IRCloggy #git 2009-05-14

Logs Search ←Prev date Next date→ Channels Documentation

Provider of IRC logs since 2005.
WARNING: As Freenode became unjoinable and lost all warnings in topics, we cannot log channels on Freenode anymore.

2009-05-14

← mithro left00:00
→ ashleyw joined00:00
mugwump you could push a new commit which uses the commit you want as a parent... :)00:00
cxreg mugwump: er, no, i mean if i dont even have the commit i'm trying to name00:00
mugwump but you know its id?00:00
cxreg yeah00:00
← RayMorgan left00:00
cxreg due to a rebase, or reset, or whatever00:00
mugwump well, you can make a commit with it as a parent then using git commit-tree00:00
← mjf left00:00
mugwump sure, it will be a broken link in your repo. you can silence that warning using .git/shallow00:01
cxreg oooh00:01
that's slick00:01
mugwump++00:01
mugwump the next problem will be tricking git send-pack into pretending that it 'has' the commit you 'want'00:02
→ techrush joined00:02
Pieter you can also cat something to the receive-pack on the remote, if you know what00:02
mugwump and all before whatever 'gc' runs on the remote end that might throw away the prunable commit :)00:03
cxreg heh00:03
Pieter you can do it locally on a fake repo with a --receive-pack="script with tee file|git-receive-pack" && cat tee | ssh git.....00:03
cxreg well, sure. mostly thinking within a day or so00:03
mugwump it's 0day for that hack though00:04
cxreg is it?00:05
← alezandro left00:05
cxreg i thought it was ~90 days00:05
mugwump well, it just occurred to me00:05
← patrikf left00:05
cxreg LOL00:05
0day git hax.rar00:05
← Bass10 left00:07
→ Modius joined00:08
loinclot_ → loincloth00:09
← jrmuizel left00:09
→ jerbear joined00:11
→ agenteo joined00:11
→ jrmuizel joined00:12
← texel left00:12
→ sergiodj joined00:15
→ altrux joined00:18
← mbrumbelow left00:19
→ bcardarella joined00:20
→ mithro joined00:22
← jrmuizel left00:23
← __iron left00:25
georgyberdyshev → gberdyshev00:26
← arohner left00:29
→ jrmuizel joined00:31
← altrux left00:32
← rettub_ left00:32
→ magic_hat joined00:32
→ rettub_ joined00:32
magic_hat hey all. I've got a working copy of my code on my local machine. I've made a bunch of changes to it, but I don't want to check them into the main repo. How do I go about creating a new branch for an existing working copy?00:33
← loincloth left00:34
cpg|away → cpg00:34
→ mbrumbelow joined00:34
← gb|ln left00:36
← johnw left00:36
peper hello00:37
is it possible to undo a rebase -i which involved a squash?00:37
← mbrumbelow left00:38
→ jmd joined00:39
← towski_ left00:39
peper seems i can just reset --hard00:40
→ joshsdc joined00:40
sebleier|away → sebleier00:40
→ JaredK joined00:43
← JaredK left00:43
← v0lksman left00:45
← dfabulich left00:46
bob2 magic_hat: you can just make a new branch and check it out - local changes will be preserved, and you can then commit them to the new branch00:48
jmd loves the word "just".00:49
bob2 'git branch yournewbranch ; git co yournewbranch ; git commit -m "I made changes!" -a'00:51
it's pretty "just" ;)00:51
← hobodave left00:51
← jfkw left00:52
← Alinermedia left00:54
← carrus85 left00:54
← ashleyw left00:56
→ Murr joined00:56
← stuffcorpse left00:56
→ ashleyw joined00:57
→ gb|work joined00:57
← jrmuizel left00:58
← gb|work left00:59
→ stuffcorpse joined00:59
← pointlessjon left00:59
← bdimcheff left01:00
← jpieper_ left01:04
← un|ogic left01:04
← Run left01:04
← bdowning left01:04
← Leemp left01:04
← codeshah left01:04
← spb left01:04
← Radar left01:04
← gberdyshev left01:04
← avaja left01:04
← giallu left01:04
← PupenoG left01:04
← Eridius left01:04
← astrocub left01:04
← madewokherd left01:04
← meder_ left01:04
← joevano left01:04
← bartman left01:04
← meyering left01:04
← Grum left01:04
← jmd left01:04
← eletuchy left01:04
← javatexan left01:04
← techrush left01:04
← k0001 left01:04
← john3909 left01:04
← will__ left01:04
← niki left01:04
← gwoo left01:04
← sfwc left01:04
← drewr left01:04
← thorkilnaur left01:04
← comp left01:04
← Gitzilla left01:04
← tedoc2000 left01:04
← Pistahh left01:04
← Weiss left01:04
← ska-fan left01:04
← droidcore left01:04
← apeiron left01:04
← Khisanth left01:04
← Derega left01:04
← jpeeler left01:04
← Arrowmaster left01:04
← EvanCarroll left01:04
← rphillips left01:04
← ferdy left01:04
← Ilari left01:04
← jspiros left01:04
← przemoc left01:04
← moh left01:04
← rettub_ left01:04
← jerbear left01:04
← zeno^^away left01:04
← sakyamuni left01:04
← tbf left01:04
← FreakGuard left01:04
← meebey left01:04
← Tsion left01:04
← decklin left01:04
← tttsssttt left01:04
← Lgb left01:04
← Helios left01:04
← a-atwood left01:04
← radii left01:04
← tokkee left01:04
← BooK_ left01:04
← eikke left01:04
← ruskie left01:04
← Kudos left01:04
← rektide left01:04
← TheDJACR left01:04
← ahf left01:04
← sr left01:04
← seb42 left01:04
← elliottcable left01:04
← joshsdc left01:04
← smashwilson left01:04
← gilimanjaro left01:04
← kuadrosx left01:04
← schlort left01:04
← lu_zero left01:04
← tazle_ left01:04
← tonyj left01:04
← DrNick left01:04
← fr left01:04
← dermoth left01:04
← harv left01:04
← dkr21 left01:04
← djpig left01:04
← tek0 left01:04
← kbingham left01:04
← betabot left01:04
← robtaylor left01:04
← r0bby left01:04
← Murr left01:04
← bcardarella left01:04
← amrox left01:04
← eigensolver_ left01:04
← dodo_the_last1 left01:04
← photon left01:04
← NoirSoldats left01:04
← krh left01:04
← dreiss left01:04
← poseidon left01:04
← sgrimm left01:04
← elik left01:04
← kold left01:04
← rjc left01:04
← tchan1 left01:04
← dmlloyd left01:04
← tjafk1 left01:04
← patmaddox left01:04
← xckpd7 left01:04
← tpope left01:04
← zpoley1 left01:04
← elarry left01:04
← mugwump left01:04
← eb_ left01:04
← ArjenL left01:04
← N0va left01:04
← osfameron left01:04
← eldragon left01:04
← mika left01:04
← sdiz-afk left01:04
← tvakah left01:04
← mpeterson left01:04
← tmz left01:04
← joeytwiddle left01:04
← pastie left01:04
← Knirch left01:04
← mitsuhiko left01:04
← jmspeex left01:04
← Blafasel_ left01:04
← soul9 left01:04
← ozzloy left01:04
← jsquared left01:04
← sebleier left01:04
← neoeinstein left01:04
← AkumaStreak left01:04
← spearce left01:04
← _Vi left01:04
← cpg left01:04
← merlin83 left01:04
← offby1 left01:04
← AAA_awright left01:04
← cthompson left01:04
← ft left01:04
← mikearr left01:04
← ethercrow left01:04
← argonel left01:04
← Nugget left01:04
← markelikalderon left01:04
← rolfb left01:04
← Dashkal left01:04
← univate left01:04
← thiago_home left01:04
← eno left01:04
← ruphy left01:04
← PugMajere left01:04
← samjam left01:04
← snuxoll left01:04
← troyt left01:04
← BlindOracle left01:04
← precision left01:04
← GyrosGeier left01:04
← sergiodj left01:04
← releod left01:04
← gaveen left01:04
← vanweerd left01:04
← doener left01:04
← jeremiah left01:04
← Strogg left01:04
← zuez left01:04
← straszheim left01:04
← jjuran left01:04
← jtong left01:04
← scientes left01:04
← jsbmsu left01:04
← robinr left01:04
← Fissure left01:04
← johan-s left01:04
← agib left01:04
← eMxyzptlk left01:04
← Salve left01:04
← agenteo left01:04
← Modius left01:04
← kukks left01:04
← jed left01:04
← pencilo left01:04
← rimenes left01:04
← flaguy48 left01:04
← curvature left01:04
← maron left01:04
← cedricv left01:04
← Aides left01:04
← cytrinox_ left01:04
← ftehw left01:04
← mbcharbonneau left01:04
← \ask left01:04
← krawek left01:04
← voker57_ left01:04
← unreal left01:04
← xkr47 left01:04
← microm left01:04
← rio left01:04
← agile left01:04
← diegoviola left01:04
← thumper left01:04
← bobesponja left01:04
← qhoxie left01:04
← seangrove left01:04
← peper left01:04
← xenoterracide left01:04
← JEEB left01:04
← jds left01:04
← ccherrett left01:04
← flippo left01:04
← RurouniJones left01:04
← hachi left01:04
← vmiklos left01:04
← pcutler left01:04
← emg-2 left01:04
← dereine[OFF] left01:04
← raz left01:04
← mithro left01:04
← ilteris__ left01:04
← aspotashev left01:04
← gambler__ left01:04
← Arafangion left01:04
← nick_h left01:04
← jkp left01:04
← fargiolas left01:04
← crankharder left01:04
← ayaka_t_ left01:04
← chrizz left01:04
← schweizer left01:04
← ashleyw left01:04
← magic_hat left01:04
← Sho_ left01:04
← wsc left01:04
← parasti left01:04
← drizzd left01:04
← nothingHappens left01:04
← nghr left01:04
← tango_ left01:04
← ia__ left01:04
← orafu left01:04
← pigeon left01:04
← kn1ttl left01:04
← cao left01:04
← eevar left01:04
← UukGoblin left01:04
← siprbaum left01:04
← zlin left01:04
← BearPerson left01:04
→ teuf joined01:10
→ Nafai joined01:10
→ mkching joined01:10
→ dirker joined01:10
→ thijso joined01:10
→ stick joined01:10
→ lll joined01:10
→ Leonidas joined01:10
→ jeek joined01:10
→ dna joined01:10
→ da5id joined01:10
→ ahall joined01:10
→ yzlin joined01:10
→ ndim joined01:10
→ piroko joined01:10
→ jjore joined01:10
→ jast joined01:10
→ jrockway joined01:10
→ MadCoder joined01:10
→ dedmajor joined01:10
→ jettero joined01:10
→ shana joined01:10
→ bob2 joined01:10
→ jtong joined01:10
→ jjuran joined01:10
→ zuez joined01:10
→ Strogg joined01:10
→ jeremiah joined01:10
→ doener joined01:10
→ vanweerd joined01:10
→ gaveen joined01:10
→ straszheim joined01:10
→ releod joined01:10
→ sergiodj joined01:10
→ numbah joined01:10
→ impulse9 joined01:10
→ impulze joined01:10
→ Ingmar joined01:10
→ hegge joined01:10
→ natlus joined01:10
→ pcgod joined01:10
→ maxb joined01:10
→ incd^ joined01:10
→ hgb joined01:10
→ dionoea joined01:10
→ Tali joined01:10
→ madduck joined01:10
→ avtobiff joined01:10
→ janl joined01:10
→ up_the_irons joined01:10
→ bmalee joined01:10
→ GyrosGeier joined01:10
→ kirillm_ joined01:10
→ bdimcheff joined01:10
→ qrush joined01:10
→ ashleyw joined01:10
→ Murr joined01:10
→ joshsdc joined01:10
→ jmd joined01:10
→ rettub_ joined01:10
→ magic_hat joined01:10
→ bcardarella joined01:10
→ agenteo joined01:10
→ jerbear joined01:10
→ Modius joined01:10
→ techrush joined01:10
→ eletuchy joined01:10
→ Sho_ joined01:10
→ markelikalderon joined01:10
→ Leemp joined01:10
→ codeshah joined01:10
→ mbcharbonneau joined01:10
→ \ask joined01:10
→ javatexan joined01:10
→ scientes joined01:10
→ AkumaStreak joined01:10
→ agile joined01:10
→ rolfb joined01:10
→ amrox joined01:10
→ spb joined01:10
→ k0001 joined01:10
→ john3909 joined01:10
→ spearce joined01:10
→ eigensolver_ joined01:10
→ Radar joined01:10
→ diegoviola joined01:10
→ gberdyshev joined01:10
→ kukks joined01:10
→ wsc joined01:10
→ unreal joined01:10
→ smashwilson joined01:10
→ dodo_the_last1 joined01:10
→ photon joined01:10
→ jsbmsu joined01:10
→ gilimanjaro joined01:10
→ thumper joined01:10
→ NoirSoldats joined01:10
→ krh joined01:10
→ zeno^^away joined01:10
→ jed joined01:10
→ will__ joined01:10
→ pencilo joined01:10
→ rimenes joined01:10
→ _Vi joined01:10
→ avaja joined01:10
→ niki joined01:10
→ dreiss joined01:10
→ Salve joined01:10
→ Dashkal joined01:10
→ troyt joined01:10
→ poseidon joined01:10
→ sakyamuni joined01:10
→ robinr joined01:10
→ Fissure joined01:10
→ parasti joined01:10
→ sgrimm joined01:10
→ elik joined01:10
→ krawek joined01:10
→ drizzd joined01:10
→ bobesponja joined01:10
→ gwoo joined01:10
→ tbf joined01:10
→ microm joined01:10
→ argonel joined01:10
→ giallu joined01:10
→ flaguy48 joined01:10
→ kuadrosx joined01:10
→ univate joined01:10
→ thiago_home joined01:10
→ sfwc joined01:10
→ curvature joined01:10
→ maron joined01:10
→ qhoxie joined01:10
→ johan-s joined01:10
→ jpieper_ joined01:10
→ nothingHappens joined01:10
→ kold joined01:10
→ eno joined01:10
→ nghr joined01:10
→ tango_ joined01:10
→ voker57_ joined01:10
→ rio joined01:10
→ bdowning joined01:10
→ drewr joined01:10
→ rjc joined01:10
→ thorkilnaur joined01:10
→ eMxyzptlk joined01:10
→ cpg joined01:10
→ comp joined01:10
→ tchan1 joined01:10
→ FreakGuard joined01:10
→ un|ogic joined01:10
→ ia__ joined01:10
→ PupenoG joined01:10
→ Eridius joined01:10
→ cedricv joined01:10
→ seangrove joined01:10
→ astrocub joined01:10
→ orafu joined01:10
→ merlin83 joined01:10
→ Aides joined01:10
→ ruphy joined01:10
→ madewokherd joined01:10
→ dmlloyd joined01:10
→ pigeon joined01:10
→ tjafk1 joined01:10
→ neoeinstein joined01:10
→ schlort joined01:10
→ offby1 joined01:10
→ patmaddox joined01:10
→ cytrinox_ joined01:10
→ peper joined01:10
→ xckpd7 joined01:10
→ xenoterracide joined01:10
→ lu_zero joined01:10
→ JEEB joined01:10
→ Gitzilla joined01:10
→ AAA_awright joined01:10
→ tpope joined01:10
→ PugMajere joined01:10
→ cthompson joined01:10
→ kn1ttl joined01:10
→ ft joined01:10
→ ftehw joined01:10
→ mikearr joined01:10
→ BlindOracle joined01:10
→ zpoley1 joined01:10
→ tedoc2000 joined01:10
→ xkr47 joined01:10
→ tazle_ joined01:10
→ emg-2 joined01:10
→ jds joined01:10
→ jsquared joined01:10
→ ccherrett joined01:10
→ raz joined01:10
→ samjam joined01:10
→ meder_ joined01:10
→ snuxoll joined01:10
→ joevano joined01:10
→ cao joined01:10
→ eevar joined01:10
→ meebey joined01:10
→ tonyj joined01:10
→ DrNick joined01:10
→ Tsion joined01:10
→ sr joined01:10
→ tokkee joined01:10
→ radii joined01:10
→ decklin joined01:10
→ a-atwood joined01:10
→ BooK_ joined01:10
→ eikke joined01:10
→ ruskie joined01:10
→ Kudos joined01:10
→ tek0 joined01:10
→ rektide joined01:10
→ djpig joined01:10
→ Pistahh joined01:10
→ Derega joined01:10
→ apeiron joined01:10
→ Ilari joined01:10
→ droidcore joined01:10
→ Arrowmaster joined01:10
→ rphillips joined01:10
→ Weiss joined01:10
→ Grum joined01:10
→ EvanCarroll joined01:10
→ ferdy joined01:10
→ ahf joined01:10
→ tttsssttt joined01:10
→ TheDJACR joined01:10
→ robtaylor joined01:10
→ fr joined01:10
→ Lgb joined01:10
→ dermoth joined01:10
→ betabot joined01:10
→ jspiros joined01:10
→ Khisanth joined01:10
→ moh joined01:10
→ r0bby joined01:10
→ ska-fan joined01:10
→ harv joined01:10
→ Helios joined01:10
→ przemoc joined01:10
→ seb42 joined01:10
→ dkr21 joined01:10
→ jpeeler joined01:10
→ elliottcable joined01:10
→ ArjenL joined01:10
→ sebleier joined01:10
→ mika joined01:10
→ pastie joined01:10
→ tmz joined01:10
→ soul9 joined01:10
→ eldragon joined01:10
→ mpeterson joined01:10
→ N0va joined01:10
→ elarry joined01:10
→ joeytwiddle joined01:10
→ Knirch joined01:10
→ jmspeex joined01:10
→ osfameron joined01:10
→ mugwump joined01:10
→ eb_ joined01:10
→ mitsuhiko joined01:10
→ tvakah joined01:10
→ Blafasel_ joined01:10
→ sdiz-afk joined01:10
→ ozzloy joined01:10
→ flippo joined01:10
→ bartman joined01:10
→ ethercrow joined01:10
→ pcutler joined01:10
→ RurouniJones joined01:10
→ agib joined01:10
→ meyering joined01:10
→ Run joined01:10
→ hachi joined01:10
→ Nugget joined01:10
→ dereine[OFF] joined01:10
→ precision joined01:10
→ vmiklos joined01:10
→ UukGoblin joined01:10
→ willb joined01:10
→ Korren joined01:10
→ imarcusthis joined01:10
→ s0ber_ joined01:10
→ sunoano joined01:10
→ Simon- joined01:10
→ dgfitch joined01:10
→ eMBee joined01:10
→ glommer joined01:10
→ aphexer joined01:10
→ tim_jenssen joined01:10
→ xeno joined01:10
→ Alien_Freak joined01:10
→ godog joined01:10
→ serfurj joined01:10
→ smtms joined01:10
→ csgeek joined01:10
→ tswicegood joined01:10
→ corecode joined01:10
→ jacobat joined01:10
→ thirsteh joined01:10
→ magnus_ joined01:10
→ cebewee joined01:10
→ rvsjoen joined01:10
→ beeble joined01:10
→ Zaba_ joined01:10
→ binjured joined01:10
→ mlins joined01:10
