gitattributes.html

   1 <?xml version="1.0" encoding="UTF-8"?>
   2 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN"
   3     "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd">
   4 <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en">
   5 <head>
   6 <meta http-equiv="Content-Type" content="application/xhtml+xml; charset=UTF-8" />
   7 <meta name="generator" content="AsciiDoc 10.2.0" />
   8 <title>gitattributes(5)</title>
   9 <style type="text/css">
  10 /* Shared CSS for AsciiDoc xhtml11 and html5 backends */
  11
  12 /* Default font. */
  13 body {
  14   font-family: Georgia,serif;
  15 }
  16
  17 /* Title font. */
  18 h1, h2, h3, h4, h5, h6,
  19 div.title, caption.title,
  20 thead, p.table.header,
  21 #toctitle,
  22 #author, #revnumber, #revdate, #revremark,
  23 #footer {
  24   font-family: Arial,Helvetica,sans-serif;
  25 }
  26
  27 body {
  28   margin: 1em 5% 1em 5%;
  29 }
  30
  31 a {
  32   color: blue;
  33   text-decoration: underline;
  34 }
  35 a:visited {
  36   color: fuchsia;
  37 }
  38
  39 em {
  40   font-style: italic;
  41   color: navy;
  42 }
  43
  44 strong {
  45   font-weight: bold;
  46   color: #083194;
  47 }
  48
  49 h1, h2, h3, h4, h5, h6 {
  50   color: #527bbd;
  51   margin-top: 1.2em;
  52   margin-bottom: 0.5em;
  53   line-height: 1.3;
  54 }
  55
  56 h1, h2, h3 {
  57   border-bottom: 2px solid silver;
  58 }
  59 h2 {
  60   padding-top: 0.5em;
  61 }
  62 h3 {
  63   float: left;
  64 }
  65 h3 + * {
  66   clear: left;
  67 }
  68 h5 {
  69   font-size: 1.0em;
  70 }
  71
  72 div.sectionbody {
  73   margin-left: 0;
  74 }
  75
  76 hr {
  77   border: 1px solid silver;
  78 }
  79
  80 p {
  81   margin-top: 0.5em;
  82   margin-bottom: 0.5em;
  83 }
  84
  85 ul, ol, li > p {
  86   margin-top: 0;
  87 }
  88 ul > li     { color: #aaa; }
  89 ul > li > * { color: black; }
  90
  91 .monospaced, code, pre {
  92   font-family: "Courier New", Courier, monospace;
  93   font-size: inherit;
  94   color: navy;
  95   padding: 0;
  96   margin: 0;
  97 }
  98 pre {
  99   white-space: pre-wrap;
 100 }
 101
 102 #author {
 103   color: #527bbd;
 104   font-weight: bold;
 105   font-size: 1.1em;
 106 }
 107 #email {
 108 }
 109 #revnumber, #revdate, #revremark {
 110 }
 111
 112 #footer {
 113   font-size: small;
 114   border-top: 2px solid silver;
 115   padding-top: 0.5em;
 116   margin-top: 4.0em;
 117 }
 118 #footer-text {
 119   float: left;
 120   padding-bottom: 0.5em;
 121 }
 122 #footer-badges {
 123   float: right;
 124   padding-bottom: 0.5em;
 125 }
 126
 127 #preamble {
 128   margin-top: 1.5em;
 129   margin-bottom: 1.5em;
 130 }
 131 div.imageblock, div.exampleblock, div.verseblock,
 132 div.quoteblock, div.literalblock, div.listingblock, div.sidebarblock,
 133 div.admonitionblock {
 134   margin-top: 1.0em;
 135   margin-bottom: 1.5em;
 136 }
 137 div.admonitionblock {
 138   margin-top: 2.0em;
 139   margin-bottom: 2.0em;
 140   margin-right: 10%;
 141   color: #606060;
 142 }
 143
 144 div.content { /* Block element content. */
 145   padding: 0;
 146 }
 147
 148 /* Block element titles. */
 149 div.title, caption.title {
 150   color: #527bbd;
 151   font-weight: bold;
 152   text-align: left;
 153   margin-top: 1.0em;
 154   margin-bottom: 0.5em;
 155 }
 156 div.title + * {
 157   margin-top: 0;
 158 }
 159
 160 td div.title:first-child {
 161   margin-top: 0.0em;
 162 }
 163 div.content div.title:first-child {
 164   margin-top: 0.0em;
 165 }
 166 div.content + div.title {
 167   margin-top: 0.0em;
 168 }
 169
 170 div.sidebarblock > div.content {
 171   background: #ffffee;
 172   border: 1px solid #dddddd;
 173   border-left: 4px solid #f0f0f0;
 174   padding: 0.5em;
 175 }
 176
 177 div.listingblock > div.content {
 178   border: 1px solid #dddddd;
 179   border-left: 5px solid #f0f0f0;
 180   background: #f8f8f8;
 181   padding: 0.5em;
 182 }
 183
 184 div.quoteblock, div.verseblock {
 185   padding-left: 1.0em;
 186   margin-left: 1.0em;
 187   margin-right: 10%;
 188   border-left: 5px solid #f0f0f0;
 189   color: #888;
 190 }
 191
 192 div.quoteblock > div.attribution {
 193   padding-top: 0.5em;
 194   text-align: right;
 195 }
 196
 197 div.verseblock > pre.content {
 198   font-family: inherit;
 199   font-size: inherit;
 200 }
 201 div.verseblock > div.attribution {
 202   padding-top: 0.75em;
 203   text-align: left;
 204 }
 205 /* DEPRECATED: Pre version 8.2.7 verse style literal block. */
 206 div.verseblock + div.attribution {
 207   text-align: left;
 208 }
 209
 210 div.admonitionblock .icon {
 211   vertical-align: top;
 212   font-size: 1.1em;
 213   font-weight: bold;
 214   text-decoration: underline;
 215   color: #527bbd;
 216   padding-right: 0.5em;
 217 }
 218 div.admonitionblock td.content {
 219   padding-left: 0.5em;
 220   border-left: 3px solid #dddddd;
 221 }
 222
 223 div.exampleblock > div.content {
 224   border-left: 3px solid #dddddd;
 225   padding-left: 0.5em;
 226 }
 227
 228 div.imageblock div.content { padding-left: 0; }
 229 span.image img { border-style: none; vertical-align: text-bottom; }
 230 a.image:visited { color: white; }
 231
 232 dl {
 233   margin-top: 0.8em;
 234   margin-bottom: 0.8em;
 235 }
 236 dt {
 237   margin-top: 0.5em;
 238   margin-bottom: 0;
 239   font-style: normal;
 240   color: navy;
 241 }
 242 dd > *:first-child {
 243   margin-top: 0.1em;
 244 }
 245
 246 ul, ol {
 247     list-style-position: outside;
 248 }
 249 ol.arabic {
 250   list-style-type: decimal;
 251 }
 252 ol.loweralpha {
 253   list-style-type: lower-alpha;
 254 }
 255 ol.upperalpha {
 256   list-style-type: upper-alpha;
 257 }
 258 ol.lowerroman {
 259   list-style-type: lower-roman;
 260 }
 261 ol.upperroman {
 262   list-style-type: upper-roman;
 263 }
 264
 265 div.compact ul, div.compact ol,
 266 div.compact p, div.compact p,
 267 div.compact div, div.compact div {
 268   margin-top: 0.1em;
 269   margin-bottom: 0.1em;
 270 }
 271
 272 tfoot {
 273   font-weight: bold;
 274 }
 275 td > div.verse {
 276   white-space: pre;
 277 }
 278
 279 div.hdlist {
 280   margin-top: 0.8em;
 281   margin-bottom: 0.8em;
 282 }
 283 div.hdlist tr {
 284   padding-bottom: 15px;
