{"thread":{"id":"26259","subject":"[RFC] Support for arbitrary tags in commits","startedAt":"2011-01-11T12:23:32Z","lastAt":"2011-01-11T13:53:08Z","messageCount":4,"participants":["Philipp Marek","Thomas Rast"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"159323","messageId":"74b0628dffbd2bc0adabe5e8b0a10960.squirrel@webmail.hitco.org","threadId":"26259","inReplyTo":null,"subject":"[RFC] Support for arbitrary tags in commits","fromName":"Philipp Marek","fromEmail":"philipp@marek.priv.at","sentAt":"2011-01-11T12:23:32Z","receivedAt":"2011-01-11T12:23:32Z","isPatch":false,"sender":{"key":"philipp@marek.priv.at","avatar":null},"body":"Hello everybody,\n\nI'd like to attach some additional data to commits.\n\nThe best way I've found (so far) is to put an additional header line in the\ncommit header that references an additional blob.\n\n\nCurrently I've found \"tree\", \"parent\", \"author\", \"committer\", and so on; git\naccepts additional lines before the \"\\n\\n\" separator (and the commit message),\nso that's where I'd like to put the line.\n\n\"git fsck\" allows that, too; only \"rebase -i\" and similar don't know about them\n(yet), and discard them.\n\n\nNow I want to ask:\n * Is this acceptable?\n * If yes, I'd try to provide a patch to make \"amend\" and similar keep the\n   header-lines (by collecting them in order - obviously with editing commits\n   a header could exist multiple times).\n * Perhaps there should be an option for \"git-commit\" that inserts arbitrary\n   header lines?\n * Is there some documentation for the commit-file format so that this can be\n   written down?\n * How should namespace-issues be handled?\n   I can imagine that multiple people want to put data there, how should the\n   tag be allocated?\n   This could be solved by the previous point, ie. there's some page that\n   header tags should be registered at to avoid collisions.\n\n\nThank you for all feedback.\n\n\nRegards,\n\nPhil\n\n\n-- \nVersioning your /etc, /home or even your whole installation?\n             Try fsvs (fsvs.tigris.org)!\n"},{"id":"159326","messageId":"201101111340.50508.trast@student.ethz.ch","threadId":"26259","inReplyTo":"74b0628dffbd2bc0adabe5e8b0a10960.squirrel@webmail.hitco.org","subject":"Re: [RFC] Support for arbitrary tags in commits","fromName":"Thomas Rast","fromEmail":"trast@student.ethz.ch","sentAt":"2011-01-11T12:40:50Z","receivedAt":"2011-01-11T12:40:50Z","isPatch":false,"sender":{"key":"tr@thomasrast.ch","avatar":"https://avatars.githubusercontent.com/u/153510?v=4"},"body":"Philipp Marek wrote:\n> \n> The best way I've found (so far) is to put an additional header line in the\n> commit header that references an additional blob.\n\nThis comes up every few months.  The last large discussion about this\nthat we had IIRC was\n\n  http://thread.gmane.org/gmane.comp.version-control.git/138848\n\nCan you please look through that thread and state in what way your\nuse-case invalidates the previous reasoning?\n\nIn particular, in what way do notes (as in git-notes(1)) fail to solve\nyour problem?\n\n-- \nThomas Rast\ntrast@{inf,student}.ethz.ch\n"},{"id":"159335","messageId":"edec2e42a531b00ffd6a2689095460d5.squirrel@webmail.hitco.org","threadId":"26259","inReplyTo":"201101111340.50508.trast@student.ethz.ch","subject":"Re: [RFC] Support for arbitrary tags in commits","fromName":"Philipp Marek","fromEmail":"philipp@marek.priv.at","sentAt":"2011-01-11T13:48:59Z","receivedAt":"2011-01-11T13:48:59Z","isPatch":false,"sender":{"key":"philipp@marek.priv.at","avatar":null},"body":"Hello Thomas,\n\nthank you very much for the quick answer.\n\n>> The best way I've found (so far) is to put an additional header line in the\n>> commit header that references an additional blob.\n>\n> This comes up every few months.  The last large discussion about this\n> that we had IIRC was\n>\n>   http://thread.gmane.org/gmane.comp.version-control.git/138848\n>\n> Can you please look through that thread and state in what way your\n> use-case invalidates the previous reasoning?\nThank you, just reading it.\n\n> In particular, in what way do notes (as in git-notes(1)) fail to solve\n> your problem?\nWell, my biggest concerns are that users might trash them, and that they are\nvoided by amend, rebase etc. (which I tried to address by collecting header\nlines).\n\n\nWell, from this thread I suppose the \"best\" (or at least easiest) way forward\nis to use notes, if necessary.\n\n\nThank you!\n\n\nRegards,\n\nPhil\n\n\n-- \nVersioning your /etc, /home or even your whole installation?\n             Try fsvs (fsvs.tigris.org)!\n"},{"id":"159336","messageId":"201101111453.08178.trast@student.ethz.ch","threadId":"26259","inReplyTo":"edec2e42a531b00ffd6a2689095460d5.squirrel@webmail.hitco.org","subject":"Re: [RFC] Support for arbitrary tags in commits","fromName":"Thomas Rast","fromEmail":"trast@student.ethz.ch","sentAt":"2011-01-11T13:53:08Z","receivedAt":"2011-01-11T13:53:08Z","isPatch":false,"sender":{"key":"tr@thomasrast.ch","avatar":"https://avatars.githubusercontent.com/u/153510?v=4"},"body":"Philipp Marek wrote:\n> > In particular, in what way do notes (as in git-notes(1)) fail to solve\n> > your problem?\n> Well, my biggest concerns are that users might trash them, and that they are\n> voided by amend, rebase etc. (which I tried to address by collecting header\n> lines).\n\nThey aren't lost any more (since 1.7.1) if you properly configure\nnotes.rewriteRef.\n\n-- \nThomas Rast\ntrast@{inf,student}.ethz.ch\n"}]}