 ojacobson They don't compress well, and obviously 'git diff' doesn't like them much, but most operations will happily coexist with even truly large blobs. 'git status' can get slow, but that's fixable.
03:14 tlolczyk joined #git
03:14 mehola joined #git
03:16 Limes_ joined #git
03:16 RustyShackleford other thing I'm trying to figure out is how to manage branching for three separate releases that are developed simultaneously
03:16 RustyShackleford I don't think Git Flow is going to work
03:17 RustyShackleford it would mean merging a month of work into master after a release, which another team found extremely frustrating
03:18 RustyShackleford their solution was that 0.2 branches from 0.1, 0.3 branches from 0.2 and so on. No master branch
03:19 xcesariox joined #git
03:21 Literphor joined #git
03:22 treaki_ joined #git
03:24 pioiytr_ joined #git
03:25 romerocesar joined #git
03:28 MattMaker joined #git
03:28 mischat joined #git
03:30 lugzer joined #git
03:32 tlolczyk joined #git
03:33 Sasazuka joined #git
03:34 markmarkmark joined #git
03:35 Aces_Charles joined #git
03:36 chrisshattuck joined #git
03:38 Literphor joined #git
03:47 johngilbrough Newbie - when one does a "git fetch repository" - what is it expecting for the repository?  My github account name?
03:48 johngilbrough The repository name I've used thus far is "scripts", but that doesn't work here.
03:49 azerus joined #git
03:49 shachaf johngilbrough: Probably from a remote, as configured in .git/config
03:54 rewt johngilbrough, if you created the local repo by cloning from github, it's most likely named origin
03:54 dmto joined #git
03:58 xaviergmail joined #git
03:58 Geo joined #git
03:58 phanimahesh joined #git
04:00 texinwien_ joined #git
04:00 clemf joined #git
04:02 bgamari joined #git
04:04 zivester joined #git
04:05 netj joined #git
04:06 hubub joined #git
04:08 Literphor joined #git
04:09 TheIdeaMan joined #git
04:09 xcesariox joined #git
04:10 rscata joined #git
04:10 ochorocho__ joined #git
04:12 TheIdeaMan I have an idea.
04:12 TheIdeaMan and i need help rounding the edges
04:12 stantonk joined #git
04:13 nilg joined #git
04:13 t-mart_ joined #git
04:13 TheIdeaMan If anyone here is willing to help
04:13 TheIdeaMan http://boards.4chan.org/g/thread/55463995
04:15 Geo left #git
04:16 boombatower joined #git
04:17 hahuang61 joined #git
04:18 stantonk joined #git
04:21 lordjancso joined #git
04:24 hahuang61 joined #git
04:25 fuchstronaut joined #git
04:25 MattMaker joined #git
04:28 hubub left #git
04:29 shgysk8zer0 joined #git
04:30 mischat joined #git
04:30 Literphor joined #git
04:34 thiago joined #git
04:35 romerocesar joined #git
04:36 d4rklit3 joined #git
04:37 DX1 joined #git
04:38 rkazak joined #git
04:38 watabou joined #git
04:41 Macaveli joined #git
04:43 subhojit777 joined #git
04:44 aidalgol joined #git
04:46 dsantiago joined #git
04:46 _ikke_ RustyShackleford: But how does that solve the problem of the 'big merge'?
04:46 ensyde__ joined #git
04:46 RustyShackleford _ikke_, merge changes every day
04:47 RustyShackleford another team there is using it with success, or so they claim
04:47 _ikke_ RustyShackleford: Yes, which is basically not much different from git-flow
04:47 _ikke_ But the goal is to regularly merge
04:47 RustyShackleford how would you adapt three simultaenous merges
04:48 RustyShackleford er, releases is what I mean
04:48 RustyShackleford at any given time, three separate releases are being developed
04:49 _ikke_ How much do they depend on eachother?
04:49 Cabanoss- joined #git
04:49 RustyShackleford can we merge everything up to a common dev branch without a huge headache?
04:49 Literphor joined #git
04:49 _ikke_ RustyShackleford: If you do it regularly, and have a process to reduce too much work being done on the same files, then yes
04:50 RustyShackleford release candidate (which might have some last minute hotfixes, may or may not be merged upstream), next release, following release
04:50 _ikke_ right
04:51 RustyShackleford so is there something I'm missing, or does git-flow not exactly fit this kind of development cycle?
04:52 rbern_ joined #git
04:52 _ikke_ git flow is a model, you don't have to use it exactly like described
04:53 _ikke_ RustyShackleford: Is the main development always done on the releases?
04:53 RustyShackleford well yeah
04:54 RustyShackleford theres two teams who work on separate features for three sprints
04:54 RustyShackleford its staggered for some reason
04:54 clemf joined #git
04:54 RustyShackleford i dunno, wasn't my idea. I just do what they tell me :p
04:55 _ikke_ hehe
04:55 MattMaker joined #git
04:55 Andrew_K joined #git
04:55 RustyShackleford but the release is definitely central to the dev cycle. Release X will target features Y and Z if that makes sense
04:56 _ikke_ Right, but do you never have dependencies of features? Feature X in Release 2 depending on feature Y in Release 1?
04:58 stonerfish joined #git
05:00 _ikke_ Thinking about it
05:01 Corosus joined #git
05:01 _ikke_ But, perhaps focussing more in feature branches then release branches might work better
05:01 _ikke_ And merge those features in the release branches when ready
05:01 RustyShackleford well the problem with that is "the big merge" again
05:02 RustyShackleford but yeah, others peoples features do break my build all the time
05:03 MattMake_ joined #git
05:06 moriko joined #git
05:06 MattMaker joined #git
05:08 _ikke_ Well, you do need some branch (either per release, or global) that's stable
05:10 zacts joined #git
05:10 Literphor joined #git
05:11 MattMaker joined #git
05:11 Vinnie_win_t joined #git
05:11 _ikke_ TheIdeaMan: What prevents someone from 24/7 logging what's being sent?
05:11 quetz joined #git
05:11 MattMaker joined #git
05:12 _ikke_ RustyShackleford: I'm off now
05:12 RustyShackleford alright thanks
05:12 RustyShackleford to answer your question, stable would be the current release
05:13 t-mart joined #git
05:13 RustyShackleford there's a code freeze, some hotfixes occaisonally, then it becomes stable
05:14 tvw joined #git
05:14 dsdeiz joined #git
05:15 izabera hi, should this command print the tag?   git describe --always --dirty
05:16 phaleth joined #git
05:16 romerocesar joined #git
05:17 yitz left #git
05:19 romerocesar joined #git
05:21 izabera oh i have to make an annotated tag
05:22 mariuscc joined #git
05:23 pioiytr_ joined #git
05:23 freimatz joined #git
05:27 pioiytr_ joined #git
05:28 pioiytr_ joined #git
05:29 pioiytr_ joined #git
05:31 freimatz joined #git
05:32 mischat joined #git
05:34 englishstudent joined #git
05:34 NeverDie joined #git
05:35 Motoko joined #git
05:36 MattMaker joined #git
05:36 Literphor joined #git
05:37 stuh84 joined #git
05:39 romerocesar_ joined #git
05:40 stantonk joined #git
05:40 shgysk8zer0 joined #git
05:40 MattMaker joined #git
05:43 zeroed joined #git
05:43 zeroed joined #git
05:43 MattMaker joined #git
05:44 fscala joined #git
05:45 a_thakur joined #git
05:48 sangy joined #git
05:49 muthu joined #git
05:50 lindenle joined #git
05:51 Literphor joined #git
05:54 dopesong joined #git
05:56 vuoto joined #git
06:02 nnyk joined #git
06:06 zacts joined #git
06:10 R_B98 joined #git
06:11 ochorocho__ joined #git
06:14 ManikM joined #git
06:15 pioiytr_ joined #git
06:19 acetakwas joined #git
06:19 dreiss joined #git
06:19 xaviergmail joined #git
06:20 rkazak joined #git
06:21 dopesong joined #git
06:21 zeroed joined #git
06:21 zeroed joined #git
06:23 freimatz joined #git
06:23 ome joined #git
06:25 sagerdearia joined #git
06:25 hahuang61 joined #git
06:26 fuchstronaut joined #git
06:26 cd-rum joined #git
06:27 djb-irc joined #git
06:28 MattMaker joined #git
06:31 ahr3n joined #git
06:32 englishstudent joined #git
06:32 toogley joined #git
06:33 texinwien_ joined #git
06:33 mischat joined #git
06:33 Softeisbieger joined #git
06:33 zacts joined #git
06:34 oich joined #git
06:35 Sceorem joined #git
06:36 moriko joined #git
06:37 romerocesar joined #git
06:37 howdoi joined #git
06:39 shgysk8zer0_ joined #git
06:40 nnyk joined #git
06:41 Oebele joined #git
06:41 maroloccio excuse me, how do i get a path from a previous version of a tracked file?
06:42 maroloccio for example, if i want to `wc -l` across some commits
06:42 MrMojito1 joined #git
06:42 maroloccio so far, i am just using git rebase --exec.. but can i do this from a single command for a single file?
06:43 maroloccio i suppose i could <(git show commit-id path/to/resource) using the shell...
06:43 maroloccio is there a better way?
06:44 pioiytr_ joined #git
06:45 maroloccio sha1:path, of course, that's what i meant, not sha1<space>path
06:46 circ-user-A26a8 joined #git
06:50 muthu joined #git
06:52 romerocesar_ joined #git
06:57 Bengt__ joined #git
06:57 netj joined #git
06:58 Macaveli joined #git
06:59 romerocesar joined #git
07:01 cyan__ joined #git
07:02 irco joined #git
07:05 dionysus69 joined #git
07:06 romerocesar_ joined #git
07:07 ahr3n joined #git
07:09 fracting joined #git
07:11 eijk_ joined #git
07:11 pioiytr_ joined #git
07:16 vuoto joined #git
07:17 xeviox_ exit
07:17 xeviox_ left #git
07:17 TomyWork joined #git
07:18 jagob joined #git
07:18 romerocesar joined #git
07:21 lugzer joined #git
07:22 dopesong joined #git
07:25 Balliad joined #git
07:25 zeemz joined #git
07:26 DevAntoine joined #git
07:28 fees joined #git
07:30 dreiss joined #git
07:31 Balliad joined #git
07:31 n-st joined #git
07:33 m0viefreak joined #git
07:33 mkoskar joined #git
07:33 vuoto joined #git
07:33 kedare joined #git
07:33 styx_ joined #git
07:34 SimonNa joined #git
07:34 mkoskar joined #git
07:34 mischat joined #git
07:35 sanct joined #git
07:35 titanium17 joined #git
07:39 cdown joined #git
07:40 mkopriva joined #git
07:41 stantonk joined #git
07:42 dan2k3k4 joined #git
07:46 Oebele joined #git
07:46 kurkale6ka joined #git
07:48 vuoto joined #git
07:49 LiohAu joined #git
07:49 multi_io joined #git
07:49 GavinMagnus joined #git
07:51 romerocesar joined #git
07:52 lindenle joined #git
07:52 MrMojito1 left #git
07:52 ferr joined #git
07:54 MrMojito1 joined #git
07:55 FunkyAss joined #git
07:57 vuoto joined #git
07:57 aax joined #git
07:59 ilogger2 joined #git
08:02 texinwien_ joined #git
08:04 heinrich5991 joined #git
08:08 chigang_ joined #git
08:09 m0viefreak joined #git
08:11 hahuang61 joined #git
08:12 ochorocho__ joined #git
08:12 ahmedelgabri joined #git
08:12 Cata|Work joined #git
08:12 zacts joined #git
08:13 ximvan joined #git
08:16 rasmusto joined #git
08:17 axisys joined #git
08:17 ix007 joined #git
08:17 timthowtdi joined #git
08:18 bongjovi joined #git
08:19 vuoto_ joined #git
08:20 Cata|Work joined #git
08:20 bsanford joined #git
08:21 romerocesar joined #git
08:22 Beams joined #git
08:24 jmd joined #git
08:24 texinwien_ joined #git
08:24 Isla_de_Muerte joined #git
08:26 davisonio joined #git
08:27 fuchstronaut joined #git
08:28 bittin joined #git
08:28 vuoto joined #git
08:30 dan2k3k4 joined #git
08:34 wrouesnel2 joined #git
08:34 cyan__ joined #git
08:36 mischat joined #git
08:42 Cata|Work joined #git
08:42 Repox joined #git
08:42 drbean joined #git
08:43 tinanoo1 joined #git
08:45 styx_ joined #git
08:45 Cata|Work joined #git
08:46 yonatankoren joined #git
08:49 PCatinean joined #git
08:49 ifra201 joined #git
08:52 Sonderblade joined #git
08:52 SimonNaa joined #git
08:55 dwmw2_gone joined #git
08:56 fracting joined #git
08:57 MrMojito joined #git
08:57 MrMojito joined #git
08:59 Neobenedict joined #git
09:00 notebox joined #git
09:02 dan2k3k4 joined #git
09:05 svartberg joined #git
09:07 nnyk joined #git
09:09 johnsno joined #git
09:09 svartberg Hi guys, i'm pretty new to git. I have a remote repository (added with git add remote live ssh://......) which I pulled to my local dir (git pull live master). Now I made some changed in the local folder that I DID NOT commit. How Can I revert these changes now? I want to get the current state of the remote repository. "git pull live master" didnt work. The new files that arent committed yet are still there. How to get rid?
09:09 dwmw2_gone joined #git
09:10 tokkee joined #git
09:11 jrabe left #git
09:11 heftig svartberg: did you add the files?