 285 }
 286 dt.hdlist1.strong, td.hdlist1.strong {
 287   font-weight: bold;
 288 }
 289 td.hdlist1 {
 290   vertical-align: top;
 291   font-style: normal;
 292   padding-right: 0.8em;
 293   color: navy;
 294 }
 295 td.hdlist2 {
 296   vertical-align: top;
 297 }
 298 div.hdlist.compact tr {
 299   margin: 0;
 300   padding-bottom: 0;
 301 }
 302
 303 .comment {
 304   background: yellow;
 305 }
 306
 307 .footnote, .footnoteref {
 308   font-size: 0.8em;
 309 }
 310
 311 span.footnote, span.footnoteref {
 312   vertical-align: super;
 313 }
 314
 315 #footnotes {
 316   margin: 20px 0 20px 0;
 317   padding: 7px 0 0 0;
 318 }
 319
 320 #footnotes div.footnote {
 321   margin: 0 0 5px 0;
 322 }
 323
 324 #footnotes hr {
 325   border: none;
 326   border-top: 1px solid silver;
 327   height: 1px;
 328   text-align: left;
 329   margin-left: 0;
 330   width: 20%;
 331   min-width: 100px;
 332 }
 333
 334 div.colist td {
 335   padding-right: 0.5em;
 336   padding-bottom: 0.3em;
 337   vertical-align: top;
 338 }
 339 div.colist td img {
 340   margin-top: 0.3em;
 341 }
 342
 343 @media print {
 344   #footer-badges { display: none; }
 345 }
 346
 347 #toc {
 348   margin-bottom: 2.5em;
 349 }
 350
 351 #toctitle {
 352   color: #527bbd;
 353   font-size: 1.1em;
 354   font-weight: bold;
 355   margin-top: 1.0em;
 356   margin-bottom: 0.1em;
 357 }
 358
 359 div.toclevel0, div.toclevel1, div.toclevel2, div.toclevel3, div.toclevel4 {
 360   margin-top: 0;
 361   margin-bottom: 0;
 362 }
 363 div.toclevel2 {
 364   margin-left: 2em;
 365   font-size: 0.9em;
 366 }
 367 div.toclevel3 {
 368   margin-left: 4em;
 369   font-size: 0.9em;
 370 }
 371 div.toclevel4 {
 372   margin-left: 6em;
 373   font-size: 0.9em;
 374 }
 375
 376 span.aqua { color: aqua; }
 377 span.black { color: black; }
 378 span.blue { color: blue; }
 379 span.fuchsia { color: fuchsia; }
 380 span.gray { color: gray; }
 381 span.green { color: green; }
 382 span.lime { color: lime; }
 383 span.maroon { color: maroon; }
 384 span.navy { color: navy; }
 385 span.olive { color: olive; }
 386 span.purple { color: purple; }
 387 span.red { color: red; }
 388 span.silver { color: silver; }
 389 span.teal { color: teal; }
 390 span.white { color: white; }
 391 span.yellow { color: yellow; }
 392
 393 span.aqua-background { background: aqua; }
 394 span.black-background { background: black; }
 395 span.blue-background { background: blue; }
 396 span.fuchsia-background { background: fuchsia; }
 397 span.gray-background { background: gray; }
 398 span.green-background { background: green; }
 399 span.lime-background { background: lime; }
 400 span.maroon-background { background: maroon; }
 401 span.navy-background { background: navy; }
 402 span.olive-background { background: olive; }
 403 span.purple-background { background: purple; }
 404 span.red-background { background: red; }
 405 span.silver-background { background: silver; }
 406 span.teal-background { background: teal; }
 407 span.white-background { background: white; }
 408 span.yellow-background { background: yellow; }
 409
 410 span.big { font-size: 2em; }
 411 span.small { font-size: 0.6em; }
 412
 413 span.underline { text-decoration: underline; }
 414 span.overline { text-decoration: overline; }
 415 span.line-through { text-decoration: line-through; }
 416
 417 div.unbreakable { page-break-inside: avoid; }
 418
 419
 420 /*
 421  * xhtml11 specific
 422  *
 423  * */
 424
 425 div.tableblock {
 426   margin-top: 1.0em;
 427   margin-bottom: 1.5em;
 428 }
 429 div.tableblock > table {
 430   border: 3px solid #527bbd;
 431 }
 432 thead, p.table.header {
 433   font-weight: bold;
 434   color: #527bbd;
 435 }
 436 p.table {
 437   margin-top: 0;
 438 }
 439 /* Because the table frame attribute is overridden by CSS in most browsers. */
 440 div.tableblock > table[frame="void"] {
 441   border-style: none;
 442 }
 443 div.tableblock > table[frame="hsides"] {
 444   border-left-style: none;
 445   border-right-style: none;
 446 }
 447 div.tableblock > table[frame="vsides"] {
 448   border-top-style: none;
 449   border-bottom-style: none;
 450 }
 451
 452
 453 /*
 454  * html5 specific
 455  *
 456  * */
 457
 458 table.tableblock {
 459   margin-top: 1.0em;
 460   margin-bottom: 1.5em;
 461 }
 462 thead, p.tableblock.header {
 463   font-weight: bold;
 464   color: #527bbd;
 465 }
 466 p.tableblock {
 467   margin-top: 0;
 468 }
 469 table.tableblock {
 470   border-width: 3px;
 471   border-spacing: 0px;
 472   border-style: solid;
 473   border-color: #527bbd;
 474   border-collapse: collapse;
 475 }
 476 th.tableblock, td.tableblock {
 477   border-width: 1px;
 478   padding: 4px;
 479   border-style: solid;
 480   border-color: #527bbd;
 481 }
 482
 483 table.tableblock.frame-topbot {
 484   border-left-style: hidden;
 485   border-right-style: hidden;
 486 }
 487 table.tableblock.frame-sides {
 488   border-top-style: hidden;
 489   border-bottom-style: hidden;
 490 }
 491 table.tableblock.frame-none {
 492   border-style: hidden;
 493 }
 494
 495 th.tableblock.halign-left, td.tableblock.halign-left {
 496   text-align: left;
 497 }
 498 th.tableblock.halign-center, td.tableblock.halign-center {
 499   text-align: center;
 500 }
 501 th.tableblock.halign-right, td.tableblock.halign-right {
 502   text-align: right;
 503 }
 504
 505 th.tableblock.valign-top, td.tableblock.valign-top {
 506   vertical-align: top;
 507 }
 508 th.tableblock.valign-middle, td.tableblock.valign-middle {
 509   vertical-align: middle;
 510 }
 511 th.tableblock.valign-bottom, td.tableblock.valign-bottom {
 512   vertical-align: bottom;
 513 }
 514
 515
 516 /*
 517  * manpage specific
 518  *
 519  * */
 520
 521 body.manpage h1 {
 522   padding-top: 0.5em;
 523   padding-bottom: 0.5em;
 524   border-top: 2px solid silver;
 525   border-bottom: 2px solid silver;
 526 }
 527 body.manpage h2 {
 528   border-style: none;
 529 }
 530 body.manpage div.sectionbody {
 531   margin-left: 3em;
 532 }
 533
 534 @media print {
 535   body.manpage div#toc { display: none; }
 536 }
 537
 538
 539 </style>
 540 <script type="text/javascript">
 541 /*<![CDATA[*/
 542 var asciidoc = {  // Namespace.
