{"thread":{"id":"17914","subject":"Wierd git over http behaviour","startedAt":"2009-02-20T06:13:20Z","lastAt":"2009-02-24T05:45:05Z","messageCount":4,"participants":["John Williams","Junio C Hamano","Sverre Rabbelier"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"105566","messageId":"1d3f23370902192213g560020b3h7a2459c1439e5e5c@mail.gmail.com","threadId":"17914","inReplyTo":null,"subject":"Wierd git over http behaviour","fromName":"John Williams","fromEmail":"john.williams@petalogix.com","sentAt":"2009-02-20T06:13:20Z","receivedAt":"2009-02-20T06:13:20Z","isPatch":false,"sender":{"key":"john.williams@petalogix.com","avatar":null},"body":"Hi,\n\nWe have a git tree that accesses fine via the git protocol, however I have\nalso opened up http support for it as some of our users are behind\ngit-unfriendly corporate firewalls.\n\nI know http transport is not encouraged, but not much I can do about that.\n\nStrange thing is, when doing an http fetch of said tree, it gets no updates\nto the tree later than about Jan 13 this year, even though the tree itself\nhas many, many commits since then (its origin is Linus's tree and we pull\nregularly). Heads created later than that date are also not visible (doing\ngit branch -a) in the clone pulled over http. Clones made using git: are\nfine.\n\nAnything I can look for in the .git repo on the server, permissions or\nsimilar that might be relevant? It's almost like there's a single commit in\nthere that somehow is making the http fetch puke. If you browse the http\ntree with a web browser, it's all there as expected.\n\nFor reference here are the URLs:\n\ngit://developer.petalogix.cmo/linux-2.6-microblaze\n\nhttp://developer.petalogix.com/git2/linux-2.6-microblaze\n\nInitially the http reference was a symlink on the server to the actual git\ntree, but I also tried just rsync'ing it across to /var/www/html/git2 to\nmake sure it wasn't the symlink breaking things - no change.\n\nAm I misunderstanding something or is this just flakey behaviour of the http\nprotocol support?\n\nThanks,\n\nJohn\n"},{"id":"105567","messageId":"7vab8hiq3o.fsf@gitster.siamese.dyndns.org","threadId":"17914","inReplyTo":"1d3f23370902192213g560020b3h7a2459c1439e5e5c@mail.gmail.com","subject":"Re: Wierd git over http behaviour","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2009-02-20T06:23:23Z","receivedAt":"2009-02-20T06:23:23Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"John Williams writes:\n\n> Strange thing is, when doing an http fetch of said tree, it gets no updates\n> to the tree later than about Jan 13 this year, even though the tree itself\n> has many, many commits since then (its origin is Linus's tree and we pull\n> regularly). Heads created later than that date are also not visible (doing\n> git branch -a) in the clone pulled over http. Clones made using git: are\n> fine.\n\nMost likely you would want to run git-update-server-info from\nproject.git/hooks/post-update hook. A sample shipped with git looks like:\n\n #!/bin/sh\n #\n # An example hook script to prepare a packed repository for use over\n # dumb transports.\n #\n # To enable this hook, rename this file to \"post-update\".\n\n exec git-update-server-info\n\nmake sure it is executable.\n\n $ cd linux-2.6-microblaze.git/hooks\n $ ls -l post-update\n $ chmod +x post-update\n\nAnd then run \"git update-server-info\" just once in the repository. I see\nit was run on Jan 22nd and not since then, but the latest of your refs\nhave Feb 19th timestamp. Every time somebody pushes into this repository,\nyou would need to run the command, and post-update hook is a standard way\nto do so easily.\n"},{"id":"105569","messageId":"bd6139dc0902192239i3c5f196qf890c8b0a6f9ff46@mail.gmail.com","threadId":"17914","inReplyTo":"1d3f23370902192224g71270880p84e150af9640dec6@mail.gmail.com","subject":"Re: Wierd git over http behaviour","fromName":"Sverre Rabbelier","fromEmail":"srabbelier@gmail.com","sentAt":"2009-02-20T06:39:36Z","receivedAt":"2009-02-20T06:39:36Z","isPatch":false,"sender":{"key":"srabbelier@gmail.com","avatar":"https://avatars.githubusercontent.com/u/3098?v=4"},"body":"Heya,\n\nOn Fri, Feb 20, 2009 at 07:24, John Williams\n wrote:\n> On Fri, Feb 20, 2009 at 4:19 PM, Sverre Rabbelier wrote:\n>> This url seems to be broken/dead?\n>\n> bah - append .git :(\n>\n> http://developer.petalogix.com/git2/linux-2.6-microblaze.git\n\nOk, that works :).\n\n-- \nCheers,\n\nSverre Rabbelier\n"},{"id":"105977","messageId":"1d3f23370902232145g61bccd67na0024ab81627c5e6@mail.gmail.com","threadId":"17914","inReplyTo":"7vab8hiq3o.fsf@gitster.siamese.dyndns.org","subject":"Re: Wierd git over http behaviour","fromName":"John Williams","fromEmail":"john.williams@petalogix.com","sentAt":"2009-02-24T05:45:05Z","receivedAt":"2009-02-24T05:45:05Z","isPatch":false,"sender":{"key":"john.williams@petalogix.com","avatar":null},"body":"On Fri, Feb 20, 2009 at 4:23 PM, Junio C Hamano wrote:\n> John Williams writes:\n>\n>> Strange thing is, when doing an http fetch of said tree, it gets no updates\n>> to the tree later than about Jan 13 this year, even though the tree itself\n>> has many, many commits since then (its origin is Linus's tree and we pull\n>> regularly).  Heads created later than that date are also not visible (doing\n>> git branch -a) in the clone pulled over http.  Clones made using git: are\n>> fine.\n>\n> Most likely you would want to run git-update-server-info from\n> project.git/hooks/post-update hook.  A sample shipped with git looks like:\n\nThat was it - thanks for the quick response.\n\nJohn\n"}]}