09:11 choki joined #git
09:11 heftig (i.e. git add)
09:11 clmsy joined #git
09:12 dopesong joined #git
09:12 ochorocho__ joined #git
09:13 svartberg no didnt do that either
09:13 vuoto joined #git
09:13 svartberg @heftig
09:13 keltvek joined #git
09:13 svartberg Of course I can now delete my local dir and do git clone ssh:// .... again but there must be a smarter way :D
09:14 dsdeiz joined #git
09:14 dan2k3k4 joined #git
09:15 Chiko joined #git
09:15 vuoto joined #git
09:15 PaulCapestany joined #git
09:17 vuoto joined #git
09:18 DomeMaster joined #git
09:20 Darcidride joined #git
09:24 davisonio joined #git
09:28 GeekMan joined #git
09:30 Shashikant86 joined #git
09:31 sveinse joined #git
09:31 davisonio joined #git
09:34 nettoweb joined #git
09:34 sveinse I have an old git repo that shall be migrated to github. However it contains a lot of commits and data using the wrong email, but I'd like to transfer it with history. Is it possible to alter the contents of the commits similar to filter-branch? (no other users for the repo)
09:34 jimi_ joined #git
09:36 t0by joined #git
09:39 toogley joined #git
09:40 boxmein joined #git
09:40 boxmein hiyo, bit of a logical question
09:41 heftig svartberg: git reset --hard live/master; git clean -xfd will reset your working dir *and your master branch* to the state of the live's master
09:41 boxmein I have a master branch of (C1 -> C2 -> C3 -> C4), with one branch C4-> C5, if I base another branch off of it ie, C5 -> C6, what happens if C5 gets merged to master? does C6 need some complex dance to be merged?
09:41 cdown joined #git
09:41 heftig boxmein: no
09:42 boxmein heftig, thanks. :p
09:42 boxmein have a grand day
09:42 boxmein left #git
09:42 heftig in fact, for this example there shouldn't be any merging happening at all; you get a linear history from C1->C6 in master
09:43 sanketdg joined #git
09:44 yqt joined #git
09:44 stantonk joined #git
09:49 svartberg @heftig: thx will test it
09:50 a_thakur joined #git
09:52 dopesong joined #git
09:53 Zamerick joined #git
09:55 muelli joined #git
09:55 allcentury joined #git
09:55 MrWoohoo joined #git
09:57 muelli hey hey.  I'm trying to "git am" an email that was sent with Evolution.  Evolution, it seems, uses a NO-BREAK SPACE (0xC20A) to prevent line breaks. Now that doesn't seem to please git am, because it complains about a corrupted patch.  Is there any way to make it arrive in 2003 and accept UTF-8 encoded spaces?
09:57 fscala joined #git
09:59 Ezzy joined #git
10:00 neogineercom joined #git
10:02 DieguezZ joined #git
10:07 cyberlard joined #git
10:08 jiangfb joined #git
10:09 mkoskar joined #git
10:10 fuchstronaut joined #git
10:12 Silenced joined #git
10:12 SimonNa joined #git
10:12 hahuang61 joined #git
10:12 Silenced joined #git
10:13 mischat_ joined #git
10:13 jiangfb left #git
10:13 abhvl joined #git
10:13 zacts joined #git
10:14 acetakwas joined #git
10:18 moldy left #git
10:19 notebox joined #git
10:19 Xnu17 joined #git
10:21 j416 muelli: try to mangle the patch to make it correct, maybe. It is what I would do. First I would ask sender to re-send patch properly
10:21 nnyk joined #git
10:22 jim_carry joined #git
10:22 davisonio joined #git
10:24 grawity sed $'s/\xc2\x0a/ /g' < foo.patch | git am
10:24 muelli grawity: yeah, that works... .
10:25 muelli grawity: But I consider that a bug in the diff format.
10:25 grawity patch/apply/am don't work with sentences and paragraphs, but with lines of code; so if it's not a space, then it's not a space
10:26 j416 muelli: git did not produce a faulty diff, your e-mail client did, you said
10:26 j416 muelli: can use git send-email if there is trouble
10:26 muelli yes grawity. And U+00A0 is Character.isSpaceChar().
10:27 grawity the question isn't whether it's a "space character"
10:27 grawity if it's not the exact same byte as in the original file, then it isn't
10:27 muelli j416: sure. but then I'd need install an MTA or leak my credentials to my harddisk.
10:28 muelli grawity: sure. I get that. But the diff doesn't even get to the point to compare bytes in the files.
10:28 circ-user-A26a8 joined #git
10:28 j416 mues
10:28 muelli I'd expect to handle that via --ignore-whitespace.
10:28 j416 bah
10:28 Darcidride_ joined #git
10:28 j416 muelli: use gmail?
10:28 j416 (I keep reading your nick as müesli, sorry)
10:28 j416 haven't slept for 24 hours
10:29 _ikke_ j416: Time to get some sleep then
10:29 j416 heh yes, soon
10:29 j416 not bedtime yet
10:29 _ikke_ Still in CEST timezone?
10:30 j416 japanese timezone
10:30 _ikke_ ah ok
10:30 _ikke_ +9
10:30 j416 yes
10:31 _ikke_ Then it should be bedtime soon :P
10:31 j416 it's 19.30
10:31 _ikke_ Sounds about bedtime for someone who hasn't slept for 24h ;-)
10:31 j416 but yes. Will probably doze off within the hour
10:32 j416 :)
10:35 thebloggu joined #git
10:36 tango_ a nice cup of muesli right before bedtime sounds nice
10:37 tenc joined #git
10:38 muelli hrm. is there a reference for the diff format?  I wonder whether they demand a "space" or 0x20.
10:41 thebloggu joined #git
10:41 skylite joined #git
10:41 nnyk joined #git
10:41 a3Dman joined #git
10:42 Literphor joined #git
10:43 sdothum joined #git
10:43 dopesong_ joined #git
10:43 davisonio joined #git
10:43 a3Dman joined #git
10:45 muthu joined #git
10:45 davisonio joined #git
10:46 fuchstronaut joined #git
10:48 Shashikant86 joined #git
10:51 corebits joined #git
10:51 cyan__ joined #git
10:51 jim_carry joined #git
10:51 netcarver joined #git
10:53 romerocesar joined #git
10:53 jatin30 joined #git
10:54 lindenle joined #git
10:55 yqt joined #git
10:55 jatin30 Hi I was trying to install GITWEB on my system and I got stuck here https://bpaste.net/show/a0b0a043e690 when I googled the error It said I need to have a paid Redhat subscription to enable the repos. Can it be resolved without having the paid subscription?
10:57 King_Hual joined #git
10:57 King_Hual joined #git
10:58 dangerousdave joined #git
10:58 moriko joined #git
10:58 lrvick joined #git
10:59 rnsanchez joined #git
11:01 fuchstronaut joined #git
11:02 TomyWork I'm trying to rebase one disconnected set of commits with its own root commit onto another
11:02 TomyWork "git rebase master disconnected" works but it throws away all the merges
11:03 TomyWork "git rebase --preserve-merges master disconnected" throws away everything and i end up with "disconnected" being empty to "master"
11:03 TomyWork "git rebase --preserve-merges master disconnected" throws away everything and i end up with "disconnected" being identical to "master"
11:03 Jakey3 joined #git
11:04 TomyWork shouldnt this work?
11:04 abhvl left #git
11:04 TomyWork i'll use filter-branch grafting now, but at first glance this looks like a bug
11:05 acetakwas joined #git
11:05 Jakey3 when I do a git clone will it clone the none master branches as well?
11:06 jatin30 Jakey3, yes It will clone everything in the remote repository including all the branches
11:06 Jakey3 thanks
11:06 jatin30 no problem :)
11:07 Silenced joined #git
11:07 Silenced joined #git
11:08 TomyWork Jakey3 after you clone, take a look at "gitk --all&"
11:08 ramsub07 joined #git
11:08 CheckDavid joined #git
11:09 Jakey3 sure, the thing is when I did a git clone on my github I have files in my gh-pages branch with a test index file in there I cant see on my local computer?
11:09 yanome joined #git
11:10 pioiytr_ joined #git
11:10 TomyWork i suppose you have to check out that branch before you can see it
11:10 Shashikant86 joined #git
11:10 TomyWork like i said, try "gitk --all&"
11:10 johnmilton joined #git
11:10 TomyWork you can take a look at all the branches, which commit they're pointed at and what all the commits change
11:11 Jakey3 ok
11:12 b00^wk left #git
11:12 johnmilton joined #git
11:12 skasturi joined #git
11:13 ochorocho__ joined #git
11:13 romerocesar joined #git
11:15 Jakey3 solved
11:15 ojdo joined #git
11:18 Guest39950 joined #git
11:18 cd-rum joined #git
11:19 HoierM joined #git
11:20 Macaveli joined #git
11:20 ifra201 joined #git
11:22 Shashikant86 joined #git
11:23 espindola joined #git
11:24 Lvjasha joined #git
11:25 Jakey3 how do i push to not the master branch?
11:26 Jakey3 If I'm checked out on another branch will it automatically push to that branch rather than the master?
11:27 aukeroorda2 joined #git
11:27 aukeroorda2 Hello
11:27 gitinfo aukeroorda2: 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.
11:28 rahtgaz joined #git
11:28 Jakey3 it does
11:29 Ezzy Hi, I'm looking at creating git-compatible binary patches.
11:29 Ezzy I feel uncomfortable, that I must set a header `index <hash>, <hash>` with hashes created by `git hash-object <content before/after>`. And this `hash-object` can do things that I really won't copy - `convert_to_git` method in convert.c that does CRLF conversions and some more magic.
11:29 Ezzy But according to http://alblue.bandlem.com/2011/08/git-tip-of-week-objects.html, this hash is just a `sha1 ("blob " + size + '\0' + bytes)` - and I wonder what would break if I use it in this simple form.
11:29 a3Dman_ joined #git
11:29 tuxlovesyou joined #git
11:30 aukeroorda2 I am very new to using git, and after using sourcetree for some time, I am now trying to edit some commit messages, also from the root node. I found this SO answer that I thought would do it, but I am lost on how to continue after executing that command: http://stackoverflow.com/a/14630424/3684659 . I deleted the comments and rewrote them here: htt
11:30 aukeroorda2 p://prntscr.com/bqc02k, but I don't know how to execute it now. I feel like I have to use the 'r' or 'reword' command, but I have no idea how.
11:30 aukeroorda2 Heh, my link is broken? http://prntscr.com/bqc02k
11:33 dionysus69 joined #git
11:38 Shashikant86 joined #git
11:39 jaguarmagenta joined #git
11:39 canton7 aukeroorda2, yeah, editing the commit messages in rebase -i's editor window does nothing: they're there for information only
11:39 canton7 yes, you need to change 'pick' to 'reword' (or 'r'), save and exit, let rebase do its thing
11:39 aukeroorda2 hmm
11:40 canton7 then rebase will pause, and prompt you to edit the message
11:40 aukeroorda2 Oke, I have changed them to reword, but I dont know how to save and exit?
11:40 aukeroorda2 http://prntscr.com/bqc6in
11:40 msonntag joined #git
11:40 thierryp joined #git
11:40 hk238 joined #git
11:42 canton7 aukeroorda2, the editor is vim. Simply, press ESC to exit insert mode, then :wq to save and exit
11:42 fracting joined #git
11:42 canton7 if you're not comfortable with vim, it might make sense to change the editor to something else
11:42 canton7 you can do this by setting the EDITOR env var, the GIT_EDITOR env var, or by changing the core.editor config setting
11:43 canton7 apparently 'git config --global core.editor notepad' now works on git for windows (> 2.5.3)
11:44 aukeroorda2 Ah, cool. I will try to set it to sublime
11:44 canton7 googling 'git windows sublime' seems to come up with plenty of suggestions
11:45 rbern_ joined #git
11:46 netj joined #git
11:46 stantonk joined #git
11:48 raintrap joined #git
11:49 synthroid joined #git
11:49 vuoto joined #git
11:50 Sceorem joined #git
11:51 aukeroorda2 Thanks canton7, that helped me a lot. I was totaly unfamiliar with a program using a text-editor as 'input field', if you understand what I mean. Cool
11:51 thierryp_ joined #git
11:52 bruce_lee joined #git
11:52 canton7 aukeroorda2, yeah, agreed it's not usual. It's kinda handy in the case of editing commit messages and setting up a rebase -i though
11:53 sveinse I've run git filter-branch --tree-filter, yet git grep xxx $(git rev-list --all) still show some of the pattern that has been changed. Why is that?
11:53 aukeroorda2 Yeah, I see that it also can be really neat to just use your preferred textbuffer as input field. Pretty neat
11:53 romerocesar joined #git
11:57 boingy joined #git
11:57 sveinse Hmm. git grep "pattern" $(git rev-list --all) shows a number of matches. If I do git show 1234 |grep "pattern", I do not find it. Why is that?
11:58 a_thakur_ joined #git
11:59 ramor joined #git
11:59 moritz sveinse: becase the "git show" only shows the diff
11:59 moritz sveinse: and git grep searches in the state of the repository at the point of that commit
12:00 a_thakur_ joined #git
12:00 sveinse If I do a git clone to another location and rerun git grep "pattern" $(git rev-list --all) the pattern is not found. So I can probably assume the pattern is gone. Good.
12:01 hoangtranwork joined #git
12:02 sveinse If I rewrite/filter-branch env-filter, msg-filter and tree-filter, I have covered all information in the repo, right?
