| 2010-01-11 |
| → xiong joined | 00:00 |
|
_Vi
| osmosis, If you want to limit it to root, use "/*.pyc". | 00:00 |
|
osmosis
| ok | 00:00 |
| ← daifuku left | 00:00 |
| → loinclot_ joined | 00:00 |
| ← loincloth left | 00:00 |
|
osmosis
| is there a quick command to add all modified files to the next commit, or do I have to add each of the files each time? | 00:00 |
|
_Vi
| osmosis, "git commit -a", "git add" with some parameter. | 00:01 |
| → daifuku joined | 00:01 |
|
_Vi
| osmosis, (It doesn't add new files) | 00:01 |
| → Slix joined | 00:01 |
|
osmosis
| got it | 00:01 |
| → abadr joined | 00:01 |
| ← rudi_s_ left | 00:02 |
| ← Paraselene_ left | 00:03 |
| ← robinr left | 00:04 |
|
sako
| Hey all, anyone have any comments about this setup? | 00:04 |
|
| http://joemaller.com/2008/11/25/a-web-focused-git-workflow/ | 00:04 |
|
| i am going to use it for version controlling my website projects going forward.. | 00:04 |
|
| is this the "suggested" way of using git with websites | 00:04 |
| ← DavidKlein left | 00:05 |
|
_Vi
| sako, You may also use such thing without "Prime" repository. | 00:05 |
|
RobertLaptop
| sako I don't know if it is the suggested way but it looks pretty close to the way I do it. | 00:06 |
|
_Vi
| sako, (Running "git archive" each time for push to production branch). | 00:06 |
|
sako
| yea actually, i am trying to figure out if i need these 2 repos | 00:06 |
| ← JasonWoof left | 00:06 |
| → JasonWoof1 joined | 00:06 |
|
_Vi
| sako, But if they are no separate machines then better to use that your scheme. | 00:06 |
|
| s/no /on / | 00:06 |
| JasonWoof1 → JasonWoof | 00:07 |
|
nevyn
| sako: I do something similar with an update-incomming hook and non-bare central repo | 00:07 |
| ← JasonWoof left | 00:07 |
| → JasonWoof joined | 00:07 |
| ← a7x left | 00:07 |
|
nevyn
| sako: actually my workflow is undocumented and enormous (dev -> UAT -> prod etc) | 00:07 |
|
sako
| nevyn: does this setup look decent then? | 00:08 |
|
| seems easy enough to implement | 00:08 |
|
| i just dont really know if there is redundency like the 2 repo deal | 00:09 |
| ← NetHawk left | 00:09 |
| ← lbayle left | 00:09 |
|
sako
| i'm coming from svn | 00:09 |
| → Paraselene_ joined | 00:09 |
|
_Vi
| sako, Are you running "hub" and "prime" on the same installation? | 00:09 |
|
sako
| i will be yes | 00:09 |
| → NetHawk joined | 00:09 |
|
sako
| ah i found out the reason why the author suggests this.. | 00:10 |
|
| never push into a repository that has a work tree attached to it.. | 00:10 |
|
_Vi
| sako, Do you plan storing big blobs in the repository (non-code such as images/sounds/videos)? | 00:10 |
|
sako
| possibly | 00:11 |
|
_Vi
| sako, It is bad already having big or generated file in repository and may be especially bad if they are mixed with normal code. | 00:12 |
|
| sako, But if repositories duplication is an issue, you can make them to know about each other and share some information. | 00:12 |
|
| sako, But if there are only normal code and text files (and not too big history), then repository is usually smaller that checkout. | 00:13 |
|
nevyn
| I don't get why pushing to a non-bare repo is so bad if it's the only mechinism for updates to that repo | 00:14 |
|
sako
| http://git.or.cz/gitwiki/GitFaq#head-b96f48bc9c925074be9f95c0fce69bcece5f6e73 | 00:14 |
|
nevyn
| I don't see how prime get's you anything over a hook that updates the working tree. | 00:14 |
| → catalis joined | 00:14 |
|
nevyn
| but all the problems with it exist with prime anyway | 00:15 |
|
| if you change somethign in prime and don't commit/push it to hub then you've got the same problem | 00:15 |
| ← daifuku left | 00:15 |
|
sako
| hmm | 00:15 |
|
| what is the working tree anyway | 00:15 |
|
_Vi
| sako, It is files outside repository, but tracked by it. | 00:16 |
| ← ronron2 left | 00:16 |
| ← rettub left | 00:16 |
| ← ph^_ left | 00:16 |
|
nevyn
| sako: the stuff you edit. | 00:17 |
|
| sako: right but as that says.. | 00:17 |
| → daifuku joined | 00:17 |
| ← ereslibre left | 00:17 |
| ← kmartiin left | 00:17 |
|
nevyn
| if you do need a remote that updates when you push to it use the linked hook code which deals with everything except a pre-existing conflict in the working tree on the remote | 00:17 |
| → pkw joined | 00:18 |
|
nevyn
| so if you know what you're doing and have hooks to update the working tree from pushes then there's no reason to add the bare repo | 00:18 |
| → rettub joined | 00:19 |
| → Cromulent joined | 00:20 |
| → engla joined | 00:20 |
| ← psoo left | 00:23 |
| ← jelmer left | 00:25 |
| ← catalis left | 00:26 |
| → godsflaw joined | 00:26 |
| ← Cromulent left | 00:26 |
| ← bentob0x left | 00:27 |
| → hadees joined | 00:28 |
| ← Slix left | 00:28 |
|
hadees
| i'm doing git commit -m but nothing is being checked in | 00:28 |
|
| i do git status and i still see the changes | 00:28 |
|
KiBi
| hadees: git add | 00:28 |
| → Slix joined | 00:29 |
|
hadees
| KiBi, i did that already | 00:30 |
|
| they aren't being committed for some reason | 00:30 |
|
| KiBi, Changes to be committed is what the files are under | 00:30 |
| → hobodave joined | 00:31 |
| ← daifuku left | 00:31 |
|
bob2
| what does git commit do then? | 00:32 |
|
| succeed but commit an empty changeset? | 00:32 |
| → daifuku joined | 00:32 |
|
bob2
| or fail with an error? | 00:32 |
|
| or... | 00:32 |
| ← kipras left | 00:32 |
|
hadees
| bob2, empty set i guess | 00:33 |
|
| [root@localhost gambit]# git commit -m "Checkin" | 00:33 |
|
| [ealter_paypal_oneclick 3a302c9] Checkin | 00:33 |
| ← _Vi left | 00:33 |
| ← godsflaw left | 00:33 |
|
bob2
| I don't know what that means | 00:34 |
|
| what does 'git log -p' say afterwards? | 00:34 |
|
hadees
| bob2, they are commits but no files | 00:34 |
| ← v0n left | 00:34 |
|
hadees
| *there | 00:35 |
| → catalis joined | 00:35 |
| → _Vi joined | 00:36 |
| ← thiago_home left | 00:37 |
| ← abadr left | 00:37 |
| → Chillance joined | 00:43 |
|
hadees
| i found one thing that might shed light on this | 00:44 |
|
| fatal: git diff [--no-index] takes two paths | 00:44 |
|
| when i did git diff | 00:44 |
| ← vbgunz left | 00:44 |
| ← daifuku left | 00:46 |
| → daifuku joined | 00:47 |
| → Cromulent joined | 00:47 |
| ← ereslibre_laptop left | 00:48 |
| ← loinclot_ left | 00:50 |
| → loincloth joined | 00:50 |
| ← mjf left | 00:50 |
| → [1]awnstudio joined | 00:52 |
| → r11t_ joined | 00:52 |
|
hadees
| argh that wasn't the problem | 00:52 |
|
| i really can't figure out why it isn't commiting | 00:52 |
| ← engla left | 00:53 |
| ← nickrw left | 00:53 |
| ← catalis left | 00:54 |
| ← r11t left | 00:56 |
| ← daifuku left | 00:59 |
| ← scarabx left | 01:01 |
| → daifuku joined | 01:01 |
| → oliver3 joined | 01:01 |
| → flaguy48 joined | 01:03 |
| ← awnstudio left | 01:04 |
| [1]awnstudio → awnstudio | 01:04 |
| → tekky joined | 01:04 |
| ← remi left | 01:06 |
| ← tobias__ left | 01:06 |
| → vbgunz joined | 01:06 |
| ← imajes left | 01:07 |
| ← aspotashev left | 01:10 |
| ← JayM left | 01:13 |
| → jelmer joined | 01:13 |
| ← xiong left | 01:13 |
| ← daifuku left | 01:15 |
| → daifuku joined | 01:17 |
| → numberGrey joined | 01:19 |
| ← numberGrey left | 01:19 |
| → juan_arandaalvar joined | 01:19 |
| → spacebat_ joined | 01:25 |
| → sudoer joined | 01:27 |
| ← sudoer left | 01:29 |
| ← daifuku left | 01:29 |
| → daifuku joined | 01:31 |
| → deschet joined | 01:31 |
| ← rudi_s left | 01:32 |
| ← priidu left | 01:37 |
| ← drigz left | 01:37 |
| → pheaver_ joined | 01:38 |
| ← spacebat left | 01:38 |
| → cedricv joined | 01:40 |
| ← pheaver left | 01:41 |
| pheaver_ → pheaver | 01:41 |
| ← pheaver left | 01:43 |
| → swisherdude joined | 01:47 |
| ← swisherdude left | 01:47 |
| → loinclot_ joined | 01:54 |
| ← loincloth left | 01:54 |
| ← NathanielJ left | 01:59 |
| ← osmosis left | 02:01 |
| → DrAk0 joined | 02:01 |
| ← daifuku left | 02:01 |
| → daifuku joined | 02:02 |
| ← loinclot_ left | 02:02 |
| → loincloth joined | 02:02 |
| ← jwoods_ left | 02:02 |
| ← loincloth left | 02:06 |
| → loincloth joined | 02:06 |
| → rmancy joined | 02:07 |
| ← hgimenez_ left | 02:08 |
| ← warthog9 left | 02:10 |
| ← docgnome left | 02:13 |
| → docgnome joined | 02:13 |
| → s1kx- joined | 02:15 |
| → project2501a joined | 02:15 |
| → tedoc2000 joined | 02:15 |
| ← coolcat left | 02:17 |
| ← orafu left | 02:18 |
| → orafu joined | 02:18 |
| ← smuf left | 02:20 |
| → imajes joined | 02:20 |
| → osmosis joined | 02:21 |
| ← Cromulent left | 02:22 |
| → VVSiz_ joined | 02:24 |
|
sako
| hey all is it ok to clone a git repo into another one? | 02:24 |
|
| i have a git repo setup, but i want to get a project which is version controlled by git as well | 02:24 |
|
| so i would get the repo by git clone project2 and that will live in project1 which is a git repo itself | 02:25 |
|
| problem is if i am in project1 and do a git-update, will it go and update project2? | 02:25 |
| → mahesh joined | 02:26 |
| ← S1kx left | 02:26 |
| → keystr0k joined | 02:26 |
| → hummingkid joined | 02:28 |
| ← DrAk0 left | 02:28 |
|
sako
| anyone? | 02:29 |
|
| is it possible to have a git repo inside another git repo? | 02:29 |
| ← keystr0k left | 02:29 |
| → agenteo joined | 02:29 |
| ← sako left | 02:30 |
| → mhagger joined | 02:30 |
| ← osmosis left | 02:31 |
| → volk joined | 02:31 |
|
volk
| before every commit, do i have to "git add filename" then "git commit" if I modified the filename? | 02:32 |
| ← daifuku left | 02:32 |
|
agenteo
| isn't git checkout ID FILE supposed to "update" the file with the ID version? | 02:32 |
| → sako joined | 02:32 |
|
sako
| anyone alive? :) | 02:32 |
|
Khisanth
| volk: you can also do git commit filename | 02:33 |
| → daifuku joined | 02:33 |
|
sako
| hey guys is it ok to do a git clone projectname of a git version controlled project into another repo? | 02:33 |
|
| shouldn't be an issue right? | 02:33 |
| → cloudhead joined | 02:33 |
| ← volk left | 02:34 |
| → upul joined | 02:37 |
| ← Aldwuin left | 02:39 |
| ← loincloth left | 02:40 |
| → loincloth joined | 02:40 |
| ← imajes left | 02:41 |
|
agenteo
| when I do a git show ID FILE I can see a few changes, but git checkout ID FILE, doesn't seem to actually update the file | 02:41 |
| ← VVSiz left | 02:42 |
| ← vbgunz left | 02:43 |
|
mhagger
| I have a simple patch series mh/rebase-fixup that Junio has merged to "next" but is waiting on further steps. Now I'm working on cleanup etc in the neighborhood, followed by the further steps. Is it preferred that I reroll the old patch series or that I submit an additional series? And if the latter, what should the new patch series apply against? | 02:45 |
| ← daifuku left | 02:45 |
| → rchavik joined | 02:45 |
| ← project2501a left | 02:47 |
| ← aliasre left | 02:47 |
| → daifuku joined | 02:47 |
| → kerx joined | 02:49 |
| ← parasti left | 02:51 |
| → pheaver joined | 02:52 |
| ← segher left | 02:53 |
| ← pheaver left | 02:53 |
| → remi joined | 02:53 |
| → coderdad joined | 02:53 |
|
coderdad
| I have a file that I want to revert to what I had a 4 commits back. I tried git reset --hard but it doesn't accept paths | 02:54 |
|
| how can I bring that version of the file up | 02:54 |
|
| I know the commit it was in | 02:55 |
| ← bauruine left | 02:56 |
|
agenteo
| coderdad that's similar to what I am trying to do | 02:56 |
|
| try git checkout ID FILE | 02:56 |
| ← jelmer left | 02:58 |
|
coderdad
| thanks agenteo | 02:59 |
| → ricky_ joined | 03:00 |
|
agenteo
| give it a go, I am not sure why it's not working for me. Maybe cause I did it from a branch :) let me know how it goes | 03:00 |
|
coderdad
| it ran, but it still looks to be the most recent file | 03:00 |
| ← ricky_ left | 03:00 |
|
coderdad
| haha same place here | 03:00 |
|
| but that shouldn't matter | 03:00 |
|
agenteo
| eventually I used git show ID FILE to see the diff, and changed it manually | 03:01 |
| → aliasre joined | 03:02 |
| ← daifuku left | 03:03 |
| ← simNIX left | 03:03 |
| → daifuku joined | 03:05 |
| → JayM joined | 03:05 |
|
coderdad
| mine is an xib file for iphone | 03:05 |
|
| all generated and crazy | 03:05 |
| ← hummingkid left | 03:06 |
| ← fcuk112 left | 03:06 |
| ← EricInBNE left | 03:06 |
| ← Twisted_Mentat left | 03:06 |
| ← gdaa left | 03:06 |
| ← ange left | 03:06 |
| ← ilteris_ left | 03:06 |
| ← kukks left | 03:06 |
| ← maattd left | 03:06 |
| ← cantoma left | 03:06 |
| ← tetha left | 03:06 |
| ← Fissure left | 03:06 |
| ← GodEater left | 03:06 |
| ← sdboyer left | 03:06 |
| ← Cesario left | 03:06 |
| ← vinse left | 03:06 |
| ← jast left | 03:06 |
| ← bgerber left | 03:06 |
| ← robinbowes left | 03:06 |
| ← johnjay_ left | 03:06 |
| ← __bert__ left | 03:06 |
| ← peper left | 03:06 |
| ← fualo left | 03:06 |
| ← MacGyverNL left | 03:06 |
| ← htns left | 03:06 |
| ← Khisanth left | 03:06 |
| ← bbeecher left | 03:06 |
| ← Arrowmaster left | 03:06 |
| ← diegoviola left | 03:06 |
| ← blast_hardcheese left | 03:06 |
| ← kate_afk left | 03:06 |
| ← milli left | 03:06 |
| ← madmoose left | 03:06 |
| ← ozzloy left | 03:06 |
| ← cdngfrog left | 03:06 |
| ← MIR100 left | 03:06 |
| ← drfree left | 03:06 |
| ← jaalto left | 03:06 |
| ← dermoth left | 03:06 |
| ← _rane left | 03:06 |
| ← bremner left | 03:06 |
| ← redondos left | 03:06 |
| ← xinit left | 03:06 |
| ← shad left | 03:06 |
| ← Risar left | 03:06 |
| ← binjured left | 03:06 |
| ← awx left | 03:06 |
| ← ChenKaie left | 03:06 |
| ← dominikh left | 03:06 |
| ← Sigma left | 03:06 |
| ← lamont left | 03:06 |
| ← GustavoRamos left | 03:06 |
| ← argonel left | 03:06 |
| ← Rhonda left | 03:06 |
| ← Cope left | 03:06 |
| ← Tass` left | 03:06 |
| ← Spark left | 03:06 |
| ← Tommy[D] left | 03:06 |
| ← remi left | 03:06 |
| ← cloudhead left | 03:06 |
| ← Paraselene_ left | 03:06 |
| ← ruediger left | 03:06 |
| ← ProperNoun left | 03:06 |
| ← RaceCondition left | 03:06 |
| ← riz2 left | 03:06 |
| ← jailbox left | 03:06 |
| ← keseldude left | 03:06 |
| ← bry left | 03:06 |
| ← pgalbraith left | 03:06 |
| ← cenuij left | 03:06 |
| ← cytrinox left | 03:06 |
| ← phadviger left | 03:06 |
| ← ft left | 03:06 |
| ← SolarWar left | 03:06 |
| ← DrNick left | 03:06 |
| ← sunaku left | 03:06 |
| ← thiago left | 03:06 |
| ← perk left | 03:06 |
| ← jksz left | 03:06 |
| ← harv left | 03:06 |
| ← mebus left | 03:06 |
| ← numbah left | 03:06 |
| ← thijso left | 03:06 |
| ← Bucciarati left | 03:06 |
| ← Ilari left | 03:06 |
| ← pcgod left | 03:06 |
| ← floppyears left | 03:06 |
| ← levifig left | 03:06 |
| ← ruxpin left | 03:06 |
| ← ruskie left | 03:06 |
| ← jacobat left | 03:06 |
| ← mediaslave left | 03:06 |
| ← ph^ left | 03:06 |
| → johnw joined | 03:06 |
| ← rchavik left | 03:07 |
| → remi joined | 03:07 |
| → cloudhead joined | 03:07 |
| → Paraselene_ joined | 03:07 |
| → ruediger joined | 03:07 |
| → ProperNoun joined | 03:07 |
| → RaceCondition joined | 03:07 |
| → riz2 joined | 03:07 |
| → jailbox joined | 03:07 |
| → keseldude joined | 03:07 |
| → bry joined | 03:07 |
| → Tommy[D] joined | 03:07 |
| → pgalbraith joined | 03:07 |
| → cenuij joined | 03:07 |
| → cytrinox joined | 03:07 |
| → phadviger joined | 03:07 |
| → perk joined | 03:07 |
| → SolarWar joined | 03:07 |
| → ft joined | 03:07 |
| → DrNick joined | 03:07 |
| → sunaku joined | 03:07 |
| → thiago joined | 03:07 |
| → ph^ joined | 03:07 |
| → mediaslave joined | 03:07 |
| → levifig joined | 03:07 |
| → jacobat joined | 03:07 |
| → numbah joined | 03:07 |
| → pcgod joined | 03:07 |
| → jksz joined | 03:07 |
| → ruskie joined | 03:07 |
| → floppyears joined | 03:07 |
| → ruxpin joined | 03:07 |
| → harv joined | 03:07 |
| → Ilari joined | 03:07 |
| → Bucciarati joined | 03:07 |
| → thijso joined | 03:07 |
| → mebus joined | 03:07 |
| → hummingkid joined | 03:08 |
| → fcuk112 joined | 03:08 |
| → EricInBNE joined | 03:08 |
| → Twisted_Mentat joined | 03:08 |
| → gdaa joined | 03:08 |
| → ange joined | 03:08 |
| → kukks joined | 03:08 |
| → maattd joined | 03:08 |
| → cantoma joined | 03:08 |
| → tetha joined | 03:08 |
| → Fissure joined | 03:08 |
| → GodEater joined | 03:08 |
| → sdboyer joined | 03:08 |
| → Cesario joined | 03:08 |
| → vinse joined | 03:08 |
| → jast joined | 03:08 |
| → blast_hardcheese joined | 03:08 |
| → cdngfrog joined | 03:08 |
| → bgerber joined | 03:08 |
| → robinbowes joined | 03:08 |
| → Cope joined | 03:08 |
| → diegoviola joined | 03:08 |
| → johnjay_ joined | 03:08 |
| → __bert__ joined | 03:08 |
| → peper joined | 03:08 |
| → fualo joined | 03:08 |
| → MacGyverNL joined | 03:08 |
| → htns joined | 03:08 |
| → Khisanth joined | 03:08 |
| → bbeecher joined | 03:08 |
| → xinit joined | 03:08 |
| → Arrowmaster joined | 03:08 |
| → Risar joined | 03:08 |
| → shad joined | 03:08 |
| → Sigma joined | 03:08 |
| → ozzloy joined | 03:08 |
| → _rane joined | 03:08 |
| → MIR100 joined | 03:08 |
| → argonel joined | 03:08 |
| → dominikh joined | 03:08 |
| → lamont joined | 03:08 |
| → Rhonda joined | 03:08 |
| → Spark joined | 03:08 |
| → bremner joined | 03:08 |
| → drfree joined | 03:08 |
| → kate_afk joined | 03:08 |
| → dermoth joined | 03:08 |
| → GustavoRamos joined | 03:08 |
| → milli joined | 03:08 |
| → redondos joined | 03:08 |
| → madmoose joined | 03:08 |
| → awx joined | 03:08 |
| → binjured joined | 03:08 |
| → jaalto joined | 03:08 |
| → ChenKaie joined | 03:08 |
| → Tass` joined | 03:08 |
| ← ruediger left | 03:08 |
| → taufiqkh joined | 03:10 |
| → Joe_Conyers joined | 03:12 |
| ← Joe_Conyers left | 03:12 |
| ← aziz left | 03:13 |
| ← kerx left | 03:13 |
| ← sako left | 03:14 |
| → JM joined | 03:15 |
| ← coderdad left | 03:15 |
| ← mhagger left | 03:15 |
| → coderdad joined | 03:15 |
| → simNIX joined | 03:15 |
| ← RobertLaptop left | 03:15 |
| ← JoeConyers left | 03:15 |
| → loinclot_ joined | 03:16 |
| ← loincloth left | 03:16 |
| → RobertLaptop joined | 03:17 |
| ← Bass10 left | 03:18 |
| ← aliasre left | 03:18 |
| ← loinclot_ left | 03:19 |
| → loincloth joined | 03:20 |
| ← mpr left | 03:20 |
| ← flaguy48 left | 03:22 |
| ← daifuku left | 03:25 |
| → cytrinox_ joined | 03:25 |
| → mpr joined | 03:27 |
| → daifuku joined | 03:27 |
| → loinclot_ joined | 03:28 |
| ← loincloth left | 03:29 |
| → sako joined | 03:31 |
| ← sako left | 03:31 |
| → sako joined | 03:32 |
| → jamesgolick joined | 03:33 |
|
jamesgolick
| is there any way to get git to tell me the total number of insertions and deletions in a commit range? | 03:34 |
| → jelmer joined | 03:34 |
| ← cedricv left | 03:36 |
| ← JM left | 03:36 |
| ← cytrinox left | 03:37 |
| cytrinox_ → cytrinox | 03:37 |
| ← gambler left | 03:39 |
| → vbgunz joined | 03:40 |
| ← justinweiss left | 03:44 |
| ← gdaa left | 03:44 |
| → pheaver joined | 03:44 |
| ← loinclot_ left | 03:46 |
| → loincloth joined | 03:46 |
| → candybar joined | 03:47 |
|
Ilari
| jamesgolick: 'git diff --stat <commit1> <commit2>'? | 03:48 |
|
jamesgolick
| nice | 03:48 |
|
| Ilari: thx | 03:48 |
|
| i had whipped out awk | 03:48 |
|
| with --numstat | 03:49 |
| ← LiamH left | 03:50 |
| → tjaway joined | 03:52 |
| ← daifuku left | 03:52 |
| ← remi left | 03:52 |
| → daifuku joined | 03:53 |
| ← Adlai left | 03:53 |
| → loinclot_ joined | 03:54 |
| ← swisherpop left | 03:54 |
| ← loincloth left | 03:55 |
| ← kukks left | 03:55 |
| ← tekky left | 03:55 |
| → Adlai joined | 03:55 |
| ← Haarg left | 03:56 |
| → schmukkk joined | 03:58 |
| ← rgr left | 03:58 |
| ← EricInBNE left | 03:58 |
| → rgr joined | 03:58 |
| ← pheaver left | 04:00 |
| ← daifuku left | 04:00 |
| → swisherpop joined | 04:00 |
| → daifuku joined | 04:01 |
| → EricInBNE joined | 04:01 |
| ← Twisted_Mentat left | 04:01 |
| ← swisherpop left | 04:02 |
| → stringo0 joined | 04:03 |
| ← stringo0 left | 04:03 |
| → stringo0 joined | 04:03 |
| ← stringo0 left | 04:04 |
| → stringo0 joined | 04:04 |
| ← stringo0 left | 04:04 |
| → stringo0 joined | 04:04 |
| → swisherpop joined | 04:04 |
| ← stringo0 left | 04:05 |
| → stringo0 joined | 04:05 |
| ← sako left | 04:05 |
| → khmarbaise joined | 04:07 |
| ← tjafk left | 04:08 |
| → bauruine joined | 04:09 |
| → ideogon joined | 04:11 |
| → godsflaw joined | 04:12 |
| ← khmarbaise left | 04:12 |
| ← daifuku left | 04:12 |
| ← Tureba left | 04:13 |
| → daifuku joined | 04:14 |
| ← daifuku left | 04:15 |
| → warthog9 joined | 04:21 |
| ← hummingkid left | 04:21 |
| → hummingkid joined | 04:21 |
| → catalis joined | 04:23 |
| ← amerine left | 04:25 |
| ← mahesh left | 04:28 |
| → amerine joined | 04:29 |
| → kssddy joined | 04:30 |
| → pheaver joined | 04:32 |
| ← loinclot_ left | 04:35 |
| ← coderdad left | 04:39 |
| → sumangurung joined | 04:39 |
| → coderdad joined | 04:39 |
