{"thread":{"id":"3046","subject":"What's cooking in git.git","startedAt":"2006-01-12T00:04:41Z","lastAt":"2006-01-16T10:53:56Z","messageCount":3,"participants":["Junio C Hamano","Petr Baudis"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"14513","messageId":"7vhd8al3ae.fsf@assigned-by-dhcp.cox.net","threadId":"3046","inReplyTo":null,"subject":"What's cooking in git.git","fromName":"Junio C Hamano","fromEmail":"junkio@cox.net","sentAt":"2006-01-12T00:04:41Z","receivedAt":"2006-01-12T00:04:41Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"I have queued the following for the maintenance branches.\n\nFor both 1.0.X and 1.1.X series:\n\n - update-index: work with c-quoted name\n - show-branch: handle [] globs as well.\n - name-rev: do not omit leading components of ref name.\n\nFor 1.1.X series only:\n\n - Add git-describe to .gitignore.\n - describe: do not silently ignore indescribable commits\n\nI plan to do 1.0.10 [*1*] and 1.1.2 sometime this weekend, along\nwith some other fixes that would appear until then.\n\nOn the development front, I have these in the proposed updates\nbranch:\n\n - Exec git programs without using PATH (Michal Ostrowski).\n\n   This has been discussed for the last couple of days on the\n   list.  I think C side is ready to go in, but I'd like to\n   decide what to do on the scripting side before merging it\n   into the master branch.\n\n - Disable USE_SYMLINK_HEAD by default (Pavel Roskin).\n\n   This has been on hold for almost two months -- forever in git\n   timescale.\n\n   One offender that could have been broken by this change was\n   fixed recently (gitweb), so this can go in anytime now.  Do\n   people still care about it?  The argument for the change when\n   it was proposed was \"then we only have to worry about one\n   format of .git/HEAD, not two\", and it still is a valid\n   argument from Porcelain writers' point of view.  To be\n   honest, since I do not do Porcelain, I am neutral about it.\n\n - convert-packs: futureproofing.\n - Require packfiles to follow the naming convention (preparation).\n\n   These two I have been holding for a while.  I would\n   eventually like to force people to rename the packfiles\n   created by GIT older than v0.99.9 which has unverifiable\n   hexadecimal string in their names and these are the first\n   two baby steps toward that goal.  If people object to the\n   idea of verifiable packname in general, please raise hand,\n   and I may be talked into shelving the whole thing.\n\n - Documentation: show-branch.\n - show-branch: always include the current branch\n - show-branch: make --topo-order noop and default.\n - show-branch: make the current branch stand out.\n\n   Here are mostly usability updates.\n\nI'll send out some of the above patches to the list separately\nfor discussion and review.\n\n\n[Footnote]\n\n*1* 1.0.X series maintenance can largely be ignored by general\npublic, since I do not see any reason to stick to 1.0.X series.\n1.1.X series does not have any backward incompatibility issues\nas far as I know.  I am doing 1.0.X series primarily as a\npractice for myself, and to experiment with a maintainer\nworkflow like this:\n\n - \"fixo/1.0\" topic branch queues fixes applicable for 1.0\n   series *and* *later* maintenance series.\n\n - \"fix\" topic branch queues fixes applicable only for the\n   current maintenance series.\n\n - \"mainto/1.0\" branch is used to cut releases for the 1.0.X\n   maintenance series from.  This pulls from \"fixo/1.0\" topic\n   branch.\n\n - \"maint\" branch is used to cut releases for the current\n   maintenance series from.  This pulls from both \"fix\" topic\n   branch and \"mainto/*\" topic branches.\n\nWith this setup, I can first merge \"fixo/1.0\" to \"mainto/1.0\",\nand then merge \"mainto/1.0\" and \"fix\" branches to \"maint\", to\ncome up with something like this:\n\n    $ git show-branch fix fixo/* maint mainto/*\n    ! [fix] describe: do not silently ignore indescribable commits\n     ! [fixo/1.0] name-rev: do not omit leading components of ref name.\n      * [maint] GIT 1.1.2 (not really -- just a demonstration)\n       ! [mainto/1.0] GIT 1.0.10 (not really -- just a demonstration)\n    ----\n      *  [maint] GIT 1.1.2 (not really -- just a demonstration)\n    + *  [fix] describe: do not silently ignore indescribable commits\n    + *  [fix^] Add git-describe to .gitignore.\n      *  [maint^] Merge fixes up to GIT 1.0.10 (not really)\n      *+ [mainto/1.0] GIT 1.0.10 (not really -- just a demonstration)\n     +*+ [fixo/1.0] name-rev: do not omit leading components of ref name.\n     +*+ [fixo/1.0^] update-index: work with c-quoted name\n      *+ [mainto/1.0^] GIT 1.0.9\n      *  [maint~2] GIT 1.1.1\n     +*+ [fixo/1.0~2] glossary: explain \"master\" and \"origin\"\n    + *  [maint~3] GIT-VERSION-GEN: detect dirty tree and mark the ver...