12:03 tomog999_ joined #git
12:03 hoangtranwork is there a way to show a summary of current dirty changes in working dir (compared to HEAD) by each types, something like: 3 changed, 2 added, 1 removed
12:05 fracting joined #git
12:06 raijin joined #git
12:06 allcentury joined #git
12:06 Seveas hoangtranwork: git diff --stat HEAD
12:07 Seveas and if you really only want the oneline summary: git diff --shortstat HEAD
12:09 yanome joined #git
12:09 thebloggu joined #git
12:12 hahuang61 joined #git
12:12 tabrez joined #git
12:13 romerocesar joined #git
12:15 nwkj86 joined #git
12:16 azbarcea joined #git
12:17 hoangtranwork @Seveas: thank you, shotstat almost much my need: https://www.dropbox.com/s/eop6vthxribwxx8/Screenshot_2016-07-08_19.16.39.png?dl=0, it's better if it shows something like 3 file changed, 2 file added, instead of 5 changed files total
12:18 Literphor joined #git
12:20 sveinse How to do git clone and losing reference to remote afterwards? git pull is propsed, but it does not retain the tags
12:21 osse sveinse: git clone ...; git remote rm origin;
12:21 cdown joined #git
12:23 GGflags joined #git
12:23 rahtgaz joined #git
12:25 notebox joined #git
12:30 liquidjamm joined #git
12:30 aukeroorda2 Hmm, how do I get rid of the bottom 'tree'? Or atleast make sure that I continue from the top two commits? I have rebased properly (I think), but I don't dare to push or pull, because I don't know which I should do to continue with the top tree
12:30 aukeroorda2 http://prntscr.com/bqcpso
12:30 cdown joined #git
12:34 MutantMahesh joined #git
12:35 _noblegas joined #git
12:36 fahadash joined #git
12:37 case39 joined #git
12:38 sanct joined #git
12:39 jim_carry joined #git
12:40 fuchstronaut joined #git
12:41 pioiytr__ joined #git
12:42 raininja joined #git
12:42 canton7 aukeroorda2, any local work you do will be on on top of 'master'
12:43 aukeroorda2 hmm, but they are both 'Master', and one also origin? How do I clear my push/pull queue?
12:43 canton7 you've rewritten the two commits which were pushed to github, why is they they appear duplicated. this is !rewriting history. In your case, it's probably not an issue, because it's unlikely that anyone else has fetched the commits since you pushed them, but in general it's a bad idea to rewrite commits which have been made public
12:43 gitinfo [!rewrite] 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/waqum
12:43 canton7 origin/master is the state of the 'master' branch in the 'origin' remote
12:44 romerocesar joined #git
12:44 canton7 so you pushed master to origin, then you rewrote master, which means that your local master is now different to the master on origin
12:45 kulelu88 joined #git
12:45 aukeroorda2 Hmm what I thought I did was: I commited the first Day 1 and Day 2; then I did the rebase, which created this new (blue) tree. I have yet to pull or push again
12:46 griffindy joined #git
12:46 canton7 indeed, that looks right
12:47 GodGinrai joined #git
12:47 acetakwas joined #git
12:50 fuchstronaut joined #git
12:51 RustyShackleford joined #git
12:51 pioiytr__ joined #git
12:52 hoangtranwork joined #git
12:52 aukeroorda2 joined #git
12:53 TooLmaN joined #git
12:53 kablamas joined #git
12:54 tvw joined #git
12:54 Hachiman joined #git
12:54 nettoweb joined #git
12:54 Sadeukko joined #git
12:55 jim_carry joined #git
12:55 jeffreylevesque joined #git
12:56 cuebix-wk joined #git
12:57 wyt joined #git
12:57 Faylite joined #git
12:59 ToBeCloud joined #git
13:00 hoangtranwork joined #git
13:01 xpk joined #git
13:01 nickthetait joined #git
13:02 cs02rm0 joined #git
13:02 zacts joined #git
13:04 ManikM_ joined #git
13:05 Macaveli joined #git
13:05 kablamas left #git
13:08 pioiytr__ joined #git
13:12 vuoto joined #git
13:12 raijin joined #git
13:13 vivekp joined #git
13:13 vivekp joined #git
13:14 ochorocho__ joined #git
13:14 liamonade joined #git
13:14 rgrinberg joined #git
13:16 zivester joined #git
13:17 ekinmur joined #git
13:19 jim_carry joined #git
13:21 Lunatrius` joined #git
13:21 FunkyAss joined #git
13:24 stasmakarov joined #git
13:24 romerocesar joined #git
13:26 jim_carry joined #git
13:28 DieguezZ joined #git
13:28 efdetonator joined #git
13:29 Shock- joined #git
13:33 notebox joined #git
13:33 marcogmonteiro joined #git
13:35 ash_workz joined #git
13:35 lucido-cwl joined #git
13:37 lucido-cwl HI, what does tracking mean in git?
13:37 chitopunk joined #git
13:37 rahtgaz joined #git
13:37 canton7 lucido-cwl, in what context?
13:37 code_geek joined #git
13:38 phantomtiger joined #git
13:38 lucido-cwl canton7, in git, what does it mean that a local branch is tracking a remote
13:39 ededdeddy joined #git
13:39 jaguarmagenta joined #git
13:40 canton7 lucido-cwl, for each local branch, there's an optional piece of configuration (in .git/config) which says if that local branch is related to any other remote branches. This allows commands like 'git pull' (without any other arguments) and 'git push' (in some modes) to merge or push to the correct remote branch, so that you don't have to keep specifying the same remote branch every time you want to pull / push
13:40 canton7 that piece of config is automatically set when you create a new local branch from a remote branch - e.g. if you ran 'git checkout -b feature origin/feature', a piece of config is set which says that 'feature' is related to 'origin/feature'
13:40 canton7 it's also set by 'push -u' and 'branch --set-upstreamto'
13:40 canton7 * --set-upstrea-to
13:40 canton7 * --set-upstream-to
13:40 * canton7 fails
13:41 fission6 joined #git
13:41 fission6 i have a ton of local branches, how do i delete all branches already merged to master
13:41 fission6 (assuming thats the logical thing to do)
13:41 kulelu88 what is the purpose of adding an email to an SSH key?
13:42 fission6 i guess i also want to remove branches upstream
13:42 efdetonator joined #git
13:42 lb kulelu88 reducing SPAM since every email will be send over the secure ssh connection </sarcasm>
13:43 kulelu88 lb: the real reason is? (or should I ask the question in another channel?)
13:44 lb kulelu88 in general yes, wrong channel, but there are a lot of ssh artists around here. maybe someone knows
13:44 WayToDoor joined #git
13:45 kulelu88 apparently a #ssh channel exists
13:45 epochsal joined #git
13:45 code_geek left #git
13:47 Dougie187 joined #git
13:47 al-damiri joined #git
13:48 tjbp joined #git
13:48 stantonk joined #git
13:49 tjbp joined #git
13:50 efdetonator joined #git
13:50 allcentury joined #git
13:52 efdetonator joined #git
13:52 tjbp joined #git
13:54 Rodya_ joined #git
13:54 kbs joined #git
13:55 romerocesar joined #git
13:56 lindenle joined #git
13:58 nettoweb joined #git
13:58 tjbp joined #git
14:00 unbalancedparen joined #git
14:00 ManikM joined #git
14:01 snowkidind joined #git
14:05 lucido-cwl canton7, I see. So git flow feature track FEATURE will simply allow my to use git push or pull without extra arguments when on the feature/FEATURE branch?
14:06 FunkyAss joined #git
14:06 canton7 lucido-cwl, no idea what git-flow does there
14:06 canton7 git-flow is third-party tooling - it isn't part of git itself
14:07 Waskeend joined #git
14:07 D-Boy joined #git
14:08 canton7 from the docs, it looks like it does what you described, however
14:08 pioiytr__ joined #git
14:09 concerti joined #git
14:10 pioiytr__ joined #git
14:10 fscala joined #git
14:10 ResidentBiscuit joined #git
14:10 zacts joined #git
14:13 Tim-SEM joined #git
14:13 hahuang61 joined #git
14:15 FilipNortic joined #git
14:16 fuchstronaut joined #git
14:16 fuchstronaut joined #git
14:17 Literphor joined #git
14:18 Asiajey joined #git
14:18 Asiajey joined #git
14:19 Tim-SEM left #git
14:20 phanimahesh joined #git
14:20 ResidentBiscuit joined #git
14:21 raijin joined #git
14:22 pioiytr__ joined #git
14:23 cdg joined #git
14:25 cdg joined #git
14:25 romerocesar joined #git
14:25 baraba joined #git
14:25 settermjd joined #git
14:26 harish_ joined #git
14:29 sangy joined #git
14:31 Repox joined #git
14:33 paws_ joined #git
14:35 catag87 joined #git
14:35 romerocesar joined #git
14:36 catag87 joined #git
14:37 Macaveli_ joined #git
14:37 thierryp joined #git
14:39 rgrinberg joined #git
14:39 alcohol fission6: i often fall back to this one; but always be careful and try a dry-run first (you can do that with xargs) http://stackoverflow.com/a/6127884/248104
14:40 mostlybadfly joined #git
14:41 diegoviola joined #git
14:43 kbs joined #git
14:44 ShekharReddy joined #git
14:45 pioiytr__ joined #git
14:46 ExoUNX joined #git
14:47 Blkt joined #git
14:47 Literphor joined #git
14:48 LeMike joined #git
14:51 fission6 thanks alcohol
14:51 xpk joined #git
14:51 smithbone joined #git
14:53 jstimm joined #git
14:55 SirScott joined #git
14:55 romerocesar joined #git
14:57 SirScott when i do 'git checkout develop', it says that my branch is "up-to-date with 'master/develop'", but then I immediately do 'git pull master develop' and receive down changes.  Why is it telling me that I'm up to date?
14:57 clemf joined #git
14:59 bluj SirScott, checking out a branch won't fetch any new changes that might be available upstream. it just switches your current working tree.
14:59 ochorocho__ joined #git
14:59 bluj e.g. git checkout could be done without any network connection. 'git pull' in most cases could not.
15:00 SirScott bluj: right, so what's the point of the 'your branch is up to date with $upstream' message?
15:00 bluj SirScott, because to the best of its knowledge, it was
15:00 lugzer joined #git
15:01 bluj if you want to check for any new changes, without being forced to merge them in, do 'git fetch'
15:02 SirScott bluj: is there a way to configure it so that it'd actually perform the sync check upon branch switching?
15:02 SirScott bluj: because when i read that message I think "oh, I don't need to do a fetch / pull"
15:03 bluj i don't know, it sounds like a horrible idea. then you would be forced to deal with any merge conflicts immediately
15:03 bluj i would just change your thinking ;) .. git is helpful for offline work, and what you see is a symptom of that.
15:04 zacts joined #git
15:04 FilipNortic joined #git
15:05 IWishIKnew joined #git
15:05 ResidentBiscuit ^
15:05 ResidentBiscuit It's really just a misunderstanding of the tool. master/develop is not the remote repo itself, it's what your local repo knows of the remote repo
15:05 SirScott bluj: i could swear that sometimes when i switch branches i get nice messages telling me that i've diverged from $upstream.  i.e. i'm either ahead or behind a few commits.
15:06 fracting joined #git
15:07 LionsMane joined #git
15:07 ResidentBiscuit Because you've done something to update what your local repo knows
15:07 tadz joined #git
15:08 cdown joined #git
15:08 circ-user-A26a8 joined #git
15:08 rkazak joined #git
15:08 ToxicFrog SirScott: it bases that on its local knowledge of the remote's state, i.e. the last time you did a "git fetch" or "git push".
15:09 dhollinger joined #git
15:09 xcesariox joined #git
15:10 chrisshattuck joined #git
15:10 SirScott Gotcha.  So there's no way to squelch the 'up-to-date with master/develop' unhelpful message?  I really just have to remember that it's local only?
15:11 zivester joined #git
15:12 englishstudent joined #git
15:12 fuchstronaut joined #git
15:12 tadz left #git
15:12 LionsMane joined #git
15:13 zacts_pi joined #git
15:13 InfoTest joined #git
15:14 kpease joined #git
15:14 ResidentBiscuit How is it unhelpful? It is up-to-date with what it knows
15:14 ramsub07 hello, i'm in a branch at commit C. A is a commit older to C. I've made a mistake in code, i suspect and i want to check if the same error occurs in commit A as well. How do i checkout to that commit, run the test suite and come back ?
15:15 ResidentBiscuit `git checkout <commit_A_hash>`
15:16 stantonk joined #git
15:17 subhojit777 joined #git
15:19 ramsub07 how do i check the diff done after committing ?
15:19 tvw joined #git
15:19 pioiytr__ joined #git
15:21 alkino joined #git
15:21 alkino hi o/
15:21 canton7 ramsub07, 'git diff HEAD^' or 'git show'
15:21 alkino is there any good reason to git-archive to take a tree-ish and not a commit-ish?
15:21 zacts joined #git
15:21 canton7 alkino, it will accept a commit as well
15:22 alkino canton7: it doesn't work for me.
15:22 alkino I can give a tag, but not a "sha1"
15:22 zacts_pi joined #git
15:22 artemisart joined #git
15:22 canton7 man gitglossary for treeish says "Dereferencing a commit object yields the tree object corresponding to the revision's top directory"
15:22 gitinfo the gitglossary manpage is available at http://jk.gs/gitglossary.html
15:22 alkino canton7: it was my understanding
15:23 canton7 it works here
15:23 settermjd joined #git
15:23 concerti joined #git
15:23 alkino canton7: I try it remote.
15:24 alkino I see there is a specific code for remote in archive.c. I look further
15:24 alkino /* Remotes are only allowed to fetch actual refs */
15:24 kpease joined #git
15:24 alkino Is there a reason why remotes can only fetch refs?