 543
 544 /////////////////////////////////////////////////////////////////////
 545 // Table Of Contents generator
 546 /////////////////////////////////////////////////////////////////////
 547
 548 /* Author: Mihai Bazon, September 2002
 549  * http://students.infoiasi.ro/~mishoo
 550  *
 551  * Table Of Content generator
 552  * Version: 0.4
 553  *
 554  * Feel free to use this script under the terms of the GNU General Public
 555  * License, as long as you do not remove or alter this notice.
 556  */
 557
 558  /* modified by Troy D. Hanson, September 2006. License: GPL */
 559  /* modified by Stuart Rackham, 2006, 2009. License: GPL */
 560
 561 // toclevels = 1..4.
 562 toc: function (toclevels) {
 563
 564   function getText(el) {
 565     var text = "";
 566     for (var i = el.firstChild; i != null; i = i.nextSibling) {
 567       if (i.nodeType == 3 /* Node.TEXT_NODE */) // IE doesn't speak constants.
 568         text += i.data;
 569       else if (i.firstChild != null)
 570         text += getText(i);
 571     }
 572     return text;
 573   }
 574
 575   function TocEntry(el, text, toclevel) {
 576     this.element = el;
 577     this.text = text;
 578     this.toclevel = toclevel;
 579   }
 580
 581   function tocEntries(el, toclevels) {
 582     var result = new Array;
 583     var re = new RegExp('[hH]([1-'+(toclevels+1)+'])');
 584     // Function that scans the DOM tree for header elements (the DOM2
 585     // nodeIterator API would be a better technique but not supported by all
 586     // browsers).
 587     var iterate = function (el) {
 588       for (var i = el.firstChild; i != null; i = i.nextSibling) {
 589         if (i.nodeType == 1 /* Node.ELEMENT_NODE */) {
 590           var mo = re.exec(i.tagName);
 591           if (mo && (i.getAttribute("class") || i.getAttribute("className")) != "float") {
 592             result[result.length] = new TocEntry(i, getText(i), mo[1]-1);
 593           }
 594           iterate(i);
 595         }
 596       }
 597     }
 598     iterate(el);
 599     return result;
 600   }
 601
 602   var toc = document.getElementById("toc");
 603   if (!toc) {
 604     return;
 605   }
 606
 607   // Delete existing TOC entries in case we're reloading the TOC.
 608   var tocEntriesToRemove = [];
 609   var i;
 610   for (i = 0; i < toc.childNodes.length; i++) {
 611     var entry = toc.childNodes[i];
 612     if (entry.nodeName.toLowerCase() == 'div'
 613      && entry.getAttribute("class")
 614      && entry.getAttribute("class").match(/^toclevel/))
 615       tocEntriesToRemove.push(entry);
 616   }
 617   for (i = 0; i < tocEntriesToRemove.length; i++) {
 618     toc.removeChild(tocEntriesToRemove[i]);
 619   }
 620
 621   // Rebuild TOC entries.
 622   var entries = tocEntries(document.getElementById("content"), toclevels);
 623   for (var i = 0; i < entries.length; ++i) {
 624     var entry = entries[i];
 625     if (entry.element.id == "")
 626       entry.element.id = "_toc_" + i;
 627     var a = document.createElement("a");
 628     a.href = "#" + entry.element.id;
 629     a.appendChild(document.createTextNode(entry.text));
 630     var div = document.createElement("div");
 631     div.appendChild(a);
 632     div.className = "toclevel" + entry.toclevel;
 633     toc.appendChild(div);
 634   }
 635   if (entries.length == 0)
 636     toc.parentNode.removeChild(toc);
 637 },
 638
 639
 640 /////////////////////////////////////////////////////////////////////
 641 // Footnotes generator
 642 /////////////////////////////////////////////////////////////////////
 643
 644 /* Based on footnote generation code from:
 645  * http://www.brandspankingnew.net/archive/2005/07/format_footnote.html
 646  */
 647
 648 footnotes: function () {
 649   // Delete existing footnote entries in case we're reloading the footnodes.
 650   var i;
 651   var noteholder = document.getElementById("footnotes");
 652   if (!noteholder) {
 653     return;
 654   }
 655   var entriesToRemove = [];
 656   for (i = 0; i < noteholder.childNodes.length; i++) {
 657     var entry = noteholder.childNodes[i];
 658     if (entry.nodeName.toLowerCase() == 'div' && entry.getAttribute("class") == "footnote")
 659       entriesToRemove.push(entry);
 660   }
 661   for (i = 0; i < entriesToRemove.length; i++) {
 662     noteholder.removeChild(entriesToRemove[i]);
 663   }
 664
 665   // Rebuild footnote entries.
 666   var cont = document.getElementById("content");
 667   var spans = cont.getElementsByTagName("span");
 668   var refs = {};
 669   var n = 0;
 670   for (i=0; i<spans.length; i++) {
 671     if (spans[i].className == "footnote") {
 672       n++;
 673       var note = spans[i].getAttribute("data-note");
 674       if (!note) {
 675         // Use [\s\S] in place of . so multi-line matches work.
 676         // Because JavaScript has no s (dotall) regex flag.