\n    + *  [maint~4] For release tarballs, include the proper version\n    + *  [maint~5] GIT 1.1.0\n    + *+ [mainto/1.0~2] GIT 1.0.8\n    ++*+ [fixo/1.0~3] mailsplit: allow empty input from stdin\n"},{"id":"14646","messageId":"7vvewnwjzy.fsf@assigned-by-dhcp.cox.net","threadId":"3046","inReplyTo":"7vhd8al3ae.fsf@assigned-by-dhcp.cox.net","subject":"What's cooking in git.git","fromName":"Junio C Hamano","fromEmail":"junkio@cox.net","sentAt":"2006-01-14T03:45:05Z","receivedAt":"2006-01-14T03:45:05Z","isPatch":false,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Junio C Hamano <junkio@cox.net> writes:\n\n>  - Exec git programs without using PATH (Michal Ostrowski).\n>\n>    This has been discussed for the last couple of days on the\n>    list.  I think C side is ready to go in, but I'd like to\n>    decide what to do on the scripting side before merging it\n>    into the master branch.\n\nI've done a minor fix in the Makefile (things were installed in\nthe wrong way around --- $bindir should hold git and gitk and\n$gitexecdir should hold the rest), and ran a brief test after\ninstalling things with gitexecdir set to /usr/lib/git-core/x.\ngitk needs to be taught to use git --exec-path upfront to set\nits own PATH, but other than that, everything looks peachy.\n\nAbout prepending to/appending to/not touching PATH, my current\nthinking is that prepending is the way to go.  The best\nrationale so far I have heard against mucking with PATH was\n\"/home/mo/bin/diff vs /usr/bin/diff\" issue, but I do not find\nthis to be so convincing in practice.  This \"use my diff not\nsystem diff\" is done most often when system diff is not GNU\nenough, but on such a machine git will not be installed under\n/usr/bin anyway (after all git prefers GNU diff).\n\nAnd this becomes truly a non-issue when you point $gitexecdir at\na separate location that holds *only* git stuff.  Everybody will\nfind the matching version of git that way, and when looking for\n\"diff\", it will be found from one of the directories in the rest\nof the PATH, not from $gitexecdir.\n\nSo unless there is an objection, this will graduate into the\n\"master\" branch over the weekend.\n\nWhat happens if gitexecdir != bindir?\n\nAs long as the user invokes \"git foo\" from the command line, not\n\"git-foo\", everything should work fine, and once gitk is\nupdated, binary distributions could ship git with only git and\ngitk in /usr/bin and others in /usr/lib/git-core/tools, or\nsomething like that.\n\nI will _not_ change the default gitexecdir in my Makefile nor\nchange the spec file, so RPM binary from kernel.org will still\nhave everything in /usr/bin for now.\n\nSomebody needs to audit how ready popular Porcelains are, for at\nleast Cogito, stgit, gitweb and qgit and feed patches to\nrespective authors as necessary.  I haven't looked into any of\nthem about this issue.  The rules are:\n\n    (1) you can assume \"git\" is available on PATH;\n\n    (2) you should be safe if you do not use \"git-foo\" form;\n\n    (3) you can run\n\n        PATH=\"$(git --exec-path):$PATH\"; export PATH\n\n        once at the beginning of your program and use \"git-foo\"\n        form.\n\nVolunteers?\n"},{"id":"14726","messageId":"20060116105356.GH28365@pasky.or.cz","threadId":"3046","inReplyTo":"7vhd8al3ae.fsf@assigned-by-dhcp.cox.net","subject":"Re: What's cooking in git.git","fromName":"Petr Baudis","fromEmail":"pasky@suse.cz","sentAt":"2006-01-16T10:53:56Z","receivedAt":"2006-01-16T10:53:56Z","isPatch":false,"sender":{"key":"pasky@ucw.cz","avatar":"https://avatars.githubusercontent.com/u/18439?v=4"},"body":"Dear diary, on Thu, Jan 12, 2006 at 01:04:41AM CET, I got a letter\nwhere Junio C Hamano <junkio@cox.net> said that...\n>  - Disable USE_SYMLINK_HEAD by default (Pavel Roskin).\n> \n>    This has been on hold for almost two months -- forever in git\n>    timescale.\n> \n>    One offender that could have been broken by this change was\n>    fixed recently (gitweb), so this can go in anytime now.  Do\n>    people still care about it?  The argument for the change when\n>    it was proposed was \"then we only have to worry about one\n>    format of .git/HEAD, not two\", and it still is a valid\n>    argument from Porcelain writers' point of view.  To be\n>    honest, since I do not do Porcelain, I am neutral about it.\n\nYes, I'd still be pleased to see this. Cogito should be ready for this\nfor long enough, as far as I'm concerned.\n\n-- \n\t\t\t\tPetr \"Pasky\" Baudis\nStuff: http://pasky.or.cz/\nOf the 3 great composers Mozart tells us what it's like to be human,\nBeethoven tells us what it's like to be Beethoven and Bach tells us\nwhat it's like to be the universe.  -- Douglas Adams\n"}]}