15:24 anddam joined #git
15:25 anddam hello, I committed a couple times onto master, now I realize I actually want to pull from remote and should have used a branch of my own
15:25 _ikke_ alkino: Is this sha reachable from a ref?
15:25 alkino _ikke_: it is.
15:25 thiago joined #git
15:26 romerocesar joined #git
15:26 anddam now I'm in a clean working directory, can I put this commit (or another empty one) in a new branch and pull from origin
15:26 _ikke_ iirc you should be able to fetch commits by hash as long as they're reachable
15:26 anddam s/$/?/
15:26 baraba_ joined #git
15:26 alkino https://github.com/git/git/blob/master/archive.c#L361 <= _ikke_, it seems not
15:27 _ikke_ Ok, I recall incorrectly then
15:27 alkino If I follow the glossary, a sha1 is not a ref.
15:27 _ikke_ but the reason is probably security related
15:27 jstimm joined #git
15:27 rafalcpp joined #git
15:27 alkino I can't see the problem. We're all doing open source code \o/
15:28 ramsub07 ResidentBiscuit: if i do `git checkout <commit hash> ` I go in a detached head state
15:28 anddam ah, websearch-fu to the rescue! http://stackoverflow.com/a/1628584/434206
15:28 alkino ramsub07: yes, but it's not a problem.
15:28 anddam sorry for the noise
15:28 ResidentBiscuit ^
15:28 ramsub07 alkino: i don't see the changes made in that commit ?
15:28 ResidentBiscuit You can run your tests in detached head state
15:28 _ikke_ alkino: That doesn't mean that all commits are public
15:28 canton7 alkino, you can't fetch specific shas either
15:29 ramsub07 ResidentBiscuit: would the code be at that particular commit ?
15:29 alkino it's a pity
15:29 ResidentBiscuit ramsub07: Yes
15:29 ramsub07 i don't see like that in file editor..
15:29 narendraj9 joined #git
15:29 canton7 alkino, what if someone accidentally commits something bad, then rewrites it? The commit object might still be in the repo, so people would still be able to fetch it
15:29 cdown joined #git
15:29 ResidentBiscuit ramsub07: Then you're at the wrong commit
15:29 alkino canton7: ... Does it mean, that if I compile my own git without this "security" I can do it? Let's try!
15:30 alkino This security should be in archive-upload.c
15:30 ToxicFrog I'm having a really weird issue with .gitignore. I'm trying to set up a repo with a whitelist approach to ignoring, i.e. everything is ignored except what's specifically permitted.
15:30 ToxicFrog So, I initialize the .gitignore with "/*", and now everything is ignored and I have to "add -f" to track anything. Working as expected.
15:31 ToxicFrog But there's an entire subdirectory, foo/bar/, that I want to unignore. So back to .gitignore I go and add the next line: "!/foo/bar"
15:31 ToxicFrog And this is where things get wacky.
15:32 jfran joined #git
15:32 canton7 ToxicFrog, for performance reasons, only things which are considered for ignoring can be unignored. So if you ignore /*, you git will check whether you unignored /anything, but it won't check inside any of the top-level folders to see if there's something else it should be unignoring
15:32 ToxicFrog When I add that, foo/bar/ gets unignored, but so do some files in /!
15:32 canton7 so you'll need '/*; !/foo; /foo/*; !/foo/bar'
15:32 ToxicFrog Files in foo/ other than foo/bar/ remain ignored, though.
15:33 ToxicFrog canton7: that works fine, thanks.
15:33 ToxicFrog I'm still really confused by how this could cause some files in / to stop being ignored, though.
15:34 canton7 that shouldn't by the case. !repro
15:34 gitinfo [!transcript] Please paste (using https://gist.github.com/ or https://vomitb.in/ ) a transcript ( https://gist.github.com/2415442 ) of your terminal session so we can see exactly what you see
15:34 ToxicFrog Yeah, working on a minimal repro now.
15:35 The_Phoenix joined #git
15:35 dmto joined #git
15:35 alkino canton7: ok. archive.c is used locally and remotely. And hopefully this is tested remotely.
15:35 dreiss joined #git
15:35 encod3 joined #git
15:36 rafalcpp joined #git
15:37 cdg_ joined #git
15:40 bgamari- joined #git
15:40 jaguarmagenta joined #git
15:41 EvilPenguin joined #git
15:42 ToxicFrog canton7: https://gist.github.com/anonymous/10af84f8b9bc07e30b16e4889faa107a
15:42 ToxicFrog Adding "!/foo/bar/" to .gitignore causes /a, /c, and /.gitignore to stop being ignored (and nothing else)
15:43 ejb joined #git
15:43 canton7 let me see...
15:44 canton7 ToxicFrog, what does 'git status' say at the end there?
15:44 ToxicFrog (if I leave off the trailing /, the contents of /foo/bar/ are unignored but /a, /c and /.gitignore are still unignored as well)
15:44 ToxicFrog Just plain "git status" with no arguments?
15:44 canton7 yes
15:45 ToxicFrog https://gist.github.com/anonymous/14c89f15a145dc7f0b828f0ad798b933
15:45 canton7 huh
15:45 * canton7 tests
15:45 xaviergmail joined #git
15:46 cdg joined #git
15:46 canton7 can't repro here, 2.8.1
15:47 akushner joined #git
15:47 ToxicFrog canton7: hmm. I have a 2.8.1 machine I can try it on
15:48 canton7 ToxicFrog, maybe you have some other ignores? 'git check-ignore -v a'
15:48 soLucien joined #git
15:48 ToxicFrog .gitignore:1:/* a
15:49 ToxicFrog As it should be.
15:49 ToxicFrog Trying it on 2.8.1 it behaves as expected, so I guess this is a bug in 2.8.0 that was fixed in 2.8.1
15:50 soLucien i am using GitExtensions in order to manage Git (since the Visual Studio pplugin is quite rudimentary). I have created a brach called feature/myfeature a week ago. I have not finished the feature, and want to get rid of the branch by pushing the changes to Master. In order to do this, i have merged master into feature/myfeature, and fixed the merge conflict
15:50 jjasinski joined #git
15:50 canton7 the release notes don't list it, and there are no relevant commits
15:50 canton7 https://github.com/git/git/commits/v2.8.1
15:51 soLucien i now want to make all the 20 commits into a single one
15:51 soLucien before making a pull request
15:51 segment-fault joined #git
15:51 soLucien what is the command/sequence of commands that i need to perform ?
15:51 javlens joined #git
15:52 soLucien i have now finished the feature *
15:52 GodGinrai https://git-scm.com/book/en/v2/Git-Tools-Rewriting-History
15:52 rafalcpp joined #git
15:52 Electrometro joined #git
15:53 soLucien i have tried rebase, but it seems it does not work if i have already merged master into the branch
15:53 sveinse left #git
15:54 GodGinrai well duh
15:54 b3f0r3 joined #git
15:54 b3f0r3 left #git
15:54 quetz joined #git
15:54 soLucien i am not a git expert, so it is not "duh" to me .. what have i done wrong ?
15:55 grawity `git rebase -i master` should work I think
15:55 soLucien should i not merge before rebase ? Revert to the point before merging ?
15:55 notebox joined #git
15:56 GodGinrai Basically, rewriting history is not advised in the first place.  If you try to rewrite something that already exists in master (which you would if you merged master commits in) then you will risk causing problems for other devs
15:56 soLucien i have a lot of commits where i was trying stuff out
15:56 GodGinrai if you intend to rewrite history, then you need to rebase your commits onto the latest master instead of merging
15:56 _ikke_ Yeah, rebase before merge
15:56 soLucien changed foo to bar. Changed foo to boo. removed foo. added foo back
15:56 lindenle joined #git
15:57 soLucien i don't want that to be in history, it was my shit-work
15:57 GodGinrai But honestly, you shouldn't really commit things you are "trying out"
15:57 _ikke_ rebase by default flattens history, which means that merge commits are not recreated
15:57 _ikke_ GodGinrai: Doesn't really matter, as long as you clean it up
15:57 _ikke_ git makes it even easy with fixup commits and autosquash
15:57 soLucien i don't need to see what trial-and-error stuff i did, just want the PR to contain the result, and a list of what was done
15:58 thierryp joined #git
15:58 GodGinrai _ikke_: You shouldn't be rewriting history, so you shouldn't be dirtying it in the first place
15:58 netcarver joined #git
15:58 soLucien GodGinrai  i was working against a unknown API, so i committed every time i made something work
15:58 _ikke_ GodGinrai: You should not be rewriting published history
15:58 FunkyAss joined #git
15:58 _ikke_ GodGinrai: It could be very valuable to be able to look at your process so far
15:58 GodGinrai _ikke_: if you get in the habit of doing it locally, you are more likely to make that error
15:59 _ikke_ GodGinrai: git push will reject it, and you can easily undo it
15:59 _ikke_ !perfect !sausage
15:59 gitinfo [!postproduction] So, you want to make your commit history look pretty before pushing? http://sethrobertson.github.com/GitPostProduction talks you through how to use 'rebase -i' to do this.
15:59 gitinfo [!sausage_making] Some developers like to "hide the sausage making", transforming their commits before presenting them to the outside world. See http://sethrobertson.github.com/GitBestPractices/#sausage and !perfect
15:59 soLucien i see it as a benefit for someone reviewing my PR
15:59 _ikke_ yes
16:00 _ikke_ But not by default always squashing everthing in a single commit
16:00 GodGinrai soLucien: proper comments and commit messages should be helpful enough to the reviewer
16:00 _ikke_ GodGinrai: And rewriting local history helps with that
16:00 soLucien "updated package version to 1.1.1-rev1" "updated package version to 1.1.1-rev2"
16:00 _ikke_ GodGinrai: It lets you look back and tidy everything up
16:01 GodGinrai _ikke_: assuming you didn't use good commit messages in the first place
16:01 _ikke_ GodGinrai: It's not always about the messages
16:01 _ikke_ GodGinrai: Also about the changes
16:01 elementalest joined #git
16:01 elementalest joined #git
16:01 soLucien i did use good commit messages, but some of them are irrelevant
16:01 segment-fault left #git
16:01 soLucien if i could keep some of them
16:01 _ikke_ You can
16:01 soLucien while hiding others
16:01 soLucien that would be amazing
16:01 _ikke_ interactive rebase lets you squash whatever commits you want
16:01 GodGinrai that's what interactive rebase allows you to do
16:01 stantonk joined #git
16:01 _ikke_ or fixup if you don't care about the messages
16:02 lvns joined #git
16:02 leprechau joined #git
16:03 _ikke_ git rebase -i <base_branch>
16:03 _ikke_ In the editor you can reorder lines, change pick to fixup or squash for commits you don't care about
16:04 soLucien http://pasteboard.co/U8SwulH1.png
16:04 soLucien some of them are really irrelevant
16:04 ekinmur joined #git
16:04 soLucien "moved folder to bla" .. "updated package" .. "Committed in order to lock changes"
16:04 soLucien no one cares
16:05 romerocesar_ joined #git
16:06 refried_ joined #git
16:07 stantonk joined #git
16:07 aep-shoutlet joined #git
16:07 romeroce_ joined #git
16:08 thierryp joined #git
16:09 stantonk_ joined #git
16:09 ploop joined #git
16:09 lindenle joined #git
16:10 Guest23954 joined #git
16:10 The_Phoenix1 joined #git
16:11 ekinmur joined #git
16:12 stantonk joined #git
16:13 The_Phoenix joined #git
16:13 stantonk joined #git
16:14 stefanha joined #git
16:14 hahuang61 joined #git
16:15 zacts joined #git
16:15 romerocesar joined #git
16:16 madewokherd joined #git
16:16 The_Phoenix1 joined #git
16:17 MoPac joined #git
16:18 stefanha I'm confused by git blame --reverse START..END output: https://paste.fedoraproject.org/389011/67994619/
16:18 kbs joined #git
16:18 stefanha Either I misunderstand what --reverse does or there is a bug.  It is showing a commit that doesn't touch the file.
16:18 stefanha git 2.7.4
16:18 stefanha Any ideas?
16:19 romerocesar_ joined #git
16:19 ToxicFrog stefanha: rename following?
16:19 stefanha ToxicFrog: No, cf7cc929 is a completely different file.
16:20 cdown joined #git
16:20 pur3eval joined #git
16:20 ToxicFrog canton7: huh. That makes it extra weird. The 2.8.1 machine is SUSE and the 2.8.0 one is Ubuntu, so I wonder if this is something introduced by an Ubuntu maintainer's patch (or fixed by a SUSE maintainer's one)
16:20 stefanha ToxicFrog: I'm running git-blame(1) on a .c file and the reported commit changes the ./configure script.
16:21 ToxicFrog ...that is super weird
16:21 _ikke_ stefanha: "this shows the last revision in which a line has existed."
16:21 GodGinrai stefanha: but did the commit after that change the file?
16:21 davidosomething joined #git
16:21 thierryp joined #git
16:21 stefanha ToxicFrog: It has identified that the line was changed by a commit, but it seems to be showing the wrong commit.  The *unchanged* lines are correct (i.e. it's showing HEAD aka master)
16:22 netcarver joined #git
16:22 nettoweb joined #git
16:22 ToxicFrog _ikke_: oooo, that makes sense.
16:22 _ikke_ stefanha: --reverse does not show commits that change a line
16:22 GodGinrai stefanha: reverse shows you, as _ikke_ highlighted, the last revision in which the line existed.  Btw, his quote is straight from the manpage that you should be reading
16:22 stefanha Thank you folks, that makes sense now :)
16:23 stefanha I thought it would still include the commit that changed the line because the line still "existed" in that commit.