 677         note = spans[i].innerHTML.match(/\s*\[([\s\S]*)]\s*/)[1];
 678         spans[i].innerHTML =
 679           "[<a id='_footnoteref_" + n + "' href='#_footnote_" + n +
 680           "' title='View footnote' class='footnote'>" + n + "</a>]";
 681         spans[i].setAttribute("data-note", note);
 682       }
 683       noteholder.innerHTML +=
 684         "<div class='footnote' id='_footnote_" + n + "'>" +
 685         "<a href='#_footnoteref_" + n + "' title='Return to text'>" +
 686         n + "</a>. " + note + "</div>";
 687       var id =spans[i].getAttribute("id");
 688       if (id != null) refs["#"+id] = n;
 689     }
 690   }
 691   if (n == 0)
 692     noteholder.parentNode.removeChild(noteholder);
 693   else {
 694     // Process footnoterefs.
 695     for (i=0; i<spans.length; i++) {
 696       if (spans[i].className == "footnoteref") {
 697         var href = spans[i].getElementsByTagName("a")[0].getAttribute("href");
 698         href = href.match(/#.*/)[0];  // Because IE return full URL.
 699         n = refs[href];
 700         spans[i].innerHTML =
 701           "[<a href='#_footnote_" + n +
 702           "' title='View footnote' class='footnote'>" + n + "</a>]";
 703       }
 704     }
 705   }
 706 },
 707
 708 install: function(toclevels) {
 709   var timerId;
 710
 711   function reinstall() {
 712     asciidoc.footnotes();
 713     if (toclevels) {
 714       asciidoc.toc(toclevels);
 715     }
 716   }
 717
 718   function reinstallAndRemoveTimer() {
 719     clearInterval(timerId);
 720     reinstall();
 721   }
 722
 723   timerId = setInterval(reinstall, 500);
 724   if (document.addEventListener)
 725     document.addEventListener("DOMContentLoaded", reinstallAndRemoveTimer, false);
 726   else
 727     window.onload = reinstallAndRemoveTimer;
 728 }
 729
 730 }
 731 asciidoc.install();
 732 /*]]>*/
 733 </script>
 734 </head>
 735 <body class="manpage">
 736 <div id="header">
 737 <h1>
 738 gitattributes(5) Manual Page
 739 </h1>
 740 <h2>NAME</h2>
 741 <div class="sectionbody">
 742 <p>gitattributes -
 743    Defining attributes per path
 744 </p>
 745 </div>
 746 </div>
 747 <div id="content">
 748 <div class="sect1">
 749 <h2 id="_synopsis">SYNOPSIS</h2>
 750 <div class="sectionbody">
 751 <div class="paragraph"><p>$GIT_DIR/info/attributes, .gitattributes</p></div>
 752 </div>
 753 </div>
 754 <div class="sect1">
 755 <h2 id="_description">DESCRIPTION</h2>
 756 <div class="sectionbody">
 757 <div class="paragraph"><p>A <code>gitattributes</code> file is a simple text file that gives
 758 <code>attributes</code> to pathnames.</p></div>
 759 <div class="paragraph"><p>Each line in <code>gitattributes</code> file is of form:</p></div>
 760 <div class="literalblock">
 761 <div class="content">
 762 <pre><code>pattern attr1 attr2 ...</code></pre>
 763 </div></div>
 764 <div class="paragraph"><p>That is, a pattern followed by an attributes list,
 765 separated by whitespaces. Leading and trailing whitespaces are
 766 ignored. Lines that begin with <em>#</em> are ignored. Patterns
 767 that begin with a double quote are quoted in C style.
 768 When the pattern matches the path in question, the attributes
 769 listed on the line are given to the path.</p></div>
 770 <div class="paragraph"><p>Each attribute can be in one of these states for a given path:</p></div>
 771 <div class="dlist"><dl>
 772 <dt class="hdlist1">
 773 Set
 774 </dt>
 775 <dd>
 776 <p>
 777         The path has the attribute with special value "true";
 778         this is specified by listing only the name of the
 779         attribute in the attribute list.
 780 </p>
 781 </dd>
 782 <dt class="hdlist1">
 783 Unset
 784 </dt>
 785 <dd>
 786 <p>
 787         The path has the attribute with special value "false";
 788         this is specified by listing the name of the attribute
 789         prefixed with a dash <code>-</code> in the attribute list.
 790 </p>
 791 </dd>
 792 <dt class="hdlist1">
 793 Set to a value
 794 </dt>
 795 <dd>
 796 <p>
 797         The path has the attribute with specified string value;
 798         this is specified by listing the name of the attribute
 799         followed by an equal sign <code>=</code> and its value in the
 800         attribute list.
 801 </p>
 802 </dd>
 803 <dt class="hdlist1">
 804 Unspecified
 805 </dt>
 806 <dd>
 807 <p>
 808         No pattern matches the path, and nothing says if
 809         the path has or does not have the attribute, the
 810         attribute for the path is said to be Unspecified.
 811 </p>
 812 </dd>
 813 </dl></div>
 814 <div class="paragraph"><p>When more than one pattern matches the path, a later line
 815 overrides an earlier line.  This overriding is done per
 816 attribute.</p></div>
 817 <div class="paragraph"><p>The rules by which the pattern matches paths are the same as in
 818 <code>.gitignore</code> files (see <a href="gitignore.html">gitignore(5)</a>), with a few exceptions:</p></div>
 819 <div class="ulist"><ul>
 820 <li>
 821 <p>
 822 negative patterns are forbidden
 823 </p>
 824 </li>
 825 <li>
 826 <p>
 827 patterns that match a directory do not recursively match paths
 828     inside that directory (so using the trailing-slash <code>path/</code> syntax is
 829     pointless in an attributes file; use <code>path/**</code> instead)
 830 </p>
 831 </li>
 832 </ul></div>
 833 <div class="paragraph"><p>When deciding what attributes are assigned to a path, Git
 834 consults <code>$GIT_DIR/info/attributes</code> file (which has the highest
 835 precedence), <code>.gitattributes</code> file in the same directory as the
 836 path in question, and its parent directories up to the toplevel of the
 837 work tree (the further the directory that contains <code>.gitattributes</code>
 838 is from the path in question, the lower its precedence). Finally
 839 global and system-wide files are considered (they have the lowest
 840 precedence).</p></div>
 841 <div class="paragraph"><p>When the <code>.gitattributes</code> file is missing from the work tree, the
 842 path in the index is used as a fall-back.  During checkout process,
 843 <code>.gitattributes</code> in the index is used and then the file in the
 844 working tree is used as a fall-back.</p></div>
 845 <div class="paragraph"><p>If you wish to affect only a single repository (i.e., to assign
 846 attributes to files that are particular to
 847 one user&#8217;s workflow for that repository), then
 848 attributes should be placed in the <code>$GIT_DIR/info/attributes</code> file.