| ← hummingkid left | 04:40 |
| ← m104 left | 04:42 |
| → mdel joined | 04:42 |
| → sako joined | 04:43 |
|
Slix
| Alright.. so I put some changes into my master branch. | 04:44 |
|
| I have a feature branch though. My feature branch needs those changes. | 04:44 |
|
| What's the best way to get those onto my feature branch? | 04:45 |
|
| I've heard of rebasing a lot, but am unsure of what that exactly means. | 04:45 |
| → m104 joined | 04:45 |
| ← rettub left | 04:50 |
| → ricky joined | 04:50 |
| ← d2dchat left | 04:51 |
| → rettub joined | 04:53 |
|
offby1
| Slix: time to read various tutorials :-| | 04:53 |
|
| in the meantime ... | 04:53 |
|
| just "git checkout feature; git merge master" | 04:53 |
|
| actually, the answer is, "it depends" :-| | 04:53 |
|
| probably not what you wanted to hear | 04:54 |
| → herlo joined | 04:54 |
| ← rgr left | 04:55 |
| → rgr joined | 04:55 |
|
Slix
| Yeah, I know I have to read more tutorials. | 04:55 |
|
| But it's 10:55pm and I can't go to sleep until I finish some hw. :P | 04:55 |
|
herlo
| hello all, would I ask gitosis config questions here? Or is there a better channel?? | 04:56 |
|
Slix
| Will the tutorials explain the "it depends"? | 04:56 |
|
Ilari
| herlo: Gitosis is obsolete, but yes, ask here. | 04:57 |
| → pinnacle joined | 04:57 |
|
herlo
| Ilari: what do you use in place of gitosis then? | 04:58 |
|
| anyway, I have a fun error that I haven't been able to identify a quick fix | 04:58 |
|
| http://pastebin.ca/1746490 | 04:58 |
|
Ilari
| herlo: Gitolite. But if you have working gitosis install, no real reason to change. | 04:58 |
|
herlo
| Ilari: I don't... | 04:58 |
|
offby1
| Slix: it depends on the tutorial :) | 04:58 |
|
| lemme see if I can help | 04:58 |
|
| Slix: is this repository just for you, or are you sharing it? | 04:59 |
|
Slix
| offby1: Just for me. | 04:59 |
|
offby1
| good! | 04:59 |
|
| then just merge. | 04:59 |
|
herlo
| Ilari: either way, any suggestions on the pastebin? | 04:59 |
|
Slix
| And then commit the merge? | 04:59 |
| → g3d joined | 04:59 |
|
offby1
| no need; merging commits automatically, unless you tell it not to | 04:59 |
|
Slix
| Ahh. | 04:59 |
|
offby1
| (that's the opposite of bzr, and maybe hg too) | 05:00 |
|
Ilari
| herlo: That kind of stuff would look like some user/repo name is misspelled. | 05:00 |
|
herlo
| yeah, don't think so... | 05:00 |
| coderdad → coderdad|away | 05:00 |
|
herlo
| Ilari: I've checked and doublecheked that many many times, since it's just gitosis-admin | 05:00 |
|
Ilari
| herlo: Can you also pastebin the gitosis configuration (~git/.gitosis.conf)? | 05:01 |
|
| kssddy: (That is the "active" configuration"). | 05:01 |
|
| Oops. | 05:01 |
|
| herlo: (That is the "active" configuration"). | 05:01 |
|
herlo
| Ilari: sure | 05:02 |
|
| Ilari: http://pastebin.ca/1746496 | 05:04 |
|
Ilari
| herlo: The right user is "herlo@herlo-lap", not "herlo". | 05:05 |
|
herlo
| I've tinkered with the 'herlo' and 'herlo@herlo-lap' | 05:05 |
|
| I knew you were going to say that :) | 05:05 |
|
| but I can change it back | 05:06 |
|
| though it doesn't seem to make a diff | 05:06 |
|
Ilari
| herlo: Edit the active configuration as you can't clone or push properly? | 05:06 |
|
herlo
| yeah, trying that again | 05:07 |
|
| strange, got it working. I swear I tried that | 05:07 |
|
| Ilari: thanks | 05:07 |
| ← schmukkk left | 05:08 |
|
Ilari
| herlo: BTW: On this computer, I run gitolite on top of git-daemon2 (no sshd at all!) | 05:09 |
|
herlo
| nice | 05:09 |
|
| Ilari: I'll have a look at gitolite | 05:09 |
|
Ilari
| herlo: gitolite is designed for ssh operation, but if underlying daemon looks like sshd to it... | 05:10 |
| ← vbgunz left | 05:10 |
| ← sumangurung left | 05:10 |
|
Ilari
| herlo: BTW: What wasn't nice was to have to security-patch it yesterday. :-/ | 05:11 |
| ← BigMack83 left | 05:11 |
| → d2dchat joined | 05:11 |
|
herlo
| ooh | 05:12 |
|
| not nice | 05:12 |
|
Ilari
| herlo: One could authenticated with forged keypair if one obtained the public key of user/server. | 05:14 |
|
Adlai
| is it possible to use git-svn to get just one branch of an SVN repo? | 05:14 |
|
Ilari
| *authenticate | 05:14 |
|
herlo
| Ilari: not good! | 05:15 |
| ← tedoc2000 left | 05:15 |
| ← rgr left | 05:15 |
| → DrAk0 joined | 05:16 |
| ← aresnick left | 05:17 |
|
sitaram
| Ilari: security patch gitolite? or ssh-daemon2? | 05:17 |
| → [1]awnstudio joined | 05:17 |
| → jceb joined | 05:21 |
| → project2501a joined | 05:22 |
| ← cloudhead left | 05:22 |
| → cloudhead joined | 05:22 |
| → amouge1 joined | 05:25 |
| coderdad|away → coderdad | 05:26 |
|
sitaram
| errm , I mean git-daemon2 | 05:27 |
| ← godsflaw left | 05:27 |
|
offby1
| Adlai: yep | 05:28 |
|
Adlai
| offby1, what if it involves externals? | 05:28 |
|
offby1
| I doubt externals work | 05:28 |
|
| the man page probably explains | 05:28 |
|
Adlai
| it just lists one command, show-externals. I guess it's worth a shot :\ | 05:30 |
| → cedricv joined | 05:30 |
| ← LRN left | 05:31 |
| ← jelmer left | 05:33 |
| → LRN joined | 05:33 |
| ← awnstudio left | 05:33 |
| [1]awnstudio → awnstudio | 05:33 |
| ← amerine left | 05:33 |
| ← sagsousuke left | 05:34 |
| ← pkw left | 05:34 |
| → jelmer joined | 05:35 |
| ← kssddy left | 05:35 |
| → hummingkid joined | 05:35 |
| ← sunaku left | 05:36 |
|
cxreg
| i'm still a little surprised to find that git-notes arent meant to be clone | 05:38 |
|
| d | 05:38 |
| ← amouge left | 05:40 |
|
Ilari
| sitaram: git-daemon2. | 05:41 |
|
sitaram
| Ilari: thanks; saw that bit about forged keypair later (juggling multiple things sorry) | 05:42 |
| → krynnotaur joined | 05:42 |
| → pkw joined | 05:43 |
|
Ilari
| sitaram: Actually not a proper fix, just tightens checks enough so that it can't be exploited. | 05:43 |
| → tekky joined | 05:43 |
| ← tekky left | 05:45 |
|
Adlai
| ok, I cloned a single SVN branch, worked ok. Tried show-externals, and it only lists the externals, but doesn't fetch them. Is there any way to fetch them into the right parts of the tree? | 05:45 |
| → tekky joined | 05:46 |
| → savant joined | 05:47 |
| ← savant left | 05:47 |
| ← jfkw left | 05:48 |
| ← marc_m left | 05:48 |
| ← blast_hardcheese left | 05:49 |
| → munichlinux joined | 05:51 |
|
wereHamster
| Adlai: nope | 05:52 |
| → Akkernight joined | 05:53 |
| → kssddy joined | 05:53 |
| ← hummingkid left | 05:54 |
|
Akkernight
| I'm trying to build a GIT repository of Wine, and after it says "=> Starting make..." it checks alot of stuff, but when it checks whether gcc -m32 works, it says no, and it says "Cannot build a 32-bit program, you need to install 32-bit development libraries", does anyone know how to help? | 05:54 |
|
cehteh
| what has that to do with git? | 05:55 |
|
johnw
| what OS? | 05:55 |
|
| cehteh build wine on 64bit debian some time ago .. you need to isntall some 32 bit devel stuff | 05:56 |
|
cehteh
| but thats bit of topic here, better ask in #wine | 05:56 |
|
johnw
| Akkernight: you'll need to look up how to install 32-bit devel libs for your platform | 05:56 |
| ← cloudhead left | 05:56 |
|
Akkernight
| johnw, I'm on Arch Linux and I have done pacman -S lib32. cehteh I thought it had something to do with Git because I'm using Git to get the repository :P | 05:57 |
|
cehteh
| nope thats not a git thing | 05:57 |
| ← jamesgolick left | 05:58 |
|
SeveredCross
| Arch FTL. | 05:58 |
|
wereHamster
| Akkernight: you are trying to build package X, so ask the people who are developing X. (head over to #winehq or #arch) | 05:59 |
|
Akkernight
| wereHamster, yeah I was already on both channels :P just thought this was a Git thing, sorry | 06:00 |
|
divVerent
| in a bare repo, how can I list which files exist? | 06:00 |
|
| git-ls-files shows nothing | 06:00 |
| ← lcpfnvc left | 06:01 |
|
divVerent
| (want to see why it's still 1.8GB after lots of branch filtering and deleting all the big stuff) | 06:01 |
| ← fishey left | 06:02 |
|
wereHamster
| git ls-tree -r | 06:02 |
|
divVerent
| ah, thanks | 06:03 |
|
| all the files appear to still be there :( | 06:03 |
|
| but well | 06:03 |
|
| seems like I got farther by that | 06:03 |
|
| another thing... what is recommended commit workflow when using git-svn and having the git repo published? | 06:03 |
|
| git svn dcommit rebases the branch, which I'd like to avoid as I then cannot continue working on it | 06:03 |
|
| what about git checkout master; git merge branch; git svn dcommit | 06:03 |
| → roop joined | 06:04 |
|
divVerent
| that still rebases, but would it be safe (as in, not rewriting PUBLISHED history)? | 06:04 |
|
| or do I then get into trouble when doing the next merge? | 06:04 |
| ← giallu left | 06:04 |
|
divVerent
| basically, assume this: two branches be fully merged, i.e. merging A to B does nothing and B to A does nothing. | 06:05 |
|
| then I rebase A. | 06:05 |
|
| and merge A into B | 06:05 |
|
| will that do anything nasty? | 06:06 |
| → solofight joined | 06:06 |
|
Ilari
| divVerent: Rebase what branch? And with what as base? | 06:06 |
|
divVerent
| well | 06:06 |
|
| assume I have a development branch xyz | 06:06 |
|
| and a master branch, which is a mirror of svn trunk | 06:06 |
|
| assuem that, in branch xyz, I did git merge master | 06:07 |
|
| then in branch master, I do git merge xyz, then git svn dcommit (which rebases master on the previous head of master) | 06:07 |
| → Chillance_ joined | 06:07 |
| ← solofight left | 06:07 |
|
Ilari
| divVerent: That would mess up git svn if you want to commit development back... | 06:07 |
|
divVerent
| what if, then in branch xyz, I do git merge master? | 06:07 |
|
| Ilari: why would it? | 06:07 |
|
Ilari
| divVerent: Merges and SVN don't mix. | 06:07 |
|
divVerent
| okay, so the only way to commit to svn is to git svn dcommit from a SEPARATE branch, then throwing away the branch? | 06:08 |
| → solofight joined | 06:08 |
|
divVerent
| note that the development branches are not kept in svn, only the master/trunk is | 06:08 |
|
Ilari
| divVerent: Don't dcommit any merges unless you really know what you are doing. | 06:09 |
|
divVerent
| which also means I really don't want to rebase the development branches | 06:09 |
|
| well, to know what I am doing, what breakage should I expect? | 06:09 |
|
| I assume it rebases all the commits from the merge on top of HEAD | 06:09 |
|
Ilari
| divVerent: Dcommits to wrong branch, git svn refusing to dcommit, etc... | 06:09 |
|
divVerent
| which by itself sounds fine to me | 06:09 |
| → jackdempsey joined | 06:09 |
|
divVerent
| the thing I don't like about dcommit is the rebasing... | 06:10 |
|
| what if I want to dcommit a branch to master/trunk | 06:10 |
|
Ilari
| divVerent: wrong commits being dcommitted, commits being lost, etc... | 06:10 |
|
divVerent
| but want to keep working on the branch | 06:10 |
|
| if I dcommit it, I rebased it, which means I really shouldn't push it | 06:10 |
|
| so assume I reset back to origin/branchname after dcommit, undoing the rebase | 06:11 |
|
Ilari
| divVerent: Basically, if you push some commit, it must be already dcommitted, or it can never be dcommitted. | 06:11 |
|
| divVerent: Sorrt, SVN is just so limited. | 06:11 |
|
divVerent
| why? | 06:11 |
|
| I don't get it | 06:11 |
|
| I want to keep development branches in git | 06:11 |
|
| and the main one in svn | 06:11 |
|
| so the only branch I ever dcommit _to_ is svn's trunk | 06:11 |
|
Ilari
| divVerent: git-svn is not distributed. | 06:12 |
|
divVerent
| sure, I know that | 06:12 |
|
johnw
| divVerent: I do what you just described all the time | 06:12 |
|
divVerent
| but how does it hinder me from doing rbanches in a published git repository? | 06:12 |
|
johnw
| you want to publish your topic branches? | 06:12 |
|
divVerent
| exactly | 06:12 |
|
johnw
| that won't work with git-svn | 06:12 |
|
divVerent
| basically... I know that git svn dcommit from a branch pushes on top of svn trunk | 06:12 |
|
johnw
| because it rebases after every dcommit | 06:12 |
|
| in order to record the metadata with the commit | 06:13 |
|
divVerent
| sure, which basically means I currently dcommit only if I know I dont need the topic branch any more | 06:13 |
|
| i.e. after a dcommit, I delete the topic branch | 06:13 |
| → numberGrey joined | 06:13 |
|
johnw
| either that, or rebase your topic branches and tell everyone who consumes from you to use git pull --rebase | 06:13 |
|
divVerent
| I just wonder... can I maybe avoid that? | 06:13 |
|
| assume I do this: | 06:13 |
|
| 1. dcommit from topic branch to svn trunk | 06:13 |
|
| 2. fetch and merge this into the master branch (containing the svn metadata) | 06:14 |
| ← sako left | 06:14 |
|
divVerent
| 3. reset the topic branch to the previous, un-rebased, head | 06:14 |
|
| 4. merge master into the topic branch | 06:14 |
|
| what will happen then? | 06:14 |
|
johnw
| you'll see "duplicated" commits in your history | 06:14 |
|
| one version with metadata, and one without | 06:14 |
|
divVerent
| fine with that | 06:14 |
|
| will it cause any trouble? | 06:14 |
|
| like millions of merge conflicts? | 06:14 |
|
johnw
| not post-merge, no | 06:14 |
| → ankit9 joined | 06:14 |
|
divVerent
| okay, so that means I can actually do that | 06:15 |
|
| clumsy workflow, maybe, but sounds like a plan | 06:15 |
|
johnw
| it should work as you imagine. your consumers will pull your merge commit and then be updated | 06:15 |
|
| it's gonna make for some pretty funky trees in gitk though | 06:15 |
|
divVerent
| so basically,a duplicate commit usually causes no real harm when merging | 06:15 |
|
| but is basically skipped | 06:15 |
|
| i.e. git merge notices the change is already in, and skips it | 06:15 |
|
johnw
| yeah, it should be ok | 06:15 |
|
divVerent
| and the other side of the question... | 06:15 |
|
| git svn dcommit will know that the commits to push next start at the last dcommit | 06:16 |
|
| despite the merge? | 06:16 |
|
johnw
| there may be cases where, if a merge conflict does occur for some reason, you could end up having to fix it twice | 06:16 |
|
| but I can't think of a case offhand | 06:16 |
| ← jackdempsey left | 06:16 |
|
divVerent
| well, I can mostly prevent that by, before pushing, pulling | 06:16 |
|
| (and making sure the git-svn-managed master branch is current too) | 06:16 |
|
johnw
| dcommit should get things right, because the metadata will be on the first parent | 06:16 |
|
divVerent
| also... for this weird git-svn workflow... I'd like to add a commit hook on server side that prevents pushing to any git-svn managed branch | 06:17 |
|
| just in case I mess up :P | 06:17 |
|
| how to know if git-svn manages a branch? | 06:17 |
|
johnw
| good idea | 06:17 |
|
| i'd list them in a file if I were you | 06:17 |
|
divVerent
| as nobody ever creates new branches on svn anyway, I can safely do that and get no conflicts | 06:17 |
|
| so there is no command do do it, okay | 06:18 |
| ← Swimming_Bird left | 06:18 |
|
divVerent
| actually... will do _positive_ matching then | 06:20 |
|
| basically, I will require branches to be prefixed with a username and a slash for being pushable | 06:20 |
| ← onigiri_ left | 06:20 |
| ← __bert__ left | 06:20 |
| → __bert__ joined | 06:21 |
| ← __bert__ left | 06:21 |
| → hyperair joined | 06:22 |
| ← kssddy left | 06:23 |
| ← Chillance left | 06:23 |
| → hummingkid joined | 06:23 |
| ← g3d left | 06:24 |
| → girishr joined | 06:24 |
| ← girishr left | 06:24 |
| → kssddy joined | 06:28 |
| ← niki left | 06:30 |
| ← catalis left | 06:33 |
| ← jceb left | 06:34 |
| → amerine joined | 06:36 |
| → sako joined | 06:37 |
| ← ideogon left | 06:37 |
| ← madewokherd left | 06:40 |
| → robinr joined | 06:40 |
| ← hummingkid left | 06:42 |
| ← wolog left | 06:42 |
| ← juan_arandaalvar left | 06:43 |
| → marc_m joined | 06:46 |
| ← maattd left | 06:50 |
| ← robinr left | 06:52 |
| ← coderdad left | 06:53 |
| ← Chillance_ left | 06:54 |
| ← mdel left | 07:00 |
| → bugfux joined | 07:00 |
| ← deschet left | 07:00 |
| ← Octalot left | 07:01 |
| ← krynnotaur left | 07:04 |
| → juan_arandaalvar joined | 07:04 |
| ← agenteo left | 07:06 |
| ← Adlai left | 07:06 |
| ← rmancy left | 07:06 |
| r11t_ → r11t | 07:11 |
| ← Bartman007 left | 07:17 |
| ← warthog9 left | 07:18 |
| ← bugfux left | 07:19 |
| → benbalbo joined | 07:19 |
| ← Slix left | 07:22 |
| → petrux joined | 07:22 |
| → Jak_o_Shadows joined | 07:24 |
| → dolphinling joined | 07:25 |
|
dolphinling
| Hi, in the config option remote.<name>.fetch, what's the syntax for multiple <refspec>s? git help fetch doesn't say. | 07:27 |
|
Ilari
| dolphinling: Just have multiple settings by that name. | 07:27 |
|
dolphinling
| Awesome, thanks | 07:28 |
|
Ilari
| dolphinling: Some settings make sense specified multiple times (such as remote URL and fetch/push lines). | 07:29 |
| ← kristian42 left | 07:33 |
| ← SunilThaha left | 07:33 |
| → ntoll joined | 07:35 |
| → giallu joined | 07:35 |
| → ideogon joined | 07:37 |
| → psoo joined | 07:38 |
| ← project2501a left | 07:39 |
| ← kssddy left | 07:39 |
| → thiago_home joined | 07:39 |
| → project2501a joined | 07:39 |
| → hummingkid joined | 07:40 |
| → warthog9 joined | 07:41 |
| → sde_antrix joined | 07:41 |
| → bentob0x joined | 07:44 |
| → Axius joined | 07:45 |
| ← pheaver left | 07:47 |
| ← adamm_ left | 07:48 |
| → lucsky joined | 07:49 |
| ← lucsky left | 07:51 |
| → lucsky joined | 07:51 |
| → SunilThaha joined | 07:51 |
| → kssddy joined | 07:53 |
| ← benbalbo left | 07:54 |
| ← hummingkid left | 07:58 |
| ← r11t left | 07:58 |
| ← johan-s left | 07:59 |
| ← NET||abuse left | 07:59 |
| ← DrAk0 left | 08:01 |
| → nico- joined | 08:02 |
| → kkoehne joined | 08:02 |
| → akar joined | 08:03 |
| ← nico- left | 08:03 |
| → nico- joined | 08:04 |
|
divVerent
| someone know a good way to do a "cronjob" working on a git repository on sourceforge? | 08:06 |
|
| from their web server I cannot access the SCM mounts | 08:06 |
|
| so next idea would be something ugly like a git update hook | 08:06 |
|
| and a server that regularily pushes to a dummy ref | 08:06 |
|
| #just, I find that so ugly, I rather ask if anyone here knows a better way | 08:07 |
| ← kahmalo left | 08:07 |
| → Adlai joined | 08:10 |
|
roop
| divVerent: i didnt really understand. you want to pull periodically from the sf git repo? | 08:10 |
| ← akar left | 08:11 |
|
divVerent
| no | 08:11 |
|
| from a remote svn into the sf git repo | 08:11 |
|
| for a transition period | 08:12 |
|
Tass`
| how to pop the lastest commit and apply the changes to the working tree? | 08:12 |
|
roop
| divVerent: oh, like call git svn dcommit now and then? | 08:13 |
|
divVerent
| actually, just the other way round | 08:13 |
|
| commands I want are git svn fetch, then git fetch | 08:13 |
| ← project2501a left | 08:14 |
|
roop
| divVerent: the cron runs in your server? or does sf provide ssh access to its servers or something? | 08:14 |
| ← nico- left | 08:14 |
|
divVerent
| cron runs on another server | 08:14 |
|
| because sf allows no cron+ | 08:14 |
| → akar joined | 08:15 |
|
divVerent
| and has some weird "shell creation" that doesn't work with public key login | 08:15 |
|
| so I can't do a pubkey ssh command from another host either | 08:15 |
|
roop
| divVerent: so the cron has to git svn rebase && git push, right? | 08:15 |
|
divVerent
| no | 08:15 |
|
| git svn fetch && git fetch | 08:16 |
|
| inside the git-svn managed repo | 08:16 |
| ← Akkernight left | 08:16 |
|
divVerent
| the git fetch is a trick to fix up the refs | 08:16 |
| → Kelly_zhou joined | 08:16 |
|
divVerent
| [remote "origin"] | 08:16 |
|
| url = . | 08:16 |
|
| fetch = +refs/remotes/trunk:refs/heads/master | 08:16 |
| ← Axius left | 08:16 |
|
divVerent
| this is what it does | 08:16 |
| ← ideogon left | 08:16 |
|
divVerent
| result is a nicely clonable read only (supporting extra branches though) git svn mirror | 08:17 |
| → suzak joined | 08:17 |
| → _gpg_ joined | 08:17 |
|
divVerent
| will get combined with an update hook that prevents pushing into git svn-managed branches | 08:17 |
|
roop
| divVerent: ok, it's a readonly git repo that's a mirror of the svn repo? | 08:17 |
| ← amouge1 left | 08:17 |
|
divVerent
| mostly readonly, yes | 08:17 |
| → nonken joined | 08:17 |
|
divVerent
| allows creating topic branches in git-only | 08:17 |
|
_gpg_
| hi | 08:18 |
|
divVerent
| and then users can dcommit these into svn | 08:18 |
|
roop
| so this 'git svn fetch && git fetch' fails eh? | 08:18 |
|
divVerent
| no | 08:18 |
|
| it normally works | 08:18 |
|
| I just am looking for a way to call it regularily | 08:18 |