16:23 stefanha but I'm thinking in terms of diffs and not the commits/trees that git stores.
16:23 dsdeiz_ joined #git
16:24 GodGinrai stefanha: I believe a line "existing" means existing "as is".  If it is changed, it becomes a different line
16:24 stefanha GodGinrai: Yep, indeed, the next commit changes the line and is the one I'm looking for.  Thank you!
16:24 GodGinrai np
16:25 MoPac joined #git
16:25 LiohAu i7c: I think that what I was looking for yersteday was "git checkout develop -- SomeFolder/"
16:25 LiohAu without any git merge
16:25 dminca joined #git
16:25 SirScott left #git
16:27 alkino to come back on my problem of git archive remote of a commit-ish. I think there is a problem of design. We can read "if (remote && !remote_allow_unreachable) {
16:28 alkino what if I want to archive reachable commit but not unreachable, without giving a ref but a sha1?
16:28 zumba_ad_ joined #git
16:30 FunkyAss joined #git
16:31 shinnya joined #git
16:31 canton7 ToxicFrog, hmm, perhaps
16:34 thierryp_ joined #git
16:36 yqt joined #git
16:36 freimatz joined #git
16:37 dionysus69 joined #git
16:37 netcarver joined #git
16:40 TomyLobo joined #git
16:41 FuzzySockets joined #git
16:41 bgamari joined #git
16:41 alkino is there a C function in git code, to get all refs of the current repository. Can't found it in refs.h
16:42 slayerjain joined #git
16:42 d4rklit3 joined #git
16:42 HoierM joined #git
16:42 thierryp joined #git
16:43 fuchstronaut joined #git
16:43 fracting joined #git
16:44 Eugene alkino - look into for-each-ref
16:44 alezandro joined #git
16:45 freimatz joined #git
16:45 PCatinean_ joined #git
16:46 netcarver joined #git
16:46 durham joined #git
16:47 rgrinberg joined #git
16:49 alkino Eugene: thanks
16:49 refried_ joined #git
16:51 Motoko joined #git
16:56 mwhooker joined #git
16:57 alkino Eugène: is there any "is_reachable_by_any_ref" function? ^^
16:57 thierryp_ joined #git
16:57 zacts joined #git
16:58 devbug joined #git
16:58 encod3 joined #git
17:00 durham_ joined #git
17:02 Peuc joined #git
17:04 romerocesar joined #git
17:06 tuxlovesyou joined #git
17:06 eijk joined #git
17:07 ochorocho__ joined #git
17:09 eijk_ joined #git
17:09 ejb joined #git
17:12 steveballmer joined #git
17:14 concerti joined #git
17:15 choki joined #git
17:18 kaldoran joined #git
17:19 inflames joined #git
17:19 Asiajey joined #git
17:20 SteffanW joined #git
17:21 eijk_ joined #git
17:22 texinwien_ joined #git
17:23 clemf joined #git
17:24 wiky joined #git
17:24 romerocesar_ joined #git
17:25 cs02rm0 joined #git
17:25 sangy joined #git
17:28 LeMike joined #git
17:29 romerocesar joined #git
17:31 UTAN_dev joined #git
17:31 Peuc joined #git
17:31 pioiytr joined #git
17:31 pur3eval joined #git
17:32 EvilPeng1 joined #git
17:32 pioiytr joined #git
17:33 rgrinberg joined #git
17:33 simonlinj joined #git
17:34 bjoe2k4 joined #git
17:35 RustyShackleford joined #git
17:35 Seveas follow whatever git branch --contains does
17:36 Seveas or git for-each-ref --contains
17:36 romerocesar_ joined #git
17:36 alkino Seveas: trying, but it's my second time in git code.
17:38 troulouliou_div2 joined #git
17:40 jstimm joined #git
17:41 jaguarmagenta joined #git
17:41 Faylite_ joined #git
17:42 cs02rm0 joined #git
17:45 stantonk joined #git
17:47 pur3eval joined #git
17:47 srcerer joined #git
17:48 NeverDie joined #git
17:49 bikky22 joined #git
17:52 mish4 joined #git
17:52 mish4 joined #git
17:53 freinn joined #git
17:54 jimi_ joined #git
17:55 holodoc Would cherry picking be the right tool for reverting to one of the previous commits without resorting to such harsh actions like git reset --hard?
17:55 yqt joined #git
17:55 Silenced joined #git
17:56 alkino holodoc: no, I don't think so
17:56 Silenced joined #git
17:57 holodoc Any alternative? I've done some work on a feature branch and merged it into my master. However, I would like to revert master to a point before the merge. Is git reset --hard my only option?
17:58 holodoc Problem is I've done work in master after the merge.
17:59 vuoto joined #git
17:59 tyreld joined #git
17:59 clemf joined #git
18:00 thiago holodoc: reverting merges is not possible without rewriting history
18:00 thiago you can apply the reverse of the changes that the merge brought in, but you can't merge again
18:02 freinn left #git
18:02 elastix joined #git
18:02 alkino holodoc: do you push it remotely?
18:02 Xnu17 joined #git
18:02 durham joined #git
18:02 holodoc No everything is done purely locally.
18:03 alkino so you need to rewrite your history. Do you understand what it means?
18:03 freinn joined #git
18:04 A5101 joined #git
18:05 sddasdasd joined #git
18:05 holodoc Yes I know what it means. I am basically just looking for a way to remove the feature branch merge from master.
18:05 holodoc I generally know what my options are.
18:05 neilthereildeil joined #git
18:05 holodoc But still there might be something I am now aware of.
18:05 neilthereildeil hi
18:06 Dougie187 holodoc: you could just revert it.
18:06 holodoc *not
18:06 dionysus69 joined #git
18:06 holodoc But reverting would destroy all commits after the merge as far as I know.
18:06 Dougie187 `git revert -m 1 <merge_hash>`
18:06 Asiajey joined #git
18:06 Asiajey joined #git
18:06 Dougie187 No, reverting just undoes the changes brought in by the merge. But what you do really depends on what you're trying to accomplish.
18:07 Dougie187 Removing a merge means several different things
18:07 freinn hi! I'm having problems understanding the output of the git status -s command. http://pastebin.com/9jXZGSSJ
18:07 Dougie187 But the options for removing a merge are: revert, rebase, and reset.
18:07 Dougie187 revert removes the changes brought in by the merge, but not the merge itself.
18:07 netj joined #git
18:07 neilthereildeil my git HEAD points to MASTER, but i want to create a branch at an earlier commit. how can i do this?
18:07 Dougie187 rebase and reset are more destructive in that they will remove actual commits and rewrite history.
18:07 Arban joined #git
18:08 ochorocho__ joined #git
18:08 holodoc Yeah I am trying to avoid rebase and reset for those very reasons.
18:08 neilthereildeil and i want to make sure the current state of the working dir (and all uncommitted changes) are left untouched
18:08 holodoc Might be worth looking into revert
18:08 Dougie187 neilthereildeil: `git checkout -b <branch_name> <branch_point>`
18:08 neilthereildeil but wouldnt that touch my working tree?
18:08 Dougie187 holodoc: you can try `git revert` locally, and see if it looks like what you want. And if not, just undo it locally`
18:08 Sasazuka joined #git
18:09 Dougie187 neilthereildeil: Yes, but if you don't want to change your working tree, do `git branch <branch_name> <branch_point>`
18:09 holodoc It's not that much of an issue because this repo is for me exclusively. It's not shared.
18:09 romerocesar joined #git
18:09 chrisshattuck joined #git
18:10 thiago holodoc: if you haven't pushed yet, you should just do reset --hard
18:10 neilthereildeil thanks. also, whats the difference between checkout and reset --hard?
18:10 holodoc But isn't reset --hard going to destroy every commit I've made after the merge and put those into a detached HEAD state?
18:10 thiago neilthereildeil: to create a branch without checking it out, git branch branchname startingpoint
18:11 thiago neilthereildeil: git checkout with no file paths will switch HEAD to point to that branch
18:11 pioiytr joined #git
18:11 thiago neilthereildeil: git reset --hard will change the branch to point to that commit
18:11 alkino holodoc: put a ref on it, before "git branch foo"
18:12 neilthereildeil thiago: doesnt "change the branch to point" mean the same as switching HEAD?
18:13 durham joined #git
18:13 thiago neilthereildeil: if you're on branch master, then HEAD is "ref: refs/heads/master" and refs/heads/master has a commit SHA-1
18:13 Limes_ joined #git
18:13 Jackneill joined #git
18:13 thiago neilthereildeil: if you then do git checkout other, HEAD becomes "ref: refs/heads/other"
18:13 thiago neilthereildeil: if you had done git reset --hard other, instead, then HEAD stays as it was, but refs/heads/master moves to point to the commit that otherbranch pointed to
18:13 thiago checkout switches to another branch; reset --hard moves the current branch
18:14 neilthereildeil what does "move a branch" mean?
18:14 neilthereildeil my understanding is that we create a commit tree, and HEAD can be used to traverse it
18:15 hahuang61 joined #git
18:16 neilthereildeil1 joined #git
18:17 thiago neilthereildeil1: it changes the branch to point at a new commit
18:17 pioiytr joined #git
18:18 stantonk joined #git
18:19 impermanence joined #git
18:19 circ-user-A26a8 joined #git
18:21 freinn hi! I'm having problems understanding the output of the git status -s command. http://pastebin.com/9jXZGSSJ
18:21 afuentes joined #git
18:22 cyan__ joined #git
18:22 satifant joined #git
18:24 impermanence Not git but I feel you all have the insight: I have an SVN repo that was intended originally to store WARs and RPMs, but over time has accumulated related data files, as well, such as jks's, certs, tgz's, etc.  I've been tasked with evaluating other possible storage solutions and of course we're looking for something cheap.  As I understand it git is not meant for binaries.  So...anyone have any suggestions?  I probably need to
18:24 impermanence together.
18:24 Eugene neilthereildeil1 - !HEAD
18:24 gitinfo neilthereildeil1: HEAD is a 'pointer' to the currently checked out branch (or commit, if HEAD is !detached). In bare repositories it tells clients which branch to checkout initially after cloning. Unlike commonly believed, HEAD is *not* something that exists separately for every branch. It also is *not* necessarily the newest commit in the repo (that's hard to define in a DVCS, anyway...)
18:24 impermanence I mean, yes there is Artifactory, which is great for binaries and really meant for binaries and is slick but it is super pricy.
18:25 impermanence And I suppose we could do a yum repo...with a basic front end...like you would see in a public linux repo or something...
18:25 impermanence but my point is...I mean...what is the problem with SVN in the first place...
18:25 rpd freinn: https://git-scm.com/docs/git-status#_output
18:26 steveballmer joined #git
18:26 rpd That status tells me that the file was deleted from git's cache and that there is currently an untracked file of the same name that exists on the filesystem
18:26 nwkj86 joined #git
18:26 impermanence I know nothing about SVN, but the general consensus seems to be: move away.  So...how is everyone else managing their binaries and related stuff?
18:27 pioiytr joined #git
18:27 rpd impermanence: svn is good at storing binaries
18:27 impermanence rpd: would you consider svn legacy?
18:28 xaviergmail joined #git
18:28 rpd not really? it's still being actively developed
18:28 rpd A lot of people are excited about DVCS but it doesn't mean they're outdated/legacy. It always depends on what operations are useful for your tool and company
18:28 impermanence rpd: but nowadays everyone that needs code-management uses git and not svn, right?  I mean, this I thought was true...
18:29 ojacobson legacy is a dog whistle phrase for anything that works well enough that you can't mock it on its faults but don't want to praise it on its successes either
18:29 ojacobson fight me :)
18:29 impermanence ojacobson: lol.
18:29 rpd impermanence: everyone uses git because it is the simplest (other than mercurial) to set up for small, distributed projects
18:29 impermanence ojacobson: and, you'll win, I'm quite sure.
18:29 neilthereildeil joined #git
18:30 rpd there are probably plenty of stories of people moving from git to svn after the needs of their project changes
18:30 impermanence rpd: right, right, and see, I feel pretty clear on that...for code management that is.  But when we're talking storage of things like WARs, RPMs and their little buddies...is there like a tool that everyone is using that I'm missing here...?
18:30 rpd and git is terrible for storing binaries and other large artifacts; so if that's something you need, git might not provide the right tools for storing those
18:30 impermanence rpd: and that's what I heard.
18:30 fission6 joined #git
18:30 impermanence rpd: that it's terrib.
18:31 rpd The tool you're missing is called svn I'm pretty sure
18:31 cdown joined #git
18:31 rpd anyways, it's hard to evaluate what you need since I don't know and probably shouldn't know the details
18:32 neilthereildeil thiago: ok, so you are saying branches can be changed. thats not a normal operation, is it? usually people commit to branches and then checkout verious commits, right?
18:32 rpd svn is not legacy nor outdated, but its model doesn't fit many modern ways to develop and distribute code
18:32 rgrinberg joined #git
18:32 freinn thanks rpd, but my question is, there are two lines describing the status, and D_ means: deleted on index and deleted (same state (space) than the index state)
18:33 impermanence rpd: yeah yeah, of course on the details, and obviously I'm not looking or paying for consulting.  But the "sitch" is essentially this: bunch of files, maybe 20 GBs in total, maybe 50 at most, half are RPMs, WARs, the other are their little buddies.  Looking for a place to stick and retrieve them all.
18:33 Jason__ joined #git
18:33 Eugene neilthereildeil - nope; branches are where everything happens.
18:33 impermanence rpd: And it doesn't really seem like there is anything but SVN, Artifactory and Nexus.