 849 Attributes which should be version-controlled and distributed to other
 850 repositories (i.e., attributes of interest to all users) should go into
 851 <code>.gitattributes</code> files. Attributes that should affect all repositories
 852 for a single user should be placed in a file specified by the
 853 <code>core.attributesFile</code> configuration option (see <a href="git-config.html">git-config(1)</a>).
 854 Its default value is $XDG_CONFIG_HOME/git/attributes. If $XDG_CONFIG_HOME
 855 is either not set or empty, $HOME/.config/git/attributes is used instead.
 856 Attributes for all users on a system should be placed in the
 857 <code>$(prefix)/etc/gitattributes</code> file.</p></div>
 858 <div class="paragraph"><p>Sometimes you would need to override a setting of an attribute
 859 for a path to <code>Unspecified</code> state.  This can be done by listing
 860 the name of the attribute prefixed with an exclamation point <code>!</code>.</p></div>
 861 </div>
 862 </div>
 863 <div class="sect1">
 864 <h2 id="_effects">EFFECTS</h2>
 865 <div class="sectionbody">
 866 <div class="paragraph"><p>Certain operations by Git can be influenced by assigning
 867 particular attributes to a path.  Currently, the following
 868 operations are attributes-aware.</p></div>
 869 <div class="sect2">
 870 <h3 id="_checking_out_and_checking_in">Checking-out and checking-in</h3>
 871 <div class="paragraph"><p>These attributes affect how the contents stored in the
 872 repository are copied to the working tree files when commands
 873 such as <em>git switch</em>, <em>git checkout</em>  and <em>git merge</em> run.
 874 They also affect how
 875 Git stores the contents you prepare in the working tree in the
 876 repository upon <em>git add</em> and <em>git commit</em>.</p></div>
 877 <div class="sect3">
 878 <h4 id="_code_text_code"><code>text</code></h4>
 879 <div class="paragraph"><p>This attribute marks the path as a text file, which enables end-of-line
 880 conversion: When a matching file is added to the index, the file&#8217;s line
 881 endings are normalized to LF in the index.  Conversely, when the file is
 882 copied from the index to the working directory, its line endings may be
 883 converted from LF to CRLF depending on the <code>eol</code> attribute, the Git
 884 config, and the platform (see explanation of <code>eol</code> below).</p></div>
 885 <div class="dlist"><dl>
 886 <dt class="hdlist1">
 887 Set
 888 </dt>
 889 <dd>
 890 <p>
 891         Setting the <code>text</code> attribute on a path enables end-of-line
 892         conversion on checkin and checkout as described above.  Line endings
 893         are normalized to LF in the index every time the file is checked in,
 894         even if the file was previously added to Git with CRLF line endings.
 895 </p>
 896 </dd>
 897 <dt class="hdlist1">
 898 Unset
 899 </dt>
 900 <dd>
 901 <p>
 902         Unsetting the <code>text</code> attribute on a path tells Git not to
 903         attempt any end-of-line conversion upon checkin or checkout.
 904 </p>
 905 </dd>
 906 <dt class="hdlist1">
 907 Set to string value "auto"
 908 </dt>
 909 <dd>
 910 <p>
 911         When <code>text</code> is set to "auto", Git decides by itself whether the file
 912         is text or binary.  If it is text and the file was not already in
 913         Git with CRLF endings, line endings are converted on checkin and
 914         checkout as described above.  Otherwise, no conversion is done on
 915         checkin or checkout.
 916 </p>
 917 </dd>
 918 <dt class="hdlist1">
 919 Unspecified
 920 </dt>
 921 <dd>
 922 <p>
 923         If the <code>text</code> attribute is unspecified, Git uses the
 924         <code>core.autocrlf</code> configuration variable to determine if the
 925         file should be converted.
 926 </p>
 927 </dd>
 928 </dl></div>
 929 <div class="paragraph"><p>Any other value causes Git to act as if <code>text</code> has been left
 930 unspecified.</p></div>
 931 </div>
 932 <div class="sect3">
 933 <h4 id="_code_eol_code"><code>eol</code></h4>
 934 <div class="paragraph"><p>This attribute marks a path to use a specific line-ending style in the
 935 working tree when it is checked out.  It has effect only if <code>text</code> or
 936 <code>text=auto</code> is set (see above), but specifying <code>eol</code> automatically sets
 937 <code>text</code> if <code>text</code> was left unspecified.</p></div>
 938 <div class="dlist"><dl>
 939 <dt class="hdlist1">
 940 Set to string value "crlf"
 941 </dt>
 942 <dd>
 943 <p>
 944         This setting converts the file&#8217;s line endings in the working
 945         directory to CRLF when the file is checked out.
 946 </p>
 947 </dd>
 948 <dt class="hdlist1">
 949 Set to string value "lf"
 950 </dt>
 951 <dd>
 952 <p>
 953         This setting uses the same line endings in the working directory as
 954         in the index when the file is checked out.
 955 </p>
 956 </dd>
 957 <dt class="hdlist1">
 958 Unspecified
 959 </dt>
 960 <dd>
 961 <p>
 962         If the <code>eol</code> attribute is unspecified for a file, its line endings
 963         in the working directory are determined by the <code>core.autocrlf</code> or
 964         <code>core.eol</code> configuration variable (see the definitions of those
 965         options in <a href="git-config.html">git-config(1)</a>).  If <code>text</code> is set but neither of
 966         those variables is, the default is <code>eol=crlf</code> on Windows and
 967         <code>eol=lf</code> on all other platforms.
 968 </p>
 969 </dd>
 970 </dl></div>
 971 </div>
 972 <div class="sect3">
 973 <h4 id="_backwards_compatibility_with_code_crlf_code_attribute">Backwards compatibility with <code>crlf</code> attribute</h4>
 974 <div class="paragraph"><p>For backwards compatibility, the <code>crlf</code> attribute is interpreted as
 975 follows:</p></div>
 976 <div class="listingblock">
 977 <div class="content">
 978 <pre><code>crlf            text
 979 -crlf           -text
 980 crlf=input      eol=lf</code></pre>
 981 </div></div>
 982 </div>
 983 <div class="sect3">
 984 <h4 id="_end_of_line_conversion">End-of-line conversion</h4>
 985 <div class="paragraph"><p>While Git normally leaves file contents alone, it can be configured to
 986 normalize line endings to LF in the repository and, optionally, to
 987 convert them to CRLF when files are checked out.</p></div>
 988 <div class="paragraph"><p>If you simply want to have CRLF line endings in your working directory
 989 regardless of the repository you are working with, you can set the
 990 config variable "core.autocrlf" without using any attributes.</p></div>
 991 <div class="listingblock">
 992 <div class="content">
 993 <pre><code>[core]
 994         autocrlf = true</code></pre>
 995 </div></div>
 996 <div class="paragraph"><p>This does not force normalization of text files, but does ensure
 997 that text files that you introduce to the repository have their line
 998 endings normalized to LF when they are added, and that files that are
 999 already normalized in the repository stay normalized.</p></div>
1000 <div class="paragraph"><p>If you want to ensure that text files that any contributor introduces to
1001 the repository have their line endings normalized, you can set the
1002 <code>text</code> attribute to "auto" for <em>all</em> files.</p></div>
1003 <div class="listingblock">
1004 <div class="content">
1005 <pre><code>*       text=auto</code></pre>
1006 </div></div>
1007 <div class="paragraph"><p>The attributes allow a fine-grained control, how the line endings
1008 are converted.