→ joevandyk joined01:10
→ uau joined01:10
→ tomaw joined01:10
→ cbreak joined01:10
→ wagle joined01:10
→ bgerber joined01:10
→ m104 joined01:10
→ daleglas1 joined01:10
→ frsk joined01:10
→ dfr|work joined01:10
→ kloeri joined01:10
→ Woody2143 joined01:10
→ ilteris_ joined01:10
→ Risar joined01:10
→ ReAn joined01:10
→ Damm joined01:10
→ josh^^ joined01:10
→ cods joined01:10
→ mutex joined01:10
→ Griffon26 joined01:10
→ wanders joined01:10
→ context joined01:10
→ beantmt joined01:10
→ pcc1 joined01:10
→ koki joined01:10
→ jpic_pro1 joined01:10
→ sgronblo joined01:10
→ PsychoTrahe joined01:10
→ milli joined01:10
→ metze_away joined01:10
→ krig joined01:10
→ jasonmay joined01:10
→ dwmw2_gone joined01:10
→ jegade joined01:10
→ hatseflats joined01:10
→ simosx joined01:10
→ reaVer joined01:10
→ lamont joined01:10
→ ph8 joined01:10
→ mstormo joined01:10
→ bdrewery joined01:10
→ Bartman007 joined01:10
→ bleything joined01:10
→ warthog9 joined01:10
→ jayne joined01:10
→ jonmasters joined01:10
→ mpr joined01:10
→ jarh joined01:10
→ hiffy joined01:10
→ shad joined01:10
→ cxreg joined01:10
→ dominikh joined01:10
→ paparent joined01:10
→ KiBi joined01:10
→ deavid joined01:10
→ pasky joined01:10
→ Gitbot joined01:10
→ Rhonda joined01:10
→ tro joined01:10
→ orospakr joined01:10
→ Garoth joined01:10
→ kelvie joined01:10
→ nomadium joined01:10
→ jefftriplett joined01:10
→ rholden joined01:10
→ Pieter joined01:10
→ coopsh joined01:10
→ andres joined01:10
→ Jc2k joined01:10
→ bgamari joined01:10
→ andreaja joined01:10
→ bla joined01:10
→ nadim joined01:10
→ Louie` joined01:10
→ nil joined01:10
→ thiago joined01:10
→ Tekerson joined01:10
→ Whtiger joined01:10
→ jepler joined01:10
→ majoh joined01:10
→ static^ joined01:10
→ levifig joined01:10
→ nickm joined01:10
→ xblaster joined01:10
→ Ratler joined01:10
→ Bombe joined01:10
→ elmex joined01:10
→ ashH joined01:10
→ pilif|away joined01:10
→ alip joined01:10
→ zaius joined01:10
→ bwalk|uni joined01:10
→ awx joined01:10
→ kirillm joined01:10
→ drfreeze joined01:10
→ PuffTheMagic joined01:10
→ joshthecoder joined01:10
→ BeerSerc joined01:10
→ AlexB joined01:10
→ tarbo joined01:10
→ charon joined01:10
→ omega joined01:10
→ kusma joined01:10
→ cardioid joined01:10
→ R_Dday joined01:10
→ navik joined01:10
→ halfline joined01:10
→ sejo joined01:10
→ laj joined01:10
→ TML joined01:10
→ nomis joined01:10
→ redondos joined01:10
→ trodrigues joined01:10
→ xmux joined01:10
→ Kasperle joined01:10
→ ElAngelo joined01:10
→ harryman100 joined01:10
→ wereHamster joined01:10
→ fonseca joined01:10
→ mattikus joined01:10
→ Tommy[D] joined01:10
→ bremner joined01:10
→ CIA-27 joined01:10
→ felipe joined01:10
→ jones- joined01:10
→ EgS joined01:10
→ flexd joined01:10
→ ReinH joined01:10
→ emias joined01:10
→ azuk joined01:10
→ siprbaum joined01:10
→ BearPerson joined01:10
→ zlin joined01:10
→ blizz joined01:10
→ exg joined01:10
→ ciaranm joined01:10
← zuez left01:11
→ zuez joined01:11
← cpg left01:11
→ cpg joined01:11
EvanCarroll how do you see a file at revision X01:11
not just the diff, but the whole file01:11
mugwump git show rev:filename01:11
or git cat-file blob rev:filename01:11
← sgrimm left01:11
← zuez left01:11
EvanCarroll where blob is the word blob?01:11
→ zuez joined01:11
mugwump EvanCarroll: see man git-cat-file01:11
Gitbot EvanCarroll: the git-cat-file manpage can be found at http://git.or.cz/man/git-cat-file01:11
EvanCarroll git show worked01:12
← AAA_awright left01:14
← Dashkal left01:14
← Murr left01:15
→ aspotashev joined01:16
→ chrizz joined01:16
→ nothingmuch joined01:16
→ Mutru joined01:16
→ elspeth joined01:16
→ nick_h joined01:16
→ fanf2 joined01:16
→ PerlJam joined01:16
→ reuss joined01:16
→ Nilbus joined01:16
→ jkp joined01:16
→ Arafangion joined01:16
→ crankharder joined01:16
→ fargiolas joined01:16
→ kadoban joined01:16
→ gambler__ joined01:16
→ ayaka_t_ joined01:16
← aspotashev left01:16
→ mithro joined01:16
→ segher joined01:16
→ aspotashev joined01:17
→ ilteris__ joined01:17
← kirillm left01:17
→ SamB joined01:18
← javatexan left01:18
← releod left01:19
→ IRSeekBot joined01:19
← ashleyw left01:19
← dreiss left01:24
← wsc left01:26
→ ezyang joined01:27
ezyang Heya, lets say I have a directory with unversioned file, and a repository that corresponds to the unversioned files (think of the former as a untarred git export result). Is there any way to magically turn the former into a git repository, with the working copy setup correctly?01:28
mugwump sure, but why?01:29
I mean, if you know that the repository corresponds to a tree you can just check it out01:29
ezyang We maintain a bunch of autoinstalled web scripts on 1000+ user accounts, and we want to convert them into something that uses Git as the update infrastructure01:29
Doing a fresh checkout has too many failure modes.01:30
mugwump should be safe with: git clean -x -d -f; git checkout commit01:30
maybe add a git reset --hard in there too01:30
ezyang Um, wouldn't that blow away local changes?01:31
mugwump right01:31
ezyang That's bad. We want to *keep* any local changes01:31
→ RandalSchwartz joined01:31
mugwump ezyang: well, go read man git-read-tree01:32
Gitbot ezyang: the git-read-tree manpage can be found at http://git.or.cz/man/git-read-tree01:32
ezyang I suppose I could make a checkout and then copy the .git over to it.01:32
mugwump you can also set GIT_DIR01:32
ezyang: and GIT_INDEX_FILE - see ENVIRONMENT section of man git01:32
ezyang Hmm. That's an interesting approach, but I don't think we can actually modify all of the environments necessary.01:33
git-read-tree... don't I have to be in a repos for that to work?01:34
mugwump git read-tree by default modifies only the index01:34
→ th3lib joined01:34
mugwump by setting GIT_DIR and/or GIT_WORK_TREE (iirc) you don't have to copy things around01:35
by using a temporary index you can use the same git repo for multiple checkouts01:35
ezyang Ah, I see. So only our scripts twiddle the the environment variable and actually see the repository; everyone else just thinks they've still got a normal checkout.01:35
Hmm. I wonder if that's a good idea.01:36
mugwump for deployments it's a good idea - keep your .git well away from your webroots01:36
← bcardarella left01:37
ezyang Fair enough.01:37
← qhoxie left01:37
→ johnw joined01:38
th3lib Hi, I'm getting a compile error with 1.6.3 (and .1) on OpenBSD, and was wondering if anyone else here had come across this issue.01:40
← wycats left01:42
ezyang What's the error?01:42
th3lib CC reachable.o01:42
CC read-cache.o01:42
th3lib read-cache.c: In function `fill_stat_cache_info':01:42
th3lib read-cache.c:72: error: structure has no member named `st_ctim'01:42
th3lib read-cache.c:73: error: structure has no member named `st_mtim'01:42
th3lib read-cache.c: In function `read_index_from':01:42
th3lib read-cache.c:1302: error: structure has no member named `st_mtim'01:42
th3lib read-cache.c: In function `write_index':01:42
th3lib read-cache.c:1567: error: structure has no member named `st_mtim'01:42
th3lib gmake: *** [read-cache.o] Error 101:42
Damm sighs01:42
th3lib This is on OpenBSD 4.4 AMD6401:42
← bdowning left01:42
Damm th3lib, pastie.org or gist.github.com01:42
th3lib d'oh!01:43
sorry about that01:43
Damm th3lib, your more then likely going to need the patched version from pkgsrc or ports or whatever openbsd uses now01:44
th3lib The whole thing can be seen here: http://gist.github.com/11142001:44
damm: I've not had this error on any previous version of git - it has always compile fine from the official tarballs01:45
mugwump th3lib: looks like the #ifdefs in git-compat-util.h are wrong for your platform01:45
if you have a known good and a known bad version then you can bisect it01:46
Damm th3lib, should be able to now adays... but when it comes to openbsd01:46
heh01:46
mugwump patches for platform support should be included01:46
→ ashleyw joined01:46
mugwump it's just not every platform is tested for every release so regressions can sneak in01:46
← stuffcorpse left01:47
th3lib mugwump: I'll clone the repo and run a bisect01:47
→ stuffcorpse joined01:47
th3lib The version I'm currently running on the OpenBSD box is 1.6.2.2 - which obviously compile OK01:48
Damm mugwump, oh it's not my lack of faith in git, it's openbsd.01:48
mugwump, a quick google on openbsd + st_ctime made me remember why i stopped touching obsd01:48
RandalSchwartz what's wrong with st_ctime on obsd?01:49
← parasti left01:50
RandalSchwartz I did a google, didn't notice anything odd01:50
← kukks left01:51
← rimenes left01:51
← kuadrosx left01:51
→ kuadrosx joined01:52
← poseidon left01:55
← bdimcheff left01:58
→ bdimcheff joined02:00
→ bnferguson joined02:02
← amrox left02:02
← bnferguson left02:06
→ hyperair joined02:08
← ezyang left02:08
← techrush left02:08
→ amrox joined02:09
← smashwilson left02:10
→ scarabx joined02:10
← pencilo left02:14
→ pencilo joined02:14
th3lib OK - I've done the bisect, and c06ff is the first bad commit02:14
"Record ns-timestamps if possible, but do not use it without USE_NSEC"02:14
Damm RandalSchwartz, nothing really02:15
→ hahdhekt3 joined02:15
Damm it's just the attitudes i saw is all02:15
mugwump th3lib: that looks highly relevant02:15
← Modius left02:15
← hahdhekt3 left02:15
← magic_hat left02:15
mugwump time to write to the list! Even better, time to see if you can patch it and then write to the list! :)02:15
→ bdowning joined02:15
→ Modius joined02:16
th3lib OK, I'll see what's going on. If I can't figure it out, I'll just send a report02:16
I've not had much C programming experience, so I wouldn't hold out much hope :)02:16
→ bpalmer joined02:17
→ fujin joined02:18
mugwump #ifdefs are pretty simple preprocessor stuff02:18
but figuring out the right thing can still be tricky02:18
th3lib I'll see what I can see. But first, time for lunch :)02:19
← rolfb left02:19
← drewr left02:19
→ Bass10 joined02:19
th3lib BTW, this is the first time I've used bisect - it's very nice.02:20
→ cytrinox joined02:20
← th3lib left02:21
→ texel joined02:21
→ bcardarella joined02:21
→ arohner joined02:22
→ jereme joined02:24
← bdimcheff left02:27
→ bdimcheff joined02:27
→ jrmuizel joined02:29
→ STiAT joined02:32
→ AAA_awright joined02:33
STiAT Hi. Is there any serious way to get a "normal" patchfile out of git between master and branch of the current directory?02:33
← qrush left02:33
STiAT like a patch somebody else could import to svn?02:33
← zeno^^away left02:34
johnw you could just use git diff02:35
"patch" can read that02:36
← cytrinox_ left02:36
bpalmer you may also want to use git-format-patch , which can break it into many small patchfiles that can be applied to svn02:36
johnw Git patch files are meant to be consumed by git, as they contain information relevant to the commit itself, as well as the changes made by it02:36
yes, what bpalmer said02:37
I hand off Git patch files (made by git-format-patch) to co-workers who use CVS all the time02:37
STiAT but patch can only read that with -p1 or -p2 option due to the a/ and b/ added by git02:37
johnw is that a problem?02:37
STiAT depends on, for people not, for kdes reviewboard it is02:37
→ wmoxam joined02:37
← Bass10 left02:39
→ sgrimm joined02:39
← jerbear left02:40
→ diegovio1a joined02:41
→ eno__ joined02:44
← flaguy48 left02:45
← bcardarella left02:47
← kold left02:48
← eno left02:48
eno__ → eno02:48
← amrox left02:48
→ tjafk2 joined02:52
photon → Guest5257502:52
← cytrinox left02:56
→ photon joined02:57
← Guest52575 left02:58
← bdimcheff left02:58
← diegoviola left02:58
→ bnferguson joined03:00
← texel left03:00
← warthog9 left03:01
← hyperair left03:06
→ tansell joined03:09
← tjafk1 left03:09
tansell anyone here familiar with how gitk build's it's merge tree graphic thing?03:09
← fujin left03:10
→ hyperair joined03:10
← spearce left03:10
→ fujin joined03:11
← markelikalderon left03:11
← seb42 left03:12
→ bcardarella joined03:12
← fujin left03:14
→ fujin joined03:15
→ releod joined03:15
← fujin left03:15
← cedricv left03:16
→ gberdyshev_ joined03:19
← jereme left03:25
sdiz-afk → sdiz03:26
← photon left03:27
← k0001 left03:29
→ amrox joined03:30
← bcardarella left03:30
→ dfabulich joined03:31
← gberdyshev left03:33
→ Paf joined03:33
← SamB left03:36
← ashleyw left03:37
cpg → cpg|biab03:37
← STiAT left03:38
→ ajonat joined03:40
→ m104_ joined03:43
← m104 left03:44
← jmd left03:44
→ mfilipe joined03:45
→ offby1` joined03:48
→ wycats joined03:49
← orafu left03:49
→ orafu joined03:49
→ dreiss joined03:52
→ WALoeIII joined03:54
← dfabulich left03:55
→ mfilipe_ joined03:57
→ girishr joined03:58
← maron left04:00
← scarabx left04:00
← offby1 left04:02
→ schacon joined04:02
← agenteo left04:05
← jrmuizel left04:06
← offby1` left04:07
→ ashleyw joined04:08
← Sho_ left04:09
→ mfilipe__ joined04:12
← mfilipe left04:14
← gaveen left04:14
→ mfilipe joined04:15
→ offby1 joined04:16
← mfilipe_ left04:16
→ gaveen joined04:17
→ mfilipe_ joined04:17
← schacon left04:19
→ mfilipe___ joined04:22
cpg|biab → cpg04:24
← joshsdc left04:26
→ mfilipe____ joined04:27
← mfilipe__ left04:30
← mfilipe left04:33
→ mfilipe joined04:34
← mfilipe____ left04:35
← mfilipe_ left04:38
→ mfilipe_ joined04:39
← gaveen left04:39
→ wsc joined04:39
→ gaveen joined04:40
→ eddyp joined04:41
← mfilipe_ left04:41
← mfilipe___ left04:43
→ mfilipe_ joined04:43
→ altrux joined04:43
← wsc left04:49
→ wsc joined04:49
→ girishr_ joined04:50
← wsc left04:52
→ abstrusething joined04:52
→ wsc joined04:52
→ jdahm joined04:53
← mfilipe left04:54
← mpr left04:55
← giallu left04:59
→ mpr joined05:00
← wsc left05:00
→ wsc joined05:00
← curvature left05:01
→ curvature joined05:01
← wsc left05:01
→ psankar joined05:02
← mithro left05:05
→ mlins_ joined05:05
← girishr left05:05
fr → fr|05:07
← altrux left05:09
→ vals_ joined05:10
← girishr_ left05:11
← tchan1 left05:11
→ tchan1 joined05:12
← abstrusething left05:12
← wmoxam left05:12
→ Weaselweb joined05:14
← mlins left05:16
→ ezyang joined05:17
ezyang I want to host lots of working copies (which may be on different commits) out of a single .git (this is distinct from a single object repository). Is this possible?05:17
Fissure possible? probably. a good idea? unlikely05:18
though you might want to look into the git-new-workdir script in contrib/05:19
→ joseph_ joined05:19
ezyang Ok. My next question, then, is how can I convert a plain directory of files (unversioned, a bit like an untarred git export) into a git repository in place?05:20
RandalSchwartz git init :)05:20
done05:20
that's what git init does05:20
ezyang We have a different git repository that we'd like to "stick" the files on.05:20
RandalSchwartz makes a .git05:20
← ajonat left05:20
joseph_ → twas_05:20
RandalSchwartz git init in place05:20
then add all files05:20
commit05:20
and then push it05:20
→ priidu joined05:20
ezyang I am familiar with the basic use-case of Git.05:21
twas_ → twas__05:21
ezyang Hmm, how to explain better.05:21
RandalSchwartz git init; git commit -a -m "tarball from Jerry"05:21
Nugget "basic use-case of git" sounds like an oyxmoron :)05:21
RandalSchwartz git remote add /path/to/that/other/repo05:21
git push that05:21
yeah - basic use-case05:21
← tango_ left05:21
RandalSchwartz there's no such beast05:21