18:33 GodGinrai rpd: "doesn't fit modern ways" is generally what is referred to as "outdated"
18:33 freinn rpd, my question was about this being two lines, and one of them, the first, is contradictory
18:34 pioiytr joined #git
18:34 neilthereildeil Eugene: nope the what?
18:34 thiago neilthereildeil: yes, you can move branches laterally or backwards in history
18:34 thiago neilthereildeil: whenever you do that, other people feel pain
18:34 neilthereildeil nope to what*
18:34 rpd GodGinrai: That's fine, but some problems don't have modern solutions.
18:34 Eugene neilthereildeil - no, people don't checkout commits directly(usually).
18:34 EvilPeng1 joined #git
18:35 aard_ joined #git
18:35 thiago neilthereildeil: if you just want to obtain the repository at a given previous commit for (for example) testing, you should use git checkout
18:35 neilthereildeil thiago: ok, so git reset it not a preferred way to do things because it causes pain?
18:35 Eugene A commit is just a snapshot at a point in time. A branch is used to organize these snapshots as they evolve over time. A tag is used to save a specific point(usually a software release)
18:35 rpd impermanence: I'm naive about Artifactory, never heard of Nexus. I do think svn does that job well.
18:35 thiago neilthereildeil: moving history backwards or laterally is the source of pain
18:35 neilthereildeil thiago: and that is done with reset, not checkout, right?
18:35 thiago neilthereildeil: the normal commands people use to obtain updates (git merge and git rebase) do not cope with those
18:36 impermanence rpd: Both are updated svn with varying degrees of pretty GUIs over top and expensive licensing.  So for 20 - 50 GBs of data versus 5K/yr it's like, uhhh...okay, how much do we really need this stuff...
18:36 thiago neilthereildeil: resetting or checking out, by themelves, are not bad commands. It's pushing the rewritten history for others, that is the source of problems.
18:36 GodGinrai impermanence: generally, git is for source, not the artifacts themselves.
18:36 Eugene rpd - I would disagree that git is the simplest; it's actually really complex, not user-friendly, has some performance problems(though is still much faster than previous-gen SCMs), and generally sucks, at least as far as I can tell from being in #git for a long time. git is used because 1) Linus uses it and 2) github is popular. That's really about it, but those are both very good reasons.
18:36 thiago neilthereildeil: you can do whatever you want with your copy
18:36 impermanence GodGinrai: right.  We're past this.
18:36 thiago neilthereildeil: but sharing code with others usually implie a "contract" of what you're going to do and what not to do
18:36 rpd Eugene: I actually agree with all that.
18:37 impermanence GodGinrai: I never asked about what source code management tool to use.
18:37 dreiss joined #git
18:37 rpd freinn: I'm not sure if I'm reading `D ` the same way. To be that means: deleted in the index; unchanged in the work tree
18:38 neilthereildeil whats the correct way to discard all working tree changes and have the working tree be commit X?
18:38 Eugene The original intent of git(from Linus) is that other people would build a friendly UI that uses the git protocol. This never really materialized; Github/Bitbucket/Gitlab/etc arguably come close
18:38 GodGinrai impermanence: Yea, I was just pointing out that if you are trying to use it for something that isn't source code management, it would be best to consider a different tool
18:38 rpd freinn: which sort of makes the ?? redundant because that means the same thing. You deleted it from git but it still exists in your work tree
18:38 neilthereildeil i would usually use reset --hard, but you guys are saying that causes problems
18:38 Eugene neilthereildeil - !resets
18:38 gitinfo neilthereildeil: tl;dr of man git-reset: --soft moves HEAD, --mixed moves HEAD+index, --hard moves HEAD+index+work-tree, -- foo.txt will !unstage
18:38 impermanence GodGinrai: Use "what" for something that isn't source code management?
18:39 freinn thanks rpd
18:39 Eugene Just be aware that --hard will trash uncommitted work. If you don't want that, consider git-stash
18:39 Eugene "I want to put away what I'm doing and work on something else" would be: git stash; git checkout newbranch <somecommit>.
18:40 Eugene Sorry, typo. `git checkout -b newbranch <somecommit>`. My brain isn't all the way on this morning
18:40 rpd freinn: Oh I see. You deleted it from you cached files. I think that it should be more clear; it is confusing.
18:40 Topic for #git is now Welcome to #git, the place for git help and ancient riddles | Public logs at http://goo.gl/BuUi5o | Current stable version: 2.9.0 | First visit? Read: http://jk.gs/git | Getting "cannot send to channel"? /msg gitinfo .voice | Why did the commit cross the rebase? To git to the other repo
18:41 rpd You can also just not --hard, if you don't want to trash your uncommitted changes I guess
18:41 pioiytr joined #git
18:41 GodGinrai impermanence: It depends on what you're storing.  There's a lot of solutions.  I don't store build artifacts all that often, so I couldn't really tell you objectively which is the best to use.
18:41 rpd it seems like `git stash; git reset --hard; git stash pop` would be the same thing as just doing `git reset`
18:42 Eugene Yup, TMTOWTDI
18:42 pioiytr joined #git
18:42 impermanence GodGinrai: okay.  Yes, I suppose build artifacts is the most correct way to describe what I'm storing.  Build artifacts and related bits that SysOps puts in there...
18:43 romerocesar_ joined #git
18:43 Bombe rpd, stash/reset/pop is pretty much a no-op.
18:44 fuchstronaut joined #git
18:45 GodGinrai impermanence: It looks like Apache has its own repo software for build artifacts: https://archiva.apache.org/index.cgi
18:46 impermanence GodGinrai:  Juuuuust looking at that now, coincidentally.  Though it seems somewhat inactive?
18:46 GodGinrai impermanence: Yea, I just found it, so I'm not sure how active it is
18:46 impermanence GodGinrai: wait, no it doesn't.
18:46 impermanence GodGinrai: Last rel was may 30th.
18:47 impermanence GodGinrai: But doooeeees it do binaries well?  Because that would be crazy and definitely the sleeper hit of the year!
18:47 GodGinrai I like the "our source code is now using git" post XD
18:47 EduardoMartins joined #git
18:48 Balliad joined #git
18:48 GodGinrai impermanence: well, if it can't do binaries well, then I don't see why *anyone* would use it
18:49 romerocesar joined #git
18:49 GodGinrai Since binaries are the most common build artifact
18:50 circ-user-A26a8 joined #git
18:50 thiago why would you store artifacts in the first place?
18:50 fracting joined #git
18:51 thiago if they can be recreated from source code, keep the source, not the artifacts
18:51 stantonk joined #git
18:51 GodGinrai thiago: I'm generally of this opinion, but sometimes it is useful to keep at least some history of build artifacts for quick reverts
18:53 diegoviola joined #git
18:53 cdg joined #git
18:53 GavinMagnus left #git
18:54 impermanence thiago: You know, that is a good question in the first place.  I think that for some (big) companies, teams get spread apart and important files sometimes get stored by other teams and blah blah blah...I think that's what's going on.  Because you're absolutely right...why are we storing RPMs in the first place?  Or WARs...
18:54 zacts joined #git
18:54 cdg joined #git
18:55 netj joined #git
18:56 meLon joined #git
18:56 GavinMagnus joined #git
18:57 shaggycat joined #git
19:00 neilthereildeil thiago: why do you say git reset changes a branch? i am reading an article, and it says reset only changes head, index,working dir. none of those have been committed yet and are therefore not in the branch
19:00 ekinmur joined #git
19:00 impermanence thiago: Uhh...do you consider a jar a build artifact?
19:01 editshare-ericf joined #git
19:01 zumba_ad_ joined #git
19:02 ojacobson can you regenerate it as-is from other tracked assets (source code, build scripts, etc)?
19:02 impermanence mm hm.
19:03 GavinMagnus left #git
19:03 ynanm joined #git
19:03 impermanence See I asked Thiago that because he was saying he prefers to not store build artifacts.  So is that typical...?  You just store source and then build everything you need at whatever time you need it?
19:03 ojacobson There are good reasons to archive and publish rebuildable assets like build output, but not in source control directly
19:04 GavinMagnus joined #git
19:04 impermanence totally
19:04 ojacobson Java folks tend to use maven's repo system as a distribution mechanism for built artefacts
19:04 ojacobson which makes sense; a "maven repo" is a dumb tree of files, so you can throw something up on a web server pretty easily if your needs are modest
19:04 impermanence Right.  And...I guess for me: DevOps, it's nice to have builds around.  Helps deployment...I think, lol...
19:04 ojacobson (there's metadata in the tree, but you don't need sophisticated software to host it)
19:05 impermanence to your "thow something up on a web server"...exactly...
19:05 impermanence see...I don't handle the build tools...
19:05 impermanence Don't know them really...
19:06 impermanence So...I guess sometimes we get builds from people and store them.
19:06 impermanence We don't do this with stuff we're deploying every week or whatever, obviously.
19:06 impermanence Shit, I don't know.  I need to dig into our binary storage much more deeply to resolve this.
19:07 impermanence Lunch.
19:07 impermanence I genuinely appreciate the help and insight.
19:08 romerocesar_ joined #git
19:09 Fosha joined #git
19:10 zeemz joined #git
19:10 thiago impermanence: yes
19:10 thiago neilthereildeil: --hard changes the branch
19:11 thiago neilthereildeil: git reset has three modes, and those changes what it changes
19:11 thiago neilthereildeil: --soft changes the commit, but not the index or working tree
19:11 thiago neilthereildeil: --mixed (default) changes the commit and the index, but not the working tree
19:11 thiago neilthereildeil: --hard changes all three
19:12 impermanence thiago: so you wouldn't store any jars?  If you were to need one you would just compile one?
19:13 thiago impermanence: yes
19:13 impermanence thiago: I see.
19:13 impermanence thiago: thank you.
19:13 thiago that's different from making a release. When we do that, we upload the built binaries to the download server. Those are meant to stay there forever.
19:13 thiago but that has nothing to do with Git
19:15 impermanence thiago: I see.  So a build does not necessarily equal a release.  I'm fairly ignorant of software engineering as an enterprise practice and system, in all honesty.  So that makes sense...
19:16 WaReZ joined #git
19:16 dreiss joined #git
19:16 thiago impermanence: right
19:16 impermanence thiago: okay.
19:16 thiago impermanence: now, your project may have a need to keep some artifacts from non-release builds for some time, like the ability to track down bugs, sharing among multiple systems, etc.
19:17 thiago but the point is that all of that is beyond Git. Git wasn't designed for that and will do a poor job at it if you try.
19:17 thiago there are some extra commands that might be useful, like git-lfs or git-annex. I've never tried them, though.
19:18 impermanence thiago: I'm chatting with an engineer not far from my team now and here's why he says he's using Artifactory (artifact storage): "i am using it to host all our maven artifacts we produce and cache the one we use from public repo."
19:18 thiago ok, there you go
19:18 thiago even if I have no idea who maven is
19:18 impermanence hilarious.
19:19 impermanence thiago: Right.  I know that.  I came here though because I sensed that this would be a fitting place to gain some insight about how people are storing things like binaries, etc.
19:19 thiago I'm guessing that's a software project named after a person name, like travis, jenkins and gerrit do
19:19 crayon joined #git
19:19 crayon joined #git
19:20 impermanence I don't know why the people behind it called it Maven...
19:20 impermanence There is a reason, though...
19:20 thiago impermanence: I know we do need some temporary artifact storage in our project. When building module B, it needs module A to have been compiled.
19:20 thiago one option is to build it again, but that takes time (over an hour for the slower machines in the farm)
19:21 thiago another is to cache the build artifacts and simply unpack them when doing B's integration
19:21 impermanence thiago: hm.
19:21 thiago I don't know what's used, though. And considering that it goes across VMs too.
19:21 GavinMagnus left #git
19:21 thiago you could ask the QA team tomorrow. It's late now in Finland.
19:22 impermanence thiago: Are you in Finland yourself?
19:22 Limes_ joined #git
19:22 thiago no, US West coast
19:23 impermanence Mm.  Finland: good socialism anyways.
19:23 cs02rm0 joined #git
19:23 thiago great schools
19:23 impermanence Well, maybe I will ask the QA team tomorrow/tonight whenever the appropriate Finland time is.
19:23 impermanence yeah, no kidding.  At least from what I have seen/heard/wathced
19:23 impermanence watched
19:24 zumba_ad_ joined #git
19:24 impermanence I'm asking the engineer on my end about why he stores all those jars.  He says: "because they are dependent on each other"...
19:24 thiago ping tosaraja during Finnish working hours (GMT+3 this time of the year)
19:24 Balliad joined #git
19:24 romerocesar joined #git
19:24 impermanence thiago: I will do that indeed.
19:24 flaguy48 joined #git
19:24 thiago wait, tomorrow is Saturday
19:25 impermanence right, doh.  But Monday.  Monday it's on.
19:26 impermanence thiago: Anyway, the RPM my engineer spits out at the end is like almost 2 GBs so it has tons of dependencies.  He says he wants them all cached and immediately available.  "We are not going to rebuild everything each time we need them."
19:26 pioiytr joined #git
19:27 EvilPeng1 joined #git
19:29 zumba_ad_ joined #git
19:29 frem_ joined #git
19:30 catag87 joined #git
19:30 ajf- joined #git
19:31 catag87 joined #git
19:31 ajf- I wonder if I can do "git checkout BRANCH -- $FILES" where FILES holds files that are in BRANCH but were changed within the last month
19:32 thiago ajf-: are you asking if that could be scripted? Yes
19:32 sangy joined #git
19:33 ajf- the pickle is that
19:33 thiago step 1: find the commit that corresponds to "one month ago"
19:33 phantomtiger joined #git
19:33 thiago then git checkout BRANCH -- `git diff --name-only $ONE_MONTH_AGO BRANCH`
19:34 thiago if your reflog is good, you can use BRANCH@{one.month.ago}
19:35 anddam left #git
19:35 ajf- error: pathspec '...'  did not match any file(s) known to git.