1009 Here is an example that will make Git normalize .txt, .vcproj and .sh
1010 files, ensure that .vcproj files have CRLF and .sh files have LF in
1011 the working directory, and prevent .jpg files from being normalized
1012 regardless of their content.</p></div>
1013 <div class="listingblock">
1014 <div class="content">
1015 <pre><code>*               text=auto
1016 *.txt           text
1017 *.vcproj        text eol=crlf
1018 *.sh            text eol=lf
1019 *.jpg           -text</code></pre>
1020 </div></div>
1021 <div class="admonitionblock">
1022 <table><tr>
1023 <td class="icon">
1024 <div class="title">Note</div>
1025 </td>
1026 <td class="content">When <code>text=auto</code> conversion is enabled in a cross-platform
1027 project using push and pull to a central repository the text files
1028 containing CRLFs should be normalized.</td>
1029 </tr></table>
1030 </div>
1031 <div class="paragraph"><p>From a clean working directory:</p></div>
1032 <div class="listingblock">
1033 <div class="content">
1034 <pre><code>$ echo "* text=auto" &gt;.gitattributes
1035 $ git add --renormalize .
1036 $ git status        # Show files that will be normalized
1037 $ git commit -m "Introduce end-of-line normalization"</code></pre>
1038 </div></div>
1039 <div class="paragraph"><p>If any files that should not be normalized show up in <em>git status</em>,
1040 unset their <code>text</code> attribute before running <em>git add -u</em>.</p></div>
1041 <div class="listingblock">
1042 <div class="content">
1043 <pre><code>manual.pdf      -text</code></pre>
1044 </div></div>
1045 <div class="paragraph"><p>Conversely, text files that Git does not detect can have normalization
1046 enabled manually.</p></div>
1047 <div class="listingblock">
1048 <div class="content">
1049 <pre><code>weirdchars.txt  text</code></pre>
1050 </div></div>
1051 <div class="paragraph"><p>If <code>core.safecrlf</code> is set to "true" or "warn", Git verifies if
1052 the conversion is reversible for the current setting of
1053 <code>core.autocrlf</code>.  For "true", Git rejects irreversible
1054 conversions; for "warn", Git only prints a warning but accepts
1055 an irreversible conversion.  The safety triggers to prevent such
1056 a conversion done to the files in the work tree, but there are a
1057 few exceptions.  Even though&#8230;</p></div>
1058 <div class="ulist"><ul>
1059 <li>
1060 <p>
1061 <em>git add</em> itself does not touch the files in the work tree, the
1062   next checkout would, so the safety triggers;
1063 </p>
1064 </li>
1065 <li>
1066 <p>
1067 <em>git apply</em> to update a text file with a patch does touch the files
1068   in the work tree, but the operation is about text files and CRLF
1069   conversion is about fixing the line ending inconsistencies, so the
1070   safety does not trigger;
1071 </p>
1072 </li>
1073 <li>
1074 <p>
1075 <em>git diff</em> itself does not touch the files in the work tree, it is
1076   often run to inspect the changes you intend to next <em>git add</em>.  To
1077   catch potential problems early, safety triggers.
1078 </p>
1079 </li>
1080 </ul></div>
1081 </div>
1082 <div class="sect3">
1083 <h4 id="_code_working_tree_encoding_code"><code>working-tree-encoding</code></h4>
1084 <div class="paragraph"><p>Git recognizes files encoded in ASCII or one of its supersets (e.g.
1085 UTF-8, ISO-8859-1, &#8230;) as text files. Files encoded in certain other
1086 encodings (e.g. UTF-16) are interpreted as binary and consequently
1087 built-in Git text processing tools (e.g. <em>git diff</em>) as well as most Git
1088 web front ends do not visualize the contents of these files by default.</p></div>
1089 <div class="paragraph"><p>In these cases you can tell Git the encoding of a file in the working
1090 directory with the <code>working-tree-encoding</code> attribute. If a file with this
1091 attribute is added to Git, then Git re-encodes the content from the
1092 specified encoding to UTF-8. Finally, Git stores the UTF-8 encoded
1093 content in its internal data structure (called "the index"). On checkout
1094 the content is re-encoded back to the specified encoding.</p></div>
1095 <div class="paragraph"><p>Please note that using the <code>working-tree-encoding</code> attribute may have a
1096 number of pitfalls:</p></div>
1097 <div class="ulist"><ul>
1098 <li>
1099 <p>
1100 Alternative Git implementations (e.g. JGit or libgit2) and older Git
1101   versions (as of March 2018) do not support the <code>working-tree-encoding</code>
1102   attribute. If you decide to use the <code>working-tree-encoding</code> attribute
1103   in your repository, then it is strongly recommended to ensure that all
1104   clients working with the repository support it.
1105 </p>
1106 <div class="paragraph"><p>For example, Microsoft Visual Studio resources files (<code>*.rc</code>) or
1107 PowerShell script files (<code>*.ps1</code>) are sometimes encoded in UTF-16.
1108 If you declare <code>*.ps1</code> as files as UTF-16 and you add <code>foo.ps1</code> with
1109 a <code>working-tree-encoding</code> enabled Git client, then <code>foo.ps1</code> will be
1110 stored as UTF-8 internally. A client without <code>working-tree-encoding</code>
1111 support will checkout <code>foo.ps1</code> as UTF-8 encoded file. This will
1112 typically cause trouble for the users of this file.</p></div>
1113 <div class="paragraph"><p>If a Git client that does not support the <code>working-tree-encoding</code>
1114 attribute adds a new file <code>bar.ps1</code>, then <code>bar.ps1</code> will be
1115 stored "as-is" internally (in this example probably as UTF-16).
1116 A client with <code>working-tree-encoding</code> support will interpret the
1117 internal contents as UTF-8 and try to convert it to UTF-16 on checkout.
1118 That operation will fail and cause an error.</p></div>
1119 </li>
1120 <li>
1121 <p>
1122 Reencoding content to non-UTF encodings can cause errors as the
1123   conversion might not be UTF-8 round trip safe. If you suspect your
1124   encoding to not be round trip safe, then add it to
1125   <code>core.checkRoundtripEncoding</code> to make Git check the round trip
1126   encoding (see <a href="git-config.html">git-config(1)</a>). SHIFT-JIS (Japanese character
1127   set) is known to have round trip issues with UTF-8 and is checked by
1128   default.