|
| because SF has no cron | 08:18 |
|
_gpg_
| i have pushed my change to a public repository, and id like to change one commit "log", is it possible ? | 08:18 |
| ← xlistener left | 08:18 |
|
roop
| but "cron runs on another server" anyway, right? | 08:18 |
|
divVerent
| yes | 08:19 |
| → amouge joined | 08:19 |
|
divVerent
| but I want the fetch and stuff to run on sf's server | 08:19 |
|
| as I want to forbid pushing to svn-managed refs | 08:19 |
|
| I can't do this git svn stuff on another machine | 08:19 |
| ← sako left | 08:20 |
|
roop
| divVerent: so why cant you push from yourserver to the sf git repo? | 08:21 |
|
| divVerent: (and disable push access for all but this cronjob user) | 08:22 |
|
| (... in the sf git server) | 08:22 |
|
divVerent
| because my "yourserver" is not that reliable | 08:22 |
|
| and I don't really want to clone the full repo there | 08:22 |
| ← sde_antrix left | 08:22 |
|
divVerent
| also... I want to allow push from everyone | 08:22 |
|
| just not to the git-svn managed branches | 08:22 |
|
| ugh... the commit hook way fails too | 08:23 |
|
| sh-3.2$ git push origin :refs/svn/update | 08:23 |
|
| [email@hidden.address] password: | 08:23 |
|
| git: 'svn' is not a git-command. See 'git --help'. | 08:23 |
|
| looks like sourceforge is unsuitable for my needs, then | 08:23 |
|
roop
| divVerent: allowing push selectively can be done with hooks, i guess, but i dont know what level you can play with them at the sf repo | 08:23 |
|
| divVerent: oops. they dont have git svn installed, eh? :) | 08:24 |
|
divVerent
| not on the git server | 08:24 |
|
| they have on the shell server | 08:24 |
|
| but using that one cannot be automated | 08:24 |
| → sde_antrix joined | 08:25 |
| → sroehle joined | 08:25 |
|
divVerent
| I wish icculus.org weren't so low on HDD space | 08:25 |
|
| they'd otherwise be suitable for it | 08:25 |
|
| but, one of the repos is 2.3 GB | 08:25 |
|
| and df says 1.7 GB free | 08:25 |
| ← sroehle left | 08:25 |
| → sroehle joined | 08:26 |
| ← sroehle left | 08:26 |
| ← amerine left | 08:26 |
|
divVerent
| well, maybe I can arrange with the guy to put in a larger HD for a donation | 08:26 |
|
| yet another attempt could be asking github if they'd want to support this open source project | 08:26 |
|
| but, normaly they allow 300MB only for public repos | 08:26 |
|
| I wish there were good free git hosters allowing >3GB of space | 08:27 |
|
blazer
| divVerent: what do you need all that space for? | 08:28 |
|
bob2
| 3gb is a looooot of code | 08:28 |
| → wolog joined | 08:28 |
|
roop
| divVerent: i think gitorious.org just asks you to tell them beforehand if you exceed some limit. you can check with them too. | 08:28 |
|
| divVerent: you have binaries too in the repo, eh? | 08:29 |
|
divVerent
| yes | 08:29 |
|
| images | 08:29 |
|
| and music files | 08:29 |
|
| it's a game | 08:29 |
|
| and game data has to stay consistent wiht the code | 08:29 |
|
| and be branchable | 08:29 |
|
| so you can, e.g., add a branch that adds a new weapon to the game | 08:29 |
|
| and adding as new weapon means new textures, new models, AND code | 08:29 |
| ← bentob0x left | 08:30 |
| → Polysics joined | 08:30 |
|
divVerent
| but well... given the size, I'd also want branch-based access control | 08:30 |
| ← juan_arandaalvar left | 08:30 |
|
Polysics
| hello | 08:30 |
|
divVerent
| so I can e.g. add a user who can only push to branches that start with hisusername/ | 08:30 |
|
Polysics
| slightly OT but i don't know where else to ask | 08:30 |
|
divVerent
| that'd save HDD space on the server for extra clones :P | 08:30 |
|
Polysics
| is github having issues today? | 08:30 |
|
divVerent
| and would allow him to do forks | 08:30 |
|
Polysics
| i need to clone a rep to read the docs, but i can't get to it :-) | 08:31 |
|
bob2
| there's a #github | 08:31 |
|
Polysics
| bob2, thanks and sorry | 08:31 |
| → ph^_ joined | 08:31 |
| ← samcarter left | 08:31 |
| → pheaver joined | 08:32 |
|
divVerent
| if grouped by file type, most space is used by .tga files (images) | 08:32 |
|
| then by .ogg files | 08:32 |
| ← SunilThaha left | 08:32 |
|
roop
| divVerent: mmm, makes sense. so sf git repos dont have a limit eh? | 08:32 |
|
divVerent
| they have no limit indeed | 08:32 |
| → SunilThaha joined | 08:32 |
|
divVerent
| and sf is a reliable server | 08:32 |
|
| so two conditions fulfilled | 08:32 |
|
| now I only need that svn mirror for a transition period | 08:32 |
|
| (until everyone is using git) | 08:32 |
|
roop
| divVerent: can you point me to the repo. just curious to see a git repo with lots of binaries. :) | 08:33 |
| → niki joined | 08:33 |
|
divVerent
| http://141.2.16.23/cgi-bin/gitweb.cgi?p=nexuiz.git;a=summary | 08:33 |
|
| it's not on SF yet, still push-mirror-ing :P | 08:33 |
|
| and yes, it once was cvs too | 08:33 |
|
| it was on SF cvs... | 08:33 |
|
| then got imported into icculus svn | 08:33 |
|
| and this is a git-svn clone of it | 08:34 |
| ← bauruine left | 08:34 |
|
divVerent
| this server is my not-so-reliable server at university | 08:34 |
|
| actually, my old desktop machine | 08:34 |
| → imajes joined | 08:34 |
|
divVerent
| wouldn't want to bet the project on that | 08:34 |
|
| as it may fail any time | 08:34 |
|
| to explain the structure... | 08:35 |
|
| and why submodules wouldn't work well | 08:35 |
|
| to add a weapon | 08:35 |
| ← Polysics left | 08:35 |
|
divVerent
| you need files in data/qcsrc (game code) | 08:35 |
|
| in textures/ | 08:35 |
|
| and in models/weapons/ | 08:35 |
|
| and textures is the largest subdirectory | 08:35 |
|
| we gave up trying to tell modellers to save their textures in the same dir as the model | 08:36 |
|
| as apparently blender exporters don't get then path names right then | 08:36 |
| → d0ugal joined | 08:36 |
|
divVerent
| and artists usually aren't fluent with a hex editor :P | 08:36 |
|
| also, the balance*.cfg files contain weapon specific stuff | 08:37 |
| → anno^da joined | 08:37 |
|
divVerent
| the other stuff that's bound to the code is gfx/* (although not THAT bad if it's desynced), and all the cfg fiels | 08:37 |
|
| I am also currently trying to separate it using submodules, but this is getting hard too | 08:37 |
|
| as whatever may be part of the same branch should be in one submodule | 08:38 |
|
roop
| divVerent: just thinking aloud: maybe you can make icculus remember only the last few commits and forget the ones before | 08:38 |
|
| so you wouldnt need a lot of space | 08:38 |
|
divVerent
| that would work initially | 08:38 |
|
| but not later :P | 08:38 |
|
| sure, instead of using a git-svn import, we could clear history | 08:39 |
|
| but I don't really want to | 08:39 |
|
roop
| divVerent: i mean, keep forgetting | 08:39 |
|
divVerent
| that works "reliably"? | 08:39 |
| → galderz joined | 08:39 |
|
divVerent
| but even then... often people want a 1 year old version | 08:39 |
|
| can't forget anything less than 2 years old | 08:39 |
|
roop
| divVerent: the sf git repo will still have all the commits, so there's nothing to worry | 08:39 |
|
divVerent
| still, could maybe bring the 2.3GB down to 1.5GB or like that | 08:39 |
|
| well, I want to end up wiht ONE good official repo | 08:40 |
|
| and preferably on icculus, except their HDD is small | 08:40 |
|
roop
| divVerent: i dont mean let's forget the history. let me try and reword that. | 08:40 |
|
cxreg
| man, schacon is going to have to do a version 2 of his book already | 08:40 |
|
| git changes so fast | 08:40 |
|
roop
| divVerent: we pull 100 commits from svn, push to sf git. then pull 100 more, forget the first 99, push the 101 commit \s we have to git, and so on. | 08:41 |
|
| divVerent: and a similar strategy for the cron even after they're synced | 08:41 |
|
divVerent
| oh, you mean for keeping the mirror up to date | 08:41 |
| ← dolphinling left | 08:41 |
|
divVerent
| not for publishing the icculus.org repo? | 08:41 |
|
| that may work then | 08:41 |
|
roop
| divVerent: right. | 08:41 |
|
divVerent
| just, how? | 08:42 |
|
roop
| divVerent: it's hack and a lot of work still :) | 08:42 |
|
divVerent
| also, I can do the INITIAL setup from 141.2.16.23, my uni host | 08:42 |
|
| only the keeping-it-updated would have to be on something more reliable, like icculus | 08:42 |
|
| so how to forget history? :P | 08:43 |
|
roop
| divVerent: one sec. just checking some man pages. :) | 08:43 |
|
divVerent
| but well | 08:43 |
|
| no matter how well it works | 08:43 |
|
| icculus will still need about 1G HDD space | 08:43 |
|
| because that's the current sizes of all media, zipped | 08:43 |
|
| and this should be the size of the git history "without past" | 08:43 |
|
roop
| yeah, something line 1.5x to 2x of current repo size would be good | 08:44 |
|
divVerent
| also, the art files (tgas etc.) tend to not have much history | 08:44 |
|
| well | 08:44 |
|
| that is how big the git history is :P | 08:45 |
|
| actually, it is 2.0GB | 08:45 |
|
| the large fiels are usually added once, then never touched again | 08:45 |
|
| (except for maybe moving, renaming) | 08:45 |
|
| still I'd like to see how big that would be wiht all history forgotten :P | 08:45 |
|
roop
| divVerent: mm, so the repo size just increases as the binaries got added | 08:46 |
|
divVerent
| yes | 08:46 |
|
| mostly | 08:46 |
| → Twisted_Mentat joined | 08:46 |
|
divVerent
| sometimes binaries get deleted or replaced | 08:46 |
|
| but that is seldom | 08:46 |
|
| although, we MAY have duplicates | 08:46 |
|
| as I don't know if git-svn tracks "svn mv" correctly | 08:46 |
|
| or creates a new copy, and a delete for the old fiel | 08:46 |
| ← akar left | 08:47 |
|
divVerent
| will be back soon | 08:47 |
| → pantsman joined | 08:48 |
| ← stringo0 left | 08:49 |
| marc_m → marc_kdab | 08:49 |
| ← hobodave left | 08:50 |
| → jmk2 joined | 08:51 |
| ← thiago_home left | 08:53 |
| → nico- joined | 08:53 |
|
roop
| divVerent: git svn init <svnpath>; git svn fetch -r<N>; git push <sfgitpath>; rm -rf . #hehe | 08:53 |
|
| divVerent: so we dont forget, but we reinitialize every time from the svn repo. | 08:54 |
| → Haarg joined | 08:55 |
| ← johnw left | 08:56 |
|
jmk2
| two questions are driving nuts: #1 whats with the .git in project names (eg: website.git) and is it important to have it? #2 is it okay to leave my .git repo directory on a websites document_root, more exactly, i have a repo for a website and pull/push to it for updates, is that okay? | 08:57 |
|
_rane
| jmk2: it's not in the project name | 08:57 |
| → ashH joined | 08:57 |
|
Ilari
| jmk2: #1: Its just convention. | 08:57 |
|
| jmk2: $faq non-bare | 08:57 |
|
Gitbot
| jmk2: Pushing to non-bare repositories is discouraged. See http://git.or.cz/gitwiki/GitFaq#non-bare | 08:57 |
| → johan-s joined | 08:58 |
| ← johan-s left | 08:59 |
|
jmk2
| Ilari: thx | 09:01 |
|
| now if you can just point me to the FAQ "my back hurts after sleeping the wrong way" that would be awesome :) | 09:02 |
| ← nico- left | 09:03 |
| → johan-s joined | 09:03 |
| ← GustavoRamos left | 09:03 |
| → HG` joined | 09:04 |
| → becomingGuru joined | 09:11 |
|
divVerent
| roop: that is interesting approach... | 09:11 |
| VVSiz_ → VVSiz | 09:14 |
| ← giallu left | 09:15 |
| → ixti joined | 09:16 |
|
jmk2
| roop: i like that concept as well. | 09:17 |
| ← kssddy left | 09:20 |
| → hummingkid joined | 09:20 |
| ← nadim left | 09:20 |
| ← pcgod left | 09:21 |
| → pcgod joined | 09:22 |
| ← robinbowes left | 09:22 |
|
_gpg_
| is it possible to change some commit logs ? | 09:22 |
|
jmk2
| so question #2: is it okay for my website directory (document_root) to have the repository .git directory ? | 09:23 |
|
Ilari
| jmk2: It should not. Or if it does, you should setup hooks to update it (because push does not update working copy). | 09:24 |
| ← hummingkid left | 09:24 |
|
jmk2
| _gpg_: quick glance read, http://stackoverflow.com/questions/179123/how-do-i-edit-an-incorrect-commit-message-in-git | 09:24 |
|
_gpg_
| jkm2 thank you | 09:25 |
|
jmk2
| Ilari: good to know, i guess what i'm asking is for security purposes, is it find to keep it there. | 09:26 |
| → robinbowes joined | 09:27 |
|
jmk2
| _gpg_: try this site, its been my swiss-army-knife to learning Git, http://www-cs-students.stanford.edu/~blynn/gitmagic/ | 09:27 |
| → xiong joined | 09:28 |
| → robinbowes_ joined | 09:28 |
|
Ilari
| jmk2: http://toroid.org/ams/git-website-howto | 09:28 |
| ← m104 left | 09:28 |
| ← munichlinux left | 09:29 |
| ← robinbowes_ left | 09:29 |
|
_gpg_
| jmk2, thank you, really interesting | 09:29 |
| → GarethAdams|Work joined | 09:30 |
|
_gpg_
| i have seen in the past some website propising free git video cast, i can't find them again, any one know them ? | 09:30 |
| → priidu joined | 09:31 |
| → nickrw joined | 09:31 |
| → weev joined | 09:32 |
| → gitte joined | 09:33 |
| → curtana joined | 09:33 |
| ← AkumaStreak left | 09:34 |
| → bentob0x joined | 09:34 |
| → shruggar joined | 09:36 |
| ← mathrick left | 09:37 |
| ← jmk2 left | 09:38 |
| → project2501a joined | 09:38 |
| ← lemonchicken left | 09:39 |
| → AkumaStreak joined | 09:39 |
| ← weev left | 09:40 |
| ← pheaver left | 09:41 |
| → tvw joined | 09:42 |
| → tbf joined | 09:45 |
| dwmw2_ → dwmw2 | 09:57 |
| → gebi joined | 09:57 |
| ← solofight left | 09:59 |
| → solofight joined | 10:00 |
| ← Twisted_Mentat left | 10:00 |
| → hummingkid joined | 10:05 |
| ← ianmcorvidae left | 10:06 |
| → ianmcorvidae joined | 10:06 |
| ← RaceCondition left | 10:09 |
| ← tetha left | 10:10 |
| → tetha joined | 10:10 |
| ← jelmer left | 10:14 |
|
yhager
| git rebase, conflict, resolved it, ran 'git add', trying git rebase --continus, but I get "No changes - did you forget to use 'git add'" | 10:14 |
| → jelmer joined | 10:15 |
| → g0 joined | 10:16 |
|
g0
| really need help here... I had a few image files in my working directory, untracked in .gitignore, I switched back to a previous commit where .gitignore didn't contain anything, then did git add . forgetin that i'm not at the last commit, then I thought 'duh' and did git checkout master now the images are gone! | 10:19 |
| → TBAA joined | 10:20 |
|
TBAA
| hi | 10:20 |
|
| i like to switch back to an old commitstate | 10:20 |
|
| what is the command to do that if i have the hash of this commit? | 10:20 |
|
vkroll_
| git checkout hash | 10:21 |
| ← Kelly_zhou left | 10:21 |
|
TBAA
| thx | 10:21 |
|
vkroll_
| or git revert hash | 10:21 |
| → __bert__ joined | 10:22 |
|
g0
| actually I was lying, I did a git reset --hard befre trying git checkout master | 10:23 |
|
| I'm totally f*d right? | 10:23 |
|
TBAA
| hmm git push says i cannot fast forward. is there a way to force a push? | 10:23 |
| → aspotashev joined | 10:24 |
|
wereHamster
| g0: the images are saved within the git object database | 10:24 |
|
TBAA
| -f thx that this is so easy | 10:24 |
| ← hummingkid left | 10:24 |
|
wereHamster
| g0: you just have to find them (git fsck --lost-found and then inspect 'tree's and see which one contains references to your images) | 10:24 |
| → hummingkid joined | 10:24 |
|
g0
| wereHamster I did not commit the after git add ., i instead did git reset --hard | 10:25 |
|
| will they still be there? | 10:25 |
|
wereHamster
| yes | 10:25 |
|
g0
| ok ok... | 10:25 |
|
| g0 needs to calm down | 10:25 |
| → Araneidae joined | 10:25 |
| → psankar joined | 10:26 |
|
g0
| wereHamster git fsck --lost-found gave me a few dangling commits and dangling trees.. um.. how to i see what's in them? | 10:27 |
|
wereHamster
| git show <tree> | 10:27 |
|
g0
| thanks | 10:27 |
|
wereHamster
| or better: git cat-file -p <tree> | 10:28 |
|
Araneidae
| I'm in the middle of a rebase conflict: how can I get git-diff to show *all* the changes on the conflicting file? | 10:28 |
| → lemonchicken joined | 10:28 |
|
Araneidae
| As far as I can tell git-diff by default only shows me the differences it hasn't merged ... not sure I understand it, to be honest | 10:28 |
|
| In particular, I'm confused at the horribly things the automatic merge has done to my file! | 10:29 |
| FauxFaux → AAAAAAAAAAAAAAAA | 10:29 |
|
wereHamster
| g0: nah, use git ls-tree -r <tree> | grep .jpg or so and the git the images with 'git cat-file -p <blob> > path/to/image.jpg' | 10:29 |
| AAAAAAAAAAAAAAAA → FauxFaux | 10:29 |
|
broonie
| git diff <something>.. to diff from <something> to your working directory. | 10:29 |
|
Araneidae
| broonie, sorry, I don't quite get that. I've tried `git diff $FILENAME..` and `git diff $FILENAME ..` without success | 10:30 |
|
g0
| wereHamster: Actually I had 571 images, do I have to do that for each one of them? | 10:30 |
|
Araneidae
| ".. is outside repository" | 10:30 |
|
wereHamster
| Araneidae: <something> is a commit, not a filename | 10:31 |
|
g0
| My sad life :'( | 10:31 |
|
Araneidae
| Well, that's not useful: I haven't commited the conflict yet... | 10:31 |
|
g0
| I think I found the tree | 10:31 |
|
broonie
| Araneidae: git diff <something>.. filename | 10:31 |
|
wereHamster
| g0: you can script taht ('git ls-tree -r <tree> | grep ... | while read ...; do git cat-file ...) | 10:31 |
|
broonie
| Araneidae: where something is something like HEAD or one of the branches you're merging. | 10:31 |
|
TBAA
| hmm Not currently on any branch. | 10:31 |
|
Araneidae
| Ahh... actually, `git diff HEAD <filename>` --- simple as that! | 10:32 |
|
wereHamster
| Araneidae: if you just had a merge conflict, you can simply run 'git diff' to see what caused the conflict | 10:32 |
|
TBAA
| i want that my code now should be the HEAD | 10:32 |
|
wereHamster
| TBAA: it is | 10:32 |
|
Araneidae
| wereHamster, my problem was that `git diff` shows both sides of the merge and I wasn't trusting it... | 10:32 |
|
wereHamster
| HEAD is whatyou have currently checked out | 10:32 |
|
TBAA
| ok, but this message is new | 10:32 |
| ← froschi left | 10:34 |
|
TBAA
| ok instead of HEAD i think i mean MASTER | 10:34 |
|
Araneidae
| In particulara, when merging a conflict, git-diff doesn't show the already merged part. Anyhow, thanks for pointing me at `git diff HEAD <filename>` | 10:34 |
|
wereHamster
| TBAA: 'master' (lower case) | 10:34 |
|
TBAA
| ok :-D | 10:34 |
|
wereHamster
| TBAA: git checkout master; git reset --hard HEAD@{1} | 10:34 |
|
| TBAA: but read faq detached first! | 10:35 |
|
Gitbot
| TBAA: You are on a detached HEAD and might lose commits.. See http://git.or.cz/gitwiki/GitFaq#detached | 10:35 |
| → simonpe joined | 10:35 |
|
TBAA
| lose commits is what i want to do :-D | 10:35 |
|
Ilari
| TBAA: But the commits to lose in this case are not likely the ones you want to lose. | 10:36 |
|
TBAA
| yeah maybe :-D | 10:36 |
|
simonpe
| I have branched from master and tagged that branch, without changing any files. When I try to push I am told that everything is up to date, i.e. I can't push. Why is this? | 10:36 |
|
TBAA
| i read the faq | 10:36 |
| → munichlinux joined | 10:36 |
| ← munichlinux left | 10:36 |
|
wereHamster
| simonpe: do you want to push the tag or the new branch? | 10:37 |
|
simonpe
| wereHamster: yes | 10:37 |
|
Ilari
| simonpe: 'git push <remote> <tag>'? | 10:37 |
|
simonpe
| Im checking the man page... 1 sec | 10:38 |
|
TBAA
| wereHamster: your command helped thanks! | 10:38 |
|
simonpe
| Ilari: why do I need to specify <remote>? When I do only 'git push', <remote> is implicit | 10:39 |
| ← Jak_o_Shadows left | 10:41 |
|
wereHamster
| simonpe: by default only tags are pushed which are reachable from branches already on the remote | 10:42 |
| ← TBAA left | 10:42 |
|
Ilari
| simonpe: If its implicit, the right value is origin. The first non-option for fetch/push always designates remote, rest specify what to transfer. | 10:43 |
|
| *likely is origin | 10:43 |
| ← hummingkid left | 10:43 |
|
simonpe
| i see, so if I base a tag 0.0.1 on a branch 0.0.1, that would also push the branch? | 10:45 |
|
wereHamster
| no. 'git push' would only push the tag if branch-0.0.1 was already on the remote | 10:45 |
| → mastro joined | 10:46 |
|
wereHamster
| but pushing the tag doesn't also push the branch | 10:46 |
|
simonpe
| man this is confusing :P | 10:46 |
|
wereHamster
| 'git push' by default pushes branches which are common to both local and remote repository (IOW no new branches are pushed by default) | 10:47 |
|
| that behaviour can be changed with the config option push.default | 10:48 |
|
simonpe
| ah so I explicitly need to push the branch first, then the tag | 10:49 |
| → kwinz joined | 10:50 |
| aptituz_ → aptituz | 10:51 |
|
simonpe
| should I prefix my tags and branches with tag- and branch- ? | 10:53 |