← mfilipe_ left05:22
ezyang Ok, so we have N users who each of have an autoinstalled web script. These users may have different versions (we know which version they have, though), and may have made local changes.05:22
→ girishr joined05:22
ezyang These directories are NOT versioned.05:22
RandalSchwartz each time you talk, this gets more confusing05:22
ezyang Let's say I create a Git repository, with history v1.0 -> v1.1 -> v1.205:22
RandalSchwartz why didn't you start with your real question05:22
RandalSchwartz waits for the next version of the question05:22
RandalSchwartz issues "git pull from_the_future"05:23
ezyang I want to know convert the unversioned directories into checkouts of the new Git repository05:23
RandalSchwartz "unversioned directories"?05:23
ezyang Yep, unversioned directories05:23
This is a sort of migration thing.05:23
RandalSchwartz can you say that again without using those words05:24
ezyang Ok.05:24
RandalSchwartz all directories in git are unversioned :)05:24
ezyang s/directory/working copy/05:24
RandalSchwartz the working copy is unversioned as well05:24
so that doesn't add anything05:24
ezyang s/working copy/directory that is not a working copy/05:25
RandalSchwartz are you saying you want commit 1 to represent one tree05:25
→ diegoviola joined05:25
RandalSchwartz then commit 2 to represent a second tree05:25
ezyang Let's try again.05:25
RandalSchwartz with commit 1 being its parent05:25
like that?05:25
ezyang I want to do something that would be functionally equivalent to copying .git from one directory to another, and then twiddling with .git/HEAD05:26
RandalSchwartz no you don't05:26
RandalSchwartz grins05:26
ezyang Why not?05:26
→ jceb joined05:26
RandalSchwartz well, you don't want to copy .git around05:26
I think maybe you mean, keep updating the tree05:26
brb05:26
ezyang :-/05:27
An alternate phrasing of the problem is I want to git clone a repository, and then cp -r a directory full of local changes on top of that. However, I want to do this all in place.05:27
vals_ → tango_05:27
ezyang Since the original directory has some special ACLs and files in it, and I'd rather not "touch" all of them if I don't have to.05:28
johnw you could use rsync -cr05:29
this copies only files whose contents have changed05:29
it pretty much ignores metadata05:29
of course, it's slow for that very reason too05:29
→ fujin joined05:30
ezyang The trouble is that the directory I'd be coping from is considered the "original" copy. Whatever git setup for me is after the fact.05:30
It's a long and sordid migration story.05:30
johnw i still don't know exactly what you want then05:30
→ Tuomas joined05:31
ezyang Let's say you downloaded some software and untarred it into a directory, and made some changes. At some later point, you took that software and stuck it in a Git repository. You want to turn the original directory into a git working copy with local changes, but you don't want to mv the original directory out of the way, do a git clone, and then cp -R the original directory over that.05:31
I'm not sure how clear that is.05:32
→ peshalto joined05:32
johnw you know, you can just copy the .git directory into the directory you want to become a working tree05:33
then "git status" will show you whatever changes it has05:33
relative to the .git repository that you just copied into it05:33
ezyang I mentioned that earlier.05:33
johnw or even better, just set GIT_DIR05:33
ezyang Is that the best way to do it?05:33
johnw like, GIT_DIR=/path/to/.git git status05:33
ezyang Setting GIT_DIR was also considered.05:33
→ nighthwk1 joined05:34
johnw well, I can't know if it's the best way, since I'm still not entirely clear what you want05:34
ezyang The problem with that is when you have many of these plain directories you suddenly want to version.05:34
johnw but it "sounds" like what you're asking for05:34
→ davidad joined05:34
ezyang Yep, it does.05:34
← nighthwk1 left05:34
← davidad left05:34
ezyang Sharing a .git directory was very attractive, since it meant we wouldn't have to scribble in N unversioned directories.05:36
However, since each of these copies could be on different commits of the original, you'd also need a way to specify different HEADs, which wasn't possible.05:37
Ilari ezyang: You want way to specify the states to declare derivation from and what ref to update?05:38
→ djgera joined05:39
← niki left05:39
ezyang Possibly!05:39
Or maybe I just want separate clones for all of my copies and put them all on distinct branches.05:39
xenoterracide hmm is there a way to remove a file from git but not the filesystem?05:40
johnw git rm --cached05:40
← codeshah left05:40
Ilari ezyang: Well, using 'write-tree + commit-tree + update-ref' comoo, you can control lot more (apart that commit-tree wants message from stdin, which is bit nasty).05:40
*combo05:40
ezyang I'm not quite sure I want to use plumbing commands for this :-/05:41
→ agenteo joined05:41
← diegovio1a left05:42
ezyang Back to the other question; is there any problem with me copying in a .git directory and frobbing HEAD?05:42
→ codeshah joined05:42
Ilari ezyang: The index has parts which don't copy properly. You would need 'git update-index --refresh' after copying.05:43
ezyang: There are stuff like inode numbers there.05:43
← ilteris__ left05:44
ezyang Hmm... that wouldn't be too much of a problem.05:44
Also, I would try to enforce the index to be clean before copying .git.05:44
Ilari ezyang: Not enough. You still need to fix the index after copy.05:44
ezyang Understood.05:45
Ilari ezyang: And that command would fix dirty index just as well as clean one.05:45
ezyang Are there any other problems with this approach?05:45
→ jsbmsu1 joined05:45
→ ilteris__ joined05:46
← jsbmsu1 left05:47
Ilari ezyang: BTW: AFAIK, one can use 'ln' to frob HEAD, as symlinks as symrefs is still supported.05:48
ezyang Oh wow.05:48
So we use a relative symlink and store the head in the other directory?05:49
That's trippy.05:49
Ilari ezyang: You could also create .git directory that symlinks in the common stuff (refs, objects and logs) and has its own copy of HEAD and index.05:50
RandalSchwartz goes away, and comes back to find this to be impossibly hard instead of trivial05:51
Ilari ezyang: Except that 'logs/HEAD' should also be private to that .git dir.05:51
RandalSchwartz wonders what went wrong05:51
Ilari ezyang: the symlinks as symrefs thing is that if git finds that HEAD is symlink to 'refs/heads/foo', it treats it as symref to 'refs/heads/foo'.05:52
ezyang Ilari: I'd rather not do that.05:52
Oh.05:52
That doesn't actually work, since we still need a unique HEAD for all checkouts.05:52
I suppose what we really need is an environment variable akin to GIT_DIR just for HEAD :-)05:53
Well, you have some great suggestions.05:53
Ilari ezyang: That would require messing with low-level commands.05:54
ezyang: Probably from script.05:54
mugwump /last/lastlog clear05:54
/fail05:54
ezyang Heh05:55
→ ph^ joined05:56
← robinr left05:57
← ph^ left05:59
→ cyberglitch joined06:00
← offby1 left06:01
← stuffcorpse left06:01
→ stuffcorpse joined06:02
← releod left06:02
→ ankit9 joined06:03
→ markelikalderon joined06:04
← peshalto left06:04
→ wshimmy joined06:04
→ peshalto joined06:05
→ ilteris___ joined06:06
fr| → fr06:06
← johan-s left06:09
← jsbmsu left06:13
→ kraymer joined06:20
← ilteris___ left06:20
← wycats left06:23
← ilteris__ left06:25
→ giallu joined06:26
→ psychoschlumpf joined06:28
psychoschlumpf why does "git describe" return "fatal: cannot describe" when i manually checked out a tagged commit?06:29
Fissure lightweight tag?06:29
← ankit9 left06:30
Fissure you need to use --tags for lightweight tags to be counted06:30
psychoschlumpf whats the difference between a tag and a lightweight tag?06:30
or is there a difference?06:31
mugwump a lightweight tag has no 'tag' object, it's just a ref06:31
ie, no message, no signature, etc06:31
psychoschlumpf and how do i add a tag with a tag reference?06:31
mugwump make a tag with a message06:31
psychoschlumpf ahhh06:33
thank you :-)06:33
← peshalto left06:35
← tchan1 left06:36
→ zwu joined06:38
→ tal67 joined06:39
→ WALoeIII_ joined06:40
→ jony joined06:42
← tal67 left06:42
← gberdyshev_ left06:42
→ tal67 joined06:45
→ ankit9 joined06:45
← jceb left06:47
← Radar left06:47
→ bentob0x joined06:48
→ ceej joined06:48
← hyperair left06:48
→ hyperair joined06:49
← codeshah left06:49
→ marc_m joined06:49
← argonel left06:52
Fissure it's funny watching linus rant about how much he hates OS X filesystems :)06:53
bpalmer watch where?06:54
→ Grahack joined06:55
Fissure the big "cross-platform version control" thread on the mailing list06:55
→ Bogh joined06:56
← WALoeIII left06:56
→ tchan1 joined06:56
snuxoll well, HFS+ isn't exactly what I'd call great06:57
it's certainly not terrible by any strech of the word though06:57
Fissure well, the filename you pass to open() and the one you get back from readdir() are not the same in some cases06:59
← ankit9 left06:59
→ ankit9 joined06:59
Fissure which can be annoying when you're writing a version control system meant to work with multiple platforms06:59
← kuadrosx left07:00
→ solydzajs joined07:01
→ johan-s joined07:02
→ ciskje joined07:02
← markelikalderon left07:03
→ abstrusething joined07:04
← avaja left07:04
→ neunon joined07:05
→ mikeando joined07:05
→ tchan joined07:05
→ ph^ joined07:06
→ galderz joined07:07
→ jsbmsu joined07:08
→ kbingham joined07:08
Ilari Not to mention the one you get from readdir isn't exactly sane...07:10
Fissure they do normalization all crazy-like?07:12
→ peshalto_ joined07:13
→ cilly joined07:14
← Grahack left07:14
→ mc_plectrum joined07:15
← tchan1 left07:15
→ lohrun joined07:17
← sakyamuni left07:19
→ dohzya joined07:22
→ the_zeno joined07:24
joshthecoder → joshthecoder_afk07:24
← Paf left07:24
Ilari IIRC, not only use NFD, which in itself is crazy (cross-platform compatiblity problems!), but IIRC it was some variant of it.07:24
→ jmesnil joined07:25
← ceej left07:25
thiago_home Mac?07:25
Ilari thiago_home: Yes.07:25
← WALoeIII_ left07:26
Ilari thiago_home: Or to be more specific, HFS+. UFS doesn't do it.07:26
thiago_home in other words, to be cross-platform, you need a full unicode implementation07:26
to do proper NFC/NFD and to be able to tell if two filenames differ only by case07:27
Ilari thiago_home: More fun: Unicode has three cases for some characters: 'all-lowercase', 'first-upper' and 'all-uppercase'. And there probably are distinct codepoints that uppercase in the same way.07:28
thiago_home lowercase, titlecase and uppercase07:28
yes07:28
and remember that all those operations are *string* operations, not character operations07:28
toupper('ß') is invalid07:28
toupper("ß") is "SS"07:29
thiago_home wonders if Windows and Mac FS honour that07:29
← stuffcorpse left07:29
→ stuffcorpse joined07:30
← ezyang left07:30
→ angerman joined07:30
→ ceej joined07:32
← emias left07:32
→ emias joined07:32
bpalmer the eszett has a somewhat complicated history. unicode 5.1 includes "U+1E9E LATIN CAPITAL LETTER SHARP S"07:32
← ceej left07:34
← solydzajs left07:34
← agenteo left07:35
← emg-2 left07:35
→ solydzajs joined07:36
← neunon left07:37
← johnw left07:38
→ shruggar joined07:39
← eddyp left07:46
Ilari Based on Unicode character case conversion routines I once did: There are following three-case codepoints: 454-453-452, 457-456-455, 460-459-458 and 499-498-497. Then there are folowing characters that can result from upper/titlecasing multiple codepoint: 73, 83, 914, 917, 920, 921, 922, 924, 928, 929, 931, 934 and 7776. Then there are characters that result from multiple lowercasings: 105, 107, 229, 952 and 969.07:46
← thiago_home left07:47
Ilari Quite fun if you are looking to do case-insensitive string compare. Plus those multi-char cases...07:48
→ codeshah joined07:49
→ ceyusa joined07:50
← ashleyw left07:51
→ psoo joined07:53
Ilari I think those routines were based on Unicode 5.0... Updating them to Unicode 5.1 isn't easy as it compresses the tables in some really weird format I would need to figure out first...07:54
→ tono joined07:54
← ia__ left07:54
abstrusething → abstrusething|aw07:54
← psoo left07:54
→ ia__ joined07:55
→ psoo joined07:56
the_zeno → zeno^^away07:58
← diegoviola left07:59
→ mtkd joined07:59
abstrusething|aw → abstrusething08:00
← mikeando left08:01
→ ashleyw joined08:01
← jsbmsu left08:02
← sunoano left08:04
← merlin83 left08:08
→ argonel joined08:09
→ dwave joined08:11
→ dwave_ joined08:11
← krawek left08:17
→ Ogedei joined08:18
→ MisterN joined08:18
Ogedei we have a project with two branches that were both imported from CVS, but when I merge them every change is seen as a conflict, because git doesn't seem to know that the same change is also part of the merged branch08:20
is there a merge strategy that is the inverse of 'ours'?08:21
← ph^ left08:21
Ogedei or, even better, a way to reset the master branch to some other branch that can then safely be pushed to other clones08:22
← Bogh left08:23
zeno^^away → the_zeno08:24
→ merlin83 joined08:25
← dwave left08:27
→ solitar joined08:29
← peshalto_ left08:30
← sergiodj left08:33
→ sako joined08:35
← dreiss left08:37
sako whats a good ide with git support? i was thinking of using eclipse does it have a good plugin for git?08:37
← codeshah left08:37
un|ogic sako: it does, whether it's any good or not I don't know... I use the command line because at least I know what's happening to my repo08:38
→ intripoon joined08:39
sako yea thats true.. prob end up using cli08:39
wshimmy sako: IIRC, anjuta just recently got one as well08:39
sako unlogic what do you develop?08:39
← m104_ left08:43
← gaveen left08:43
→ seanius joined08:44
→ sunoano joined08:45
un|ogic sako: what do you mean? language wise?08:45
sako yea08:46
un|ogic perl, python, php, c++, are the more commong ones08:46
sako nice08:46
same here minus perl08:46
i gotta pick that up :P08:46
un|ogic there's the occassional javascript too.. :P08:47
sako what ide do you use?08:47
un|ogic vim08:47
doener vim08:47
psychoschlumpf what command do i need to get the short version of a git release?08:47
s/release/commit/08:47
sako see the thing is though im working with magentocommerce so the code is in different files..08:47
un|ogic if I'm doing web stuff on a Mac I use Coda...08:48
sako eh, i guess i gotta use the plugins for vim ;P08:48
un|ogic sako: vim 7.0 has tabs08:48
sako i just cant figure out yet how to search in other files08:48
← bobesponja left08:48
un|ogic I use grep on the command line :)08:48
doener sako: grep?08:48
sako oh ya08:48
un|ogic is there an echo here?08:48
:)08:49
doener sako: git grep is pretty powerful, and you can make vim use that for its grep support08:49
sako nice08:49
you guys have any other recommendations for vim plugins08:49
un|ogic don't use any myself...08:49
doener sako: http://git.pastebin.com/m7ac1ca04 -- stolen from someone from vim.org ;-)08:50
psychoschlumpf how do i get the short (32bit) name of a git commit?08:50
sako what does that do?08:51
doener psychoschlumpf: it's not necessarily 32bit, but just a unique prefix. I usually just copy/paste the for few characters of the object name08:51
