CINXE.COM

<?xml version="1.0" encoding="UTF-8" standalone="yes"?> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>LKML: Linus Torvalds: Re: more git updates..</title><link href="/css/message.css" rel="stylesheet" type="text/css" /><link href="/css/wrap.css" rel="alternate stylesheet" type="text/css" title="wrap" /><link href="/css/nowrap.css" rel="stylesheet" type="text/css" title="nowrap" /><link href="/favicon.ico" rel="shortcut icon" /><script src="/js/simple-calendar.js" type="text/javascript"></script><script src="/js/styleswitcher.js" type="text/javascript"></script><link rel="alternate" type="application/rss+xml" title="lkml.org : last 100 messages" href="/rss.php" /><link rel="alternate" type="application/rss+xml" title="lkml.org : last messages by Linus Torvalds" href="/groupie.php?aid=1" /><script> var _paq = window._paq = window._paq || []; /* tracker methods like "setCustomDimension" should be called before "trackPageView" */ _paq.push(["setDoNotTrack", true]); _paq.push(["disableCookies"]); _paq.push(['trackPageView']); _paq.push(['enableLinkTracking']); (function() { var u="//m.lkml.org/"; _paq.push(['setTrackerUrl', u+'matomo.php']); _paq.push(['setSiteId', '1']); var d=document, g=d.createElement('script'), s=d.getElementsByTagName('script')[0]; g.async=true; g.src=u+'matomo.js'; s.parentNode.insertBefore(g,s); })(); </script></head><body onload="es.jasper.simpleCalendar.init();" itemscope="itemscope" itemtype="http://schema.org/BlogPosting"><table border="0" cellpadding="0" cellspacing="0"><tr><td width="180" align="center"><a href="/"><img style="border:0;width:135px;height:32px" src="/images/toprowlk.gif" alt="lkml.org" /></a></td><td width="32">聽</td><td class="nb"><div><a class="nb" href="/lkml"> [lkml]</a> 聽 <a class="nb" href="/lkml/2005"> [2005]</a> 聽 <a class="nb" href="/lkml/2005/4"> [Apr]</a> 聽 <a class="nb" href="/lkml/2005/4/10"> [10]</a> 聽 <a class="nb" href="/lkml/last100"> [last100]</a> 聽 <a href="/rss.php"><img src="/images/rss-or.gif" border="0" alt="RSS Feed" /></a></div><div>Views: <a href="#" class="nowrap" onclick="setActiveStyleSheet('wrap');return false;">[wrap]</a><a href="#" class="wrap" onclick="setActiveStyleSheet('nowrap');return false;">[no wrap]</a> 聽 <a class="nb" href="/lkml/mheaders/2005/4/10/70" onclick="this.href='/lkml/headers'+'/2005/4/10/70';">[headers]</a>聽 <a href="/lkml/bounce/2005/4/10/70">[forward]</a>聽 </div></td><td width="32">聽</td></tr><tr><td valign="top"><div class="es-jasper-simpleCalendar" baseurl="/lkml/"></div><div class="threadlist">Messages in this thread</div><ul class="threadlist"><li class="root"><a href="/lkml/2005/4/9/103">First message in thread</a></li><li><a href="/lkml/2005/4/9/109">Linus Torvalds</a><ul><li><a href="/lkml/2005/4/9/110">Linus Torvalds</a><ul><li><a href="/lkml/2005/4/9/136">Linus Torvalds</a><ul><li><a href="/lkml/2005/4/9/155">Petr Baudis</a><ul><li><a href="/lkml/2005/4/10/72">Petr Baudis</a></li></ul></li><li><a href="/lkml/2005/4/10/25">Christopher Li</a><ul><li><a href="/lkml/2005/4/10/41">Ralph Corderoy</a></li><li><a href="/lkml/2005/4/10/125">Paul Jackson</a></li><li><a href="/lkml/2005/4/11/103">(H. Peter Anvin)</a></li></ul></li><li><a href="/lkml/2005/4/11/74">Ingo Molnar</a><ul><li><a href="/lkml/2005/4/11/110">Paul Jackson</a></li></ul></li><li><a href="/lkml/2005/4/12/32">David Eger</a><ul><li><a href="/lkml/2005/4/12/67">Petr Baudis</a></li></ul></li></ul></li></ul></li><li><a href="/lkml/2005/4/9/150">Paul Jackson</a><ul><li><a href="/lkml/2005/4/9/152">Paul