19:35 ajf- for all
19:35 ajf- do I need to prepend a ./ to the paths for the git diff ?
19:35 thiago no
19:35 thiago but you need to run that command from the root of the repostory
19:36 ajf- I am in the root
19:36 thiago start by looking at what git diff --name-only printed
19:36 thiago is it just a series of file names?
19:36 ajf- yes and they all seem right to me
19:37 thiago ok, and if you do git checkout BRANCH -- <paste the name of one of the files from the output>
19:37 thiago does it print that error message?
19:37 fission6 joined #git
19:38 oich if I change the commit that a submodule is based on and then checkout a different branch in the same working tree, git sees submodule as changed in the branch that I checked out. How can I recover from this so that the submodule stays at the commit it was based on, other than by creating a new working tree?
19:38 ajf- yes thiago and it works well
19:38 thiago ajf-: then there's just something wrong with the scripting
19:39 thiago oich: your description most likely does not match reality
19:39 thiago oich: please describe the sequence of commands. I can't tell what you did in either repository from your description.
19:40 moritz ajf-: do any of your file names contain spaces?
19:40 ajf- nope
19:40 SwiftMatt joined #git
19:41 thiago does the list include files that were deleted?
19:41 thiago that script won't work with deletions
19:41 lrp joined #git
19:41 ajf- it could. yes that must be it
19:42 jaguarmagenta joined #git
19:43 ajf- alright let me try all this again. with diff I should be able to get a list of files that changed between two commits in the branch stage
19:43 thiago add --diff-filter=AM
19:43 ajf- ah ok
19:43 ajf- but wait this is not right ..
19:43 fission6 how can i see what a file looked like 10 days ago
19:44 moritz thiago, ajf-: if you're checking out the older version, shouldn't added file be a problem?
19:45 ajf- I think I am not planning this right
19:45 moritz fission6: git show HEAD{10 days ago}:the/file
19:45 thiago fission6: git show <commit of ten days ago>:<filename>
19:45 moritz erm, forgot ao @
19:45 moritz fission6: git show HEAD@{10 days ago}:the/file
19:45 oich thiago thanks. I'll do it from memory if it doesn't make sense I'll do it again an describe the sequence. In repository A, git checkout mainA; git submodule add -b somebranch someremote here/there; git branch mainB; git checkout mainB; (elsewhere chnages are made to someremote branch somebranch); git submodule update --remote; git commit -a -m 'message'; git checkout mainA; git status (at...
19:45 ekinmur joined #git
19:45 oich ...this point git says that here/there has been modified
19:45 moritz uhm, needs quotes
19:46 moritz git show "HEAD@{10 days 10}":thefile
19:46 moritz now it actually works
19:46 EvilPeng1 joined #git
19:47 thiago oich: the part you glossed over in "elsewhere changes are made" is important
19:47 thiago oich: from your description without it, the submodule was unchanged until that git submodule update
19:48 yqt joined #git
19:48 oich thiago ok git clone someremote; git checkout somebranch; echo x > x; git add x; git commit -m 'message' (then back to repository A and the git submodule update --remote
19:48 romerocesar_ joined #git
19:51 flaguy48 joined #git
19:52 thiago oich: ok, thanks
19:53 tvw joined #git
19:53 thiago oich: that git submodule update should modify the submodule to point back at the commit that the super module expects it to be
19:53 thiago oich: git status, at the end, should not say "changed"
19:54 neilthereildeil joined #git
19:54 neilthereildeil thiago: u there?
19:54 oich thiago yes. if I clone into a new working tree selecting the branch without the updated submodule, it doesn't say th esubmodule is changed
19:54 romerocesar joined #git
19:55 DolphinDream i'm tying to define a config alias to list the common files modified by two branches.. but my alias is malformed .. any ideas what's the problem here:  lcf= "!f() { A=$1; C=$2; B=`git merge-base $A $C`; comm -12 <(git diff --name-only $B...$A | sort) <(git diff --name-only $C...$A | sort); }; f"
19:55 oich if I checkout the branch without the updated module, in the same working tree, then it says the submodule is changed
19:56 neilthereildeil thiago: what i was asking is how git reset causes pain to others if it doesnt change branches. you said it only changes HEAD ptr (just like checkout), index and working tree
19:56 zumba_ad_ joined #git
19:57 thiago oich: ok, that's probably right. If you check out another commit without changing the submodule, the commit you checked out may be asking for a different commit in teh submoudle
19:57 thiago neilthereildeil: it changes branches
19:57 neilthereildeil you mean the branches stored in history?
19:57 thiago neilthereildeil: it changes history too
19:57 neilthereildeil i read this article: https://git-scm.com/blog
19:57 neilthereildeil about git reset
19:58 ShekharReddy joined #git
19:58 neilthereildeil it said reset just changes HEAD ptr, index and working tree, all of which get changed during normal operation also
19:58 DolphinDream Seveas, howdy.. any ideas why my alias does not work ? (list common files touched by two branches topic)
19:58 thiago it changes the commit, not just HEAD
19:58 thiago I mean, it changes the commit that the branch points to
19:59 rpd It changes the head to point to a different commit
19:59 rpd which you can see in your reflog
20:00 jfran joined #git
20:00 neilthereildeil thiago: ohh so if a branch pointed to 0xABC, reset would make it point to 0xKLM?
20:00 neilthereildeil like a different commit on which it is based?
20:01 m0viefreak joined #git
20:01 neilthereildeil rpd: yea, but commit changes HEAD also. that shouldnt be problematic for others
20:01 Balliad joined #git
20:02 rpd If you think of every commit as a different state of the index and HEAD@{0} as a pointer to the latest index, git reset changes HEAD@{0} to point to a different state
20:02 rpd I also don't think git reset is problematic for others, just clarifying what it does
20:02 oich thiago so, starting with supermodule branch A I update the submodule with changes, so it's hash is AAAAAAA. before that it's hash was BBBBBB. now if I checkout branch B git says the submodule is modified. if Instead, I clone the supermodle specifying -b B, git doesn't complain and the submodule's hash is the original: BBBBBB
20:03 davisonio joined #git
20:03 ekinmur joined #git
20:04 neilthereildeil rpd: ohh ok, so when i change HEAD@{0} and do git log, it would show the latest commit being different in the history?
20:04 rpd correct and you can look at the real chain of commits by using git reflog
20:04 neilthereildeil ahh ok
20:04 Asiajey joined #git
20:05 neilthereildeil rpd: whats the best way to throw away all changes (tracked and untracked) in my working tree and make it be whatever a certain commit was?
20:05 thiago oich: ok
20:06 thiago neilthereildeil: git reset --hard thatcommit
20:06 neilthereildeil thiago: but doesnt that change the git branch history, causing trouble for others?
20:06 thiago neilthereildeil: with history rewrite. You can't do that if you want to share this change with others.
20:06 thiago neilthereildeil: yes
20:07 neilthereildeil ahh
20:07 neilthereildeil whats a better way to do it if i am sharing?
20:07 thiago neilthereildeil: if you want to share, then do git reset --hard (discards all your changes) then git checkout -f thatcommit
20:07 thiago not sure how well that copes with deleted files, though. You should test.
20:07 vuoto joined #git
20:08 neilthereildeil thiago: whats the point of the reset step there?
20:10 ochorocho__ joined #git
20:10 jstimm joined #git
20:12 jstimm joined #git
20:12 IWishIKnew joined #git
20:12 AfroThundr joined #git
20:13 mablae joined #git
20:14 boombatower joined #git
20:15 hahuang61 joined #git
20:16 refried_ joined #git
20:17 jeffreylevesque joined #git
20:17 oich thiago I tried it again to make sure. is there a way to recover so that the submodule uses the commit it was added as in that branch, aside from cloneing a new working tree or without deleting all files and doing git checkout -- `git ls-files -d`. The reason being that people other than me have to deal with this repository
20:19 lindelabs joined #git
20:20 tonyj joined #git
20:22 romerocesar_ joined #git
20:22 flaguy48 joined #git
20:22 pur3eval joined #git
20:23 neilthereildeil also, how can i switch to another branch without losing my uncomitted/untracked changes?
20:23 thiago oich: git submodule update
20:23 neilthereildeil ill need to switch back later
20:23 thiago neilthereildeil: commit
20:23 GodGinrai neilthereildeil: `git stash`
20:23 neilthereildeil ok
20:23 neilthereildeil how can i unstash?
20:23 GodGinrai `git stash pop`
20:25 troulouliou_dev joined #git
20:25 neilthereildeil if someone copied my git tree, will all the stashing/unstashing be copied also?
20:25 neilthereildeil or does pop leave git .git as it was before?
20:26 mwhooker joined #git
20:26 osse the latter
20:26 osse in practice, but it's recoverable for a while
20:26 oich thiago that doesn't work. It updates to the same commit as in the other branch
20:26 Electrometro joined #git
20:26 gkatsev joined #git
20:27 ensyde joined #git
20:28 muelli joined #git
20:28 jimi_ joined #git
20:28 stantonk joined #git
20:29 GodGinrai neilthereildeil: pop will "pop" the stashed changes off of the stack and reapply them
20:29 GodGinrai neilthereildeil: if you want to apply the changes, but leave them on the stack, you use `git stash apply`
20:29 neilthereildeil ok thanks
20:29 neilthereildeil thiago: in your command earlier, can i just use the git checkout -f? why do i need the reset --hard?
20:29 GodGinrai although there's not many reasons to want to keep changes on the stack you've already applied
20:30 ekinmur joined #git
20:31 stantonk joined #git
20:31 LeMike joined #git
20:33 smithbone joined #git
20:34 notcharlie joined #git
20:35 neilthereildeil also, im trying to branch
20:35 neilthereildeil i used git branch HASH Name
20:35 neilthereildeil why is this wrong?
20:35 ahmedelgabri joined #git
20:36 _ikke_ git branch <name> [starting-point]
20:36 DolphinDream howdy _ikke_
20:37 DolphinDream any  ideas why my alias does not work ? (list common files touched by two branches topic)
20:37 DolphinDream i'm tying to define a config alias to list the common files modified by two branches.. but my alias is malformed .. any ideas what's the problem here:  lcf= "!f() { A=$1; C=$2; B=`git merge-base $A $C`; comm -12 <(git diff --name-only $B...$A | sort) <(git diff --name-only $C...$A | sort); }; f"
20:38 DieguezZ joined #git
20:39 _ikke_ DolphinDream: note that with 3 dots, you don't explicitly need merge-base
20:40 DolphinDream when adding this as an alias... i get Syntax error: "(" unexpected
20:41 Balliad joined #git
20:42 neilthereildeil osse: were you saying the reflog periodically cleaned up?
20:42 _ikke_ what kind of alias?
20:42 DolphinDream git config alias
20:42 Balliad joined #git
20:42 DolphinDream lcf= "!f() { A=$1; C=$2; B=`git merge-base $A $C`; comm -12 <(git diff --name-only $B...$A | sort) <(git diff --name-only $C...$A | sort); }; f"
20:43 DolphinDream to be used as git lcf branchA branchB
20:43 dopesong joined #git
20:44 flaguy48 joined #git
20:44 shinnya joined #git
20:45 fuchstronaut joined #git
20:47 _ikke_ I'm not sure why it does not like the <() part
20:47 _ikke_ https://stackoverflow.com/questions/31827146/how-to-use-a-parenthesis-character-inside-a-command-substitution-inside-a-bash-s
20:48 Literphor joined #git
20:48 thiago <() is process substitution
20:48 thiago make sure you have that feature enabled in your shell
20:48 _ikke_ yes
20:48 _ikke_ right, and the default one git uses (sh) does not support it
20:49 DolphinDream hm..
20:49 romerocesar joined #git
20:49 lindelabs joined #git
20:49 _ikke_ git config alias.test '!bash -c "f() { cat <(git show HEAD --name-only); }; f"'
20:49 dopesong joined #git
20:49 _ikke_ That works for me
20:49 ekinmur joined #git
20:50 neilthereildeil i did git checkout on an older commit, and now i want to go back to a different commit, but i dont see that commit in git log. how can i find the name of the newer commits?
20:51 _ikke_ git log --all
20:51 dopesong joined #git
20:51 _ikke_ git log by default only shows commits reachable from the currently checked out commit
20:51 lindelabs left #git
20:51 lindelabs joined #git
20:52 The_Phoenix joined #git
20:52 ahmed_elgabri joined #git
20:53 DolphinDream lcf='!bash -c "f() { A=$1; C=$2; B=`git merge-base $A $C`; comm -12 <(git diff --name-only $B...$A | sort) <(git diff --name-only $C...$A | sort); }; f"'
20:53 lindelabs left #git
20:53 lindelabs joined #git
20:53 DolphinDream tells me that it does not recognises this as a git command
20:53 englishstudent joined #git
20:53 DolphinDream Expansion of alias 'lcf' failed
20:54 dopesong joined #git
20:54 GodGinrai what the hell is that monster alias?