← patmaddox left08:51
doener psychoschlumpf: "git rev-parse --short $commit" would also produce that08:52
psychoschlumpf: or use e.g. --abbrev-commit with "git log" to make it show the short version08:52
psychoschlumpf okay08:52
thanks08:53
doener sako: allows you to type e.g. ":G foobar" to use git grep to search for foobar08:53
→ SRabbelier joined08:53
doener sako: the results will be available in vim's quickfix buffer (:cl, :cn, :cp, ...)08:53
→ fluxusps joined08:54
← fluxusps left08:54
doener sako: for C/C++, I _really_ like using cscope and its vim support08:54
psychoschlumpf doener: is there a way to get the short-version of the last commit?08:55
→ Grahack joined08:56
doener psychoschlumpf: define "last commit", chronologically? On some branch? The checked out commit is always reachable through HEAD, so that would be e.g. "git rev-parse --short HEAD"08:56
psychoschlumpf ah thanks :-)08:57
yeah, i meant HEAD08:57
← Weaselweb left09:00
sako hmmm09:00
well eclipse is taking forever to install09:00
and pdt09:00
do you guys use any debuggers09:00
→ Anantvijay joined09:00
sako for php? i know gdb is nice for c09:00
wshimmy doener: that vim snippet is AWESOME09:02
muchos gracias09:02
dedmajor → dedm09:02
doener sako: xdebug looks quite ok, though I used the debugging support only once, I usually just stick to using my brain ;-)09:02
sako ya09:03
not really necessary09:03
doener sako: http://www.xdebug.org/09:03
dedm → dedmajor09:03
doener sako: the profiling support is quite nice though09:03
← angerman left09:03
→ Radar joined09:05
→ tchan1 joined09:06
← tchan left09:09
→ seb42 joined09:13
← mtkd left09:13
← Arafangion left09:13
→ Arafangion joined09:13
← solydzajs left09:15
→ seb42_ joined09:17
← fujin left09:19
← cilly left09:20
← KiBi left09:20
→ fujin joined09:20
← seb42_ left09:24
→ __iron joined09:24
→ peshalto joined09:25
sdiz → sdiz-afk09:27
→ gitte joined09:29
shruggar what is the best way to determine whether dangling blobs/commits should be removed, and to remove those which should?09:32
← kn1ttl left09:33
→ cilly joined09:36
← gilimanjaro left09:38
gitte shruggar: you should not determine that yourself, but run "git gc" every once in a while.09:39
→ kraymer_ joined09:39
shruggar gitte: I just did, but there are a lot of dangling commits and blobs, with no explanation of where they come from :/09:41
→ jaeckel joined09:42
← sako left09:43
→ JNRowe joined09:43
← kraymer left09:47
doener shruggar: dangling blobs are trivial to get. Just do "edit foo; git add foo; edit foo; git add foo"09:47
gitte shruggar: it is a safety valve that "git gc" does not remove dangling blobs right away.09:47
doener shruggar: each "git add" might have created a blob, the first one will not be referenced by anything09:47
gitte shruggar: and continuing in the line of what doener said, some "dangling" blobs will become "undangling" when you call "git commit".09:48
doener: there are other sources of dangling blobs, too: merge conflicts, stashes, deleted branches...09:48
doener gitte: sure, just wanted to give a trivial exmaple09:48
s/maple/ample/09:49
gitte: regarding "git commit", at least fsck treats blobs referenced by the index as non-dangling. In which case does "git commit" turn blobs to non-dangling?09:49
← Grahack left09:50
→ jfkw joined09:50
→ mtkd joined09:50
→ j416 joined09:52
← jaeckel left09:52
→ jaeckel joined09:53
← mc_plectrum left09:53
Ilari Committing alternate index?09:53
doener ah yeah, that'd work, when fsck and commit operate on different things09:54
← jaeckel left09:55
→ jaeckel joined09:55
→ dwt joined09:56
dwt Hi there! Is there a german support channel available for git?09:56
rio i'd learn english ;)09:57
dwt Yeah well... I had a small hope that I could explain my problems there in more detail09:58
Ilari dwt: Well, why not just try to describe those problems here (in english)?10:00
dwt Ah well. I'll have a go at it10:00
This is the thing. I'm trying to contribute to an open source project (cappuccino)10:01
all the wile trying to synchronize our internal workflow with svn with that10:01
and having an easy time merging if possible10:01
so - there is the cappuccino repo (git) with some patches that are not in yet10:01
I tried just commiting them in git, but got a hell of a bad experience merging chagnes from upstream10:02
so I just switched to using mercurial queues to manage them10:02
now the merging is easy - but the cappuccino guys want git commits on github. :/10:02
Ilari dwt: Can't Hg export the branch as patches?10:03
dwt yes it does - converting them would not be so hard10:03
doener isn't that queues stuff like using rebase -i or stgit or so?10:03
gitte doener: you're correct... I was wrong. Well, actually, I was right, but that stopped a long time ago.10:03
dwt more like stgit actually10:04
← abstrusething left10:04
dwt but trying stgit I killed my repository thrice already10:04
so I'd like to avoid it if possible10:04
so the original problem is this: I want to get clean patches - yet I do have to do incremental development as I'm not perfect10:05
the first workflow that I found that might support this10:05
was to create branches for each patch10:05
but10:05
how do I make sure there is one branch (master?) that integrates them all10:05
so I can provide that to other people in my company10:05
Ilari dwt: Stgit was mostly from time before Git version 1.5.3...10:06
doener gitte: heh, good old "historical knowledge". Even I start to get affected by outdated knowledge, and you certainly know a lot more about the old behaviour of git10:06
Ilari dwt: It can do some extra stuff, but 1.5.3 introduced interactive rebase.10:06
dwt Well. I tried that actually10:06
its actually not as bad as it looks at first hand10:07
but still10:07
doener dwt: your use case sounds a bit like what I've heard about topgit, but I've never used it10:07
dwt :-/ I had a look at that10:07
it looks _really_ complicated10:07
and I fear that it repeats my experience with stgit - that is that I can't mix commands to the two tools10:08
or else I will kill my repository again (to a state where it's easier to just re-check-out than repair)10:08
← wagle left10:08
← jed left10:09
← JNRowe left10:10
← j416 left10:11
dwt So on the one hand I understand that the cappuccino people want github commit urls to merge from10:11
but on the other I totally don't get how I can provide that with very small overhead10:12
(something that is comparable to hg queues)10:12
and not a safety disaster like stgit. :/10:12
I mean, I must be missing something10:12
→ savonarola joined10:13
doener well, I'm pretty happy with rebase -i, but your "downstream" needs to be aware that you rewrite history10:13
← savonarola left10:13
dwt well - since I will be pushing my changes to github regularly for other people to check out and try10:13
thats pretty much out of the question, right?10:14
→ ignas joined10:14
doener dwt: git.git's next and pu branches are also regularly rewritten. Users of those branches now that and deal with it10:15
→ solydzajs joined10:16
→ ntoll joined10:16
dwt hm. so how does this help me keep patches in development seperated10:17
← psankar left10:20
dwt still - regularly rebasing also means that I can't provide stable commit-urls in github for bugreports10:21
← przemoc left10:22
→ bcarlyon|laptop joined10:23
gitte doener: it's a curse.10:25
← hyperair left10:31
→ Octalot joined10:33
← Radar left10:34
← ciskje left10:34
→ Radar joined10:34
dwt so, to summarize my thinking about this: really developing patches in parallel (no dependency usually) is really, really hard with git10:35
so I'm going with just commiting everything to my mainline10:35
which makes it really easy for me10:35
but (probably?) much harder for upstream to integrate again10:35
and also my history isn't pretty10:35
but - who cares10:35
→ parasti joined10:36
dwt right?10:36
doener dwt: one branch per patch (series), use rebase -i to keep those clean, have one or more integration branches that merge the individual topic branches together10:36
dwt: that's how I would do it10:37
dwt: the integration branches are recreated from scratch for each new "release"10:37
← peshalto left10:37
← EgS left10:37
→ EgS joined10:38
dwt but this is one hell of an overhead10:40
and I still don't have real nice commit urls10:40
that stay10:40
well... I could push all the branches and refer upstream to their tip10:41
except that they don't want branches.10:41
Argh10:41
and I have to update them all individually, rebase them, deal with keeping track of 10 or so branches10:42
→ jrmuizel joined10:42
dwt and merge them continually to my integration branch10:42
gitte doener: FWIW I found Peff's script: http://article.gmane.org/gmane.comp.version-control.git/9927810:42
→ Ryback_ joined10:42
→ countd joined10:48
← Radar left10:50
→ gaveen joined10:51
dwt ah well, thanks for you guys trying to help me10:52
good buy10:52
← dwt left10:52
→ ciskje joined10:53
← mtkd left10:54
→ tono_ joined10:54
→ wagle joined10:55
→ psankar joined10:55
→ drewr joined10:56
→ dwave joined10:56
→ hyperair joined10:57
→ remo joined11:02
← remo left11:03
→ tono__ joined11:05
← tono left11:05
→ naeu joined11:06
← cilly left11:07
→ cilly joined11:08
← scientes left11:09
← dwave_ left11:11
← tal67 left11:12
→ javatexan joined11:18
Anantvijay hi.. i'm getting an error when i run "sudo -u user git-daemon --base-path=/home/user/repositories/ --export-all". Error - "fatal: unable to allocate any listen sockets on host (null) port 9418"11:18
has any one experienced the same11:19
pls anyone help me debug the issue :(11:20
← tono_ left11:23
drizzd Anantvijay: ps ax | grep '[g]it-daemon'11:25
Anantvijay: also, lsof -i -n | grep git11:26
→ cytrinox joined11:28
← _Vi left11:30
← Ogedei left11:32
→ kn1ttl joined11:34
→ cryo_ joined11:35
← dodo_the_last1 left11:35
→ dodo_the_last joined11:35
← jrmuizel left11:36
← Anantvijay left11:36
→ Anantvijay joined11:38
Anantvijay http://www.mibbit.com/pb/m7T3FO11:39
sorry my connection got closed automatically11:39
Grum you already have a git-daemon running11:41
which binds to the same port11:41
so the error is completely as it should11:41
FreakGuard I'm picking a bunch of commits form another fork at github.com - cherry-pick is the tool to use?11:41
Ilari Anantvijay: Based on command line, it seems to use base path '/pub/gittrees'.11:41
FreakGuard: Or if the extra commits form a branch, you could also merge.11:42
FreakGuard Ilari: not all commits ;)11:42
bremner FreakGuard: you could use rebase --interactive on the branch and then merge. not sure what is easier11:44
→ craigp123 joined11:44
bremner or there is some way to do it directly with rebase11:44
craigp123 hi all11:44
how can i delete old files from the repo history?11:45
i have a bunch of crap that got committed, no longer required, but making the repo size huge11:45
→ grc joined11:45
MisterN craigp123: git rebase --interactive might help you11:46
craigp123 thanks mate, will give it a try11:47
Anantvijay grum, llari: thanks for the responses11:47
FreakGuard bremner: well, cherry-pick then ;)11:47
grc git novice doing somethig silly here: git clone ssh://10.44.12.2/home/grc/orgfiles11:47
Initialized empty Git repository in /home/grc/orgfiles/.git/11:47
[email@hidden.address] password:11:47
remote: Counting objects: 6, done.11:47
remote: Compressing objects: 50% (3/6)11:47
hangs at compressing objects. any suggestions?11:48
MisterN grc: wait a little.11:48
grc MisterN: sure11:49
MisterN grc: not for me! for git... it might just be slow11:49
→ mithro joined11:50
grc ah!11:50
→ struberg joined11:52
← grc left11:53
← girishr left11:54
→ mbroeker joined11:55
Arafangion Hey, suppose I have a repository, how do I "push" it to a bare, just-initialized git repository on a remote server? (local is the msysgit version, server is a bare repo running on linux)11:56
(I'm expecting a url here, not a dozen lines of irc. :) )11:57
← cyberglitch left11:57
→ lorandi joined11:58
Ilari Arafangion: push?11:58
Anantvijay grum, llari: 'm trying to mirror a reopsitory... how do i export this path "/home/user/repos" on the server11:58
← ashleyw left11:58
Grum Anantvijay: euuh? why do you feel that you need to do that?11:58
Arafangion Ilari: To a bare repository that does not yet have an initial commit?11:58
Grum Arafangion: yes11:59
→ tal67 joined11:59
Grum unless you are using a *REALLY* old git ... you can11:59
Arafangion Aren't there some subtle things to keep in mind with the first commit?11:59
Grum: How old is "really old"?11:59
struberg 0.99 ;)11:59
Grum Arafangion: latest version will do11:59
and no, you dont have to keep anything in mind for the first commit12:00
craigp123 hmm, git rebase --interactive does nothing...12:00
Anantvijay grum: its basically setting up a local repo for android... so i dont want to download 2.1gb everytime i want to clone12:00
craigp123 stupid repo is 300mb, need to get rid of a bunch of documents that were committed to it12:00
think i will just recreate it12:00
→ cannonball joined12:01
struberg Arafangion: git-push ssh://youruser@yourserver.com/yourdirectory/gitrepo.git or something12:01
or if you have it locally, then simply file:///yourdir...12:01
→ jrmuizel joined12:01
Arafangion struberg: Nice. :)12:03
MisterN struberg: the dashed form is deprecated, isn't it? :)12:04
Arafangion MisterN: Definetly. :)12:04
struberg yup, but I'm used to it ;)12:04
and it's neat because you easily can man git-push with the dashed for12:05
Gitbot struberg: the git-push manpage can be found at http://git.or.cz/man/git-push12:05
MisterN struberg: i just use git help push12:05
- which calls man git-push12:05
Gitbot MisterN: the git-push manpage can be found at http://git.or.cz/man/git-push12:05
Arafangion MisterN: But that's one extra letter!12:05
struberg ah this is so svn like ;P12:05
Grum AAAAAAAAARGH!!!12:06
MisterN Arafangion: you can change git's autocompletion to include help in the list of autocompleted commands12:06
Arafangion: then it's the same number of letters12:06
actually one less12:06
← mbroeker left12:06
→ gaveen_ joined12:06
Arafangion MisterN: I think that'd become /way/ too slow on windows.12:06
Grum i push/clone from: git:projectname12:06
bpalmer alias gh="git help"12:07
Grum just set it up properly and there is no problem :p12:07
MisterN Arafangion: on windows i don't typically have man either12:07
struberg I think we start going OT? ;)12:07
← jrmuizel left12:07
MisterN bpalmer: that's even better:)12:07
Arafangion MisterN: msysgit has html help pages, which are pretty good.12:07
MisterN Arafangion: you mean those you can open with git help -w push?12:07
← ankit9 left12:08
Grum git help on windows opens the html page in your default browser12:08
Arafangion Indeed.12:09
← orafu left12:09
MisterN ah, so they have help.format=web by default12:09
→ orafu joined12:10
Arafangion THe latest git seems to be 1.6.3. (on windows)12:10
→ mbroeker joined12:12
← cilly left12:13
→ stuffcor1se joined12:13
MisterN Arafangion: don't say anyone but i still use 1.5.4.3 :D12:13
→ cilly joined12:13
Arafangion MisterN: That's the one that's in Debian Stable, isn't it? :)12:13
MisterN Arafangion: ubuntu 8.04, actually12:13