Jackson</a></li></ul></li><li><a href="/lkml/2005/4/9/151">Paul Jackson</a></li><li><a href="/lkml/2005/4/10/8">Junio C Hamano</a><ul><li><a href="/lkml/2005/4/10/13">Christopher Li</a><ul><li><a href="/lkml/2005/4/10/18">Junio C Hamano</a><ul><li><a href="/lkml/2005/4/10/23">Petr Baudis</a></li><li><a href="/lkml/2005/4/10/27">Christopher Li</a></li></ul></li><li><a href="/lkml/2005/4/10/21">Wichert Akkerman</a></li><li><a href="/lkml/2005/4/10/22">Petr Baudis</a><ul><li><a href="/lkml/2005/4/10/28">Christopher Li</a></li></ul></li></ul></li><li><a href="/lkml/2005/4/10/37">Rutger Nijlunsing</a></li><li class="origin"><a href="/lkml/2005/4/10/76">Linus Torvalds</a><ul><li><a href="/lkml/2005/4/10/76">Rutger Nijlunsing</a></li><li><a href="/lkml/2005/4/10/118">Paul Jackson</a><ul><li><a href="/lkml/2005/4/10/142">Linus Torvalds</a></li></ul></li></ul></li></ul></li><li><a href="/lkml/2005/4/10/44"> tony.luck@intel ...</a><ul><li><a href="/lkml/2005/4/10/71">Linus Torvalds</a><ul><li><a href="/lkml/2005/4/12/306">Helge Hafting</a></li></ul></li><li><a href="/lkml/2005/4/10/103">Paul Jackson</a><ul><li><a href="/lkml/2005/4/10/165">Bernd Eckenfels</a><ul><li><a href="/lkml/2005/4/11/50">Anton Altaparmakov</a></li></ul></li></ul></li></ul></li></ul></li></ul></td><td width="32" rowspan="2" class="c" valign="top"><img src="/images/icornerl.gif" width="32" height="32" alt="/" /></td><td class="c" rowspan="2" valign="top" style="padding-top: 1em"><table><tr><td><table><tr><td class="lp">Date</td><td class="rp" itemprop="datePublished">Sun, 10 Apr 2005 08:44:56 -0700 (PDT)</td></tr><tr><td class="lp">From</td><td class="rp" itemprop="author">Linus Torvalds <></td></tr><tr><td class="lp">Subject</td><td class="rp" itemprop="name">Re: more git updates..</td></tr></table></td><td></td></tr></table><pre itemprop="articleBody"><br /><br />On Sun, 10 Apr 2005, Junio C Hamano wrote:<br />> <br />> But I am wondering what your plans are to handle renames---or<br />> does git already represent them?<br /><br />You can represent renames on top of git - git itself really doesn't care. <br />In many ways you can just see git as a filesystem - it's content-<br />addressable, and it has a notion of versioning, but I really really<br />designed it coming at the problem from the viewpoint of a _filesystem_<br />person (hey, kernels is what I do), and I actually have absolutely _zero_<br />interest in creating a traditional SCM system.<br /><br />So to take renaming a file as an example - why do you actually want to <br />track renames? In traditional SCM's, you do it for two reasons:<br /><br /> - space efficiency. Most SCM's are based on describing changes to a file, <br /> and compress the data by doing revisions on the same file. In order to <br /> continue that process past a rename, such an SCM _has_ to track <br /> renames, or lose the delta-based approach.<br /><br /> The most trivial example of this is "diff", ie a rename ends up <br /> generating a _huge_ diff unless you track the rename explicitly.<br /><br /> GIT doesn't care. There is _zero_ space efficiency in trying to track <br /> renames. In fact, it would add overhead to the system, not lessen it. <br /> That's because GIT fundamentally doesn't do the "delta-within-a-file" <br /> model.<br /><br /> - annotate/blame. This is a valid concern, but the fact is, I never use <br /> it. It may be a deficiency of mine, but I simply don't do the per-line <br /> thing when I debug or try to find who was responsible. I do "blame" on <br /> a much bigger-picture level, and I personally believe (pretty strongly) <br /> that per-line annotations are not actually a good thing - they come not <br /> because people _want_ to do things at that low level, but because <br /> historically, you didn't _have_ the bigger-picture thing.