| ← anno^da left | 10:53 |
|
Ilari
| simonpe: they implicitly have prefixes tags/ and heads/ | 10:53 |
|
| simonpe: And in many instances, those are guessed correctly. | 10:53 |
|
| simonpe: So you shouldn't prefix them manually. | 10:54 |
|
g0
| wereHamster Yay!! I got them all back, thank you! thanks indeed! :) | 10:54 |
|
wereHamster
| but you shouldn't use the same names for tags and branches - it's only confusing | 10:54 |
|
Ilari
| simonpe: The push can be done in either order (or both can be pushed in the same transaction). | 10:54 |
| ← tango_ left | 10:54 |
|
wereHamster
| git.git prefixes all tags with 'v' (v1.6.6) for example | 10:54 |
|
g0
| for all those who might get into this mess here's a line of code to resurrect something you added but reset and lost | 10:55 |
|
| after finding <tree> run `git ls-tree -r <tree>|sed 's/.*blob \([a-f0-9]*\)\(.*\)/git cat-file -p \1 \> \2/'` | 10:55 |
|
Ilari
| simonpe: If you have both branch and tag 'foo', then one way to specify which one you mean for push is to use 'heads/foo' and 'tags/foo'. | 10:55 |
|
wereHamster
| git checkout <tree> -- <path> would probably have been better.. | 10:56 |
|
simonpe
| Ilari: ah that's what I was looking for :P | 10:56 |
|
wereHamster
| g0: ^^^ | 10:56 |
|
simonpe
| Ilari: thanks for the help, u've been very kind | 10:56 |
|
Ilari
| simonpe: Actually, even if you push them in the same transaction, one gets created/updated after other. | 10:57 |
|
| simonpe: And tags aren't on any branch, so no problem whatever the order is. | 10:57 |
|
simonpe
| ok, so a tag is just a state, not a permutation of a branch? | 10:58 |
|
Ilari
| simonpe: Its kind of read-only branch by itself (possibly with extra message). | 10:58 |
| → tango_ joined | 10:59 |
|
g0
| wereHamster Why didn't you tell that before? Shift man :D | 10:59 |
|
| How do I drop files I have currently added using git add for the first time? git rm? | 11:01 |
|
Ilari
| g0: 'git rm --cached <file>'? | 11:02 |
| ← simonpe left | 11:02 |
| → simonpe joined | 11:03 |
| → nadim joined | 11:03 |
| → eletuchy joined | 11:04 |
|
g0
| is there a particular mishap that can happen by moving around git repository to different directories? | 11:04 |
| → _daniel_ joined | 11:04 |
|
_daniel_
| hi guys | 11:04 |
|
| how do I revert a file to an older version | 11:04 |
|
_rane
| git checkout <revision> -- path/to/file | 11:05 |
| → hummingkid joined | 11:06 |
| → jogla joined | 11:07 |
| → jogla_ joined | 11:08 |
| ← jogla_ left | 11:08 |
| jogla → foooo | 11:08 |
| → __iron joined | 11:08 |
| → jogla joined | 11:09 |
| ← foooo left | 11:09 |
| ← jogla left | 11:10 |
|
TTilus
| how do i make my local branch go to origin also? | 11:10 |
|
| i've got a clone, created a branch and stuff to it | 11:10 |
|
| now i wanna "publish" that branch | 11:10 |
|
| the repo i cloned from is public | 11:11 |
| → kraymer joined | 11:11 |
|
TTilus
| and i can push | 11:11 |
| → jogla joined | 11:11 |
|
Ilari
| TTilus: 'git push <remote> <branchname>'. | 11:12 |
|
simonpe
| TTilus: that's where I'm at now too, I can push but when I remove my clone and re-clone the branch isn't there | 11:12 |
|
Ilari
| simonpe: You shouldn't remove your clone. | 11:13 |
|
simonpe
| Ilari: I just wanted to check what others get when they checkout | 11:14 |
| → RaceCondition joined | 11:14 |
| → Geralt joined | 11:14 |
|
simonpe
| sort of verify my branch is there | 11:14 |
| ← project2501a left | 11:14 |
|
Ilari
| simonpe: 'git ls-remote <URL>'. | 11:14 |
|
simonpe
| Ilari: ah sweet :P thaks | 11:14 |
|
Ilari
| simonpe: It directly queries the remote repo. All refs contained are shown. | 11:15 |
|
simonpe
| I'm getting the hang of it soon :P | 11:17 |
| → juan_arandaalvar joined | 11:17 |
|
simonpe
| I'm used to svn and thought it'd be similar but it isn't o_O | 11:17 |
|
Ilari
| simonpe: This isn't SVN. :-) | 11:18 |
| ← imajes left | 11:18 |
| → ideogon joined | 11:18 |
|
GarethAdams|Work
| THIS. IS. SPARTAAAAAAAAA | 11:18 |
| ← RaceCondition left | 11:20 |
| ← juan_arandaalvar left | 11:23 |
| → krynnotaur joined | 11:26 |
| ← JackWinter left | 11:27 |
| → dtrott joined | 11:28 |
| ← ideogon left | 11:28 |
| → augustl joined | 11:30 |
| → phadviger1 joined | 11:30 |
|
augustl
| is there a way to ignore a file with configs in my .git dir? I.e. repository local ignore. | 11:30 |
|
| so that my NOTES doesn't show up in either .gitignore or `git status` ;) | 11:31 |
|
wereHamster
| .git/info/exclude | 11:31 |
|
augustl
| wereHamster: thanks! | 11:31 |
| ← _gpg_ left | 11:32 |
|
divVerent
| Is there a way to list all file names that ever existed in a git repo? | 11:32 |
| ← upul left | 11:33 |
| → unapiedra joined | 11:33 |
|
unapiedra
| How do I remove all untracked files in the folder where I use git? | 11:35 |
|
| git clean -f or -n or??? | 11:36 |
|
dtrott
| divVerent: git diff --name-only FIRST_COMMIT LAST_COMMIT | 11:36 |
|
divVerent
| does that also list ones created and deleted again in between? | 11:37 |
| ← phadviger left | 11:37 |
| → warthog19 joined | 11:37 |
| ← warthog9 left | 11:37 |
|
dtrott
| divVerent: sorry don't think so | 11:37 |
|
TTilus
| i've got a local tracking branch and i rebased it | 11:37 |
|
| how do i push the rebase | 11:37 |
|
| git push (in rebased branch) refuses saying its non-fastforward | 11:38 |
|
divVerent
| well | 11:38 |
|
| you know what you are doing? | 11:38 |
|
TTilus
| i want the tracked branch in origin to be rebased too | 11:39 |
|
| like push the rebase operation | 11:39 |
| ← psankar left | 11:39 |
|
TTilus
| so rephrasing, how do i rebase a branch in origin? | 11:39 |
|
divVerent
| git push --force... but for this you REALLY need to know what you're doing | 11:40 |
|
| because this will harm anyone tracking the branch | 11:40 |
|
TTilus
| i dont wanna --force | 11:40 |
|
| say i have people tracking a branch which needs to be rebased | 11:40 |
|
| how do i rebase? | 11:40 |
|
dtrott
| TTilus: Basically you don't rebase origins unless you KNOW no-one is tracking it. | 11:40 |
|
divVerent
| you can't | 11:40 |
|
TTilus
| ok | 11:41 |
|
divVerent
| rebasing IS non-fastforward | 11:41 |
| → imajes joined | 11:41 |
|
divVerent
| these people can of course pull the new base using git pull --force | 11:41 |
|
| so not all is lost | 11:41 |
|
dtrott
| Better to just merge changes and push them | 11:41 |
|
divVerent
| normally yes | 11:41 |
| ← jas4711 left | 11:42 |
|
divVerent
| e.g. if you do push a rebase, others need to trust you that you included all commits on upstream | 11:42 |
|
TTilus
| create a new branch, merge/cherrypick the old there and push the new to origin? | 11:42 |
|
divVerent
| if someone pushed right before you, his changes would get lost | 11:42 |
|
TTilus
| divVerent: uhh, thats scary, dont wanna go that way | 11:42 |
|
divVerent
| for example, yes - just why do you want to rebase anyway? | 11:42 |
| → froschi joined | 11:42 |
| ← xiong left | 11:43 |
|
divVerent
| I ask, because there may be a better way to achieve that goal | 11:43 |
|
TTilus
| there's a wip which has been worked with and i just found out it's been branched from "devel" instead of "stable" | 11:43 |
|
| its basically a bugfix which need to be done against "stable" | 11:44 |
| s1kx- → S1kx | 11:44 |
|
TTilus
| and they stay in wip branch until they are tested and accepted | 11:44 |
|
| thats why | 11:44 |
| → kesor joined | 11:44 |
|
divVerent
| then cherry-pick probably is the way to go | 11:44 |
|
| make a new separate branch | 11:44 |
|
| cherry-pick the bugfix into it | 11:44 |
|
TTilus
| ok | 11:44 |
| ← imajes left | 11:44 |
|
divVerent
| and | 11:44 |
|
| merge or push that onto stable | 11:45 |
|
| is that branch ONLY the bugfix? | 11:45 |
|
kesor
| can someone help please, i want to get a list of all the commits that were pushed in a post-receive hook ... the regular "$oldrev..$newrev" is good, but I am concerned about merge commits - will "git log $oldrev..$newrev" show the commits that were merged or not? | 11:45 |
| → imajes joined | 11:45 |
|
divVerent
| if yes, you can indeed try rebasing the branch onto the head of stable, then push that into stable | 11:45 |
|
| (THAT would be fastforward) | 11:46 |
|
| but don't push that to the wip branch then | 11:46 |
|
dtrott
| Has the wip branch been pushed to the origin already ? | 11:46 |
| → JackWinter joined | 11:46 |
|
TTilus
| dtrott: yes | 11:46 |
|
dtrott
| I would rebase the wip branch locally then rename the branch and push that up then. | 11:47 |
| ← ph^_ left | 11:51 |
| → ph^_ joined | 11:52 |
| → vadi joined | 11:52 |
|
vadi
| How to create a new branch from changed files without committing them to the current branch? | 11:52 |
| → ike__ joined | 11:52 |
|
ike__
| hi .. I've commited a project unto github, but I inadvertedly committed the .svn folder as well | 11:53 |
| → pfig joined | 11:53 |
| ← pfig left | 11:53 |
|
ike__
| how do I remove it and have it ignored? | 11:53 |
| → pfig joined | 11:53 |
| ← kesor left | 11:54 |
|
dtrott
| vadi: git checkout -b NEWBRANCH | 11:55 |
| → d0k joined | 11:55 |
|
vadi
| thanks dtrott | 11:56 |
| → RaceCondition joined | 11:57 |
| ← simonpe left | 11:58 |
|
dtrott
| ike__: git rm -r .svn | 11:59 |
|
| then commit that change and push it | 11:59 |
|
ike__
| dtrott, no, i dont want to DELETE it from my disk. I just want git to ignore it | 11:59 |
|
| and delete it from github | 11:59 |
|
| the project is a fork from a svn | 11:59 |
|
| http://github.com/ikolar/ghostpluplus-plus | 12:00 |
|
| this is it, it's messy | 12:00 |
|
| .gitignore only works for future files, not for existing commings, yes? | 12:00 |
| → Guest96666 joined | 12:02 |
|
dtrott
| just add --cached to the command then: git rm -r --cached .svn | 12:02 |
| ← _daniel_ left | 12:03 |
| → smuf joined | 12:03 |
|
wereHamster
| ike__: why didn't you use git-svn to import the svn repo into git? | 12:03 |
|
TTilus
| (i can feel the facepalms, but) how do i delete a branch from origin? already deleted the tracking branch, but wanna delete the tracked too | 12:04 |
|
wereHamster
| git push origin :branch | 12:04 |
|
Grum
| git push origin :branch | 12:04 |
|
ike__
| dtrott, does --caches leave the files on the disk? | 12:04 |
|
wereHamster
| yes | 12:04 |
|
ike__
| ty | 12:04 |
|
| i uset git svn | 12:05 |
| ← Sonderblade left | 12:05 |
|
wereHamster
| interesting, so how did the .svn folder end up being tracked? | 12:05 |
| → Sonderblade_ joined | 12:05 |
|
wereHamster
| git-svn doesn't add/commit those folders... not even leave them in the filesystem | 12:05 |
|
dtrott
| ike__: no but --cached (with a d does ) | 12:06 |
|
ike__
| kk | 12:06 |
| → ProLoser1Trmnl joined | 12:06 |
|
ProLoser1Trmnl
| hello | 12:06 |
| → capndiesel joined | 12:06 |
|
ProLoser1Trmnl
| so i'm trying to setup a system where git publishes to my webserver, is there a way to make it use a specific username? | 12:06 |
| ← froschi left | 12:07 |
|
ike__
| ok, one more q ... I've accidentaly git -rm -ed some files, but did not commit yet | 12:08 |
|
| how do i get them back | 12:09 |
| ← capndiesel left | 12:09 |
| → capndiesel joined | 12:10 |
|
dtrott
| you typed caches didn't you | 12:10 |
| → project2501a joined | 12:11 |
|
dtrott
| git checkout LIST_OF_FILES YOU WANT BACK | 12:12 |
|
divVerent
| how do I git push ALL branches, but all TAGS too? | 12:13 |
|
| --all is just branches | 12:13 |
|
| --tags is just tags | 12:14 |
|
| can I somehow push both? | 12:14 |
|
gebi
| sure, git push --all; git push --tags | 12:15 |
|
divVerent
| git push mirror 'refs/heads/*' 'refs/tags/*' trying this, i.e. passing wildcards | 12:15 |
|
| well, I want it to be atomic :P | 12:15 |
|
dtrott
| because ? | 12:15 |
|
divVerent
| cleaner | 12:15 |
| ← mithro left | 12:16 |
| ← dtrott left | 12:16 |
| → vbgunz joined | 12:16 |
|
divVerent
| just, are wildcards supported in push | 12:16 |
|
| or will it tell me at the end of the operation that it didn't do it? | 12:17 |
| ← hyperboreean left | 12:17 |
| → H4ns joined | 12:17 |
| → Pupeno joined | 12:18 |
| → rgr joined | 12:21 |
| ← pcgod left | 12:21 |
| → pcgod joined | 12:21 |
| → segher joined | 12:21 |
| → ^^raven^^ joined | 12:22 |
| → kristian42 joined | 12:22 |
| → xiong joined | 12:23 |
| ← H4ns left | 12:23 |
| ← galderz left | 12:24 |
|
ProLoser1Trmnl
| fatal: 'iwdsa@iwdsa.org/website.git' does not appear to be a git repository | 12:24 |
|
| that isn't true, what am i doing wrong? | 12:24 |
| → galderz joined | 12:25 |
|
divVerent
| you may be missing the ssh://? | 12:28 |
|
| try ssh://iwdsa@iwdsa.org/website.git | 12:28 |
|
ProLoser1Trmnl
| i did | 12:28 |
|
curtana
| can someone tell me what determines what shows up in the diff of a merge commit? | 12:28 |
|
ProLoser1Trmnl
| does adding the user at the beginning cause the problem? | 12:28 |
|
divVerent
| ProLoser1Trmnl: no | 12:28 |
| ← nullvariable left | 12:28 |
|
divVerent
| where exactly is the repo on iwdsa.org? | 12:28 |
|
| possibly you mean ~/website.git and not /website.git? | 12:29 |
|
| try ssh://iwdsa@iwdsa.org/~/website.git, then | 12:29 |
|
ProLoser1Trmnl
| git remote add web ssh://iwdsa@iwdsa.org/website.git | 12:29 |
|
| ohh | 12:29 |
|
| good point :) | 12:29 |
|
| i love you :) | 12:30 |
| ← simNIX left | 12:32 |
| → simNIX joined | 12:34 |
|
VVSiz
| hi folks, a quick question about git push and problems that it creates merge commits | 12:36 |
|
| for our project, we tend to use "linear" history, no merges, etc | 12:37 |
|
| but, sometimes, when two folks do pushes about the same time, we end up with merge commit which makes history not linear | 12:37 |
|
| is there a way to prevent git doing those merge commits? | 12:38 |
|
teuf
| isn't it the pull before the dev pushes which create the merge commits? | 12:38 |
|
VVSiz
| so that if somebody tries to push, he/she would fail | 12:38 |
| → Pata joined | 12:39 |
|
bremner
| VVSiz: I don't think push makes new commits | 12:39 |
|
VVSiz
| teuf: hmm, I'm not sure. but probably, yeah! git pull is the culprit | 12:39 |
|
teuf
| VVSiz: git pull --rebase would avoid that | 12:40 |
|
| I think it can be made the default | 12:40 |
| → lohrun joined | 12:40 |
|
Pata
| Hey All, I have a branch which had 4 commits, #2 was ver bad but got pushed anyway along with 3 + 4. I've used cherry-pick to make my local branch good and working, but now when I try to push it's saying that we've diverged, what's the correct course of action now? | 12:40 |
|
VVSiz
| teuf, bremner: thanks, this clears my confusion! | 12:41 |
| → froschi joined | 12:41 |
| ← vbgunz left | 12:41 |
| ← augustl left | 12:41 |
|
Pata
| Any clues/keywords for me to search by? | 12:43 |
| → christophsturm joined | 12:44 |
|
teuf
| if other people are working on the branch with 2 + 3 + 4, pushing the "fixed" commit would break things for them | 12:44 |
|
| so maybe it's better to fix #2 on top of 3 + 4 ? | 12:45 |
| → psankar joined | 12:45 |
|
Pata
| Nobody else is working on anything, #2 I think came from a mis guided revert last thursday, it basically un-does everything from #1 | 12:45 |
|
| (and perhaps a couple before #1) | 12:46 |
|
teuf
| git push origin +master should push the fixed branch I think | 12:46 |
|
| or git push -f | 12:46 |
|
Pata
| Can I rebase and skip #2? (Perhaps? I've just understood rebase I believe) | 12:46 |
|
| I pondered using -f | 12:46 |
|
| but couldn't find too much info about how it worked | 12:47 |
|
| will it say 'What I have is correct, so take it anyway"? | 12:48 |
|
| ooh, and if so what's the difference between that and +master | 12:50 |
|
teuf
| I don't think there's a difference between -f and +master | 12:50 |
|
| and yeah, it basically means that | 12:50 |
|
| "take what I have, don't ask questions" | 12:50 |
|
| and also means "I know I'm screwing with other people who are using this branch" | 12:51 |
| → halfline_ joined | 12:51 |
|
Pata
| hm ok, so it then 'deletes' the commits on the other end? | 12:51 |
| → halflin__ joined | 12:51 |
|
Pata
| back up till the last merge? | 12:52 |
| → kukks joined | 12:52 |
| ← pfig left | 12:53 |
| ← halfline_ left | 12:53 |
| → stamina_ joined | 12:54 |
| → Axius joined | 12:55 |
| ← halfline left | 12:55 |
| ← ike__ left | 12:57 |
| → loincloth joined | 12:57 |
| → cannonball joined | 12:58 |
| → hicham joined | 12:59 |
|
hicham
| you are in a middle of a merge conflict | 12:59 |
|
| what should i do in that case ? | 12:59 |
| → auslander joined | 13:00 |
| ← ProLoser1Trmnl left | 13:00 |
|
Rhonda
| Take a look at the conflict and git add the files after you resolved the conflict. | 13:01 |
|
| I think git diff is handy here, and git mergetool | 13:01 |
| ← loincloth left | 13:03 |
| → loincloth joined | 13:03 |
| → aziz joined | 13:06 |
| ← awnstudio left | 13:10 |
| ← mnyon left | 13:11 |
| ← xiong left | 13:11 |
| → tstclair joined | 13:13 |
| ← capndiesel left | 13:15 |
|
Pata
| So if I "git push -f origin master", all the commits on 'origin/master' up until which point get overwritten? | 13:16 |
| → NET||abuse joined | 13:16 |
|
wereHamster
| they will be overwritten with your local master | 13:16 |
|
Pata
| So just any files with a merge conflict will get overwritten by my local master? | 13:17 |
|
| and the rest is as normal | 13:17 |
|
wereHamster
| push doesn't solve any conflict | 13:21 |
|
| if you have a merge conflict, and want to overwrite any conflicting files with the versinos from HEAD, use git checkout --ours -- . | 13:21 |
|
divVerent
| is there a git command to call a specified hook? | 13:22 |
| → parasti joined | 13:22 |
|
divVerent
| basically, in a bare repo, I'd do | 13:22 |
|
| [ -x hooks/post-update ] && hooks/post-update | 13:22 |
|
Pata
| I have a commit which I want to ignore, I've used cherry-pick to get my local in the state in which I want it but am having trouble pushing now | 13:22 |
|
divVerent
| what is the generic case? | 13:22 |
|
Pata
| teuf has suggested using "git push origin +master " which sounds like it's what I want, but I wanted to understand what it was doing | 13:24 |
|
wereHamster
| divVerent: the hooks are called automatically, why would you want to force it? | 13:26 |
|
divVerent
| because I want to share the code from post-update and in a script called after git svn fetch | 13:26 |
|
wereHamster
| if you have a hook which you want to run both from post-update and also at other times, make it a separate script and exec it from post-update. that way it's much clearer which scripts are when executed (IMO) | 13:27 |
|
divVerent
| (in both cases, pushing to a backup server) | 13:27 |
|
wereHamster
| btw, you can also 'git push . branch' | 13:27 |
|
| which would run the post-update script | 13:27 |
| → ehabkost joined | 13:28 |
| → PN joined | 13:29 |
|
divVerent
| hm... how to create the first branch after a git init? | 13:29 |
|
| first time I actually CREATE a repo... before I always cloned | 13:29 |
|
_rane
| master is automatically created | 13:29 |
|
divVerent
| git branch begs to differ | 13:29 |
|
_rane
| I might be wrong :P | 13:30 |
|
divVerent
| its output is empty | 13:30 |
|
Ilari
| divVerent: First commit would create it. | 13:30 |
|
_rane
| maybe it's not there until the first ocmmit | 13:30 |
|
divVerent
| ah, okay then | 13:31 |
| → sumangurung joined | 13:31 |
| ← loincloth left | 13:31 |
| → Chaser_ joined | 13:32 |
| → catalis joined | 13:33 |
| ← catalis left | 13:34 |
| ← g0 left | 13:35 |
| → gergap joined | 13:35 |
| → lericson joined | 13:35 |
|
lericson
| How do I switch branch without touching my index? | 13:35 |
|
Ilari
| lericson: That doesn't make much sense. | 13:35 |
|
gergap
| hi | 13:36 |
|
lericson
| Ilari: The stuff I have on the filesystem, I want that to remain intact. | 13:36 |
| ← simNIX left | 13:36 |
| → loincloth joined | 13:37 |
|
gergap
| can msysgit create symlinks on windows? | 13:37 |
|
Ilari
| gergap: AFAIK, symlink creation is priviledged operation on windows. | 13:37 |
| ← sumangurung left | 13:38 |
|
Ilari
| lericson: Leaving that intact would show up as changes. | 13:38 |
|
lericson
| Ilari: Yes, my point exactly. | 13:38 |
|
gergap
| Ilari: ok, does this mean 'no'? | 13:38 |
|
Ilari
| gergap: I don't think it can create symlinks, but I'm not sure. | 13:39 |
|
| lericson: Should the staging area also be left as is? | 13:40 |
|
lericson
| Ilari: Optional, don't mind - whichever is easier. | 13:40 |
| → gunmuc1 joined | 13:41 |
|
Ilari
| lericson: 'git symbolic-ref HEAD refs/heads/branchtoswitchto'. | 13:41 |
|
| lericson: Note the refs/heads/ prefix. | 13:41 |
|
lericson
| I see. Thank you. | 13:42 |
|
gergap