Arafangion (The best distro ever)12:13
← stuffcorpse left12:13
Arafangion MisterN: Haven't upgraded yet, I see.12:14
MisterN: When you do, I strongly advise you to back up everything first, as Ubuntu's not as good at managing an upgrade as Debian is.12:14
MisterN Arafangion: i'm a bit wary of upgrading.12:14
yeah for that reason...12:14
maybe it's actually better to do a fresh install12:14
Arafangion MisterN: I'm inclined to agree, for Ubuntu. It's stupid, but that's how it is.12:15
→ ashleyw joined12:15
→ UUncia joined12:15
Arafangion We're OT, anyway.12:15
→ davidroderick joined12:15
MisterN well, slightly. it's about why i still use an old git12:15
Arafangion So... For git 1.6.3... Can one just push to a freshly initialised bare git repository, or does one have to keep certain things in mind?12:16
MisterN: git's actually insanely easy to compile yourself.12:16
MisterN Arafangion: well... i haven't felt the need to really upgrade git. 1.5.4.3 works like a charm12:17
it lacks git log --graph, sure... and some other things12:17
→ jed joined12:17
Arafangion MisterN: On windows one tends to stay on the edge, simply because there's alot of noticible improvements.12:17
← orafu left12:17
← gaveen left12:17
Arafangion MisterN: For instance, the latest version supports repos of size greater than 2GB. (total).12:18
MisterN Arafangion: on ubuntu, my biggest problem is that gnome-power-manager crashes about once a day :D12:18
Arafangion MisterN: That's Ubuntu for ya. ;)12:18
MisterN on windows?12:18
Grum Arafangion: euuh why wouldnt it befoer?12:18
Ilari Arafangion: Pushing has been the official way to populate bare repositories for a long time...12:18
MisterN on linux i suppose big repose were supported before12:18
→ Laibsch joined12:18
← UUncia left12:18
Ilari Arafangion: The 'recent feature' presumably refers to cloning empty repositories.12:19
→ mobhawk joined12:19
→ quoin joined12:22
← mbcharbonneau left12:22
gaveen_ → gaveen12:23
→ orafu joined12:23
→ cmarcelo joined12:24
Arafangion Grum: If I recall correctly, it used to barf on account of having a detached or broken HEAD, because it didn't actually point to anything.12:24
Or somethign.12:24
→ angerman joined12:25
← angerman left12:26
→ angerman_ joined12:26
Ilari Arafangion: Clone did barf on those (until somewhat recently).12:26
→ mtkd joined12:27
← craigp123 left12:28
davidroderick I have problems with git pull /foo and it says something about altering the config file at destination about branching. I am not using different branches. is there a quick fix?12:29
Ilari davidroderick: 'git pull /foo <branch-to-pull>'.12:30
davidroderick Ilari: fatal cannot be used without a working tree12:31
Ilari davidroderick: Pull invokes merge on local end. And merge requires working tree.12:32
davidroderick: If you want to copy branches, perhaps look at git fetch?12:32
→ mfilipe joined12:33
→ bremner_ joined12:36
davidroderick so how do I use git fetch? git fetch /foo achieves nothing12:36
→ elmex_ joined12:36
← elmex left12:36
Ilari davidroderick: 'git fetch /foo remote-branch-name:local-branch-name'.12:37
davidroderick I was using git pull until this stopped working12:37
elmex_ → elmex12:37
davidroderick how do I discover the branch names?12:37
→ beket joined12:38
← Laibsch left12:38
Ilari davidroderick: If there's only one branch, both sides of : are likely 'master'.12:38
davidroderick when I do git status is says index file corrupt12:38
beket Hi people! I'd like to generate a tar archive of my src tree to send it to a remote machine which doesn't have git installed. What git cmd do I need ?12:38
Ilari beket: man git-archive12:39
Gitbot beket: the git-archive manpage can be found at http://git.or.cz/man/git-archive12:39
beket Thank you Ilari :)12:39
→ mikl joined12:39
davidroderick worked thanks happy joy12:39
← ignas left12:39
Ilari davidroderick: Status should complain about no working tree if run in bare repo.12:39
davidroderick I am learning. where is the best place to learn other than man pages?12:40
→ bobmcw joined12:41
← psankar left12:41
Ilari davidroderick: There are various guides around the net..12:42
← [intra]lanman left12:42
davidroderick thanks12:44
← kn1ttl left12:44
← davidroderick left12:45
→ d0k joined12:45
← cryo_ left12:45
← bobmcw left12:46
← cmarcelo left12:46
kirillm_ → kirillm12:47
→ UUncia joined12:47
→ SteveC_ joined12:48
← amrox left12:49
→ khmarbaise joined12:51
→ bx2 joined12:59
→ gh34 joined12:59
→ psankar joined13:00
← vmiklos left13:00
→ vmiklos joined13:01
← UUncia left13:02
← vmiklos left13:02
→ vmiklos joined13:05
→ Bass10 joined13:06
→ edw joined13:08
→ schweizer_ joined13:08
schweizer_ salaam13:08
→ leanucci joined13:09
→ bobmcw joined13:11
→ scarabx joined13:14
tswicegood → tswicegood-afk13:17
→ cu3edweb joined13:18
cu3edweb I am getting this error? --> ! [rejected] master -> master (non-fast forward) error: failed to push some refs to?13:19
→ aser joined13:19
← bremner_ left13:19
aser can I perform a fully manual merge? i.e. to disable automerging but to run difftool for all files ?13:20
cu3edweb I git pushed from one server and I didn't mean to now when I tried to git push from the right one I get that error?13:20
dwmw2_gone → dwmw213:22
→ ChrisCauser joined13:23
→ qrush joined13:23
→ outworlder joined13:26
outworlder so, let's say I git push hundreds of megabytes, for some reason. meanwhile, I have other people pushing changes so that the big push may not be a fast forward. will it always be rejected in that case?13:28
Ilari cu3edweb: The 'master' of remote has commits that your 'master' doesn't.13:29
→ mw___ joined13:29
→ [intra]lanman joined13:29
← stuffcor1se left13:30
→ kn1ttl joined13:30
Ilari aser: Why you would want that?13:30
cu3edweb: Maybe fetch that branch and then merge it into local master.13:31
→ JhonnyCash_ joined13:31
cu3edweb Ilari: I got it. I was trying to go back in history or something along those lines. I was able to force my push and correct it13:31
Thanks13:32
← JhonnyCash_ left13:32
Grum cu3edweb: you know that doing a push -f messes it up for every person who has already got the commits you just tossed away?13:32
cu3edweb Grum: yes it is my own private repo. No one else is using it.13:33
→ voker57__ joined13:33
Grum then no-one will care =D13:33
cu3edweb Thnaks13:34
→ stuffcorpse joined13:34
→ bremner_ joined13:35
← flippo left13:37
→ schacon joined13:37
← scarabx left13:37
→ przemoc joined13:38
← psankar left13:40
← parasti left13:41
← kraymer_ left13:41
← nickm left13:41
← rio left13:41
← beantmt left13:41
← Kasperle left13:41
← unreal left13:41
← microm left13:41
← uau left13:41
← xblaster left13:41
← context left13:41
← xkr47 left13:41
→ amrox joined13:44
→ parasti joined13:44
→ kraymer_ joined13:44
→ unreal joined13:44
→ microm joined13:44
→ rio joined13:44
→ xkr47 joined13:44
→ uau joined13:44
→ context joined13:44
→ beantmt joined13:44
→ nickm joined13:44
→ xblaster joined13:44
→ Kasperle joined13:44
← cu3edweb left13:45
← twas__ left13:46
← bwalk|uni left13:46
← voker57_ left13:46
→ wmoxam joined13:47
→ bwalk joined13:48
bwalk → bwalk|uni13:49
aser Ilari: I think I got the answer, merge driver13:49
→ j416 joined13:50
→ jrmuizel joined13:50
→ altrux joined13:50
→ DavidKlein joined13:52
← bremner_ left13:54
← Leemp left13:57
← Lgb left13:59
→ Lgb joined14:00
→ offby1` joined14:02
← aser left14:02
→ cmarcelo joined14:03
← solydzajs left14:07
→ ph^ joined14:07
→ spearce joined14:07
→ kn1tt1 joined14:08
→ hpoydar joined14:09
← bx2 left14:09
← agile left14:09
← impulse9 left14:10
→ hugoxrosa joined14:11
→ helo joined14:11
offby1` → offby114:11
helo man... that "Git for Computer Scientists" article is the most helpful thing i've read re: git14:11
offby1 indeed14:12
it brought enlightenment to me!14:12
offby1 nods gravely14:12
→ brett_h joined14:12
→ kukks joined14:13
← hyperair left14:13
→ hyperair joined14:13
← johan-s left14:14
← schacon left14:14
← rjc left14:15
→ rjc joined14:15
→ ankit9 joined14:15
→ alezandro joined14:15
tswicegood-afk → tswicegood14:15
will__ → throughnothing14:19
← rjc left14:19
← kn1ttl left14:19
→ sakyamuni joined14:21
← beket left14:21
← j416 left14:22
csgeek how would I generate a patch of the between HEAD and the previous commit14:24
drizzd git format-patch HEAD~14:25
offby1 git show14:25
csgeek btw.. I saw HEAD before.. what does the ~ mean?14:26
drizzd previous commit14:26
or parent in git terms14:26
← nick_h left14:26
→ rylar joined14:27
→ mozinator joined14:27
mozinator hi all, got just 1 thing to say, THANKS GIT ! you saved me sooo much time14:27
Anantvijay how to share a git repo locally.. so tat others can clone fom it?14:28
psychoschlumpf is there a plan to add gpg signed commits/signed-by commits to git?14:28
→ bcardarella joined14:29
offby1 mozinator: aw, shucks14:30
→ hydrogen joined14:31
hydrogen fatal: object 5ebc2873b7e90cec7d71b6659d9d271ec6a9fccd is corrupted14:32
how do I uncorrupt it :)14:32
offby1 a retreat to the lake, with other cleancut youths?14:33
hydrogen tried that14:33
michael jackson wanted to come along though, so I think it only made it worse :(14:33
offby1 no kidding14:34
seriously, I am sure there's an example of this sort of thing somewhere in the docs, but I can't find it14:35
→ patrikf joined14:35
Anantvijay how to share a git repo locally.. so tat others can clone from it... has anyone tried this out???14:35
offby1 Anantvijay: by "locally", do you mean that all the others use the same multi-user computer?14:36
hydrogen looks at thiago, who knows everything, and wonders if he knows this14:36
← PupenoG left14:36
offby1 usually doener or Ilari pipes up around now14:37
must be asleep, or clueless14:37
hydrogen: my best advice is: see http://git.or.cz/gitwiki/GitFaq, and search for "How to fix a broken repository", about 60% of the way down14:39
→ rjc joined14:40
mbroeker Anantvijay, take a look at the git-init manpage, exspecially --shared option14:40
→ nick_h joined14:41
mbroeker the rest is easy: git clone /var/lib/git/test.git and everybody can push and pull from the local bare repo...14:42
→ agile joined14:42
← solitar left14:42
doener offby1: you gave the answer, nothing left for me to do ;-)14:43
bentob0x how do you 'clone' a repository with a local copy of _all_ the repository's remote branches?14:44
in one go14:44
git pull-it-all style14:44
Weiss what you really mean is, "how do you automatically create a local branch tracking every remote branch"?14:45
doener bentob0x: _pull_-it-all? Merge it all to the checked out branch? ;-)14:45
← ashleyw left14:45
Weiss nastiness: something like: cp .git/refs/remotes/origin/* .git/refs/heads/ ?14:45
bentob0x yes Weiss but having the luxury to do a full pull (meaning, with all the branches) in one go14:45
doener bentob0x: maybe give an example of the refs the existing repo has, and which branches you expect to exist in the "clone"14:46
offby1 doener: watch your back; I'll be stealing your wife, car, dog, and house next14:46
→ peshalto joined14:46
doener offby1: I guess you're done then, I have none of these14:47
→ Sho_ joined14:47
bentob0x repo: master, dev, live, branch-test-controller and locally: master with refs on origin/dev, origin/live and origin/branch-test-controller14:47
offby1 doener: nuts, someone got there before me14:47
bentob0x but the repo will be deleted all together14:47
offby1 and no thanks, I don't want your cat14:47
→ Yuuhi joined14:47
bentob0x so I want to make sure I have it all locally14:47
→ bhearsum joined14:47
← nick_h left14:47
bhearsum i'm coming from lots of Mercurial experience and looking to use git - can anyone recommend good documentation for me?14:48
→ nick_h joined14:48
doener bentob0x: master, dev, live and branch-test-controller look like branch heads, you were talking about "the repository's remote branches" (which I assumed to mean remote tracking branches)14:48
bentob0x: and "master with refs on ..." makes no sense to me14:48
bentob0x ok, I'll explain differently14:49
→ bx2 joined14:49
bentob0x I want to delete a repository on github to replace with a new one14:49
I just want to make sure I have it all locally before I delete the repository14:49
doener bentob0x: git clone --mirror?14:49
bentob0x ahaaaaaaaaaa14:50
let me man that14:50
→ mbcharbonneau joined14:50
offby1 bhearsum: the problem with git is that there's _too much_ documentation. I'm semi-serious. Anyway, start at git-scm.com14:52
bhearsum hah14:52
offby1 bhearsum: also a bunch of us swear that "git for computer scientists" is the best thing we've ever read about it14:52
bhearsum i'll check it out - thanks!14:52
→ johan-s joined14:55
→ solydzajs joined14:56
→ mbcharbonneau_ joined14:59
Anantvijay offby1: one desktop machine(say mach1) will clone from the main android repo. All other members will clone from this desktop machine(mach1).15:00
offby1: so any special setup needed for this?15:01
← mbcharbonneau_ left15:02
→ bdiego joined15:03
← sakyamuni left15:03
jdahm bhearsum: it's not _just_ for computer scientists, I find it easy to use, and it's good for most coding projects15:03
← peshalto left15:03
→ _dave_h_d joined15:04
offby1 Anantvijay: certainly nothing _special_; your situation sounds typical15:05
← rjc left15:05
_dave_h_d right i did a git pull and there are a lot of conflicts, now i know the stuff i am pulling is the right stuff and i want the pull to use those files, how do i do it?15:05
i have already done a pull and the conflicts are there15:05
eb_ you have local changes?15:06
_dave_h_d had some local changes which i comitted and did not push15:06
but i don't need them15:06
offby1 sounds like "git reset --hard origin"15:06
Anantvijay offby1: any pointers on how to setup??15:07
eb_ yea, reset, then properly track the remote branch, and pull again15:07
← bx2 left15:07
countd how come git appears to insist on add a symlinked directory that is in gitignore - is there some way to stop this?15:08
→ ceyusa_ joined15:09
← mbcharbonneau left15:09
← solydzajs left15:09
← arohner left15:09
→ rjc joined15:10
← hyperair left15:11
bentob0x doener: that did it thx15:11
so it is simple a perfect copy of the remote repository?15:11
*simply15:11
patrikf countd: what does the gitignore entry look like?15:11
→ hyperair joined15:12
countd patrikf: various attempts include /images images images/ images/*15:13
patrikf: can't really seem to nail the correct syntax!15:13
patrikf countd: it should just be "images"15:14
(read man gitignore)15:14
Gitbot patrikf: the gitignore manpage can be found at http://git.or.cz/man/gitignore15:14
countd patrikf: thanks i will retry15:14
← ceyusa left15:15
countd patrikf: i was just worried that a symlink was the issue - didn't get much on that from googling about15:15
patrikf well, ignoring a symlink works fine for me here15:16
offby1 Anantvijay: all the docs are at git-scm.com15:16
→ ceyusa joined15:16
straszheim doener: thanks for that example15:17
→ jwshen joined15:17
straszheim i'm curious about15:17
git rev-list --objects --all | egrep '.{42}' | cut -d' ' -f2 | sort | uniq | egrep "(^$1|/$1)$"15:17
why the egrep '.{42}' ?15:18
→ ceej joined15:18
countd patrikf: yes - it appears to work fine - thank you :)15:19
