{"thread":{"id":"2611","subject":"What's in master tonight.","startedAt":"2005-11-20T08:40:21Z","lastAt":"2005-11-20T14:28:38Z","messageCount":2,"participants":["Junio C Hamano","Yaacov Akiba Slama"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"12365","messageId":"7v7jb31zm2.fsf@assigned-by-dhcp.cox.net","threadId":"2611","inReplyTo":null,"subject":"What's in master tonight.","fromName":"Junio C Hamano","fromEmail":"junkio@cox.net","sentAt":"2005-11-20T08:40:21Z","receivedAt":"2005-11-20T08:40:21Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"I've merged http updates (mostly code refactoring and cleanups)\nfrom Nick Hengeveld, config-set from Johannes Schindelin, and\nmost of user-relative paths series from Andreas Ericsson.\nCooking in proposed updates are the final part of user-relative\npaths series that changes git-daemon, git:// proxy support from\nPaul Collins, and repository format futureproofing from Martin\nAtukunda.\n\nI'd like to do the next 1.0rc when at least the repository\nformat futureproofing is ready.  I personally can live without\nuser-relative paths or git:// proxy, although both are probably\nnice to have.  Here are the current status of these three\ntopics.\n\nrepository format futureproofing::\n\n\tI think the current code can copy template/config that\n\twas prepared for wrong version of repository format\n\twithout complaining.  It should barf and fail.\n\n        I am unsure if the check in setup_git_repository() is\n        the right thing to do.  It doesn't check when $GIT_DIR\n        environment is set, which is wrong; receive-pack and\n        upload-pack are run with this environment set to \".\".\n\n\tAlthough majority of important core programs do check\n\tthe repository format before doing any harmful\n\toperations, some higher level scripts should be told to\n\tdo the checking upfront.  The config-set command by\n\tJohannes is probably handy for this purpose.\n\nuser relative paths::\n\n\tWhat's left from the master is the part that touches the\n\tgit-daemon.  We need to be careful about being strict\n\twhile validating the path against whitelist _after_\n\tDWIM, so the version in the proposed updates need to be\n\tchecked to make sure it is safe.\n\nproxy for git:// transport:\n\n\tI do not have any pending issues on this patch, and I\n\tthink it is ready to go in when somebody, preferably\n\tother than the original author, says it is useful and\n\tdoes not break things.\n\nThe master branch is fairly ahead of the last 1.0rc as of this\nwriting, and I'll be attending a wedding of a coworker, so there\nwill not be maint updates this weekend.\n"},{"id":"12373","messageId":"43808816.4050501@slamail.org","threadId":"2611","inReplyTo":"7v7jb31zm2.fsf@assigned-by-dhcp.cox.net","subject":"Simple proxy test (was Re: What's in master tonight)","fromName":"Yaacov Akiba Slama","fromEmail":"ya@slamail.org","sentAt":"2005-11-20T14:28:38Z","receivedAt":"2005-11-20T14:28:38Z","isPatch":false,"sender":{"key":"ya@slamail.org","avatar":null},"body":"Junio C Hamano wrote:\n\n>proxy for git:// transport:\n>\n>\tI do not have any pending issues on this patch, and I\n>\tthink it is ready to go in when somebody, preferably\n>\tother than the original author, says it is useful and\n>\tdoes not break things.\n>\nI tested the proxy code from a place when the git:// port is blocked.\n\n1) I made a ssh socks tunel (ssh -D 2080 a-remote-server).\n2) I compiled connect.c from \nhttp://zippo.taiyo.co.jp/~gotoh/ssh/connect.html\n3) I created a simple script called myproxy:\n\n#!/bin/sh\n\n/path/to/connect -S localhost:2080 $@\n\n4) GIT_PROXY_COMMAND=/path/to/myproxy\ncg-clone git://git.kernel.org/pub/scm/git/git.git git2 worked perfectly\n\n4bis)\n% cat .git/config\n[core]\ngitproxy = /path/to/myproxy for kernel.org\n\ncg update worked perfectly.\n\n--yas\n"}]}