| is there a way to let msysgit checkout symlinks as regular copies of files? So changing one file and doing commit/pull would synchronize all other linked copies? | 13:43 |
|
lericson
| I just realize | 13:43 |
|
| I just realized that wasn't at all what I wanted to do. | 13:44 |
| → rrrrutdk joined | 13:44 |
|
lericson
| Let me phrase it like a normal question, how do I switch branch and keep my local changes to the previous branch, reapplying them to the new branch? | 13:44 |
| → DavidKlein joined | 13:45 |
| ← ProperNoun left | 13:46 |
|
gergap
| lericson: isn't it possible to use the stash for your purpose? git stash; git checkout newbrach; stash pop | 13:46 |
|
tango_
| lericson: are these COMMITTED changes or just worktree changes? | 13:47 |
|
| also, is that a single commit or multiple commits, in the former case? | 13:47 |
| ← psankar left | 13:47 |
|
lericson
| Changes to the working directory, which I want to keep. | 13:47 |
|
| And yes, stashing would be possible, but I'd like an easier way to do it | 13:48 |
|
| The usecase is that I often realize half-way through a change that I was on the wrong branch. | 13:48 |
|
Tass`
| how to push one branch only? | 13:48 |
|
tango_
| lericson: then git stash ; git checkout ; git stash pop is the quickest way | 13:48 |
| ← Axius left | 13:48 |
|
tango_
| unless git checkout supports -m | 13:48 |
|
| lericson: ok, git checkout -m otherbranch ought to work | 13:49 |
|
| Tass`: git push dest branch ? | 13:49 |
|
Tass`
| tango_, yeah, correct :-) | 13:49 |
|
lericson
| tango_: Nice, thank you. | 13:51 |
|
tango_
| yw | 13:52 |
| → nessundorma joined | 13:52 |
| ← EricInBNE left | 13:54 |
| → EricInBNE joined | 13:55 |
| ← gunmuc left | 13:57 |
| ← lohrun left | 13:58 |
| → aresnick joined | 13:59 |
|
divVerent
| is it normal that TortoiseGit always shows "Failed" when cloning? | 13:59 |
|
| (but actually successfully cloned) | 13:59 |
| → ereslibre joined | 14:00 |
| → Bass10 joined | 14:00 |
| → Axius joined | 14:00 |
| ← Axius left | 14:01 |
| ← univate left | 14:01 |
| → Ryback_ joined | 14:02 |
| → univate joined | 14:03 |
| ← rrrrutdk left | 14:03 |
| ← univate left | 14:04 |
| → esc joined | 14:05 |
| → nothingHappens joined | 14:05 |
| → univate joined | 14:05 |
| ← _Vi left | 14:06 |
| ← univate left | 14:07 |
| → kmartiin joined | 14:07 |
| ← ereslibre left | 14:07 |
| → illsci joined | 14:07 |
|
illsci
| if you have a git repo you've cloned and you created a bunch of different branches and you git pull from the remote master | 14:07 |
|
| how do you merge those changes into your branches from your local master | 14:08 |
| → univate joined | 14:08 |
|
illsci
| assuming you've git checkout somebranchyoumade | 14:08 |
| ← univate left | 14:09 |
| → univate joined | 14:10 |
| → LiamH joined | 14:11 |
| ← univate left | 14:12 |
| → univate joined | 14:13 |
|
jacobat
| illsci: git checkout yourbranch ; git merge master | 14:13 |
|
| if i read you correctly | 14:14 |
| ← univate left | 14:14 |
| → justinweiss joined | 14:14 |
|
fr0sty
| illsci: alternatively: git checkout yourbranch; git rebase master | 14:14 |
| ← rubbs left | 14:16 |
|
illsci
| thanks | 14:16 |
|
| i left my git book at home today... learning slowly | 14:16 |
| → reprore joined | 14:17 |
|
illsci
| does rebase automattically commit it and merge doesn't? | 14:17 |
|
jacobat
| rebase changes your history | 14:17 |
|
| you might not want that | 14:17 |
|
illsci
| to late :P | 14:17 |
|
| but yeah i was curious about that | 14:17 |
|
jacobat
| especially if you're sharing your branches with others | 14:17 |
|
illsci
| like if i pulled down some guys git branch and merged it | 14:18 |
|
| and decided it wasn't what i wanted | 14:18 |
|
| now that I've rebased im not sure how to undo that | 14:18 |
| → xiong joined | 14:18 |
|
illsci
| but if i had merged it, it might not have been committed | 14:18 |
|
| so i could just check out my branch again | 14:18 |
| → malumalu joined | 14:18 |
|
jacobat
| with a merge you get a merge commit | 14:18 |
|
| with a rebase you "replay" your commits on top of the new head of the master branch (or whatever you're rebasing onto) | 14:19 |
| → brodyberg joined | 14:19 |
| ← johan-s left | 14:20 |
|
illsci
| so its the order in the history that's different | 14:20 |
|
| one's merged in and the other happens following whatever was in head? | 14:20 |
|
| i can go read up on the differenes between a merge commit and what happens when you rebase | 14:21 |
|
| thanks for the help | 14:21 |
| ← illsci left | 14:21 |
| → ideogon joined | 14:22 |
| → univate joined | 14:25 |
| → intripoon joined | 14:25 |
| → johan-s joined | 14:27 |
| ← tstclair left | 14:27 |
| ← univate left | 14:27 |
| → disappearedng joined | 14:28 |
| → nud joined | 14:28 |
|
disappearedng
| Hey I am seeing "Your branch and origin/master have diverged and have 10 and 4 different commits each respectivly, how do I tell git to get origin/master | 14:28 |
| ← nud left | 14:29 |
|
jacobat
| what do you mean "to get origin/master"? | 14:30 |
| → altrux joined | 14:31 |
|
jacobat
| perhaps you're looking for "git pull origin master"? | 14:31 |
| ← roop left | 14:32 |
| → Mezner joined | 14:34 |
| → corni joined | 14:36 |
| → nullvariable joined | 14:36 |
| → d3xter joined | 14:37 |
|
disappearedng
| yeah but I want to completely disregard my current branch | 14:37 |
| → spearce joined | 14:37 |
| → psankar joined | 14:37 |
| → Swimming_Bird joined | 14:37 |
| → virus_found joined | 14:38 |
| → sumangurung joined | 14:39 |
| → rrrrutdk joined | 14:39 |
| ← sumangurung left | 14:40 |
| → sumangurung joined | 14:40 |
| → philsturgeon joined | 14:42 |
| ← loincloth left | 14:43 |
|
philsturgeon
| Hey guys. Having a problem with pushing to my newly created git repository. error "bash: git-receive-pack: command not found". I ran the following command with no luck $ git config --global remote.origin.receivepack="/usr/local/git/bin/git-receive-pack" | 14:43 |
| → slabbeh joined | 14:44 |
| ← virus_found left | 14:44 |
| → emmanueloga joined | 14:44 |
|
Ilari
| philsturgeon: On what computer? And it should definitely not be --global. | 14:45 |
| ← ph^_ left | 14:45 |
| → virus_found joined | 14:45 |
|
philsturgeon
| <Ilari> I am creating a git server and trying to push to it from my client. server is OSX 10.5.8 running Git 1.6.5.1 and client in question is Ubuntu 9.10 git 1.6.3.3 | 14:46 |
|
virus_found
| Hello. I need to get a log between two revisions. What am I doing wrong? git log --all --date=local f31cf6f..2648123 | 14:46 |
|
Ilari
| philsturgeon: Anyway, that setting would have to be set in repository sourcing the push (as repo-local setting). | 14:47 |
| → ph^_ joined | 14:47 |
|
Ilari
| philsturgeon: And 'git config --global --unset remote.origin.receivepack'. It breaks stuff. | 14:47 |
|
| virus_found: Are you sure f31cf6f is in past of 2648123? | 14:48 |
|
philsturgeon
| Ilari: That was set on server not the client. it was suggested in the Git FAQ | 14:48 |
|
Ilari
| philsturgeon: It must be set on client repo, not on server! | 14:48 |
| → nud joined | 14:49 |
| ← nud left | 14:50 |
| → chris_n joined | 14:50 |
|
philsturgeon
| Ilari: Gotcha, done. How can I solve this on the server? There must be a way to save having to do it for all clients, there will be many | 14:50 |
|
chris_n
| is it possible to split a single commit up into multiple commits during an interactive rebase? | 14:50 |
|
philsturgeon
| Ilari: first day on the job, they hear I have used Git before and want me to set up a nice working server for them. I have only ever made repo's on GitHub! :p | 14:51 |
|
Ilari
| philsturgeon: Set $PATH correctly. | 14:51 |
|
chris_n
| ie. I have 80 commits and want to split commit #2 up into two separate commits | 14:51 |
|
Ilari
| philsturgeon: For _non-interactive_ logins. | 14:51 |
|
| philsturgeon: Hint: gitolite. | 14:51 |
|
philsturgeon
| Ilari: non interactive logins? I currently have ssh://admin@ip/path.git | 14:52 |
|
virus_found
| Ilari: I don't know. >.< | 14:52 |
|
Ilari
| philsturgeon: 'ssh user@host echo \$PATH' should show the bindir of git install. | 14:53 |
|
virus_found
| Ilari: I know, that 2648123 is the result of git branch -v | awk '{print $3}' | 14:53 |
|
| And was about just to test it. | 14:53 |
|
Ilari
| virus_found: If it isn't, you either get complete log or no log at all. | 14:53 |
| → JohnBat26 joined | 14:53 |
|
Ilari
| virus_found: 'git merge-base f31cf6f 2648123' says what? | 14:54 |
| ← Swimming_Bird left | 14:54 |
|
virus_found
| f31cf6f9fe805539d50ed55b62b05ce12daf7cb0 | 14:54 |
| → ike__ joined | 14:54 |
| → Yuuhi joined | 14:54 |
|
virus_found
| That's it. | 14:54 |
|
Ilari
| virus_found: Drop that --all from log command. It makes no sense. | 14:54 |
| → mnyon joined | 14:55 |
|
virus_found
| Yes, that's better :) | 14:55 |
|
| Thank you :) | 14:55 |
| ← mnyon left | 14:55 |
|
virus_found
| I was always using --all for getting the recent log from server after git fetch. | 14:56 |
|
| For without --all it threw only local log. | 14:56 |
| ← solofight left | 14:56 |
|
Ilari
| virus_found: --all means all refs. | 14:56 |
|
| virus_found: Which doesn't make much sense if you already specify ranges. | 14:57 |
| → struberg joined | 14:57 |
| → solofight joined | 14:57 |
|
virus_found
| In this case ,yes. | 14:57 |
|
| Thank you again :) | 14:57 |
| ← virus_found left | 14:57 |
| → mm_work joined | 14:57 |
| → jepler joined | 15:00 |
| ← rrrrutdk left | 15:00 |
| → guardian joined | 15:01 |
|
philsturgeon
| Ilari: When logged into the server if I enter $ echo $PATH; I get "/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin:/usr/X11/bin:/opt/local/bin:/usr/local/git/bin | 15:01 |
|
| " | 15:01 |
|
Ilari
| philsturgeon: When logged does not matter. | 15:02 |
|
| philsturgeon: 'ssh user@host echo \$PATH' | 15:02 |
|
jepler
| I have a commit ff26d90c that is "merge branch 'xxx' into master". I want to see what was in the topic 'xxx'. Is there a more succinct way to refer to those commits? I found that the following will visualize what I think I'm looking for: $ gitk ff26d90c^2..ff26d90c^ | 15:02 |
|
guardian
| hello, I'm looking for a best-practice advice. I have a code base that is somehow "the core framework" and now I'm starting coding different apps based on this core stuff. Would you do 1 branch per app, or a different repository per app? | 15:02 |
|
philsturgeon
| Ilari: But when logged in I only get /usr/bin:/bin:/usr/sbin:/sbin | 15:02 |
| ← agile left | 15:03 |
|
Ilari
| philsturgeon: The $PATH can be different for both types of logins. Non-interactive doesn't read .bash_profile. | 15:03 |
|
| jepler: I think the parents should be other way around: X^..X^2 | 15:04 |
|
jepler
| Ilari: ah, you'd be right except that I lied. The commit is "Merge branch 'master' into xxx" which the developer turned around and merged into master as a fast-forward | 15:05 |
|
Ilari
| jepler: Ah, that explains it. | 15:05 |
| → gmcinnes joined | 15:06 |
| ← offby1 left | 15:06 |
| ← HG` left | 15:06 |
| ← NetHawk left | 15:06 |
|
gmcinnes
| hi all: how can I see the state of a given file in all branches? | 15:06 |
| ← __iron left | 15:06 |
| → jfkw joined | 15:07 |
| → jlilly joined | 15:08 |
|
jepler
| gmcinnes: you can combine for-each-ref and ls-tree | 15:09 |
|
| gmcinnes: for instance, this will show the ls-tree information about tree.h in ever branch in git: | 15:09 |
|
| git for-each-ref --shell --format="ref=%(refname)" refs/heads | while read entry; do eval "$entry"; printf "%20s " $ref; git ls-tree $ref tree.h; done | 15:09 |
|
gmcinnes
| hoorah! ok. | 15:09 |
|
jepler
| once you're looping over each branch with the git for-each-ref | while read (which I simply copied from the 'git --help for-each-ref') you can do a lot of things | 15:10 |
|
philsturgeon
| Lari: This is probably outside of your expertise now, but do you know where and how I edit $PATH? I can only find instructions for how to modify per-shell. | 15:10 |
| → jessep joined | 15:10 |
|
gmcinnes
| jepler: makes sense. thought there'd be something closer to a built in, but that's cool. | 15:11 |
|
Ilari
| philsturgeon: You want to edit it globally? | 15:11 |
|
jepler
| gmcinnes: there may be for all I know | 15:11 |
|
Ilari
| philsturgeon: Or it might be easier just to install to /usr, then one doesn't need to edit $PATH. | 15:13 |
| → offby1 joined | 15:13 |
| → r11t joined | 15:14 |
| → tlb joined | 15:15 |
|
philsturgeon
| i installed from DMG because I am lazy. I might have a go manually | 15:15 |
| → loincloth joined | 15:16 |
|
Ilari
| philsturgeon: Uh... Git hardcodes path to some binaries in its executable... | 15:16 |
| → mnyon joined | 15:17 |
|
philsturgeon
| Ilari: surely that would mean all downloads here are wrong? http://code.google.com/p/git-osx-installer/ | 15:18 |
|
| it installs to /usr/local/git/bin | 15:19 |
| → mattayers joined | 15:19 |
|
Ilari
| philsturgeon: Ah, then it must be hardcoded to use it. Try symlinking the binaries in local/git/bin from /usr/bin? | 15:19 |
| ← ph^_ left | 15:20 |
|
mattayers
| Hey there, when I rebase. If I have committed a file say 6 times, and one person edits it between my rebase. Do I really have to go through every change I made, can I not tell it just to look at my latest file? | 15:21 |
| → Swimming_Bird joined | 15:21 |
| ← oliver3 left | 15:23 |
| → oliver3 joined | 15:24 |
| ← gajop left | 15:25 |
| → JoeConyers joined | 15:27 |
| ← ^^raven^^ left | 15:29 |
| ← niki left | 15:31 |
| ← JohnBat26 left | 15:32 |
| → keystr0k joined | 15:34 |
| → schierbeck joined | 15:35 |
| → agile joined | 15:35 |
|
GarethAdams|Work
| mattayers: well, rebasing applies each commit sequentially the same as when it it was first committed. Or to put it another way, rebasing still leaves a way to go back to the point between every successive commit | 15:35 |
| → coolcat joined | 15:36 |
| ← wolog left | 15:36 |
|
GarethAdams|Work
| mattayers: so yes, rebasing does require you to look at each commit. But it sounds like merging might be a better option for you? | 15:36 |
| ← coolcat left | 15:37 |
| ← cedricv left | 15:37 |
| → madewokherd joined | 15:37 |
| → coolcat joined | 15:37 |
| ← coolcat left | 15:38 |
| → coolcat joined | 15:38 |
|
mattayers
| GarethAdams|Work: Yea it's just for when I have a project that I need to keep up to date, not actually merge it in to the master. If that's the case it's fine it's just a bit confusing. Like I could understand if it were commits from other people but I thought git might be clever enough to workout the next ones in line are all from me alone. | 15:38 |
| → cedricv joined | 15:38 |
|
GarethAdams|Work
| mattayers: maybe, but still git needs to know what the state of the file would be if you checked out a commit from halfway through your rebase | 15:39 |
| → Axius joined | 15:39 |
|
GarethAdams|Work
| I see what you're saying but I think merging is complicated enough a topic that it isn't that simple | 15:40 |
| → maxter joined | 15:41 |
| → TeckniX joined | 15:41 |
| ← tekky left | 15:41 |
| → tek joined | 15:41 |
| → wolog joined | 15:47 |
| ← emias left | 15:48 |
| ← brendon_work left | 15:48 |
| → sumangurung_ joined | 15:50 |
| ← trodrigues left | 15:50 |
| ← ice799 left | 15:50 |
| ← com4 left | 15:50 |
| → com4 joined | 15:50 |
| → _Vi joined | 15:50 |
| ← ben_h left | 15:51 |
| → irc2samus joined | 15:51 |
| → tedoc2000 joined | 15:51 |
| → ice799 joined | 15:52 |
| → rudi_s joined | 15:53 |
| → ben_h joined | 15:53 |
| ← EricInBNE left | 15:54 |
| → trodrigues joined | 15:54 |
|
brodyberg
| Let's say I am on branch a) of a repo with several branches. I then run 'git branch foo'. How do I tell later that I've branched foo from a) rather than any of the other branches in the repo? | 15:54 |
| → aspotashev_ joined | 15:57 |
| ← Axius left | 15:57 |
| ← aspotashev left | 15:58 |
| aspotashev_ → aspotashev | 15:58 |
| ← becomingGuru left | 15:59 |
| ← psoo left | 15:59 |
| → psoo joined | 15:59 |
| → tstclair joined | 16:00 |
| ← psoo left | 16:00 |
| ← d3xter left | 16:01 |
|
jepler
| brodyberg: if the reflog of the branch extends that far back, it will say what branch it was created from. Here's an example from my own project: | 16:01 |
|
| $ git log -g --oneline PR18995 | tail -1 | 16:02 |
|
| a9939cd PR18995@{5}: branch: Created from master | 16:02 |
| ← sde_antrix left | 16:03 |
|
brodyberg
| jepler: hmm, I just get HEAD | 16:05 |
| ← philsturgeon left | 16:06 |
|
brodyberg
| jepler: thank you though, it's a hint :-) | 16:06 |
| → jessep_ joined | 16:07 |
| ← johan-s left | 16:08 |
| ← sumangurung left | 16:08 |
| sumangurung_ → sumangurung | 16:08 |
| → trogan joined | 16:09 |
| → dfr|work joined | 16:09 |
| ← slabbeh left | 16:10 |
| trogan → slabbeh | 16:12 |
|
guardian
| is it possible to maintain a link between 2 tags in 2 different repositories: for instance AppBar v1.0 in AppBar.git is tied to libFoo v1.1 in libFoo.git ? | 16:13 |
| → [Po]lentino joined | 16:13 |
|
guardian
| to facilitate future checkout and rebuilding of the app | 16:13 |
| → mattyfo joined | 16:14 |
| ← Pata left | 16:14 |
|
offby1
| guardian: maybe submodules | 16:14 |
| ← whaley left | 16:15 |
|
guardian
| hmm investigating, didn't know about this feature | 16:15 |
| → juan_arandaalvar joined | 16:15 |
| → whaley joined | 16:15 |
|
mattyfo
| Hey I have a newbie git question. Let's say I've been building a program on my laptop and doing a bunch of commits with git. Is there a way I can package it all up and move it to my server to make it a private repository? | 16:16 |
| → mnyon1 joined | 16:16 |
| ← tvw left | 16:16 |
|
offby1
| of course | 16:16 |
|
mattyfo
| I ask because I don't think I can do git clone git://mylaptop | 16:16 |
|
offby1
| that _might_ work | 16:17 |
|
mattyfo
| offby1: Can I package it as a tarball? | 16:17 |
|
offby1
| sure | 16:17 |
| → ProperNoun joined | 16:17 |
|
offby1
| once you do that, you'll probably want to "connect" the two repositories somehow -- make one a "remote" of the other, e.g. | 16:17 |
|
mattyfo
| offby1: Ok, yeah I tried uploading the folder to the server (ftp) and when I run 'git status' it doesn't see anything | 16:17 |
|
offby1
| although it might be best to make the one on your server "bare" | 16:17 |
| ← jessep left | 16:18 |
| jessep_ → jessep | 16:18 |
|
mattyfo
| bare? | 16:18 |
|
offby1
| mattyfo: that description was far too vague for me to help you. Can you show me a transcript of what you actually typed, and how the computer actually responded? | 16:18 |
|
mattyfo
| offby1: Sorry, so I uploaded the files via ftp and when I type 'git status' I got 'fatal: Not a git repository' | 16:19 |
|
| is it possible ftp missed the .git folder? | 16:20 |
|
offby1
| uh | 16:20 |
|
| that's not a transcript | 16:20 |
| → gavinbeatty joined | 16:20 |
|
offby1
| you make a transcript by running a program like "script", or perhaps by copying the contents of your screen session | 16:20 |
| → [Po]lentino_ joined | 16:20 |
|
offby1
| and I can't answer the question about ftp, because I never use ftp | 16:20 |
| → vincent joined | 16:20 |
| vincent → Guest80280 | 16:21 |
|
mattyfo
| Understood, should I move the folder via scp? | 16:21 |
|
offby1
| I dunno! | 16:21 |
|
| I don't know what's wrong. | 16:21 |
|
| Sure, it's possible that ftp omitted the .git folder. | 16:21 |
|
mattyfo
| fair enough | 16:21 |
| → xanonus joined | 16:21 |
|
offby1
| But if that were the case, you could confirm or deny it, yes? By examining the remote system? | 16:21 |
| ← [Po]lentino left | 16:22 |
|
offby1
| I _will_ tell you this: scp will not arbitrarily lose subdirectories :) | 16:22 |
| ← [Po]lentino_ left | 16:22 |
|
mattyfo
| offby1: I think you've pointed out where I went wrong, I'm going to look into that right now | 16:22 |
|
offby1
| Ho Kay | 16:22 |
|
mattyfo
| Sorry to be so clueless! | 16:22 |
|
offby1
| :) | 16:23 |
| ← hadees left | 16:25 |
|
mattyfo
| Well for the record it looks like ftp missed the .git directory | 16:25 |
|
offby1
| pff | 16:25 |
|
| good catch, I guess | 16:25 |
|
| ftp is ancient and weird. | 16:26 |
|
| I'm surprised anyone still uses it | 16:26 |
| ← PN left | 16:26 |
|
mattyfo
| it's handy but you're right ancient | 16:27 |
| → sroehle joined | 16:27 |
|
mattyfo
| I'll try and stick to scp | 16:27 |
| → ruediger joined | 16:27 |
|
cbreak
| ssh costs extra with my provider | 16:27 |
|
mattyfo
| cbreak: Crazy, my host required a photo id for it but that was no biggie. I also have a Linode box which is just full bliss. | 16:28 |
| ← sroehle left | 16:28 |