bleything straszheim: it filters out lines that are only the hash.15:20
doener straszheim: in case that you have weird files with 40 character hash name15:21
s15:21
→ wycats joined15:21
bleything straszheim: run just the rev-list command and you'll see some entries that don't have filenames, the egrep will filter those out15:21
doener straszheim: though the script probably already fails for filenames with spaces or characters that have a special meaning for egrap15:22
← jony left15:23
→ rjc_ joined15:23
← rjc left15:23
doener straszheim: and seeing the --all there, this might be a bit faster, limiting that rev-list to the interesting commits: http://git.pastebin.com/m6ac5747e15:24
straszheim oh duh15:24
→ peshalto joined15:25
← ph^ left15:25
doener saves 10% (0.6s) for my test case here15:25
straszheim boots up his brain15:25
doener well, that --all was my fault ;-)15:25
← dwave left15:25
straszheim i was spacing out... i thought the .{42} was looking for some magic pattern in the hashes :/15:26
helo how often does code merge cleanly, but function incorrectly?15:27
doener helo: you mean no textual conflicts, but the result is crap anyway?15:27
helo yeah15:27
unit testing ftw i guess...15:28
doener depends a lot on the code base, it's dependencies, how many people work on the same stuff, etc.15:28
→ aziz joined15:28
doener s/it's/its/15:28
psychoschlumpf helo: merge, run testsuite, fix15:28
doener If you want some useless data point: I've encountered three such broken merges in the last year15:29
and about a dozen or more broken merges broken due to manual merging with svn...15:29
← ceyusa_ left15:31
→ cemerick joined15:35
→ releod joined15:36
→ aspotashev_ joined15:37
→ dwave joined15:38
→ hobodave joined15:42
← psychoschlumpf left15:44
← struberg left15:45
← khmarbaise left15:47
→ nbrunson joined15:48
jdahm can anyone help me with gitweb? I have the css file in the right place but it's not showing any text formatting on the page, and furthermore, it's calling gitweb.css a directory and trying to find files within it15:48
drizzd sounds like a webserver problem, not a gitweb problem15:50
→ corybantic joined15:51
→ markelikalderon joined15:51
FreakGuard how to add a remote branch?15:52
← aspotashev left15:52
drizzd git fetch <remote> refs/heads/<branch>:refs/remotes/<remote>/<branch>; or just git remote add <name> <remote>; git remote update;15:53
← psoo left15:53
→ dodo_the_last1 joined15:54
← dodo_the_last left15:54
← giallu left15:56
→ Necker joined15:57
→ tw123 joined15:58
← markelikalderon left15:59
← gitte left15:59
jefftriplett → jefftriplett-afk16:00
→ kuadrosx joined16:00
← mikl left16:01
straszheim from the log format docs:16:03
tw123 I'm trying to print a log message from a commit using " git log .. --pretty=tformat:%s". My problem is that the %s seems to be stripping the newlines that were in my message. I can't see any other format operator that lets me print the newlines that were in the message! (I know I can use medium or whatever, but I want it formattted). Is there a way to do this? A related question: what %b (as opposed to %s) and what's the difference between "16:03
straszheim o %cd: committer date16:03
← csgeek left16:03
straszheim o %ad: author date16:03
→ RayMorgan joined16:04
← GyrosGeier left16:04
straszheim these timestamps always appear to be the same... what's the difference?16:04
drizzd tw123: %n16:04
tw123 drizzd: in the middle of the contents of a log message??16:05
%s%n puts a newline after the message16:05
(and puts the original message that hd newlines on one line! :( )16:06
doener straszheim: when you do the editing and the commit yourself, they'll be the same, initially.16:06
drizzd tw123: %s is just the first line, it has only one newline16:06
doener straszheim: but e.g. commit --amend may change the commit time, but not the author time16:06
→ drd20 joined16:06
straszheim ah so. okay thanks.16:07
doener straszheim: and when you get a patch via email and apply that with "git am", the author time is taken from the email, but the commit time is $now16:07
tw123 drizzd: that's not what I'm seeing!16:07
drizzd tw123: %b is the body of the message (everything except the first line)16:07
straszheim okay, that makes sense16:07
← lohrun left16:07
tw123 Hmm. I'm seeing %s as the whole thing..maybe for some reason it's not finding the newlines in my message (but medium prints it with newlines..confusing!)16:08
← Modius left16:08
drizzd tw123: maybe you don't have newlines in your message16:08
← RayMorgan left16:08
→ ph^ joined16:08
→ RayMorgan joined16:08
→ romao joined16:08
tw123 Why would medim thenprint it with newlines?16:08
← naeu left16:08
doener IIRC %s is the whole first paragraph (i.e. up to the first empty line)16:09
← dodo_the_last1 left16:09
doener maybe %s collapses that to a single line, but the "medium" format just takes the whole message16:09
drizzd doener: ah, didn't know that16:10
doener yep, that's it16:10
→ dodo_the_last joined16:10
tw123 aha - paragraph!..I see now. that makes sense. SO I'm still left with the problem of formatting my "first paragraph" as it was entered (ie with newlines)..16:10
drizzd I suppose the best practice is to keep the subject line short16:10
← BlindOracle left16:10
→ Stravy joined16:10
tw123 Sounds liek we need an extra format operator to just print the whole message as it was?16:11
→ HG` joined16:11
tw123 is tehre any way to do this using format??16:12
drizzd darn, if I had only known about %f _before_ I implemented its equivalent as sed script...16:12
ElAngelo http://rafb.net/p/sXwphc67.html16:13
← ciskje left16:13
ElAngelo i don't know how to fix this16:13
those are just new files16:13
drizzd tw123: from what I can see, yes16:13
ElAngelo there is nothing i can merge in them16:13
tw123 drizzd: help me out a little more! please..how?16:14
joshthecoder_afk → joshthecoder16:14
→ solydzajs joined16:15
helo laughs at pedogit16:16
→ schmrz joined16:16
ElAngelo no one?16:17
→ markelikalderon joined16:17
drizzd tw123: oh, I was responding to your second last question. I don't think it's possible.16:18
tw123 drizzd: thanks. Is there some official way to make this a "feature request"?16:20
doener the best one is "send a patch"16:21
tw123 naturally! :)16:21
drizzd should be really trivial too16:21
→ jballanc joined16:21
tw123 ok - thanks all.16:22
← tw123 left16:22
doener hm, I was going to tell him where to look...16:24
← brett_h left16:24
← Necker left16:24
← jeremiah left16:25
→ jeremiah joined16:25
→ NoirSoldats_ joined16:25
← ChrisCauser left16:25
→ UUncia joined16:25
← NoirSoldats left16:26
→ bradly joined16:27
← johan-s left16:29
→ johan-s_ joined16:29
← mfilipe left16:30
← dwave left16:30
← markelikalderon left16:31
← shruggar left16:34
→ rick_h_ joined16:34
← tal67 left16:34
jefftriplett-afk → jefftriplett16:35
→ LiamH joined16:35
→ WALoeIII joined16:35
→ psoo joined16:36
joshthecoder → joshthecoder_afk16:36
joshthecoder_afk → joshthecoder16:36
← javatexan left16:37
→ toxx_ joined16:39
← hyperair left16:42
← galderz left16:42
→ hyperair joined16:42
→ galderz joined16:42
← amrox left16:44
← _dave_h_d left16:44
← drizzd left16:45
→ bx2 joined16:45
← nickm left16:45
← microm left16:45
← beantmt left16:45
← unreal left16:45
← rio left16:45
← Kasperle left16:45
← uau left16:45
← kraymer_ left16:45
← parasti left16:45
← xblaster left16:45
← context left16:45
← xkr47 left16:45
Zaba_ → Zaba16:46
→ jsbmsu joined16:46
→ beantmt joined16:47
→ xblaster joined16:47
→ parasti joined16:47
→ rio joined16:47
→ Kasperle joined16:47
→ microm joined16:47
← sgrimm left16:47
→ Modius joined16:48
→ Weaselweb joined16:48
← rick_h_ left16:49
→ jceb joined16:49
→ juanjoc joined16:50
→ uau joined16:50
→ unreal joined16:51
→ nickm joined16:52
→ Necker joined16:53
→ amrox joined16:54
← drd20 left16:55
→ gwoo_ joined16:55
→ towski joined16:56
→ markelikalderon joined16:56
← ph^ left16:57
← wycats left16:59
→ vipul joined17:02
vipul hi folks, I have some changes in my local git repository that I want to overwrite from HEAD17:02
only specific files though ... is git reset --hard the right command to use ?17:02
→ aspotashev|eeepc joined17:02
doener vipul: git checkout HEAD -- <paths>17:03
vipul ah, thanks doener17:03
what is git-reset used for then ?17:03
and what's the difference between that and git-revert ?17:03
← marc_m left17:04
← jmesnil left17:05
→ tatsuyao joined17:05
→ mfilipe joined17:06
← markelikalderon left17:07
voker57__ git-revert creates commit to revert another one17:08
git-reset moves HEAD17:08
← Necker left17:08
→ sekimura joined17:09
← gwoo left17:10
← mtkd left17:13
→ la_mer joined17:14
← hydrogen left17:14
bentob0x I did a commit, I did a git reset HEAD^ and I realised that my commit wasn't that bad, git checkout commitID (found with git reflog) says that one of the files in my working directory is edited so cannot switch branches17:14
→ davidc__ joined17:14
→ Dashkal joined17:14
davidc__ how do I compare a local branch's file and the remote orgin's copy of the file?17:14
← cemerick left17:15
la_mer → cemerick17:15
doener bentob0x: you probably don't want to use checkout there anyway. That would detach HEAD.17:16
bentob0x: instead, you probably want to use reset again, to reset your currently checked out branch head back to that commit17:16
→ lois[pac] joined17:17
doener davidc__: git diff $local $remote -- $file17:17
→ gitte joined17:17
← outworlder left17:17
davidc__ $local and $remote are the branches?17:18
doener yep17:18
davidc__ cheers.17:18
bentob0x thx doener but how can I get my patch in reflog now?17:19
→ wycats joined17:19
bentob0x git cherry-pick?17:20
doener bentob0x: hm? if you reset back to that commit (I assumed that you didn't do any other commits in the meantime), the branch head will reference that commit again17:20
davidc__ hmm?17:21
→ sergiodj joined17:23
bentob0x no it didn't work, the head still points to HEAD^17:23
tswicegood → tswicegood-afk17:23
→ neoeinstein_ joined17:24
doener bentob0x: what exactly did you do?17:25
→ sgrimm joined17:25
doener bentob0x: maybe pastebin the interesting part from your shell session?17:25
bentob0x git reflog gives me the following patch I want to get back: 67df426 HEAD@{2}:17:26
→ boto joined17:26
bentob0x do I just do a git cherry-pick 67df426 to get that patch back into HEAD?17:27
or into master17:27
should I say17:27
→ paulboone joined17:27
bentob0x basically how do you 'undo' a git reset HEAD^17:27
parasti bentob0x: you're asking two different things17:28
doener bentob0x: right after the reset, "git reset HEAD@{1}"17:28
bentob0x ahaa17:29
doener bentob0x: well, that's what I said before. Reset back to the old commit17:29
← leanucci left17:29
bentob0x yea but that reflog with all those HEAD@... got me confused17:29
doener and that's not a "full" undo. The index state is lost17:29
bentob0x that worked, thx doener17:29
where git reflog still keeps the work done since that git reset HEAD^17:30
I think I got it17:30
doener No, the reflog just keeps the state of the ref17:30
bentob0x gosh17:30
doener but "git reset $committish" simply doesn't touch the working tree, and the commit is obviously undoable17:31
so just the index state is lost17:31
with --hard, the working tree state would also be lost17:31
(in both cases, assuming that you had uncommitted changes)17:31
→ ph^ joined17:32
← aspotashev|eeepc left17:32
← toxx_ left17:35
→ ankit9_ joined17:35
→ k0001 joined17:36
→ alienbrain joined17:38
alienbrain Where does git cache auth data? It's using the wrong cached password17:39
→ olly joined17:42
→ ctremblay joined17:42
jaeckel is there a possibility to configure default push behavior? e.g. automatically pushing tags17:46
← HG` left17:47
→ texel joined17:47
← kadoban left17:47
→ Hendu joined17:47
Grum alienbrain: git doesnt involve itself with auth data at all17:48
alienbrain Grum: I see, I've actually remembered that I have threw a ~/.netrc which had this data :D17:49
Grum humz what uses netrc? O.o17:49
alienbrain curl17:49
Grum oh i think in the later gits they updated it so it would just prompt17:50
← ankit9 left17:51
← davidc__ left17:52
→ carrus85 joined17:52
← doener left17:54
→ doener joined17:54
→ maron3 joined17:55
← bentob0x left17:55
→ _sps joined17:55
← ankit9_ left17:56
← peshalto left17:58
→ gberdyshev joined17:58
← Hendu left18:00
→ kadoban joined18:01
→ bobesponja joined18:01
← ctremblay left18:02
→ javatexan joined18:02
dereine[OFF] → dereine18:03
← paulboone left18:04
→ paulboone joined18:04
→ blingo joined18:05
blingo hello18:05
Anyone knows if git has some sort of progress bar or way to say what file it is currently working on ?18:06
→ wsc joined18:08
→ drizzd joined18:10
→ niki joined18:10
drizzd spearce: I'm reading your git-aware http transport specification and I don't understand when the server should stop sending "status continue"s. It's supposed to stop if it has a "closed set of objects". What do you mean by that?18:12
← psoo left18:12
charon blingo: some potentially costly commands fire up a progress bar if things take longer than a few seconds, e.g., 'git checkout'. for the most part however, they don't18:12
blingo thanks <charon>18:12
this is importent to me, as I want to create a graphical front-end for git18:13
spearce drizzd: its found a complete cut across the DAG, such that every object can be partitioned onto one side, or the other. or ,put another way, from each want, there is a path to an object in the common set18:13
← alienbrain left18:13
blingo is there some way to make a progress bar/ current file appear ?18:13
← cilly left18:14
drizzd spearce: ok, so is that different from the standard git transport? If I understand correctly the client sends "done" if it runs out of PENDING commits18:14
→ ankit9_ joined18:14
meyering hello, I've aggregated glibc.git's commits into semblance of cset normality on "master", adjusted all tags to point to rewritten refs, and grafted the 16 branches onto the new tree. Now I'd like to make the grafts permanent in time proportional to the #commits on branches (there are almost 20k on master). Tried git filter-branch -f "$b" as well as master..$b, but in each case, it also rewrites all of master. Any way around that?18:15
spearce drizzd: no, its the same rule. the server can stop before the client says done. but the client says done, then the server accepts what it has in the common set, even if that set is potentially empty.18:15
← ceyusa left18:16
← jsbmsu left18:16
meyering hmm.. I guess it's inevitable, since graft affects graft-point, too18:17
maybe I can get away with a single filter-branch run, to do all 16 branches18:17
drizzd spearce: I see. And we already know that the client is done if it sends less than 256 haves.18:17
→ davidc_ joined18:18
← jaeckel left18:18
spearce drizzd: no, the server doesn't count that, the client just kicks out a "done" if it wants to give up.18:19
drizzd spearce: ok, thanks.18:20
have we ever considered sending haves for the tips of the corresponding tracking branches first?18:21
← countd left18:21
drizzd those should already give a closed set of objects in the most common cases18:22
→ countd joined18:22
→ dkerschner joined18:22
← dkerschner left18:23
→ dkerschner joined18:23