20:54 GodGinrai also, why is this an alias instead of a function
20:54 Faylite_ joined #git
20:54 GodGinrai I mean... you define a function in the alias
20:54 _ikke_ GodGinrai: arguments
20:55 GodGinrai _ikke_: functions take arguments
20:57 _ikke_ DolphinDream: I get a usage for merge-base
20:57 pioiytr joined #git
20:59 lindelabs left #git
20:59 lindelabs joined #git
20:59 samson joined #git
21:00 pioiytr joined #git
21:01 pioiytr joined #git
21:01 lindelabs left #git
21:02 LeMike joined #git
21:03 lindelabs joined #git
21:03 CheckDavid joined #git
21:04 dsdeiz joined #git
21:04 dsdeiz joined #git
21:07 Darcidride joined #git
21:07 DolphinDream _ikke_, huh? what usage?
21:07 GodGinrai DolphinDream: seriously.  Why are you defining a function in an alias instead of just using a function in the first place?
21:08 nilg joined #git
21:08 DolphinDream GodGinrai, what do you mean ?
21:09 ojacobson GodGinrai: git has no concept of a function
21:09 GodGinrai ojacobson: bash does.
21:09 ojacobson Sure. But he's not writing inside of bash.
21:09 ojacobson He's writing inside of a Git alias.
21:09 DolphinDream create a separate git script that does the thing ?
21:09 GodGinrai ojacobson: And my point is that he *should*
21:09 nopf joined #git
21:09 ojacobson mu
21:09 ojacobson Git will happily find scripts named `git-foo` as Git commands (`git foo`) if they're on $PATH or in git's libexec dir
21:09 DolphinDream perhaps i should define a bash script to do that whole shabang and createa an git alias to invoke the script
21:10 GodGinrai a separate script works as well, if you feel more comfortable with that
21:10 diegoviola joined #git
21:10 ojacobson aliases, freestanding commands, etc. are all valid ways to package behaviour
21:10 doebi joined #git
21:10 ojacobson For shell-ish git aliases, the convention of defining the alias as '!f() { …body… }; f' is useful because of the way Git expands argument lists for aliases
21:10 DolphinDream so i don't even need to create an alias if i create a script prefiexed with git-  ?
21:11 GodGinrai DolphinDream: sounds like it
21:11 ojacobson `git my-alias one two three` will expand to '…; f one two three'
21:12 DolphinDream btw ojacobson .. this alias does not work due to some git/sh exapnsion problem it seems: lcf="f() { A=$1; C=$2; B=`git merge-base $A $C`; comm -12 <(git diff --name-only $B...$A | sort) <(git diff --name-only $C...$A | sort); }; f" .. while adding !bash -c to it to enforce bash it doesn't seem to work at my end.
21:12 GodGinrai ojacobson: that's overcomplicating things.  If what you need to do is complex enough you *need* a function, you should be using a script or actual bash functions outside of git
21:12 ojacobson GodGinrai: I'm glad your life is so simple. :)
21:13 GodGinrai ojacobson: I'm sure you like making things hard for yourself :)
21:13 ojacobson I use Git.
21:13 kbs heh
21:13 ojacobson All else is trivial by comparison.
21:14 Sasazuka joined #git
21:14 ojacobson DolphinDream: There's at least a missing ! on your alias, and your quoting looks wildly wrong to me
21:15 * kbs waits for someone to write the git equivalent of doug crockford's "javascript - the good parts"
21:16 GodGinrai kbs: "the good parts" considered harmful
21:16 GodGinrai :P
21:16 liamonade joined #git
21:17 DolphinDream ojacobson, sorry.. yeah. the ! is actually in my config
21:17 kbs :-) I've slowly come to like the elegance of the underlying data model, and have convinced myself it's not Stockholm syndrome...
21:17 patarr joined #git
21:17 ojacobson heh.
21:17 Dougie187 left #git
21:18 GT4066 joined #git
21:18 GodGinrai kbs: the data model for git?
21:18 ojacobson Broadly, the worst thing about Git is the lack of abstraction, followed shortly by the highly modal UI behaviour
21:18 ojacobson (how many distinct behaviours can you get out of git-checkout? I know of at least four)
21:18 DolphinDream so for git scripts.. i added the path to my envvar ... but git still does not find the command .. i named my file git-lcf .. and i expec that git lcf branchA branchB would work
21:18 texinwien_ joined #git
21:18 GodGinrai as a vim user, a modal interface doesn't really bother me
21:18 ojacobson DolphinDream: does `which git-lcf` find your script?
21:18 kbs GodGinrai: yea - I finally grok'ed git by pretty much ignoring all the commands and more or less pawing through .git/objects
21:19 DolphinDream ojacobson, yep
21:19 GodGinrai kbs: impressive
21:19 DolphinDream ojacobson,  i can run the script directly git-lcf branchA branchB but not as git lcf
21:19 lindelabs left #git
21:20 Rodya_ joined #git
21:21 phoe6 joined #git
21:22 elastix joined #git
21:23 pk-john joined #git
21:23 Literphor joined #git
21:25 GavinMagnus joined #git
21:26 azbarcea joined #git
21:26 rgrinberg joined #git
21:26 nnyk joined #git
21:26 GodGinrai DolphinDream: so the script *is* executable and in your path?  (What you are saying sounds like it is, but I want to confirm)
21:27 ojacobson Cannot reproduce; putting a script named `git-lcf` in my PATH (in ~/Console/bin, because of how I have PATH set up) causes `git lcf` to run that script
21:28 osse neilthereildeil: i wasn't, but it is
21:28 stantonk joined #git
21:28 phoe6 left #git
21:30 xaviergmail joined #git
21:32 spudowiar joined #git
21:35 aidalgol joined #git
21:39 Rodya_ joined #git
21:39 romerocesar_ joined #git
21:40 ochorocho__ joined #git
21:43 jaguarmagenta joined #git
21:43 mischat joined #git
21:43 ahr3n joined #git
21:43 d5sx43 joined #git
21:43 m_ben joined #git
21:45 Xnu17 joined #git
21:46 stantonk joined #git
21:46 hahuang61 joined #git
21:47 akiva joined #git
21:49 romerocesar joined #git
21:49 King_Hual joined #git
21:49 King_Hual joined #git
21:49 clemf_ joined #git
21:53 cdown joined #git
21:54 minn joined #git
21:54 ferdereine joined #git
21:57 clemf joined #git
21:57 Eryn_1983_FL joined #git
21:57 minn I'm a terrible person and don't know how to use git. I cloned a git repository and started working on it without knowing what I was doing (with git, that is). Now I need to push all those changes to a new branch. Assume I will learn git over the weekend but need to do this soon. How would I go about doing it?
21:58 clemf joined #git
21:58 SwiftMatt joined #git
22:00 romerocesar_ joined #git
22:00 cs02rm0 joined #git
22:01 rpd minn: about learning or pushing the changes to a new branch?
22:01 rpd minn: not sure what the folks here think, but I really like the docs on https://git-scm.com/
22:02 minn There's lots of great resources for learning. I'm mostly concerned with creating a new branch and pushing to our gitlab repository. I think I can create a branch with checkout -b, commit, and push that, but it's not entirely clear to me that this won't have any negative effects for anybody else (because I'm not familiar with the tools yet).
22:03 Eugene That is the entirety of the process, yup.
22:03 Eugene !float talks about making a new branch with uncommited work
22:03 gitinfo If you have made a change in your working directory and have NOT YET COMMITTED, you may "float" that change over to another (`git checkout oldbranch`) or new (`git checkout -b newbranch`) branch and commit it there.  If the files you changed differ between branches, the checkout will fail.  In that case, `git stash` then checkout, and `git stash apply` and go through normal conflict resolution.
22:04 Eugene `git push origin mybranch -u` will send the new branch
22:04 joki joined #git
22:04 rpd eww who puts the -u at the end like that =P
22:05 m_ben instead of using git-mv to move some files and directories in a repository, I accidentally moved them using the shell command mv. now all previous commits of these files are lost. I already committed these changes, and also starting modifying some of the files. what is the quickest way to reverse this process so that I can view all previous commits?
22:05 minn That appears to be exactly what I need - thanks, Eugene and rpd
22:05 Eugene I like to put the dangerous flags at the end, eg `rm /etc/important.conf.backup -rf` can't accidentally break the wrong thing when my cat steps on the keybaord
22:07 minn I've always aliased dangerous flags I use all the time to avoid blanking and typing the wrong thing, but that's a much better solution.
22:08 Tndg joined #git
22:10 romerocesar joined #git
22:11 clemf joined #git
22:12 ochorocho__ joined #git
22:15 liamonade joined #git
22:18 safe joined #git
22:19 titanium17 joined #git
22:20 alkino joined #git
22:20 texinwien_ joined #git
22:21 davisonio joined #git
22:22 fishcooker joined #git
22:22 HoierM joined #git
22:24 cdown joined #git
22:26 cdown_ joined #git
22:27 ash_workz joined #git
22:27 rpd m_ben: you could just rebase back to before you mv'd the files and git-mv them instead then go ahead and re-apply your new patches
22:28 m_ben rpd: thanks. I didn't know about the --follow option. it seems with this option I do get the entire commit history of a file
22:29 fishcooker im new on SCM esp on git... yes it is the first time. Actually im on python, i've noticed the issue section on the git esp gitlab. Yes i know that when the commit message contains issue number it will close the issue... let's say if i have an old issue.. then when i try to fix it, i got another issue appears... is it possible to make the previous issue as root cause of the next issue. what is the best practice of using the git if we build the python sc
22:29 fission6 joined #git
22:30 GodGinrai fishcooker: what do you mean "as the root".  Git doesn't do issues.  Other software built around git does
22:31 neilthereildeil left #git
22:32 fishcooker GodGinrai: the first issue as root cause of the another issue.. yes im using gitlab... it means that the gitlab handles the issue, even TODO?
22:33 rpd fishcooker: like GodGinrai says, you're asking Gitlab or Github, not about Git. If another issue shows up you can mark it as a duplicate of an existing issue or w/e. The best practice is up to your team. People usually just comment on the issue with some like "related to issue ###" which will post a comment on the "root cause" that it is being referenced.
22:33 GavinMagnus left #git
22:33 fishcooker ic your poin, rpd
22:33 rpd but that's all convention, just do what you want; you can always close these issues or link them together manually
22:34 ensyde joined #git
22:34 romerocesar joined #git
22:34 liamonade joined #git
22:34 fishcooker that's why i ask the best practice of them, rpd
22:34 Literphor joined #git
22:35 pioiytr joined #git
22:35 madewokherd joined #git
22:35 fishcooker but thanks rpd, i've just realized that git there is nothing to do with issues and TODOs
22:35 saidelike joined #git
22:36 fishcooker GodGinrai: 2
22:36 rpd fishcooker: yeah it's unrelated but here's Github's conventions: https://guides.github.com/features/issues/. Read the section on Notifications, @mentions, and references
22:38 pioiytr joined #git
22:39 refried_ joined #git
22:40 cdg joined #git
22:44 BlaXpirit joined #git
22:44 titanium17 joined #git
22:45 nettoweb joined #git
22:45 spudowiar joined #git
22:46 fuchstronaut joined #git
22:46 zumba_addict joined #git
22:46 zumba_addict is it possible to bypass git hooks during git commit?
22:46 ojacobson --no-verify
22:46 zumba_addict thanks
22:47 ojacobson that'll ignore pre-commit, at least
22:47 zumba_addict cool
22:48 zumba_addict it worked. Thanks :)
22:50 d4rklit3 joined #git
22:51 Literphor joined #git
22:51 ffabi joined #git
22:53 kbs out of mostly curiosity - besides "git" the C application, are there other widely-used implementations of git-as-a-network+file-protocol? (I know of jgit, and afaict it's used primarily by gerrit.)
22:54 Seveas there's libgit2 and its binding
22:54 Seveas bindings*
22:55 kbs aah, interesting thanks for the pointer.
22:56 kbs oh I see, so github/gitlab et. al. actually use this - that makes sense, come to think of it :-)
22:56 EvilPeng1 joined #git
23:00 Garen joined #git
23:02 akushner joined #git
23:03 clemf joined #git
23:06 Cyclohexane joined #git
23:06 Eugene Dulwich is also of note
23:08 jim_carry joined #git
23:08 romerocesar joined #git
23:13 romerocesar_ joined #git
23:14 kbs Eugene: interesting thanks. Any major users of this code?
23:16 Eugene https://github.com/search?utf8=%E2%9C%93&amp;q=import+dulwich&amp;type=Code&amp;ref=searchresults
23:16 troulouliou_dev joined #git
23:18 kbs fair enough :)
23:21 DieguezZ joined #git
23:23 acetakwas joined #git
23:23 romerocesar joined #git
23:26 Gianormagantrous joined #git
23:28 lalilulelo joined #git
23:29 ffabi joined #git
23:32 segment-fault joined #git
23:32 OnkelTem joined #git
23:32 Rodya_ joined #git
23:38 romerocesar_ joined #git
23:38 dmto joined #git
23:41 pioiytr joined #git
23:42 pioiytr joined #git
23:43 pioiytr joined #git
23:43 jstimm joined #git
23:43 jaguarmagenta joined #git
23:44 jcadduono joined #git
23:44 jcadduono is there a way to make git am act like patch (resolve line # conflicts automatically)
23:45 stantonk joined #git
23:46 jcadduono oh just realized ic ould just git am, patch -p1, git am --continue
23:47 jcadduono still would be cool to have it all in 1 command
23:47 ffabi joined #git
23:47 hahuang61 joined #git
23:47 paws_ joined #git
23:49 mablae_ joined #git
23:50 johnnyfive joined #git
23:52 mafrasi2 joined #git
23:55 pioiytr joined #git
23:56 kulelu88 left #git
23:59 fracting joined #git
23:59 ffabi joined #git
23:59 xaviergmail joined #git
23:59 akushner joined #git

| Channels | #git index | Today | | Search | Google Search | Plain-Text | summary