| → sroehle joined | 16:28 |
| ← sroehle left | 16:28 |
| → jdahm joined | 16:29 |
|
jdahm
| is there a way to delete all untracked files in a git repository without going through all the directories that contains them? | 16:30 |
|
offby1
| "costs extra"? "photo ID"?! | 16:30 |
|
| geez | 16:30 |
|
| jdahm: "git clean -dxf" maybe? | 16:30 |
|
| not quite sure what you mean by "going through all the directories" | 16:30 |
| → v0n joined | 16:31 |
|
mattyfo
| offby1: Well, my host is pretty cheap so I don't expect much from them. On the otherhand Linode is a breeze and I have a full Linux box for whatever I want. | 16:31 |
| ← oliver3 left | 16:32 |
| ← sh1mmer left | 16:32 |
|
jdahm
| offby1: by the looks of man git-clean that'll do the trick. Thanks! | 16:32 |
|
Gitbot
| offby1: the git-clean manpage can be found at http://git.or.cz/man/git-clean | 16:32 |
| → oliver3 joined | 16:32 |
| ← \ask left | 16:32 |
|
offby1
| Gitbot: why, thank you, kind sir | 16:33 |
|
mattyfo
| yeah, and I'm now off reading the manpage for that ^_^ good gitbot! | 16:33 |
| ← Guest96666 left | 16:33 |
|
offby1
| limode++ | 16:33 |
|
| s/m/n/ | 16:33 |
|
SeveredCross
| Yeah, I've seen hosts that require photo ID for SSH. Never understood it. | 16:34 |
| ← gmcinnes left | 16:34 |
| ← xiong left | 16:35 |
| → ph^_ joined | 16:36 |
|
mattyfo
| SeveredCross: Maybe as security against abuse? But I guess someone that's going to abuse would know how to get around such things. | 16:36 |
|
SeveredCross
| It's also not like they're giving me root or something. | 16:37 |
| ← Geralt left | 16:37 |
| → jackdempsey joined | 16:38 |
| → alley_cat joined | 16:38 |
|
mattyfo
| Yeah, that's why I'm in the process of moving from Bluehost(ID required) to Linode(full Linux box to myself), Lindoe is challenging for me but I'm enjoying it | 16:39 |
| → kumbayo joined | 16:40 |
| ← mnyon left | 16:41 |
| → hobodave joined | 16:42 |
| ← hobodave left | 16:42 |
| ← tek left | 16:42 |
| ← r11t left | 16:42 |
| → tekky joined | 16:42 |
| → hobodave joined | 16:42 |
| → apeiros_ joined | 16:43 |
| ← Araneidae left | 16:43 |
| → illsci joined | 16:43 |
|
illsci
| if you want to clone a remote git repo and pull down all the branches how do you do that | 16:44 |
|
offby1
| "git clone url" | 16:45 |
|
| oddly enough | 16:45 |
| → marty_h joined | 16:45 |
|
illsci
| oh wait... i was just doing git branch instead of git branch -r | 16:45 |
| mnyon1 → mnyon | 16:45 |
| → maattd joined | 16:45 |
| → Suhail joined | 16:45 |
|
Suhail
| So I have git pull'd but I have directories that should be deleted hanging around | 16:46 |
|
| how do I clean that up? | 16:46 |
|
marty_h
| I've run into a strange issue where I can't apply a git stash due to a dirty state | 16:46 |
|
| but git status doesn't show the changes | 16:46 |
|
| even though git diff does | 16:46 |
|
| git reset --hard HEAD didn't help either | 16:47 |
|
curtana
| Suhail, if they show up as untracked in 'git status' you can remove them yourself. you can also use git-clean if you don't want to do it by hand. but use caution | 16:47 |
|
marty_h
| any suggestions? | 16:47 |
| → Araneidae joined | 16:47 |
|
illsci
| if you git reset --soft some_commit_in_the_past | 16:47 |
|
| and then look at it and want it to just to back to the way it was | 16:48 |
|
| how do you do that? | 16:48 |
| ← HokieTux left | 16:48 |
|
illsci
| like if you just wanted to see the way the repo looked back in the day, you see it, and now you want to go back to HEAD | 16:48 |
|
Suhail
| curtana: it's weird, even when i git status it doesn't show them as deleted--this probably because another branch uses those files | 16:48 |
| → Sigma` joined | 16:48 |
|
Suhail
| curtana: it just really throws things off | 16:48 |
|
| for me | 16:48 |
| → HokieTux joined | 16:48 |
| → sagsousuke joined | 16:49 |
| → sbiuhg joined | 16:50 |
| ← jackdempsey left | 16:51 |
| → WebDragon joined | 16:52 |
| ← project2501a left | 16:52 |
| ← hicham left | 16:52 |
| ← mattayers left | 16:53 |
| → hgpekm joined | 16:54 |
| ← marty_h left | 16:55 |
| → EricInBNE joined | 16:55 |
| ← jessep left | 16:56 |
| ← Sigma left | 16:56 |
| → jessep joined | 16:57 |
| ← candybar left | 16:57 |
| ← jessep left | 16:57 |
| → jessep joined | 16:58 |
| → amerine joined | 16:58 |
| → ecwNick joined | 16:59 |
| ← gebi left | 16:59 |
| ← Adlai left | 16:59 |
| → catalis joined | 16:59 |
| ← sbiuhg left | 16:59 |
| ← guardian left | 16:59 |
| ← ferdy left | 17:00 |
| → kalkin- joined | 17:00 |
|
kalkin-
| hi | 17:00 |
|
| how i use git svn with svn externals? | 17:00 |
|
| git fetch doesn't checks them out | 17:00 |
| → Pewpewarrows joined | 17:01 |
|
kalkin-
| (git svn fetch ofcourse) | 17:01 |
| ← wolog left | 17:01 |
| → mefesto joined | 17:02 |
| ← Guest80280 left | 17:02 |
| → smotchkkiss joined | 17:02 |
| ← numberGrey left | 17:03 |
| → numberGrey joined | 17:03 |
| → emias joined | 17:04 |
| → \ask joined | 17:04 |
|
mastro
| I use git svn. Once in a while i can't access our internal subversion server but I can ask some collaborator to send me files to update my local git repository. If I leave a git repository cloning the svn on a server that can always access the subversion how can i transfer the upgrades from that git repository to my git repository? Maybe with git bundle but i'm a little confused on how to use it | 17:05 |
|
| say there are this 3 remote (svn) branch i'm tracking: trunk version1 version2 | 17:06 |
| ← numberGrey left | 17:07 |
| → numberGrey joined | 17:07 |
|
mastro
| and i know which is the last revision (git commit hash) i had from that remote repos. which command should I give on the other git repository to create a bundle with the missing parts? | 17:07 |
| → godsflaw joined | 17:08 |
| → ferdy joined | 17:09 |
| ← godsflaw left | 17:09 |
|
mastro
| did you understood my problem? | 17:09 |
| → godsflaw joined | 17:10 |
| → gk__ joined | 17:10 |
| → Tureba joined | 17:10 |
| → DrAk0 joined | 17:10 |
| → tbuehlmann joined | 17:12 |
| ← catalis left | 17:13 |
|
gavinbeatty
| how do i get the following pastebin code to include the current refs/heads/new tree in the commit? | 17:14 |
|
| http://pastebin.com/m44bda951 | 17:14 |
|
| my commit object obv only has the new tree's contents, none of the previous tree's (if any) | 17:15 |
|
| can't figure how to use ls-tree et al to do it though | 17:15 |
|
| thanks | 17:15 |
| → schmukkk joined | 17:15 |
| ← lericson left | 17:16 |
| ← Suhail left | 17:16 |
| → rektide joined | 17:17 |
| ← gk left | 17:19 |
| → rnjqse joined | 17:19 |
| ← hgpekm left | 17:19 |
| → Pewpew joined | 17:19 |
| WebDragon → WebDragon|away | 17:20 |
| ← jessep left | 17:21 |
| ← ike__ left | 17:21 |
| → gebi joined | 17:22 |
| → kbingham_ joined | 17:22 |
| → tstclair_ joined | 17:22 |
| → gbacon joined | 17:23 |
| → x-demon joined | 17:23 |
| → sh1mmer joined | 17:24 |
| → maxter_ joined | 17:25 |
| tbf → tbf|eating | 17:26 |
| ← maxter left | 17:26 |
| ← marc_kdab left | 17:28 |
| → catalis joined | 17:28 |
| ← petrux left | 17:28 |
| → jessep joined | 17:29 |
| → nimrodg joined | 17:30 |
| → dreiss joined | 17:30 |
| WebDragon|away → WebDragon | 17:30 |
| → sako joined | 17:31 |
| → rjeq2 joined | 17:31 |
| rjeq2 → drigz | 17:31 |
| ← bbeecher left | 17:31 |
| ← kbingham left | 17:32 |
| ← unapiedra left | 17:33 |
| → remi joined | 17:33 |
| → dauerbaustelle joined | 17:33 |
| → bbbtwb joined | 17:33 |
| → pheaver joined | 17:34 |
| ← shruggar left | 17:34 |
| ← jessep left | 17:34 |
|
dauerbaustelle
| how can I make git diff show the diff the last commit brought without typing the commit's id? (I'm from hg, there you did hg diff --rev tip) | 17:34 |
| ← Pewpewarrows left | 17:34 |
|
nimrodg
| I'm maintaining a bare repository on a remote server for which I have SSH access. On that repository, there are two branches. On my local machine, I checked out the non-master branch and made commits to it. When I type "git push remotename nonmaster", I get "src refspec nonmaster does not match any.". When I run "git branch", I don't see "nonmaster". So I've been checking in stuff to "(no branch)". I don't understand what's going on. | 17:34 |
|
_rane
| dauerbaustelle: you can use ´git show´ for that | 17:35 |
| ← EricInBNE left | 17:35 |
|
_rane
| or do you insist on using git diff? | 17:35 |
| → kipras joined | 17:35 |
|
mastro
| dauerbaustelle, git show | 17:35 |
|
| dauerbaustelle, or git diff HEAD~ | 17:36 |
|
dauerbaustelle
| I don't care about the tool, so thanks ;-) | 17:36 |
|
mastro
| dauerbaustelle, or git log -p -1 | 17:36 |
|
| dauerbaustelle, HEAD is the "current" checked out branch, HEAD~ is the parent of the current | 17:36 |
|
_rane
| HEAD^ | 17:37 |
| ← kraymer left | 17:37 |
|
_rane
| more often, I think? | 17:37 |
|
mastro
| dauerbaustelle, if you write git diff <hash> it makes the diff between HEAD and <hash> | 17:37 |
| → JSeligstein joined | 17:37 |
|
dauerbaustelle
| thanks, thanks, thanks, so many ways oO | 17:37 |
|
JSeligstein
| Hey guys, how can I copy a file in git? | 17:38 |
| ← tstclair left | 17:38 |
| ← S1kx left | 17:38 |
| → coderdad joined | 17:38 |
|
lack
| JSeligstein: Depends what you mean by "copy"... from where to where, and for what purpose? | 17:39 |
|
dauerbaustelle
| two questions left: a) can I have the diff between ANY two revisions? and b) git log -n shows from -n to HEAD, can I show from -n to any other revision? | 17:39 |
| tbf|eating → tbf | 17:39 |
|
nimrodg
| I guess a smaller question would be what does "(no branch)" mean in git branch? | 17:39 |
| → ogeluj joined | 17:39 |
|
JSeligstein
| lack: lets say i want to split a file into two? or just copy a file for a new version? | 17:39 |
|
lack
| nimrodg: You have a detached head. | 17:39 |
|
| JSeligstein: Is there some reason just copying the file and doing 'git add' would not meet your needs? | 17:40 |
|
nimrodg
| lack: How do you deal with this problem? | 17:40 |
|
lack
| nimrodg: Hang on, finding you a resource... | 17:40 |
|
nimrodg
| lack: Is it a problem? | 17:40 |
|
JSeligstein
| lack: i could do that, but does git track that properly? | 17:40 |
|
mastro
| dauerbaustelle, a) git diff <first-revision> <second-revision>. b) git log -n <head-revision> | 17:40 |
|
GarethAdams|Work
| does anyone else do `git pull; git checkout otherbranch; git pull` instead of pull;checkout;merge? | 17:41 |
|
lack
| nimrodg: http://sitaramc.github.com/concepts/detached-head.html | 17:41 |
|
nimrodg
| lack: thanks. | 17:41 |
|
lack
| JSeligstein: Yes, it does! It uses the same logic as "rename detection" to detect copies. | 17:41 |
|
mastro
| _rane, HEAD^ and HEAD~ are equivalent, but HEAD^2 and HEAD~2 are not. and it is better to get used to the ~ notation because that's what you expect when you ask for parents. the HEAD^ is used for DIRECT parents so HEAD^2 is the second direct parent (from merging), instead HEAD~2 is the grand parent | 17:42 |
| → wolog joined | 17:42 |
|
lack
| JSeligstein: Though I suppose "properly" is open for interpretation. | 17:42 |
| ← sh1mmer left | 17:42 |
| ← catalis left | 17:42 |
|
JSeligstein
| lack: i would want the code in the new file to be properly traced to the old file | 17:42 |
|
lack
| GarethAdams|Work: Since 'pull' is a combination of 'fetch and then merge', you're really just adding an extra 'fetch' in there with the 1st of your 2 invocations... | 17:43 |
| → sh1mmer joined | 17:43 |
|
mastro
| anyone here available to share how to use git bundle in a simple example? say i have branch X and i want all commits after commit <myhash> how do i create that bundle? | 17:43 |
|
lack
| pull;checkout;pull = fetch;merge;checkout;fetch;merge whereas pull;checkout;merge = fetch;merge;checkout;merge | 17:43 |
| → sourcemaker_ joined | 17:43 |
|
mastro
| the git bundle documentation is not very easy to understand | 17:43 |
| ← coderdad left | 17:44 |
|
dauerbaustelle
| mastro, git diff -3 -2 doesn't work, aren't (negative) indices possible as revision specification? | 17:44 |
|
GarethAdams|Work
| lack: yes, I agree, but some of my colleagues always use 'pull' because it always uses the tracking branch and avoids them having to type the full remote/branch as required by merge | 17:44 |
| → coderdad joined | 17:44 |
|
lack
| GarethAdams|Work: Explain to them how to enable bash completion for git, and they can then tab-complete their branches ;) | 17:44 |
|
wereHamster
| dauerbaustelle: what would negative indices refer to? | 17:44 |
|
mastro
| dauerbaustelle, what do you want to do? compare the 3rd grand parent with the second? (grand grand parent with the grand parent) so you should do git diff HEAD~3 HEAD~2 | 17:45 |
|
dauerbaustelle
| number(HEAD~) - N | 17:45 |
|
mastro
| dauerbaustelle, the -n is for git log it say: show me the last N commit | 17:45 |
| → ask joined | 17:45 |
| ← rnjqse left | 17:45 |
|
GarethAdams|Work
| lack: My personal favourite solution would be for `git merge` to default to the tracked branch. I mean, that information is right there in my repository so why shouldn't it be used? Am I missing something? | 17:45 |
| ← ask left | 17:46 |
|
mastro
| dauerbaustelle, git log -5 mybranch: will show the last 5 commit for "mybranch" | 17:46 |
|
lack
| GarethAdams|Work: But what is your concern? There is no harm in doing another fetch because of a pull there, it's pretty quick and painless for git to check "Oh, nothing new to download? Okay, I'll continue on then" | 17:46 |
|
mastro
| dauerbaustelle, instead git diff show difference between two commit, so you had to specify the commit. If there is only one parameter git diff take HEAD as the second parameter | 17:46 |
|
| dauerbaustelle, so if you want to compare your current head with another commit you just do git diff <the-other-commit> | 17:47 |
|
GarethAdams|Work
| lack: well it just seems redundant. No real concern. Maybe under suboptimal network conditions it might be more of an issue? | 17:47 |
|
mastro
| dauerbaustelle, if you want to compare any two commit you do git diff <first-commit-hash/pointer> <second-commit-hash/pointer> | 17:47 |
| → jessep joined | 17:47 |
|
mastro
| dauerbaustelle, see this: http://progit.org/book/ch6-1.html | 17:48 |
| ← schmukkk left | 17:48 |
| → ringlej joined | 17:48 |
|
nimrodg
| lack:Thank you very much. Everything worked perfectly. | 17:48 |
| → ereslibre joined | 17:48 |
|
dauerbaustelle
| mastro, yes but I want to compare two non-HEAD-commits, for example "compare 3rd-last commit to 2nd-last commit" (< if you say so in english) | 17:48 |
|
mastro
| dauerbaustelle, git diff HEAD~3 HEAD~2 | 17:49 |
|
lack
| GarethAdams|Work: It's only redundant if no one pushed to that other branch between your first 'pull' and the second 'pull'... If it's a very busy branch it may be *better* to do the 2nd pull! But you're right, there is some redundancy and overhead of 'pull' as opposed to merge if nothing has changed. But it's as small as possible within the constraints of the network. | 17:49 |
| → catalis joined | 17:49 |
|
mastro
| dauerbaustelle, since the two commit are consecutive you could just do git show HEAD~2 | 17:49 |
| → juro joined | 17:49 |
|
lack
| GarethAdams|Work: I can see some value to having 'git merge' default to the tracked remote branch... Maybe suggest it on the official git mailing list (with a patch?) and see if it works out... | 17:50 |
|
juro
| hi, I am using subversion and am curious about git - is it possible to use a remote location as a repository? | 17:50 |
|
mastro
| dauerbaustelle, HEAD~3 means the parent of the parent of the parent of HEAD (the 3rd parent of HEAD) | 17:50 |
|
lack
| juro: Of course! Probably. What do you mean exactly? | 17:50 |
|
dauerbaustelle
| mastro, thanks | 17:51 |
|
| dauerbaustelle will try that with git diff | 17:51 |
|
GarethAdams|Work
| lack: I was just looking for the mailing list :) but I'm just a web programmer, not a real programmer ;) so a patch isn't especially easy to come up with | 17:51 |
|
lack
| mastro: Did you read 'man git-bundle'? It seems pretty straight-forward to me... | 17:51 |
|
Gitbot
| mastro: the git-bundle manpage can be found at http://git.or.cz/man/git-bundle | 17:51 |
|
lack
| Gitbot: Mailing list? | 17:51 |
| ← maxter_ left | 17:51 |
| ← sh1mmer left | 17:51 |
|
dauerbaustelle
| works, thank you very much :-) | 17:52 |
|
| good day | 17:52 |
| ← dauerbaustelle left | 17:52 |
|
juro
| lack, I have a hoster, where I cannot install anything but I was thinking of using that as an off-site repository | 17:52 |
|
lack
| juro: Well, you *can* use the http protocol to push to a remote site, but it's not as good as using git-over-ssh | 17:53 |
| → resmo joined | 17:54 |
|
mastro
| lack, i read it but i don't understand how to say: I already have commit 12345, make me a bundle with everything after this and I want the bundle of branch trunk | 17:54 |
|
resmo
| hi | 17:54 |
|
juro
| lack but I can use webspace where I have ftp access as a reopistory without installing anything there, right? | 17:54 |
|
mastro
| lack, and I don't understand if I could use to update the local copy of the remote repo or just for local branches | 17:54 |
|
| lack, also I hope it is compatible with git-svn | 17:55 |
|
lack
| mastro: You saw the invocation 'git bundle create <file> <git-rev-list args>', right? Did it occur to you to try 'git bundle create mybundle 12345..HEAD'? | 17:55 |
| ← gergap left | 17:55 |
| ← bbbtwb left | 17:55 |
|
juro
| lack, so I install git on my Ubuntu machine here and use my remote access as a repository space | 17:56 |
| → chalkbag joined | 17:56 |
| ← struberg left | 17:56 |
|
resmo
| I have uncommited changes in a file and want to commit not all changes, just a few? I know there is an git command where you can say, "yes, this I want to change" and so on, please give me a hint | 17:56 |
| ← Yuuhi left | 17:56 |
|
lack
| mastro: Well, git bundle is a replacement for 'git push' if there's no network. You do 'git bundle create' and it packs up all the commits you specify into a file. You then take that file to your other git repo and run 'git bundle unbundle' and it acts as if you had pushed those commits into the history... | 17:56 |
| → Yuuhi joined | 17:56 |
|
wereHamster
| resmo: git add -p | 17:56 |
|
| resmo: or use one of the GUIs which can do that on a per-line basis | 17:57 |
|
mastro
| lack, well.. since I use git svn i never (directly) do push on the repository, I do git svn rebase when i want to udpate and git svn dcommit when i want to commit | 17:57 |
|
| lack, i'll make some try anyway | 17:57 |
|
| lack, thanks | 17:57 |
|
| had to go | 17:57 |
| → BigMack83 joined | 17:57 |
| ← amerine left | 17:58 |
|
lack
| juro: Yes, with some performance hit... Though I'm not sure about FTP. Check the 'GIT URLS' section of 'man git-clone' and see what you think... | 17:58 |
|
Gitbot
| juro: the git-clone manpage can be found at http://git.or.cz/man/git-clone | 17:58 |
|
curtana
| git add -p is the bes tthing in the world | 17:58 |
|
| at least the world of revision control | 17:58 |
| → jstephenson joined | 17:58 |
| ← mastro left | 17:58 |
|
juro
| lack, Gitbot, thanx | 17:58 |
|
lack
| juro: There are probably some good docs on Google out there on what's required to host your repo over HTTP. But it's much better to use something else if possible. | 17:58 |
|
juro
| lack, ssh for instance? | 17:59 |
|
lack
| juro: There are a number of free git hosting solutions available... repo.or.cz github.com etc. | 17:59 |
| ← sourcemaker left | 17:59 |
|
curtana
| juro, remote on http works fine but is slow | 17:59 |
|
lack
| juro: Yes, I think for bi-directional git operations you will find ssh is the best and most common solution. | 17:59 |
|
resmo
| wereHamster, got it with your help, it was actaully "git add -i" :) thx | 18:00 |
| ← solofight left | 18:00 |
|
lack
| curtana: Did you see his earlier question about setting up a remote hosting solution for hosting a git repo? How does push work over http? | 18:00 |
|
curtana
| lack, didn't see that | 18:00 |
| → Chillance joined | 18:00 |
| ← nickrw left | 18:00 |
|
curtana
| maybe with webdav? doubt it tho | 18:00 |
| ← psankar left | 18:01 |
|
chris_n
| occasionally after doing repeated 'git rebase -i HEAD~X' git seems to loose track of where it is and loads up the to-do list with 500+ commits | 18:01 |
| ← \ask left | 18:01 |
| → ilmari joined | 18:02 |
|
ilmari
| is the tree object on-disk format documented anywhere? | 18:03 |
| ← tbf left | 18:05 |
| ← chalkbag left | 18:06 |
| → em-dash joined | 18:06 |
| → Haaarg joined | 18:06 |
| ← godsflaw left | 18:07 |