← dkerschner left18:23
drizzd oh, nvm. I only know the client doesn't have a commit if it's newer than the oldest have it sent. So I still have to start from the newest commit first.18:24
charon blingo: not to dissuade you, but note that there are GUI projects for almost any platform/toolkit combination, so your time might be better spent helping those efforts18:24
(i think the wiki has a list)18:24
→ leanucci joined18:24
→ dkerschner joined18:25
← olly left18:25
← sfwc left18:28
→ codeshah joined18:29
blingo Much thanks <charon> I want to create something in Java, There are some gui's that look a lot like the text mode, do you remember any examples of a mature product? I will look it up myself, regardless18:29
charon blingo: there's a fairly good java implementation called jgit and a corresponding eclipse plugin called egit18:31
dfr|work → poop18:33
← nbrunson left18:33
→ Alinermedia joined18:33
blingo Much thanks <charon>, will look it up :)18:35
→ kumbayo joined18:35
← bcarlyon|laptop left18:35
→ poh joined18:41
→ context joined18:41
poh if I started coding but forgot to branch, is there a way to create a new branch with some of the changes I have made?18:41
gwoo_ → gwoo18:43
→ mikl joined18:43
poop → dfr|work18:43
← galderz left18:44
parasti poh: just git checkout -b newbranch and fix up your previous branch if necessary18:44
→ kraymer_ joined18:45
→ yann joined18:45
poh parasti: thanks18:45
← kraymer_ left18:45
← hyperair left18:47
← gberdyshev left18:47
→ xkr47 joined18:49
← poh left18:49
cxreg is there a simple command to say "what branch does this branch have the most recent merge-base with" ?18:55
← DavidKlein left18:56
→ DavidKlein joined18:56
← xkr47 left18:57
← javatexan left18:58
Ilari cxreg: AFAIK, no.18:59
→ mjf joined19:02
→ BlindOracle joined19:03
→ crystal joined19:05
← mikl left19:07
→ David^ joined19:07
← jeremiah left19:07
→ jsbmsu joined19:07
→ jeremiah joined19:08
→ avaja joined19:08
← DavidKlein left19:08
→ patmaddox joined19:12
→ struberg joined19:16
← eletuchy left19:18
← Alinermedia left19:18
→ eddyp joined19:19
→ dreiss joined19:19
→ DavidKlein joined19:19
tswicegood-afk → tswicegood19:20
← David^ left19:20
→ matthewmcculloug joined19:25
← matthewmcculloug left19:26
← jjuran left19:27
SteveC_ http://pastie.org/47834919:28
what's going on there? it wont let me delete that folder19:28
→ eletuchy joined19:29
→ thiago_home joined19:30
mbroeker SteveC_, well, how many chars are before the .svn/ folder?19:30
SteveC_ um.. none19:31
← eletuchy left19:31
SteveC_ even if i do.. git rm -r .svn/ it doesn't work19:31
← mfilipe left19:32
→ lucsky joined19:32
→ hydrogen joined19:33
blingo Anyone installed egit?19:33
→ loincloth joined19:33
→ lightcap joined19:34
→ plexq joined19:35
plexq Hello - I'm getting the error: fatal: Invalid symmetric difference expression 1377f5d8fa7bbc8286fe691888627ee6f862cf3f...9174b990db57e78ac71a7551ee9799fd409e860719:35
after a bad push19:35
or possibly a pull19:35
→ khmarbaise joined19:37
→ lightcap_ joined19:37
← lightcap left19:37
→ jksz_ joined19:37
→ Alex_Norton joined19:39
henr_k can I set the .gitconfig to automatically add the diff in commit (as in git commit -v) ?19:39
← ceej left19:39
← lightcap_ left19:39
→ lightcap joined19:40
Alex_Norton www.YouFastPorn.com - Free,Fast,Safe & NO ADV..NO-PROFIT !!!19:40
← Alex_Norton left19:40
← lightcap left19:42
→ lightcap joined19:42
← lightcap left19:42
drizzd henr_k: I don't think so. You can make an alias.19:45
→ mfilipe joined19:45
henr_k drizzd: ok19:45
tchan1 → tchan19:46
← gitte left19:47
← d0k left19:48
hugoxrosa SteveC_, Is the folder empty?19:50
← avaja left19:51
henr_k drizzd: I guess the same goes for signoff (git commit -s) ?19:51
SteveC_ hugoxrosa: nope19:51
hugoxrosa: it has two other dot folders inside of it19:51
hugoxrosa SteveC_, hum... is the folder unstagged?19:54
← mobhawk left19:54
hugoxrosa SteveC_, I have problem only in this two situations.19:54
bleything → bleything|away19:55
→ dwave joined19:55
hugoxrosa SteveC_, maybe you should delete (with rm) and use git gui to correct the status of git.19:57
SteveC_ hugoxrosa: i ended up deleting it through the svn repository19:57
hugoxrosa SteveC_, It have been my solution in case like this one.19:57
drizzd henr_k: pretty sure, yes19:57
SteveC_ hugoxrosa: git wouldn't see it no matter what idid19:57
henr_k drizzd: ok, thanks. then I can stop looking for it :)19:59
→ dbtid joined20:03
← UUncia left20:04
→ peshalto joined20:04
→ porcnick joined20:05
→ Ademan joined20:07
→ Travis-42 joined20:08
→ ciskje joined20:08
→ UUncia joined20:09
→ kold joined20:10
← dwave left20:11
← jwshen left20:11
→ Azerthoth joined20:12
← peshalto left20:13
→ warthog9 joined20:13
→ poseidon joined20:14
← fujin left20:17
zuez I have the following listed in .git/info/exclude, yet it still shows up in git status client/myproject/sf.pbxuser20:17
does that mean that it is being versioned already?20:17
← schmrz left20:18
patrikf zuez: is it listed under "untracked files"?20:19
→ boto_ joined20:20
← ph^ left20:21
→ naeu joined20:21
→ Bagder joined20:21
tswicegood → tswicegood-afk20:21
→ schacon joined20:24
← tomaw left20:25
jksz_ → jksz20:25
→ tomaw joined20:26
→ nighthwk1 joined20:29
← Modius left20:29
→ aspotashev|eeepc joined20:30
← bdiego left20:31
← boto left20:33
→ rimenes joined20:35
→ Modius joined20:36
← wsc left20:37
← seangrove left20:37
← joevandyk left20:37
→ wsc joined20:37
→ seangrove joined20:37
→ joevandyk joined20:37
rimenes → rimenes-away20:38
← krh left20:39
← naeu left20:42
Stravy hi all : the hook for http protocol run git-update-server-info, does this means that git must be installed on the machine running the web server?20:42
← Bagder left20:43
→ javatexan joined20:43
← patrikf left20:43
Ilari Stravy: Well, you need it installed anyway for ssh push...20:43
Stravy: And HTTP push doesn't run hooks.20:44
blingo I've used "git init, git add ." on a large diractory, it took it some minutes, and no information like ending time was displayed, is there some way to see it?20:46
← javatexan left20:46
Fissure don't think so20:46
→ s0ber joined20:47
Fissure though in my experience, add is almost instantaneous20:47
→ javatexan joined20:48
→ eletuchy joined20:48
blingo Darn, thanks. <Fissure>20:50
Stravy Ilari: funny I didn't realized that git had to be installed on both sides of the push, as it was always the case when I tried20:50
← jceb left20:51
← Weaselweb left20:51
← ntoll left20:53
← DavidKlein left20:53
Ilari blingo: Not usable after-the-fact, but there is shell builtin 'time' that can be used to show real, user and system times at process exit.20:54
blingo intersting20:54
I assume can use TOP to see the time it is running, but not the time to finish :(20:55
→ ntoll joined20:55
← carrus85 left20:56
uau blingo: i'd guess it was disk activity20:56
← ntoll left20:56
← lorandi left20:56
uau IIRC 'git add' needs to read all the files you're adding20:56
blingo I'm by no way blaming git, I want to create a gui, and it's a must for it, I only see tweaking git as an option, a bad one.20:57
← bhearsum left20:57
← gh34 left20:58
blingo Maybe there's a way to see what file git is currently working on?20:59
← s0ber_ left20:59
uau if nothing else, you could use a tool like lsof to see what files the process has open (or look directly under /proc/pid)21:00
Ilari Heh... Running one program I have done (on rather big iron): It spent 5.56s in user space, 2.19s in system calls, with wallclock completion time of 2.269s... Program internal CPU usage stat showed 499% CPU usage...21:01
← cannonball left21:01
blingo 499% is a bit high :)21:02
uau you can get similar times on basic consumer processors nowadays (though more than 4 cores is still relatively rare i guess)21:02
← AkumaStreak left21:02
→ krawek joined21:02
Ilari More like bit low. That program is bottlenecked by mmap performance.21:03
dmlloyd not if you have, say, 8 CPUs :)21:03
blingo for 8 CPUs, being one CPU 100%, it's low :-)21:03
uau dmlloyd: i know there are machines with more, i meant that quad-core processors have been common consumer stuff for a while21:04
Ilari Data throughput stat was 241 028kB/s... The program could take some optimization.21:04
← Stravy left21:05
→ mithraic joined21:06
blingo I will check the lsof option, thanks, hope that git open files one by one..21:06
Ilari blingo: If you want performance, issue multiple database writes in parallel and then run index update?21:07
← lucsky left21:08
← cemerick left21:08
Ilari blingo: And since the program is issuing the commands, it presumably knows what file is currently being handled.21:08
uau blingo: if you mean you want to find the files from a _program_ (as opposed to checking it in your interactive use), then hacks like lsof are certainly not a good idea21:10
→ _Vi joined21:10
→ Suhail joined21:10
blingo I appreciate your kind efforts, I will explain, It's not the speed that worries me, I'm writing a gui for git, and I need to show it's current status, something like working and not showing any status isn't good for me. this is the reason I want to know what file it working on - to show something in the gui, maybe calc the number of files and than showing a progress bar by counting the files remaining..21:10
Suhail Is there anything special I need to do, I am changing servers and I want to move my remote repo I've been using to another21:11
place21:11
→ eletuchy_ joined21:11
Ilari Suhail: Nothing special, aside of changing the URLs. Also if you use file-level copy, you might want to use fsck on the result.21:12
← nighthwk1 left21:12
→ AkumaStreak joined21:12
Ilari blingo: Use hash-object + update-index? That way you know what is being added right now...21:13
← paulboone left21:13
Ilari blingo: The update-index is single step, but it should be very fast even with huge number of files.21:13
Fissure unless you're doing something crazy like writing to sshfs over a modem link ;)21:14
← [intra]lanman left21:14
← quoin left21:14
← qrush left21:14
→ paulboone joined21:15
← solydzajs left21:15
Ilari blingo: For even more control, write the object files yourself. It isn't as hard as it sounds. The hardest part in it is actually discovering wheither or not you actually need to write an object.21:15
← aspotashev_ left21:16
blingo I thank you <Ilari> this is huge help, for the easier part can you tell me how to combine the hash-object + update-index to the add . ?21:16
← kumbayo left21:17
Ilari blingo: Well, the idea is: use hash-object to write the changed stuff to object database and then update-index to update what the index contains. You have to extract the list of files to update yourself.21:17
→ smashwilson joined21:18
Suhail Ilari: how would you change the url?21:18
Ilari blingo: There is update-index mode that accepts complete index entries from stdin.21:19
← leanucci left21:19
Ilari Suhail: edit .git/config or use 'git config remote.<remote>.url <newURL>'.21:19
Suhail: I would go and directly edit .git/config.21:20
→ leanucci joined21:20
Suhail Ilari: so i change the url and just push the changes from my local repo?21:20
blingo o.k I assume what you're suggesting requires a higher level of understanding of git internals, I will try to achieve it, but getting this information of which file is "handled" does seem possible to you?21:20
← leanucci left21:21
→ leanucci joined21:21
Ilari Suhail: You can do that way as well, if you have the needed stuff in local repo.21:21
← eletuchy left21:21
Suhail Ilari: yeah i do21:21
← yann left21:22
← mikearr left21:22
Ilari blingo: Search for "Git for Computer Scientists" and "Git from the Bottom Up". And I don't think its possible to get info about progress of commands during invocation, only about what is being invoked.21:23
→ mikearr joined21:23
blingo So i'll have to do each command on one file at a time to make this work?21:23
Ilari blingo: So to get smooth progress even on large files, you would need to write the stuff out yourself.21:24
blingo but on many small files, your technique will work?21:25
Ilari blingo: Or batch files to windows by size/number and hand off each window at time.21:25
blingo You mean add 1.file, move progress bar, add 2.file etc' ?21:26
→ ebneter joined21:26
Ilari blingo: I mean: You have list of files that need adding. Find size of each. Now batch them into sets, such that each set either stays under given size and number limit or contains single file that blows the size limit. Then hand off each set to one command.21:27
blingo got you.21:28
and on commit, I see the list of files, but only when the commit finishes, is there something to be done the easy way ?21:29
Ilari blingo: Doing such batching to optimal number of batches is probably NP-hard, but any reasonable approximation should do.21:29
← eletuchy_ left21:29
doener blingo: committing should be fast enough not to matter at all21:29
blingo: if something is slow, then it's the hash-object part21:29
Ilari blingo: To commit, you actually need index file that contains the stuff you want to commit.21:29
→ bremner_ joined21:30
← khmarbaise left21:30
troyt Quick question about git commit messages: the syntax highlight of my $EDITOR changes colors at column 50 for the commit message (and the syntax is git commitmsg). I've noticed similar behavior with a couple GUI programs for git commits. Can anybody explain why?21:31
← priidu left21:32
→ fargiola` joined21:32
Ilari troyt: The convention is that first line of commit msg is at most 50 characters.21:32
doener Ilari, blingo: For the hash-object/update-index part, you could fork hash-object -w --stdin-paths and feed paths to it. It reports back the hash immediately, so you can write the index entry21:32
troyt OK; thanks, that was what I was wondering.21:32
doener Ilari, blingo: and as both run in batch mode, you don't have to fork like hell21:32
Ilari doener: Ah, yes that...21:32
← bremner_ left21:32
doener I think git-svn does that21:33
→ qrush joined21:33
Ilari doener: There isn't any command that could accept batch object write input from stdin complete with assumed paths and even with types? Didn't see such in git-hash-object(1).21:34
→ gberdyshev joined21:35
blingo I think It is all clear now..21:35
← pcutler left21:35
doener Ilari: no, nothing I'd know of. With --stdin it writes just a single object, and --stdin-paths just writes objects of one type.21:36
Ilari doener: And besides, the --stdin-paths presumably accepts paths, not contents.21:36
blingo You all were huge help, I will dig into this "Git for Computer Scientists" and "Git from the Bottom Up" many thanks.21:36
doener Ilari: uhm, yeah. But wasn't that the idea? To get a progress report for adding a bunch of existing files?21:37
Ilari: didn't follow the whole discussion, just skipped to the start once21:37
Ilari blingo: Just don't do anything stupid like trying to add all files every time (first see which have mods in working dir).21:38
blingo lol thanks :)21:39
← gberdyshev left21:39
← Azerthoth left21:40
← LiamH left21:41
Ilari troyt: Presumably the rest of the lines can be little longer, but don't overdo it. Even if you have 200 wide terminal doesn't mean everyone has one. And even if one has, reading that long lines is painful.21:41