<br /><br /> In other words, pretty much every SCM out there is based on SCCS <br /> "mentally", even if not in any other model. That's why people think <br /> per-line blame is important - you have that mental model. <br /><br />So consider me deficient, or consider me radical. It boils down to the <br />same thing. Renames don't matter. <br /><br />That said, if somebody wants to create a _real_ SCM (rather than my notion<br />of a pure content tracker) on top of GIT, you probably could fairly easily<br />do so by imposing a few limitations on a higher level. For example, most<br />SCM's that track renames require that the user _tell_ them about the<br />renames: you do a "bk mv" or a "svn rename" or something.<br /><br />If you want to do the same on top of GIT, then you should think of GIT as<br />what it is: GIT just tracks contents. It's a filesystem - although a<br />fairly strange one. How would you track renames on top of that? Easy: add<br />your own fields to the GIT revision messages: GIT enforces the header, but<br />you can add anything you want to the "free-form" part that follows it. <br /><br />Same goes for any other information where you care about what happens <br />"within" a file. GIT simply doesn't track it. You can build things on top <br />of GIT if you want to, though. They may not be as efficient as they would <br />be if they were built _into_ GIT, but on the other hand GIT does a lot of <br />other things a hell of a lot faster thanks to it's design.<br /><br />So whether you agree with the things that _I_ consider important probably<br />depends on how you work. The real downside of GIT may be that _my_ way of <br />doing things is quite possibly very rare.<br /><br />But it clearly is the only right way. The fact that everybody else does it <br />some other way only means that they are wrong.<br /><br /> Linus<br />-<br />To unsubscribe from this list: send the line "unsubscribe linux-kernel" in<br />the body of a message to majordomo@vger.kernel.org<br />More majordomo info at <a href="http://vger.kernel.org/majordomo-info.html">http://vger.kernel.org/majordomo-info.html</a><br />Please read the FAQ at <a href="http://www.tux.org/lkml/">http://www.tux.org/lkml/</a><br /><br /></pre></td><td width="32" rowspan="2" class="c" valign="top"><img src="/images/icornerr.gif" width="32" height="32" alt="\" /></td></tr><tr><td align="right" valign="bottom"> 聽 </td></tr><tr><td align="right" valign="bottom">聽</td><td class="c" valign="bottom" style="padding-bottom: 0px"><img src="/images/bcornerl.gif" width="32" height="32" alt="\" /></td><td class="c">聽</td><td class="c" valign="bottom" style="padding-bottom: 0px"><img src="/images/bcornerr.gif" width="32" height="32" alt="/" /></td></tr><tr><td align="right" valign="top" colspan="2"> 聽 </td><td class="lm">Last update: 2009-11-18 23:46 聽聽 [from the cache]<br />漏2003-2020 <a href="http://blog.jasper.es/"><span itemprop="editor">Jasper Spaans</span></a>|hosted at <a href="https://www.digitalocean.com/?refcode=9a8e99d24cf9">Digital Ocean</a> and my Meterkast|<a href="http://blog.jasper.es/categories.html#lkml-ref">Read the blog</a></td><td>聽</td></tr></table><script language="javascript" src="/js/styleswitcher.js" type="text/javascript"></script></body></html>