1129 </p>
1130 </li>
1131 <li>
1132 <p>
1133 Reencoding content requires resources that might slow down certain
1134   Git operations (e.g <em>git checkout</em> or <em>git add</em>).
1135 </p>
1136 </li>
1137 </ul></div>
1138 <div class="paragraph"><p>Use the <code>working-tree-encoding</code> attribute only if you cannot store a file
1139 in UTF-8 encoding and if you want Git to be able to process the content
1140 as text.</p></div>
1141 <div class="paragraph"><p>As an example, use the following attributes if your <em>*.ps1</em> files are
1142 UTF-16 encoded with byte order mark (BOM) and you want Git to perform
1143 automatic line ending conversion based on your platform.</p></div>
1144 <div class="listingblock">
1145 <div class="content">
1146 <pre><code>*.ps1           text working-tree-encoding=UTF-16</code></pre>
1147 </div></div>
1148 <div class="paragraph"><p>Use the following attributes if your <em>*.ps1</em> files are UTF-16 little
1149 endian encoded without BOM and you want Git to use Windows line endings
1150 in the working directory (use <code>UTF-16LE-BOM</code> instead of <code>UTF-16LE</code> if
1151 you want UTF-16 little endian with BOM).
1152 Please note, it is highly recommended to
1153 explicitly define the line endings with <code>eol</code> if the <code>working-tree-encoding</code>
1154 attribute is used to avoid ambiguity.</p></div>
1155 <div class="listingblock">
1156 <div class="content">
1157 <pre><code>*.ps1           text working-tree-encoding=UTF-16LE eol=CRLF</code></pre>
1158 </div></div>
1159 <div class="paragraph"><p>You can get a list of all available encodings on your platform with the
1160 following command:</p></div>
1161 <div class="listingblock">
1162 <div class="content">
1163 <pre><code>iconv --list</code></pre>
1164 </div></div>
1165 <div class="paragraph"><p>If you do not know the encoding of a file, then you can use the <code>file</code>
1166 command to guess the encoding:</p></div>
1167 <div class="listingblock">
1168 <div class="content">
1169 <pre><code>file foo.ps1</code></pre>
1170 </div></div>
1171 </div>
1172 <div class="sect3">
1173 <h4 id="_code_ident_code"><code>ident</code></h4>
1174 <div class="paragraph"><p>When the attribute <code>ident</code> is set for a path, Git replaces
1175 <code>$Id$</code> in the blob object with <code>$Id:</code>, followed by the
1176 40-character hexadecimal blob object name, followed by a dollar
1177 sign <code>$</code> upon checkout.  Any byte sequence that begins with
1178 <code>$Id:</code> and ends with <code>$</code> in the worktree file is replaced
1179 with <code>$Id$</code> upon check-in.</p></div>
1180 </div>
1181 <div class="sect3">
1182 <h4 id="_code_filter_code"><code>filter</code></h4>
1183 <div class="paragraph"><p>A <code>filter</code> attribute can be set to a string value that names a
1184 filter driver specified in the configuration.</p></div>
1185 <div class="paragraph"><p>A filter driver consists of a <code>clean</code> command and a <code>smudge</code>
1186 command, either of which can be left unspecified.  Upon
1187 checkout, when the <code>smudge</code> command is specified, the command is
1188 fed the blob object from its standard input, and its standard
1189 output is used to update the worktree file.  Similarly, the
1190 <code>clean</code> command is used to convert the contents of worktree file
1191 upon checkin. By default these commands process only a single
1192 blob and terminate. If a long running <code>process</code> filter is used
1193 in place of <code>clean</code> and/or <code>smudge</code> filters, then Git can process
1194 all blobs with a single filter command invocation for the entire
1195 life of a single Git command, for example <code>git add --all</code>. If a
1196 long running <code>process</code> filter is configured then it always takes
1197 precedence over a configured single blob filter. See section
1198 below for the description of the protocol used to communicate with
1199 a <code>process</code> filter.</p></div>
1200 <div class="paragraph"><p>One use of the content filtering is to massage the content into a shape
1201 that is more convenient for the platform, filesystem, and the user to use.
1202 For this mode of operation, the key phrase here is "more convenient" and
1203 not "turning something unusable into usable".  In other words, the intent
1204 is that if someone unsets the filter driver definition, or does not have
1205 the appropriate filter program, the project should still be usable.</p></div>
1206 <div class="paragraph"><p>Another use of the content filtering is to store the content that cannot
1207 be directly used in the repository (e.g. a UUID that refers to the true
1208 content stored outside Git, or an encrypted content) and turn it into a
1209 usable form upon checkout (e.g. download the external content, or decrypt
1210 the encrypted content).</p></div>
1211 <div class="paragraph"><p>These two filters behave differently, and by default, a filter is taken as
1212 the former, massaging the contents into more convenient shape.  A missing
1213 filter driver definition in the config, or a filter driver that exits with
1214 a non-zero status, is not an error but makes the filter a no-op passthru.</p></div>
1215 <div class="paragraph"><p>You can declare that a filter turns a content that by itself is unusable
1216 into a usable content by setting the filter.&lt;driver&gt;.required configuration
1217 variable to <code>true</code>.</p></div>
1218 <div class="paragraph"><p>Note: Whenever the clean filter is changed, the repo should be renormalized:
1219 $ git add --renormalize .</p></div>
1220 <div class="paragraph"><p>For example, in .gitattributes, you would assign the <code>filter</code>
1221 attribute for paths.</p></div>
1222 <div class="listingblock">
1223 <div class="content">
1224 <pre><code>*.c     filter=indent</code></pre>
1225 </div></div>
1226 <div class="paragraph"><p>Then you would define a "filter.indent.clean" and "filter.indent.smudge"
1227 configuration in your .git/config to specify a pair of commands to
1228 modify the contents of C programs when the source files are checked
1229 in ("clean" is run) and checked out (no change is made because the
1230 command is "cat").</p></div>
1231 <div class="listingblock">
1232 <div class="content">
1233 <pre><code>[filter "indent"]
1234         clean = indent
1235         smudge = cat</code></pre>
1236 </div></div>
1237 <div class="paragraph"><p>For best results, <code>clean</code> should not alter its output further if it is
1238 run twice ("clean&#8594;clean" should be equivalent to "clean"), and
1239 multiple <code>smudge</code> commands should not alter <code>clean</code>'s output
1240 ("smudge&#8594;smudge&#8594;clean" should be equivalent to "clean").  See the
1241 section on merging below.</p></div>
1242 <div class="paragraph"><p>The "indent" filter is well-behaved in this regard: it will not modify
1243 input that is already correctly indented.  In this case, the lack of a
1244 smudge filter means that the clean filter <em>must</em> accept its own output
1245 without modifying it.</p></div>
1246 <div class="paragraph"><p>If a filter <em>must</em> succeed in order to make the stored contents usable,
1247 you can declare that the filter is <code>required</code>, in the configuration:</p></div>
1248 <div class="listingblock">
1249 <div class="content">
1250 <pre><code>[filter "crypt"]
1251         clean = openssl enc ...