| ← Chillance left | 18:08 |
| ← tekky left | 18:08 |
| ← LiamH left | 18:08 |
|
wereHamster
| ilmari: try this: http://progit.org/book/de/ch8-8.html | 18:09 |
| → berndj joined | 18:09 |
|
wereHamster
| (or the english variant...) | 18:09 |
| → Zephyrus joined | 18:10 |
|
wereHamster
| ilmari: this is better: http://progit.org/book/ch9-2.html | 18:10 |
|
berndj
| i suppose there's no way to undo what git reset --hard <some_commit> did - specifically uncommitted / unstaged changes? not a catastrophe, more of an annoyance | 18:10 |
| → Chillance joined | 18:10 |
| → marc_kdab joined | 18:11 |
|
ilmari
| wereHamster: doesn't say anything about the actual on-disk format (I'm trying to fix the charset handling in Git::PurePerl) | 18:13 |
|
lack
| berndj: Well, if the underlying filesystem has any sort of "deleted file recovery", then *maybe*. But not really, no. | 18:14 |
|
wereHamster
| ilmari: at the very bottom, in the 'Object Storage' section | 18:14 |
|
berndj
| hehe, no, ext3 doesn't have that | 18:14 |
| → tbf joined | 18:14 |
| ← drigz left | 18:14 |
| ← Haarg left | 18:15 |
|
berndj
| it's okay, got rid of a whole lot of now-superfluous debug output | 18:15 |
| → LiamH joined | 18:15 |
| → kennethlove joined | 18:15 |
|
lack
| berndj: Ah, so it helped you ;) | 18:16 |
|
ilmari
| wereHamster: "the commit and tree content are very specifically formatted", but it doesn't say _how_ | 18:17 |
| → makeworkeasy joined | 18:17 |
| → guardian joined | 18:17 |
| coderdad → coderdad|away | 18:17 |
| → \ask joined | 18:17 |
|
ilmari
| i.e. particularly, is the filename of a tree entry utf8-encoded? | 18:17 |
|
wereHamster
| ilmari: no, it's whatever readdir() returns | 18:18 |
| ← \ask left | 18:18 |
|
ilmari
| ah | 18:18 |
|
wereHamster
| IIRC git does'nt encode anything except maybe the commit message | 18:19 |
|
berndj
| ah, i seem to have most of my nuked changes in an older stash | 18:19 |
| → WALoeIII joined | 18:19 |
| → freeride1 joined | 18:19 |
| → sockmonk joined | 18:19 |
| ← freeride1 left | 18:19 |
|
thiago
| and it doesn't encode the commit message either | 18:20 |
|
ilmari
| wereHamster: ah, git config --get i18n.commitEncoding | 18:20 |
| ← nimrodg left | 18:20 |
|
thiago
| it may save the locale name if it's not UTF-8 | 18:20 |
| → freeride joined | 18:20 |
|
thiago
| so that it may recode it when displaying somewhere else | 18:20 |
| ← illsci left | 18:21 |
| ← tchan left | 18:21 |
| → sh1mmer joined | 18:22 |
|
d-snp
| hey, my client has this error: fatal: ‘heroku’: unable to chdir or not a git archive | 18:22 |
|
| I can't seem to reproduce, when would such an error occur? | 18:22 |
|
juro
| lack, ok, then I'll start to go through the git manual .... | 18:22 |
|
d-snp
| he says he's executing: git push -f heroku master | 18:22 |
|
lack
| juro: A good plan. There is also this excellent book at http://progit.org | 18:23 |
|
d-snp
| why would it want heroku to be a git archive? | 18:23 |
|
thiago
| d-snp: it's either a remote name or it's path on disk | 18:23 |
|
juro
| lack, I live in South Africa ... it will take about 3 weeks to get anything here | 18:23 |
|
lack
| juro: Ah, but this book you can read online :) | 18:23 |
|
| For free. | 18:23 |
| ← binjured left | 18:24 |
|
juro
| oh, ok. that makes sense - thank you | 18:24 |
|
lack
| juro: Enjoy, and good luck! | 18:24 |
| → binjured joined | 18:24 |
|
juro
| lack, thanx - I'll probably come by with a few questions later :D | 18:24 |
| ← freeride left | 18:25 |
| ← curtana left | 18:26 |
| ← madewokherd left | 18:27 |
|
d-snp
| thanks thiago :) | 18:29 |
| ← vadi left | 18:29 |
|
d-snp
| I can't reproduce the error, but I just asked him to send me his .git/config | 18:29 |
|
| he probably missed the remote add step | 18:29 |
| ← bgerber left | 18:30 |
| → \ask joined | 18:33 |
| ← Pewpew left | 18:34 |
| ← froschi left | 18:35 |
| ← NET||abuse left | 18:37 |
| → psoo joined | 18:39 |
| ← sumangurung left | 18:40 |
| → sumangurung joined | 18:40 |
| ← sako left | 18:42 |
| ← apeiros_ left | 18:43 |
| → xwkovr joined | 18:46 |
| ← Chaser_ left | 18:46 |
| → godsflaw joined | 18:48 |
| → Kavu joined | 18:49 |
| → tchan joined | 18:49 |
| ← LiamH left | 18:50 |
| ← catalis left | 18:51 |
| → project2501a joined | 18:52 |
| ← \ask left | 18:57 |
| → volk joined | 18:57 |
| ← ogeluj left | 18:58 |
| → RandalSchwartz joined | 18:58 |
|
RandalSchwartz
| ok - I think I'm not crazy, but I'm sure I remember some simple hook in the default distro that would warn on conflict markers | 18:59 |
| ← volk left | 18:59 |
|
RandalSchwartz
| but the pre-commit hook .sample doesn't seem to show that | 18:59 |
| → \ask joined | 18:59 |
| → WebDragon|away joined | 19:01 |
|
RandalSchwartz
| ahh - git diff --check does it...so the pre-commit hook just needs to be moved into place | 19:02 |
| → LiamH joined | 19:02 |
| → relagil joined | 19:04 |
| ← ecwNick left | 19:05 |
| ← malumalu left | 19:05 |
|
relagil
| If I create a git repo on my comp and give access to others. They will advance the head branch while I'm on a working copy. Not messy? | 19:06 |
| ← Kavu left | 19:06 |
|
RandalSchwartz
| ? bare | 19:06 |
|
| darn it, where's the bot | 19:06 |
|
| don't push to a non-bare repo | 19:06 |
|
| unless you know why we say that, then you can. :) | 19:06 |
| → Kavu joined | 19:07 |
|
relagil
| bare = rempty? | 19:07 |
|
| bare = empty? | 19:07 |
|
RandalSchwartz
| bare = no working dir | 19:07 |
|
| just the .git part | 19:07 |
| → Kavu_ joined | 19:07 |
|
RandalSchwartz
| see the faq about "bare" | 19:07 |
| ← Kavu left | 19:07 |
| ← Kavu_ left | 19:07 |
| warthog19 → warthog9 | 19:07 |
| → Kavu joined | 19:07 |
|
RandalSchwartz
| which the bot could answer, but I can't remember the bot invocation | 19:07 |
|
| ?bare | 19:08 |
|
| nope | 19:08 |
| → RobertLaptop_ joined | 19:08 |
| → nickrw joined | 19:08 |
| ← RobertLaptop left | 19:09 |
| → thiago_home joined | 19:09 |
|
aCiD2
| Gitbot: help | 19:09 |
|
Gitbot
| aCiD2: Try '?' instead of help. | 19:09 |
|
| aCiD2: CyBot configuration directory. | 19:09 |
| ← sagsousuke left | 19:09 |
|
relagil
| iow, a git repo accessible by many should be of a special type, bare. | 19:09 |
|
fr0sty
| faq bare ? | 19:10 |
|
relagil
| The faq says how to make a bare but not why | 19:10 |
| coderdad|away → coderdad | 19:10 |
| ← Kavu left | 19:11 |
|
RandalSchwartz
| I think it motivates it somewhere | 19:12 |
| → carrus85 joined | 19:12 |
| → illsci joined | 19:14 |
|
illsci
| how would you find out what happen to a certain file that is no longer in the repo and the git log's don't seem to show where it went | 19:14 |
|
relagil
| Define submodule location using path relative to super project. Still can't be done? Mailing list mentions patch. | 19:15 |
|
wereHamster
| faq non-bare | 19:16 |
|
Gitbot
| wereHamster: Pushing to non-bare repositories is discouraged. See http://git.or.cz/gitwiki/GitFaq#non-bare | 19:16 |
| RobertLaptop_ → RobertLaptop | 19:16 |
|
wereHamster
| relagil: ^^^ | 19:16 |
| ← WebDragon left | 19:16 |
| WebDragon|away → WebDragon | 19:16 |
| → Nebukadneza joined | 19:17 |
|
Nebukadneza
| heho | 19:17 |
| ← illsci left | 19:17 |
|
Nebukadneza
| if i use a simple git:// url, on what exactly does this try to access? some http-ish stuff, ssh or a spezialized daemon? | 19:17 |
|
wereHamster
| git:// uses port 9418 | 19:18 |
| ← ilmari left | 19:18 |
| → kyron joined | 19:18 |
|
wereHamster
| and it uses a special protocol | 19:19 |
|
| Nebukadneza: man git-daemon | 19:19 |
|
Gitbot
| Nebukadneza: the git-daemon manpage can be found at http://git.or.cz/man/git-daemon | 19:19 |
|
Nebukadneza
| okay, thanks | 19:20 |
|
relagil
| Read an article on bare. Thanks. | 19:20 |
|
kyron
| Ahoy! _Q_: I've been looking for an answer to the following but I am not sure the wording is correct: "how does one push a specific commit" I have a local work branch which has many commits but I would only like to push "up to" a specific commit. I am lucky enough that the commits we want are sequential since the last pull... | 19:21 |
|
wereHamster
| kyron: you can't | 19:21 |
| ← jessep left | 19:22 |
|
wereHamster
| oh, up to a commit: git push origin <commit>:refs/heads/<branch> | 19:22 |
| ← maattd left | 19:22 |
| → brizly joined | 19:22 |
|
kyron
| wereHamster: hehe, thanks ;) | 19:22 |
|
Mage_Dude
| Is there any way to completely remove a specific commit so that it is like ti never existed, repack the repository and then pull it again to a local machine? | 19:23 |
|
kyron
| let's see how it reacts now...actually, what I would really want would be to create a new local branch _of_ that local commit to see, if pushed, if it would break all hell loose | 19:23 |
|
wereHamster
| kyron: git checkout <commit> | 19:24 |
|
kyron
| wereHamster: is it feasable to do this into a different dir (as I will most probably have to work on both branches ...sigh...) | 19:24 |
|
wereHamster
| Mage_Dude: for that you'd typically use git filter-branch (or git rebase, if your history is linear and the commit is only in one branch) | 19:24 |
|
| kyron: make a new local clone, or use the git-new-workdir script from contrib/workdir | 19:25 |
|
Mage_Dude
| wereHamster: It is linear. I pushed a 'really bad' commit to the repository that broke it and I need to undo that. I just want to take a step backwards. | 19:25 |
|
wereHamster
| is the bad commit the latest commit in your HEAD? | 19:25 |
| → maattd joined | 19:25 |
|
Mage_Dude
| wereHamster: Yes | 19:26 |
| → bousquet joined | 19:26 |
|
wereHamster
| git reset --hard HEAD~1 | 19:26 |
| → drizzd joined | 19:26 |
| ← stamina_ left | 19:27 |
| → stamina_ joined | 19:27 |
|
Mage_Dude
| wereHamster: So that will fix the local version, when I push again to the remote repository, will I get an 'out of sync' error? Because the remote is on a different commit? | 19:27 |
|
wereHamster
| push -f | 19:27 |
|
kyron
| wereHamster: ok, so I could make a local clone of a local git repo without having to push to the "origin" ? | 19:28 |
|
wereHamster
| or you can push -f origin HEAD~1:refs/heads/foo | 19:28 |
|
Mage_Dude
| Is foo the 'branch' name? ie master? | 19:29 |
|
wereHamster
| yes | 19:29 |
|
Mage_Dude
| wereHamster: Swizzle. Thanks :) | 19:30 |
| ← s4msung left | 19:31 |
| → ipmb joined | 19:31 |
|
kyron
| wereHamster: sorry for the n00b question, am reading up git-clone(1) and will ask real questions after ;) thanks! | 19:32 |
| ← project2501a left | 19:32 |
| → davr joined | 19:32 |
|
kyron
| wereHamster: although I would still like to take a look at the git-new-workdir from 'contrib/workdir' (should this be in /usr/doc/git-* ?) | 19:33 |
|
wereHamster
| <HEAD:contrib/workdir/git-new-workdir> | 19:34 |
|
Gitbot
| [git git-new-workdir]: http://tinyurl.com/yjxezrn [blob] | 19:34 |
|
kyron
| wereHamster: most obliged ;) | 19:35 |
|
| wereHamster: looking at the script, am-I mis-interpreting that the script only works with branches rather than a specific commit SHA tag? or do both work? | 19:37 |
| → munichlinux joined | 19:38 |
|
wereHamster
| you can replace <branch> with any commit (though that will detache HEAD in the othe repo) | 19:39 |
| ← Tureba left | 19:39 |
| ← juro left | 19:39 |
|
wereHamster
| or just do: git branch foo; git new-workdir . ../newworkdir | 19:39 |
|
| er, git new-workdir . ../newworkdir foo | 19:39 |
| → jrmuizel joined | 19:40 |
|
kyron
| where foo is the SHA1 ID...correct? | 19:40 |
|
wereHamster
| lemme rephrase that: git branch newbranchname <commitwherebranchshouldpointto>; git new-workdir . ../newworkdir newbranchname | 19:41 |
| ← nickrw left | 19:42 |
| → nickrw joined | 19:42 |
| → ruediger_ joined | 19:43 |
| → aliasre joined | 19:43 |
|
kyron
| oh, ok, so, I'm assigning a "new" branch name to the old commit and then calling git new-... (ie: patching that it does not permit the creation of a new local repo off an SHA1 ID....which is in fact patching the fact that I SHOULD have created a branch here anyways ;) ) | 19:44 |
| → Bartman007 joined | 19:44 |
|
wereHamster
| it allows creating a new workdir without having a branch | 19:44 |
|
| but it will detach HEAD which will probably confuse you (or maybe not) | 19:45 |
| ← ipmb left | 19:45 |
| ← jelmer left | 19:45 |
|
kyron
| well, I want to be able to have a dir wich goes back to the SHA1 ID, then try to evolve that new "dir/branch" onto a working stat without loosing the "other" work which ended up being entangled within the same branch (and should have been an entity of its own. | 19:47 |
|
| garh...wording should probably have been: how do I branch off from a local only commit given its SHA1 tag... | 19:47 |
| ← jrmuizel left | 19:48 |
|
kyron
| wereHamster: actually, what does confuse me is the following; I got: 'Note: moving to '71387cb7c40cfe11c5dc7a50972005a7c6b96d0d' which isn't a local branch' ...isn't local? | 19:48 |
| → jrmuizel joined | 19:48 |
| → jessep joined | 19:49 |
| ← relagil left | 19:49 |
|
wereHamster
| 71387 is not a branch, so it can't be local nor remote | 19:49 |
|
Nebukadneza
| i know metaquestions are evil, but did anyone of you set up a gitorious on a own box? i feel like this is hell | 19:50 |
| → razvand joined | 19:50 |
| → amerine joined | 19:50 |
|
Nebukadneza
| i've got it running, everything except checking works. and its git-daemon tells me that for git-receive-pack 'test/test.git' it couldn't find the repository/path :/ | 19:50 |
| ← xwkovr left | 19:51 |
|
kyron
| wereHamster: so, I shouldn't worry about the error message and simply give it a branch name (as it suggests) and work on? | 19:52 |
|
wereHamster
| kyron: faq detached | 19:53 |
|
Gitbot
| kyron: You are on a detached HEAD and might lose commits.. See http://git.or.cz/gitwiki/GitFaq#detached | 19:53 |
| ← bousquet left | 19:53 |
|
kyron
| thanks | 19:53 |
| → dcestari joined | 19:54 |
|
RandalSchwartz
| git checkout -b mybranch startingsha1 | 19:54 |
|
dcestari
| hello | 19:55 |
|
| can anyone help me resolve a file/directory conflict? | 19:55 |
|
| I removed some plugins from my repo and then added them as git submodules | 19:55 |
|
| but now when I pull from others I get a conflict | 19:55 |
| ← nickrw left | 19:56 |
| → nickrw joined | 19:56 |
| → Twisted_Mentat joined | 19:56 |
| ← ruediger left | 19:56 |
| ruediger_ → ruediger | 19:56 |
| → flaguy48 joined | 19:58 |
| ← jessep left | 19:59 |
| → chalkbag joined | 20:00 |
| → meppl joined | 20:00 |
| → bousquet joined | 20:01 |
| ← bousquet left | 20:01 |
| → bousquet joined | 20:01 |
|
meppl
| hello, is the git for webdav anyhow broken with some versions? | 20:01 |
|
dcestari
| anyone? | 20:01 |
|
meppl
| i type "git clone" | 20:01 |
|
| and then | 20:01 |
|
| GET /git/meppl_scrips.git/info/refs HTTP/1.1..User-Agent: git/1.5.6.5..Host: 192.168.0.7..Accept: */*.... | 20:01 |
|
| im sure it should be "/git/meppl_scrips.git/.git/info/refs" | 20:02 |
|
| actually | 20:02 |
|
lack
| Where can I find out more about the environment in which 'git filter-branch --msg-filter' is run? I want to know if I can, from that script, find the SHA1 of the current commit object that filter-branch is rewriting. | 20:02 |
|
charon
| meppl: which is quite odd since by convention, foo.git is a GIT_DIR, i.e., the contents of what foo/.git/ would hold if it were a full repository with worktree | 20:03 |
|
meppl
| git version 1.5.6.5 | 20:03 |
|
charon
| ... and upgrade. | 20:03 |
|
lack
| Basically I want to, while rewriting this tree with filter-branch, put a line at the end of the commit log message much like the one you get from 'git cherry-pick -x' | 20:04 |
| ← jstephenson left | 20:04 |
| → jstephenson joined | 20:04 |
|
charon
| lack: the sha1 is $GIT_COMMIT which is one of the very few things actually mentioned in the manpage. unfortunately, RTFS is the only recourse for many questions about g-f-b. | 20:04 |
| ← d0ugal left | 20:05 |
| ← pheaver left | 20:05 |
|
lack
| charon: Excellent ;) I love reading the fine source. | 20:06 |
|
meppl
| charon, well, if i rename meppl_scripts.git meppl_scripts, gits first request is GET /git/meppl_scrips/info/refs | 20:06 |
|
| "info/refs" is in the subdirectory ".git" of course | 20:07 |
|
lack
| charon: Ah yes, I see that bit about GIT_COMMIT now, that may be all I need. Thanks for the tip! | 20:07 |
| → jelmer joined | 20:07 |
|
meppl
| and the manual tells me to use the url like "http://host.xz/path/to/repo.git/" | 20:08 |
| ← smuf left | 20:08 |
| ← reprore left | 20:08 |
| → gunmuc joined | 20:08 |
|
wereHamster
| meppl: mv meppl_scripts.git meppl_scripts.git.bak; mv meppl_scripts.git/.git meppl_scripts.git; | 20:08 |
|
| er adjust the second mv accordingly -.- | 20:09 |
| → dauerbaustelle joined | 20:10 |
|
meppl
| cp -rp meppl_scripts.git meppl_scripts.git.bak; mv meppl_scripts.git/.git meppl_scripts.git; | 20:10 |
|
dauerbaustelle
| how do I delete unpushed commits without touching my (modified) local files? So I want nothing but that commits to disappear from the commits history | 20:11 |
|
wereHamster
| dauerbaustelle: use git stash to save the modified files | 20:11 |
| → keex joined | 20:11 |
|
dauerbaustelle
| then reset to HEAD and use stash to get the files again? | 20:11 |
|
drizzd
| if you really mean that you don't want to change anything, then git reset --soft | 20:12 |
|
wereHamster
| git stash; reset --hard HEAD; git stash pop is a noop | 20:12 |
|
dauerbaustelle
| wereHamster, apparently HEAD is my latest commit, so I have to reset to some other commit, yes? | 20:15 |
|
charon
| dauerbaustelle: perhaps you're under the (svn) assumption that HEAD designates the newest commit? it doesn't. HEAD is always what you have checked out | 20:15 |
| ← tomaw__ left | 20:16 |
|
dauerbaustelle
| weird, I haven't checked out at that revision that is HEAD right now | 20:17 |
| → toussaint joined | 20:17 |
| ← galderz left | 20:17 |
| ← ereslibre left | 20:17 |
| ← dcestari left | 20:18 |
| ← JackWinter left | 20:19 |
| ← toussaint left | 20:19 |
| → JackWinter joined | 20:19 |
| → forze` joined | 20:19 |
| ← gitte left | 20:19 |
| → ereslibre joined | 20:20 |
|
forze`
| What are the parties here mirc | 20:21 |
| ← munichlinux left | 20:21 |
| ← forze` left | 20:21 |
| ← gunmuc1 left | 20:23 |
| → brian_ joined | 20:24 |
| ← coderdad left | 20:25 |
| → tomaw__ joined | 20:25 |
| → pheaver joined | 20:26 |
|
kyron
| hmmm...can I merge a branch with a past SHA1 ID... I want to recover a past implementation of a specific file which shouldn't have been commited (yet) and is causing problems...but we want to keep the file in a current (more recent) branch... | 20:28 |
| → Corion joined | 20:29 |
| → jceb joined | 20:29 |
|
Corion
| Hi! Is there a documented problem with Git for large(r) files? I have a 900MB database dump and msysgit fails with "out of memory" for me. The system is Windows XP with 4GB RAM installed. | 20:30 |
| → komsas joined | 20:30 |
|
kyron
| or, wait, in this case, all I really want is one or two files from a past commit and revive those versions in a work branch (also based on a past commit)... I'll read up on that, that should be basic enough ;) | 20:31 |
| ← ntoll left | 20:31 |
| ← psoo left | 20:31 |
|
Corion
| I'm using the latest msysgit, 1.6.5.1 | 20:31 |
|
| Other than that, thanks for git! ;) | 20:32 |
| ← dauerbaustelle left | 20:32 |
|
kyron
| ...or..wait, is what I am describing "chrrypicking" ? | 20:32 |
|
davr
| Corion: I'd say in general git isn't really meant to hand huge files like that. It's possible git needs to keep 2 or 3 copies of a file in memory to handle it, and on a 32-bit system I think you're limited to 2GB per process | 20:33 |
| ← imajes left | 20:34 |
|
Corion
| davr: Yes, but it added larger files without problems (this is on the initial import). | 20:34 |
|
| I admit that I'm trying to abuse git as the backend for an incremental backup system, but that would make backup+restore of incremental backups well-tested at least | 20:34 |
| ← irc2samus left | 20:35 |
| ← razvand left | 20:35 |
| ← komsas left | 20:36 |
| → cytzol joined | 20:36 |
|
Corion
| But if git is designed to always have a complete file in memory/swap/address space, that limits the files that solution can backup in size... | 20:38 |
| ← rgr left | 20:40 |
|
divVerent
| we need a "git pastebin" command :P | 20:40 |
| ← aliasre left | 20:41 |
|
kyron
| divVerent: agreed... 'cuz I have something to attempt to explain which requires one to see my 'messy' tree | 20:41 |
|
divVerent
| except... IO just noticed that command shouldn't be git specific | 20:42 |
|
| the git use would be, e.g., git diff ... | pastebin | 20:42 |