← leanucci left21:42
troyt I figured that; I just thought it odd the value was < ~80 characters (ie. most terminal widths)21:42
← blingo left21:42
→ gebi_ joined21:42
→ daed455 joined21:42
daed455 does anyone here use buildbot?21:43
← fargiolas left21:43
daed455 → daed21:43
Ilari Real fun to read some pages from terminal-based web browser when the terminal width is 212 characters...21:43
← bmalee left21:43
← impulze left21:43
← up_the_irons left21:43
← madduck left21:43
← Tali left21:43
← incd^ left21:43
← numbah left21:43
← hegge left21:43
← avtobiff left21:43
← janl left21:43
← Ingmar left21:43
← dionoea left21:43
← pcgod left21:43
← natlus left21:43
← hgb left21:43
← maxb left21:43
→ bmalee joined21:45
→ up_the_irons joined21:45
→ numbah joined21:45
→ impulze joined21:45
→ Ingmar joined21:45
→ hegge joined21:45
→ natlus joined21:45
→ pcgod joined21:45
→ maxb joined21:45
→ incd^ joined21:45
→ hgb joined21:45
→ dionoea joined21:45
→ Tali joined21:45
→ madduck joined21:45
→ avtobiff joined21:45
→ janl joined21:45
→ dodo_the_last1 joined21:46
← struberg left21:46
→ seb- joined21:48
seb- When "git checkout foo" changes your branch, does it change the SOURCE CODE in the local working directory also to reflect that change?21:48
← qrush left21:49
→ btanaka joined21:49
btanaka is "sharedRepository = group" supported in git version 1.4.4.421:50
← jefftriplett left21:50
← juanjoc left21:52
← SteveC_ left21:52
← jsbmsu left21:53
← jballanc left21:53
← dodo_the_last left21:54
← hugoxrosa left21:55
← romao left21:56
→ pinholecamera joined21:56
pinholecamera hey folks21:56
→ romao joined21:56
→ sshc joined21:56
pinholecamera was wondering, how do i resolve a situation where i try to pull from master and it tells me i can't pull, because xxx file would be overwritten by merge? i just tried merging and no dice... where do i go from here?21:57
→ Azerthoth joined21:57
sshc I'm in the middle of a bisect and I made some local changes so that it could compile. How do I revert all of my local changes I made without messing up the bisect?21:57
→ fargiolas joined21:58
← ankit9_ left22:00
← Tuomas left22:00
← angerman_ left22:00
← Lgb left22:00
← seb- left22:01
← thumper left22:01
fr → fr|22:01
→ Lgb joined22:01
← hobodave left22:01
→ j416 joined22:02
pinholecamera the file is actually the same as on the repo i'm pulling from, which is the really confusing part22:02
→ thisisrobv joined22:02
← hydrogen left22:02
→ hydrogen_ joined22:02
→ thumper joined22:03
thisisrobv Quick question, does git do bit level replacement on pushes / pulls22:03
or byte level22:05
Ilari sshc: 'git reset --hard'?22:06
pinholecamera: Does 'git status' list it?22:07
← loincloth left22:07
← fargiola` left22:07
pinholecamera Ilari: unfortunately yes22:07
sshc Ilari, thanks22:07
Ilari thisisrobv: Pushes/pulls transfer complete objects (perhaps delta-compressed).22:07
sshc how do I get the current revision?22:07
I'm trying to store it in a bash variable22:08
Ilari sshc: 'git rev-parse HEAD'?22:08
sshc foo=`git bar`22:08
Ilari, yep, thanks again :)22:08
Ilari pinholecamera: 'git diff HEAD -- file' should show the diff.22:08
pinholecamera: Also, what OS?22:08
← wycats left22:08
pinholecamera Ilari: debian22:09
parasti thisisrobv: what do you mean by "replacement"?22:09
Ilari pinholecamera: So no CRLF problems...22:09
→ qrush joined22:10
Ilari pinholecamera: And presumably the underlying filesystem is sane.22:10
→ warlock_mza joined22:10
→ jefftriplett- joined22:10
pinholecamera Ilari: yes... accord to diff there is no data difference22:10
← boto_ left22:11
Ilari pinholecamera: What about 'git diff -- file' (not the same thing as 'git diff HEAD -- file')?22:11
← btanaka left22:11
comboy git clone should fetch all branches, right? if so, any idea why it may be fetching just master for me?22:11
Ilari comboy: It does fetch them all.22:11
comboy: Look at 'git branch -r'.22:11
pinholecamera Ilari: doesn't return anything22:11
comboy Ilari: ah, so I just do git fetch origin mybranch:mybranch, right?22:12
← mikearr left22:12
Ilari comboy: No. 'git checkout --track origin/mybranch' (assuming recent enough git).22:12
← bx2 left22:12
→ mikearr joined22:13
comboy Ilari: it says --track requiries -b option22:13
Ilari pinholecamera: In what section does 'git status' list the file. What it shows being done to that file.22:13
← daed left22:14
← Travis-42 left22:14
Ilari comboy: 'git checkout -b mybranch origin/mybranch' (your git isn't new enough for --track on remote branch without -b).22:14
pinholecamera Ilari: it's under "changes to be commited"22:14
← mbroeker left22:14
thisisrobv parasti: byte level transfers Ilari said no22:14
Ilari pinholecamera: New file? Modified? Unmerged?22:14
comboy Ilari: ok, thenks a lot22:15
pinholecamera Ilari: modified22:15
Ilari pinholecamera: IIRC, unmerged stuff appears as both 'modified' and 'unmerged'.22:15
→ _sps_ joined22:16
pinholecamera Ilari: hrmm, weird22:16
← _sps_ left22:16
→ rjc joined22:17
Ilari pinholecamera: What does 'git rev-parse HEAD:file :0:file' yield, the same hash two times or different hashes?22:18
→ dodo_the_last joined22:18
→ ankit9_ joined22:18
pinholecamera Ilari: same hash twice22:18
← _sps left22:19
Ilari pinholecamera: Whee... Getting even stranger. Two same hashes means the file is identical in HEAD and in staged. But why did status report it having modifications to be committed.22:20
pinholecamera Ilari: exactly, it's really weird22:20
← gaveen left22:20
← porcnick left22:20
← sergiodj left22:20
Ilari pinholecamera: Wonder what would happen if you tried to commit...22:20
pinholecamera i can give it a shot22:20
← thisisrobv left22:21
→ gaveen joined22:21
← amrox left22:21
pinholecamera Ilari: well it let me commit22:22
← ankit9_ left22:22
← Dashkal left22:22
← agile left22:22
dereine → dereine[OFF]22:22
Ilari pinholecamera: Does pull work now?22:22
Suhail Ilari: should I do git --bare init and then push to it?22:23
Ilari Suhail: Yes (assuming it is only intended to be updated by single user).22:24
Suhail Ilari: multiple users are going to commit code to this repo22:24
pinholecamera Ilari: it tells me it's up to date, which is interesting, there was a whole other commit from another machine that i was trying to pull22:24
Suhail Ilari: we're using it as a remote repo where we can commit/bring down changes for now22:24
→ hobodave joined22:25
← xeno left22:25
→ krich87 joined22:25
← krich87 left22:25
→ xeno joined22:25
← corybantic left22:25
pinholecamera Ilari: so now i just need to back track in the commits i guess22:26
Ilari pinholecamera: How many lines 'git rev-parse HEAD^!' gives?22:26
← j416 left22:26
pinholecamera Ilari: three22:26
Ilari pinholecamera: Last commit is merge then.22:26
pinholecamera Ilari: so it merged with my last commit, you mean?22:27
Ilari pinholecamera: Apparently last commit you made did merge in what you where trying to pull...22:27
→ tvw joined22:27
pinholecamera Ilari: huh, so weird, none of the changes from the previous commit are actually there22:28
Ilari pinholecamera: Git pull/merge does save information that subsequent commit picks up to make proper merge.22:28
Suhail Ilari: i got No refs in common and none specified; doing nothing22:28
Ilari pinholecamera: Fscked up merge resolution?22:29
← rjc_ left22:29
Suhail oh wait git push origin master seems to work22:29
← ciskje left22:29
pinholecamera Ilari: yeah, i think something got missed somehow, i'll take a look at the versions22:29
Ilari Suhail: That only pushes master.22:29
Suhail Ilari: yeah how do i push everything?22:29
i basically need to move my entire repo to a new place22:30
Ilari Suhail: --all22:30
Suhail: And then second push with --tags22:30
Suhail git push --al?22:30
--all22:30
← Azerthoth left22:30
→ dodo_the_last2 joined22:30
Ilari Suhail: 'git push origin --all ; git push origin --tags'.22:30
→ Azerthoth joined22:31
← dodo_the_last left22:31
← Yuuhi left22:31
hachi is there a way to test a remote?22:33
I don't want to push22:33
nor pull22:33
I just want to see if it seems to be a valid remote22:33
broonie hachi: git remote show <remote> should do the trick.22:33
Ilari hachi: Also, 'git remote' list the remotes.22:33
Run → Aleric22:33
hachi it lists them22:33
yes, but does it connect?22:34
Ilari hachi: 'git ls-remote <remote>'?22:34
broonie yes, if you ask it to display an individual remote.22:34
cpg → cpg|away22:34
hachi okay okay, thank you22:35
cpg|away → cpg|biab22:35
hachi thanks a ton :)22:35
Ilari hachi: That connects to remote repo and obtains list of refs that are present there.22:35
← qrush left22:35
hachi ilari++ broonie++22:35
Suhail Ilari: thanks, move worked perfect22:35
Ilari: I thought it was going to suck a lot more than it actually did.22:35
→ mtkd joined22:36
← Ryback_ left22:39
← dodo_the_last1 left22:40
← Azerthoth left22:40
→ ankit9_ joined22:41
← fargiolas left22:42
→ sakyamuni joined22:43
dwmw2 → dwmw2_gone22:44
→ gaveen_ joined22:45
← gaveen left22:45
hachi if I screwed up my config file, and now I have a bunch of refs/remotes/* that shouldn't be there22:46
can I just delete the files from .git/refs/remotes/22:46
or is there more cleanup involved?22:46
thiago_home yes22:46
there's .git/logs/refs/remotes22:46
← tatsuyao left22:46
hachi wow, what does that contain?22:47
← tono__ left22:47
doener it's the reflog for the ref22:47
→ rolfb joined22:48
→ jerbear joined22:48
← tvw left22:49
→ tvw joined22:49
cpg|biab → cpg22:52
← hpoydar left22:54
← helo left22:56
pinholecamera Ilari: huh, it actually did merge them correctly. not sure why it took me so long to verify that. thanks for your help22:56
sebleier → sebleier|away23:00
← mfilipe left23:01
→ porcnick joined23:01
← djgera left23:02
→ agile joined23:03
→ Dashkal joined23:03
← MisterN left23:04
← tbf left23:07
UUncia When I try to clone a remote repository I keep getting "fatal: The remote end hung up unexpectedly" (I use ssh://), but when I connect through ssh to that machine it's ok... any ideas why?23:10
← mtkd left23:11
jefftriplett- → jefftriplett-afk23:12
← Suhail left23:12
Ilari UUncia: Any other messages?23:13
← SRabbelier left23:14
← javatexan left23:16
← elik left23:17
→ javatexan joined23:19
→ Budd joined23:19
← doener left23:20
→ djgera joined23:20
Budd Is there a nice way to get a list of git repositories on a server? I couldn't find anything relevant in the man pages.23:21
I'm trying to make a script to auto-update all the x.org packages from git.23:22
← rolfb left23:22
Ilari Budd: No, unless the server offers some specific way to do it.23:22
→ loincloth joined23:23
Ilari Budd: As with some common setups, a git repo could be lurking pretty much anywhere in filesystem. And there's no registry of them.23:24
Budd OK. They do have an index web page, which will suffice - just less automatic.23:24
Thanks.23:24
→ sergiodj joined23:25
→ LiamH joined23:25
→ agenteo joined23:25
← LiamH left23:26
← countd left23:26
djgera Hi people, in my machine, i doing some weekly backups from my entire GNU/Linux system to another harddisk with a "rsync -a", in some directories I have some GIT repos, my question is: these GIT repos, are corrupted or GIT don't know about some "file stats" (like inode, timestamps, etc are changed)?23:28
← tvw left23:28
→ jballanc joined23:29
bdrewery djgera: that's fine23:29
I rsync my .gits all the same all over, they all function properly23:29
djgera bdrewery, so git only tracks only the contents of files , true?, like read in some manpages23:30
Ilari djgera: Git DOES know about those and uses them.23:30
djgera ooh OK23:30
Ilari djgera: Use 'git update-index --refresh' or 'git status' to fix them.23:30
djgera: That info being wrong can lead to Git considering clean file dirty, leading to some operations aborting.23:31
djgera thanks, let me try....23:31
will do a simple "cp -a"23:31
bdrewery yeah forgot to mention the status part, simple solution though23:31
← rektide left23:32
Ilari djgera: And those false dirties won't show up in diff.23:32
← jrmuizel left23:33
djgera mmm OK, the git status don't show any "changes" after "cp -a" under new dir.23:33
→ sergio_ joined23:34
djgera also git update-index --refresh, none output,23:34
← poseidon left23:34
← sgrimm left23:34
→ Suhail joined23:34
Suhail i am getting permission denied for my repo after adding my pub key to gitosis23:35
Ilari djgera: Git status essentially first runs 'git update-index --refresh'.23:35
← vipul left23:35
Ilari djgera: AFAIK, there is no easy way to get list of those falsely dirty files. But they mess things such as merges.23:35
→ rektide joined23:36
Ilari Suhail: You authorized that key to access that repo too?23:36
→ qrush joined23:36
Suhail Ilari: put in the keydir and added the <filename> to the members list23:36
← wmoxam left23:36
djgera ok so, for safe/secure backup, then... ignore these git repos in rsync, then do a git clone ?23:37
Ilari Suhail: If the key file is called 'foo.pub', the username is 'foo'. Also, any warnings mean it DOES NOT WORK.23:37
djgera: You can rsync them, as long as you remember that those repos need fixing before being used.23:38
← sergio_ left23:38
Suhail Ilari: yeah i did that23:39
→ sergio_ joined23:39
Suhail Ilari: I made a called test.pub and added "test" to the members list23:39
Ilari: now I am trying to clone the repo locally and it won't let me23:39
← rylar left23:40
← Ademan left23:40
← Alien_Freak left23:40
Suhail Ilari: and it asks for my password23:41
Ilari Suhail: Key problems.23:41
djgera oh sorry, then misinterpreted. You mean I can be sure that after a "git status" or the another command, all is right and ready to use without any problem, right?23:41
Ilari djgera: Yes.23:41
djgera many thanks!23:41
Ilari Suhail: That is, if it asks for password, SSH has rejected keypair authentication attempts.23:42
← meyering left23:42
Ilari Suhail: And with gitosis it means you don't either have valid key to authenticate or SSHD is configured to deny logins to the whole account.23:43
djgera recently started working with git, I had never worked with this software, just download and compile software really is impressive. Again thank :)23:43
→ poseidon joined23:43
Ilari Suhail: Where did you get the .pub file you added to keydir/?23:44
Suhail Ilari: account isn't denied23:44
Ilari: I generated it locally23:44
Ilari Suhail: You didn't have keypair before?23:45
→ rylar joined23:45
← warthog9 left23:45
Suhail Ilari: I got it to work, just have to add the private via ssh-add23:46
Ilari: think that was the problem23:46
Ilari Suhail: I consider having multiple ssh private keys per account a mistake.23:46
Suhail Ilari: just wanted to test23:46
Ilari Suhail: Unless the others are for special purpose use.23:47
← javatexan left23:47
Ilari Suhail: That is, each account has one and only one SSH private key for interactive use. Anything else is asking for trouble.23:47
← sergiodj left23:48
← romao left23:48
→ javatexan joined23:49
← cmarcelo left23:50
→ scientes joined23:53
→ warthog9 joined23:53
→ arohner joined23:55
→ peshalto joined23:56
← wshimmy left23:58
→ wshimmy joined23:59
← intripoon left23:59

Logs Search ←Prev date Next date→ Channels Documentation