|
kyron
| divVerent: its called wgetpaste under gentoo ;) | 20:42 |
| → mmitchell joined | 20:42 |
| ← mnyon left | 20:42 |
| → sykopomp joined | 20:43 |
|
divVerent
| ah, cool, thanks :P | 20:43 |
| → rgr joined | 20:43 |
| → mnyon joined | 20:43 |
|
divVerent
| not in ubuntu | 20:43 |
|
SeveredCross
| divVerent: pastebinit is. | 20:43 |
|
| git diff ... | pastebinit | 20:43 |
| ← keex left | 20:44 |
|
kyron
| divVerent: http://wgetpaste.zlin.dk/ | 20:44 |
|
divVerent
| yes, just found that one | 20:44 |
|
kyron
| ah... nm, SeveredCross answered | 20:44 |
|
divVerent
| kyron: I really don't think one can prefer one of these over the other :P | 20:44 |
|
sykopomp
| heya, I'm dealing with an older version of git, 1.5.4.3, and I don't seem to have git-remote available. Is there a different command in 1.5 that does the same thing? It would be a bit of trouble to get the sysadmin to update git :\ | 20:44 |
|
divVerent
| I mean, what user interface stuff do they do :P | 20:44 |
|
| one of these can only be better than the other if it tab completes easier :P | 20:44 |
| ← rgr left | 20:44 |
|
kyron
| divVerent: agreed | 20:44 |
|
divVerent
| which speaks against pastebinit | 20:44 |
| → rgr joined | 20:44 |
|
divVerent
| as it has paste, passwd, in its way | 20:45 |
| ← jjuran left | 20:45 |
|
lack
| sykopomp: Did you try 'git remote' instead? The format 'git-remote' is a deprecated syntax that may not work much any more. | 20:45 |
|
divVerent
| pasuspender too even :P | 20:45 |
|
| but so does wgetpaste... | 20:45 |
| ← rgr left | 20:45 |
|
| divVerent makes alias "wpaste" :P | 20:45 |
| → montylounge joined | 20:45 |
|
sykopomp
| lack: ....thank you. | 20:46 |
|
| :| | 20:46 |
|
lack
| sykopomp: Just curious, what let you to want to try 'git-remote' instead? Working off someone's old document that needs to be updated? | 20:47 |
| → rgr joined | 20:48 |
| ← gebi left | 20:48 |
| → tfarina joined | 20:48 |
| ← rgr left | 20:48 |
| → rgr joined | 20:48 |
| ← Lisimba left | 20:49 |
| → WebDragon|away joined | 20:50 |
|
sykopomp
| lack: user error. I'm embarrased to have asked :\ | 20:50 |
| ← WebDragon left | 20:50 |
| → redbaritone joined | 20:50 |
| ← rgr left | 20:50 |
| WebDragon|away → WebDragon | 20:50 |
|
lack
| sykopomp: No worries :) | 20:51 |
| → dabinich joined | 20:51 |
|
dabinich
| http://www.pandamailer.de/?bettel=pimbolli | 20:51 |
| ← redbaritone left | 20:51 |
| ← dabinich left | 20:51 |
| → catalis joined | 20:52 |
| → rgr joined | 20:52 |
| ← brizly left | 20:52 |
| → froschi joined | 20:53 |
| → mjf joined | 20:53 |
| ← TTilus left | 20:53 |
| → TTilus joined | 20:53 |
| → blaenk joined | 20:54 |
| ← DrNick left | 20:54 |
| → imajes joined | 20:54 |
| ← smotchkkiss left | 20:54 |
| → Lisimba joined | 20:54 |
| ← rgr left | 20:55 |
| → rgr joined | 20:55 |
| ← tbf left | 20:57 |
| ← rgr left | 20:58 |
| → sgh joined | 20:58 |
| → rgr joined | 20:58 |
| → aliasre joined | 20:58 |
| ← imajes left | 20:59 |
| → Modius joined | 20:59 |
| → imajes joined | 20:59 |
| → DrNick joined | 21:01 |
| ← catalis left | 21:01 |
| → sako joined | 21:02 |
| → catalis joined | 21:02 |
| ← herlo left | 21:02 |
| ← brian_ left | 21:02 |
| → coderdad joined | 21:03 |
| → virus_found joined | 21:03 |
| ← mnyon left | 21:03 |
|
virus_found
| Hello. Suppose, I have some old dated working copy (REV revision). I need to get all the changelog since REV to the topmost, most recent revision on the server. Will "%git log REV.." suffice? | 21:05 |
| rgr → ps3hamburg | 21:05 |
| ← sgh left | 21:06 |
|
virus_found
| Or %git fetch && git log REV.. ? | 21:06 |
|
| Or %git fetch && git log -all REV.. ? | 21:06 |
|
| I mean --all | 21:06 |
| → sgh joined | 21:06 |
|
virus_found
| I just can't test it :( | 21:06 |
| → Adlai joined | 21:07 |
| ← DrNick left | 21:07 |
| ← slabbeh left | 21:08 |
| ← lucsky left | 21:08 |
| → apeiros_ joined | 21:08 |
| ← flaguy48 left | 21:09 |
| → catalis_ joined | 21:10 |
| ps3hamburg → rgr | 21:10 |
|
virus_found
| %git fetch && git log REV.. downloads new sources, but shows the log as if my working copy were old. | 21:10 |
|
| %git fetch && git log --all REV.. shows the full log | 21:10 |
| ← nessundorma left | 21:11 |
|
virus_found
| How do I get new sources and show the log since REV ?? | 21:11 |
| ← catalis left | 21:11 |
| ← blaenk left | 21:11 |
| → NET||abuse joined | 21:12 |
| ← d0k left | 21:12 |
|
virus_found
| svn log SERVER -r HEAD:REV it's so simple for svn | 21:12 |
| ← ehabkost left | 21:13 |
| → tekky joined | 21:13 |
|
virus_found
| hg pull && hg log -r tip:REV so is for the mercurial | 21:13 |
| → ablemike joined | 21:13 |
|
parasti
| virus_found: learn about remote-tracking branches, e.g., "git log origin/master" | 21:15 |
| → flaguy48 joined | 21:16 |
|
virus_found
| parasti: I know something about them. But why, the changelog is the whole one. | 21:16 |
| → tfarina_ joined | 21:17 |
| ← rgr left | 21:18 |
| ← catalis_ left | 21:19 |
|
virus_found
| Well, is it that difficult? I thought, git was a very mighty tool, that could everything. | 21:20 |
| → GustavoRamos joined | 21:20 |
| ← Swimming_Bird left | 21:21 |
| → rgr joined | 21:21 |
|
parasti
| virus_found: "git log $commit..origin/master"? | 21:21 |
| → DrNick joined | 21:23 |
|
virus_found
| parasti: the won't fix anything | 21:23 |
|
| I repeat: | 21:23 |
|
| Hello. Suppose, I have some old dated working copy (REV revision). I need to get all the changelog since REV to the topmost, most recent revision on the server. | 21:23 |
| ← remi left | 21:24 |
|
DrNick
| git fetch origin; git log HEAD..origin/master | 21:24 |
| ← bentob0x left | 21:24 |
|
virus_found
| I have some old dated working copy and don't want to do anything to it. Neither fetch, nor pull. | 21:24 |
|
parasti
| virus_found: it's impossible | 21:24 |
|
virus_found
| Sorry, fetch is allowed :) | 21:24 |
|
| pull is not | 21:24 |
|
jast
| git fetch; git log HEAD..origin/master | 21:25 |
|
parasti
| HEAD can be left off in that case, "..origin/master" | 21:25 |
| → PointMan joined | 21:25 |
|
PointMan
| shrug I did a git . and added too many files, how do I remove them again? | 21:26 |
| ← mattyfo left | 21:26 |
| → methanol joined | 21:26 |
|
davr
| what's a "git ." ? | 21:26 |
|
PointMan
| sorry git add . | 21:26 |
| → dannas_ joined | 21:26 |
|
parasti
| PointMan: git status should tell you | 21:26 |
|
jast
| PointMan: git reset (without --hard!) will unstage all changes | 21:27 |
|
PointMan
| git reset HEAD <file>..." to unstage | 21:27 |
| → bentob0x joined | 21:27 |
|
PointMan
| ah | 21:27 |
|
| i wasnt familiar with the term unstage | 21:27 |
|
| thanks | 21:27 |
| ← methanol left | 21:27 |
|
jast
| now you are :) | 21:27 |
| → moosecake joined | 21:27 |
|
| davr has a git alias "unstage = reset HEAD --" | 21:27 |
|
virus_found
| That worked, thank you. git fetch origin && git log ..origin/master | 21:27 |
|
jast
| http://git.or.cz/man/gitglossary might help :) | 21:27 |
| → methanol joined | 21:27 |
|
PointMan
| but it wont undo my changes, right? | 21:28 |
|
| as in changes in files | 21:28 |
|
jast
| reset only touches files if you use the --hard switch | 21:28 |
|
PointMan
| ty | 21:28 |
|
jast
| and then it can be quite destructive if you don't know what you're doing :) | 21:28 |
| → guest39148 joined | 21:28 |
|
jast
| --merge can change files, too, but I haven't ever used it | 21:28 |
| → s4msung joined | 21:28 |
| ← jceb left | 21:28 |
| ← virus_found left | 21:29 |
| ← bousquet left | 21:29 |
| ← moosecake left | 21:30 |
| → bousquet joined | 21:30 |
| ← altrux left | 21:31 |
| ← bousquet left | 21:31 |
| → niki joined | 21:33 |
| ← Zephyrus left | 21:34 |
| ← rgr left | 21:34 |
| → rgr joined | 21:34 |
| ← tfarina left | 21:35 |
| ← diegoviola left | 21:39 |
|
kyron
| _Q_: From a branch called out of a local SHA1 commit, is it possible to checkout into this branch some files from a past commit and some from future ones?...or is my only recourse to copy the files from the "future" commit directly into the branch? | 21:40 |
| ← corni left | 21:41 |
| → s0ber joined | 21:42 |
|
teuf
| kyron: git checkout <commit_sha1> -- filename | 21:42 |
| → hipe joined | 21:42 |
| → tfarina joined | 21:42 |
|
hipe
| what's the analog to svn:externals? | 21:42 |
|
kyron
| teuf: even with commit_sha1 in the future relative to the current branche's sha1? | 21:42 |
| ← Corion left | 21:42 |
|
teuf
| kyron: git doesn't care much about time in my experience | 21:43 |
|
hipe
| submodules, looks like | 21:43 |
|
teuf
| kyron: a commit is a commit wherever and whenever it comes from | 21:43 |
| ← krynnotaur left | 21:43 |
| ← apeiros_ left | 21:43 |
|
kyron
| teuf: nice, will try that | 21:44 |
| ← spearce left | 21:44 |
| ← Zenopus left | 21:46 |
| → Zenopus joined | 21:46 |
| → spearce joined | 21:47 |
| → gunmuc1 joined | 21:47 |
| Sigma` → Sigma | 21:47 |
| ← aliasre left | 21:48 |
| ← carrus85 left | 21:48 |
| ← guest39148 left | 21:48 |
| coderdad → coderdad|away | 21:49 |
| → adamm joined | 21:51 |
|
hipe
| so i want to pull in one little project/repository of mine as a submodule to my other project, but i plan on developing and commiting fixes and features to the pulled in project. is that a use-case for submodule? or should i just plain old clone it? | 21:52 |
| coderdad|away → coderdad | 21:52 |
| ← s0ber_ left | 21:54 |
| ← ablemike left | 21:55 |
| ← mefesto left | 21:55 |
| → robinr joined | 21:55 |
| → jas4711 joined | 21:55 |
| → apeiros_ joined | 21:55 |
| → Slix_ joined | 21:57 |
| ← brodyberg left | 21:58 |
| → mnyon joined | 21:58 |
| ← tfarina_ left | 21:58 |
|
drizzd
| hipe: my general rule for submodules is: If you don't really need them, don't use them. | 21:58 |
| ← xanonus left | 21:59 |
| → dgarrett joined | 21:59 |
|
hipe
| drizzd: this is just one or two javascript files that i'm using in a different project that i planned on using across different projects. right now they live in the one project but i need them in the other project and i plan on enhancing them there | 21:59 |
|
| drizzd: i guess, what would you do in a case like that ? | 22:00 |
| ← gunmuc left | 22:02 |
| ← sykopomp left | 22:02 |
| ← Chillance left | 22:02 |
| → WebDragon|away joined | 22:03 |
| ← pheaver left | 22:03 |
| ← WebDragon left | 22:03 |
| → tedoc2000_ joined | 22:03 |
| WebDragon|away → WebDragon | 22:03 |
| ← TDJACR left | 22:04 |
| → krynnotaur joined | 22:04 |
| jotik^^ → jotik | 22:05 |
| → DerGuteMoritz joined | 22:06 |
| ← BigMack83 left | 22:06 |
| → mindworx joined | 22:06 |
| ← DerGuteMoritz left | 22:06 |
|
mindworx
| Hi everyone | 22:06 |
| ← mmitchell left | 22:06 |
|
hipe
| hi mindworx | 22:06 |
| → ablemike joined | 22:07 |
|
mindworx
| question.. if I have a website that is up and running and I want to work on phase 2 how do i do this using git | 22:07 |
|
hipe
| mindworx: is it currently under svn or something | 22:07 |
|
mindworx
| no | 22:07 |
|
| I also have someone I am working with on the project and I dont want to accidently merge into Master | 22:07 |
|
hipe
| mindworx: just curious where does the project live? do you zip it up into a tarball and ftp it to the server? | 22:08 |
| ← montylounge left | 22:08 |
| → Skinkie joined | 22:08 |
|
Skinkie
| could anyone explain me an oddity? when i checkout with http checkout i get an older version than is actually on the server present | 22:08 |
|
mindworx
| no i use capistrano | 22:08 |
|
| hipe: its a rails project | 22:08 |
| ← tedoc2000 left | 22:08 |
|
Skinkie
| so if i clone it basically gets me an older version, the log seems to be 'old' while the http server shows the latest log | 22:09 |
|
hipe
| mindworx: i'm no pro but if it were me i would get a directory tree of your project (website) on your local machine, make *sure* that you've pruded everything you don't wnat to version (log fiels, cache files temp files), "git init" in that folder, create a github project, | 22:09 |
|
| and per the instructions there, from inside your project tree to "git add remote origin git://github/your-url" or whtaever, the "git push origin master" | 22:10 |
| → maattd_ joined | 22:10 |
|
hipe
| *pruned not pruded. were it a verb | 22:10 |
|
| *git remote add (not git add remote) github will walk u thru it | 22:11 |
|
PerlJam
| Skinkie: same branch? | 22:12 |
|
mindworx
| hipe: ok how do i create a test branch off of the master branch and treat that like its a master? | 22:12 |
|
PerlJam
| Skinkie: how do you know it's "older"? | 22:12 |
|
Skinkie
| i think it is the git update-server-info error... | 22:12 |
| ← WebDragon left | 22:12 |
|
Skinkie
| confirmed | 22:12 |
| ← ablemike left | 22:13 |
|
hipe
| mindworx: i think it's "git branch mindworx-branch" but i actually haven't done this that much, sorry | 22:15 |
| → tvw joined | 22:15 |
| ← Skinkie left | 22:15 |
| ← juan_arandaalvar left | 22:15 |
|
mindworx
| hipe: no problem | 22:16 |
|
| anyone with git experience and rails? | 22:16 |
| → dtrott joined | 22:16 |
| ← cannonball left | 22:17 |
|
dtrott
| If I have a commit on my master (mainline) and I add a remote that happens to have the same code (but a different commit message) how do I join the histories ? | 22:17 |
|
teuf
| dtrott: depends what you mean by join... | 22:18 |
|
| you can use git rebase --onto to put some commits from mainline "on top" of the commit from the remote | 22:19 |
| → pheaver joined | 22:19 |
|
dtrott
| Ok I may have oversimplified the problem when I stated it one line is imported using the git-p4 (perforce import script) the other is comming from a true git remote to another git repository. | 22:20 |
|
| they are showing as seperate line despite that fact that there is no difference between some of the commits. | 22:21 |
| ← maattd left | 22:21 |
| maattd_ → maattd | 22:21 |
|
dtrott
| I am trying to show a unified view | 22:22 |
| ← keystr0k left | 22:22 |
| → catalis joined | 22:24 |
| → xiong joined | 22:25 |
| ← dgarrett left | 22:26 |
| ← jelmer left | 22:26 |
| → jelmer joined | 22:26 |
| → NetHawk joined | 22:26 |
| → balint joined | 22:27 |
| → dgarrett joined | 22:29 |
| → Sky[x] joined | 22:29 |
| ← mm_work left | 22:30 |
| ← kumbayo left | 22:31 |
| ← dannas_ left | 22:32 |
| ← catalis left | 22:32 |
| ← Whtiger left | 22:34 |
|
lack
| dtrott: Unfortunately if they are separate commit objects they have separate SHA1 IDs, so git treats them completely separate even if the actual directory contents are equivalent. | 22:40 |
| → Tiggers joined | 22:41 |
|
Tiggers
| Hey, so I'm between a rock and a hard place | 22:42 |
|
lack
| So, you have 2 choices. Either rewrite the history using grafts and/or git filter-branch and/or git rebase, or | 22:42 |
| ← ideogon left | 22:42 |
|
lack
| Just merge the 2 branches together. I think the merge should *probably* come out conflict-free if the contents are identical, so then you have a common point between the 2 | 22:42 |
|
osfameron
| how do I reorder the last 3 commits? | 22:43 |
|
lack
| Tiggers: I thought tiggers were *good* at bouncing. Just jump out. | 22:43 |
|
Tiggers
| i tried doing "git pull" on the server to update the live website from the repository on the server, and it replies "you asked me to pull w/out telling me which branch to merge with" | 22:43 |
|
makeworkeasy
| osfameron: git rebase -i HEAD~3 | 22:43 |
|
Tiggers
| I'll be jumping my way out of a job if I cant fix it | 22:43 |
|
osfameron
| makeworkeasy: ta | 22:43 |
|
lack
| Tiggers: What branch does 'git branch' say you are on? | 22:43 |
|
Tiggers
| master | 22:44 |
| ← sgh left | 22:44 |
|
davr
| git pull is just a shortcut to doing git fetch and then git merge. so you can always do it manually, "git fetch" then "git merge <remote branch>" | 22:45 |
|
Tiggers
| oh excuse me that wasn't correct | 22:45 |
|
| I was on my local machine | 22:45 |
|
osfameron
| oo, that wasn't so hard (not that I was doing anything hard) | 22:45 |
| → Whtiger joined | 22:45 |
| ← alley_cat left | 22:45 |
| ← Whtiger left | 22:45 |
|
Tiggers
| it says "no branch" | 22:45 |
| ← loincloth left | 22:45 |
| → catalis joined | 22:45 |
|
Tiggers
| yeah its kinda weird | 22:46 |
| → johnw joined | 22:46 |
| ← Mezner left | 22:46 |
|
Tiggers
| the way we have it setup is you have a /www directory that you're only supposed to pull into, but someone uploaded a file in there and now its all messed up | 22:46 |
|
davr
| Tiggers: you likely checked out a remote branch without creating a local branch first. so you need to either create a local branch or simply check out a different existing branch | 22:47 |
|
Tiggers
| theres two branches it lists on the server, (no branch) which i'm on, and master | 22:47 |
| → blaenk joined | 22:48 |
|
Tiggers
| should I do, git checkout master then? | 22:49 |
| ← balint left | 22:50 |
|
lack
| Tiggers: Well, if you're sure no one has done any commits there on the server on that detached head, yes. 'git checkout master; git pull' will get you where you want to go. | 22:50 |
|
| But be careful before you do the checkout, and read more about detached headns. | 22:50 |
|
| Tiggers: http://sitaramc.github.com/concepts/detached-head.html | 22:50 |
| ← kmartiin left | 22:52 |
| ← hipe left | 22:52 |
|
davr
| I think git would be more friendly if it required --force or similar in order to check out a detached head | 22:52 |
| ← Sky[x] left | 22:52 |
| ← Zenopus left | 22:53 |
| → TDJACR joined | 22:54 |
| → psoo joined | 22:54 |
| ← tedoc2000_ left | 22:55 |
| → Zenopus joined | 22:55 |
| → tedoc2000_ joined | 22:56 |
| → deschet joined | 22:57 |
|
Tiggers
| lack: thanks! | 22:59 |
|
| looks like I get to keep my job for another day | 22:59 |
| ← bentob0x left | 22:59 |
| → ereslibre_laptop joined | 23:00 |
| ← catalis left | 23:01 |
| ← Tiggers left | 23:02 |
|
dtrott
| lack: Thanks for the pointers, this is looking like its going to very quickly become a quagmire (due to issues with the p4 import script). | 23:03 |
| → ideogon joined | 23:04 |
| → catalis joined | 23:04 |
| → Swimming_Bird joined | 23:05 |
| → skyjumper joined | 23:06 |
| ← Zenopus left | 23:06 |
| → Zenopus_ joined | 23:06 |
| ← jrmuizel left | 23:06 |
| → jrmuizel joined | 23:06 |
| ← tlb left | 23:07 |
| AkumaStreak → AkumaStreak_CST | 23:10 |
| → EricInBNE joined | 23:10 |
| ← justinweiss left | 23:10 |
| ← kipras left | 23:13 |
| → hipe joined | 23:13 |
| → ilmari joined | 23:14 |
| → __iron joined | 23:14 |
| coderdad → coderdad|away | 23:15 |
| AkumaStreak_CST → AkumaStreak | 23:15 |
| ← psoo left | 23:18 |
| → Chaser_ joined | 23:18 |
| ← auslander left | 23:18 |
| → Theravadan joined | 23:19 |
| ← LRN left | 23:21 |
| → L_R_N joined | 23:21 |
| ← makeworkeasy left | 23:23 |
| → benjy_ joined | 23:23 |
| ← PointMan left | 23:24 |
| → hadees joined | 23:27 |
|
hadees
| how do you see all the branches on origin | 23:27 |
| → PointMan joined | 23:27 |
|
RandalSchwartz
| you mean, as it exists there, or as you've fetched them? | 23:28 |
|
| git ls-remote origin (or) git branch -a | grep origin :) | 23:28 |
| ← agile left | 23:29 |
| ← jlilly left | 23:30 |
| → squarism joined | 23:32 |
| ← apeiros_ left | 23:34 |
| ← froschi left | 23:34 |
| ← benjy_ left | 23:36 |
| coderdad|away → coderdad | 23:36 |
| → benjy_ joined | 23:36 |
| ← coderdad left | 23:36 |
| → froschi joined | 23:36 |
| ← v0n left | 23:37 |
| ← TeckniX left | 23:37 |
| ← xiong left | 23:37 |
| ← froschi left | 23:37 |
| ← vinse left | 23:38 |
|
charon
| also, for a crazily verboose listing, 'git remote show origin' | 23:41 |
|
| s/oo/o/ | 23:42 |
| → toxx joined | 23:43 |
| → kevlarman joined | 23:43 |
| ← sockmonk left | 23:43 |
|
kevlarman
| is there anything like git-svn that would let git read from/write to hg repos? (or actually, i'd be happy with read only too at this point) | 23:44 |
| → Aldwuin joined | 23:45 |
| ← chalkbag left | 23:45 |
| → rudi_s_ joined | 23:45 |
| ← jas4711 left | 23:45 |
| ← JayM left | 23:45 |
| ← jstephenson left | 23:46 |
| ← stamina_ left | 23:47 |
|
davr
| kevlarman: I think you can do it the other way around...eg push from a hg repo into a git repo | 23:50 |
| ← PointMan left | 23:51 |
|
kevlarman
| davr: that's not particularly helpful, i'm trying to deal with a project that uses hg, but i'm spending more time fighting hg than i am reading source code | 23:53 |
| ← ereslibre left | 23:55 |
| ← Chaser_ left | 23:55 |
| ← pantsman left | 23:56 |
| ← john268 left | 23:57 |
| ← jrmuizel left | 23:57 |
|
bremner
| kevlarman: in theory tailor can help | 23:59 |
| ← ereslibre_laptop left | 23:59 |
| tedoc2000_ → tedoc2000 | 23:59 |