1252         smudge = openssl enc -d ...
1253         required</code></pre>
1254 </div></div>
1255 <div class="paragraph"><p>Sequence "%f" on the filter command line is replaced with the name of
1256 the file the filter is working on.  A filter might use this in keyword
1257 substitution.  For example:</p></div>
1258 <div class="listingblock">
1259 <div class="content">
1260 <pre><code>[filter "p4"]
1261         clean = git-p4-filter --clean %f
1262         smudge = git-p4-filter --smudge %f</code></pre>
1263 </div></div>
1264 <div class="paragraph"><p>Note that "%f" is the name of the path that is being worked on. Depending
1265 on the version that is being filtered, the corresponding file on disk may
1266 not exist, or may have different contents. So, smudge and clean commands
1267 should not try to access the file on disk, but only act as filters on the
1268 content provided to them on standard input.</p></div>
1269 </div>
1270 <div class="sect3">
1271 <h4 id="_long_running_filter_process">Long Running Filter Process</h4>
1272 <div class="paragraph"><p>If the filter command (a string value) is defined via
1273 <code>filter.&lt;driver&gt;.process</code> then Git can process all blobs with a
1274 single filter invocation for the entire life of a single Git
1275 command. This is achieved by using the long-running process protocol
1276 (described in technical/long-running-process-protocol.txt).</p></div>
1277 <div class="paragraph"><p>When Git encounters the first file that needs to be cleaned or smudged,
1278 it starts the filter and performs the handshake. In the handshake, the
1279 welcome message sent by Git is "git-filter-client", only version 2 is
1280 supported, and the supported capabilities are "clean", "smudge", and
1281 "delay".</p></div>
1282 <div class="paragraph"><p>Afterwards Git sends a list of "key=value" pairs terminated with
1283 a flush packet. The list will contain at least the filter command
1284 (based on the supported capabilities) and the pathname of the file
1285 to filter relative to the repository root. Right after the flush packet
1286 Git sends the content split in zero or more pkt-line packets and a
1287 flush packet to terminate content. Please note, that the filter
1288 must not send any response before it received the content and the
1289 final flush packet. Also note that the "value" of a "key=value" pair
1290 can contain the "=" character whereas the key would never contain
1291 that character.</p></div>
1292 <div class="listingblock">
1293 <div class="content">
1294 <pre><code>packet:          git&gt; command=smudge
1295 packet:          git&gt; pathname=path/testfile.dat
1296 packet:          git&gt; 0000
1297 packet:          git&gt; CONTENT
1298 packet:          git&gt; 0000</code></pre>
1299 </div></div>
1300 <div class="paragraph"><p>The filter is expected to respond with a list of "key=value" pairs
1301 terminated with a flush packet. If the filter does not experience
1302 problems then the list must contain a "success" status. Right after
1303 these packets the filter is expected to send the content in zero
1304 or more pkt-line packets and a flush packet at the end. Finally, a
1305 second list of "key=value" pairs terminated with a flush packet
1306 is expected. The filter can change the status in the second list
1307 or keep the status as is with an empty list. Please note that the
1308 empty list must be terminated with a flush packet regardless.</p></div>
1309 <div class="listingblock">
1310 <div class="content">
1311 <pre><code>packet:          git&lt; status=success
1312 packet:          git&lt; 0000
1313 packet:          git&lt; SMUDGED_CONTENT
1314 packet:          git&lt; 0000
1315 packet:          git&lt; 0000  # empty list, keep "status=success" unchanged!</code></pre>
1316 </div></div>
1317 <div class="paragraph"><p>If the result content is empty then the filter is expected to respond
1318 with a "success" status and a flush packet to signal the empty content.</p></div>
1319 <div class="listingblock">
1320 <div class="content">
1321 <pre><code>packet:          git&lt; status=success
1322 packet:          git&lt; 0000
1323 packet:          git&lt; 0000  # empty content!
1324 packet:          git&lt; 0000  # empty list, keep "status=success" unchanged!</code></pre>
1325 </div></div>
1326 <div class="paragraph"><p>In case the filter cannot or does not want to process the content,
1327 it is expected to respond with an "error" status.</p></div>
1328 <div class="listingblock">
1329 <div class="content">
1330 <pre><code>packet:          git&lt; status=error
1331 packet:          git&lt; 0000</code></pre>
1332 </div></div>
1333 <div class="paragraph"><p>If the filter experiences an error during processing, then it can
1334 send the status "error" after the content was (partially or
1335 completely) sent.</p></div>
1336 <div class="listingblock">
1337 <div class="content">
1338 <pre><code>packet:          git&lt; status=success
1339 packet:          git&lt; 0000
1340 packet:          git&lt; HALF_WRITTEN_ERRONEOUS_CONTENT
1341 packet:          git&lt; 0000
1342 packet:          git&lt; status=error
1343 packet:          git&lt; 0000</code></pre>
1344 </div></div>
1345 <div class="paragraph"><p>In case the filter cannot or does not want to process the content
1346 as well as any future content for the lifetime of the Git process,
1347 then it is expected to respond with an "abort" status at any point
1348 in the protocol.</p></div>
1349 <div class="listingblock">
1350 <div class="content">
1351 <pre><code>packet:          git&lt; status=abort
1352 packet:          git&lt; 0000</code></pre>
1353 </div></div>
1354 <div class="paragraph"><p>Git neither stops nor restarts the filter process in case the
1355 "error"/"abort" status is set. However, Git sets its exit code
1356 according to the <code>filter.&lt;driver&gt;.required</code> flag, mimicking the
1357 behavior of the <code>filter.&lt;driver&gt;.clean</code> / <code>filter.&lt;driver&gt;.smudge</code>
1358 mechanism.</p></div>
1359 <div class="paragraph"><p>If the filter dies during the communication or does not adhere to
1360 the protocol then Git will stop the filter process and restart it
1361 with the next file that needs to be processed. Depending on the
1362 <code>filter.&lt;driver&gt;.required</code> flag Git will interpret that as error.</p></div>
1363 </div>
1364 <div class="sect3">
1365 <h4 id="_delay">Delay</h4>
1366 <div class="paragraph"><p>If the filter supports the "delay" capability, then Git can send the
1367 flag "can-delay" after the filter command and pathname. This flag
1368 denotes that the filter can delay filtering the current blob (e.g. to
1369 compensate network latencies) by responding with no content but with
1370 the status "delayed" and a flush packet.</p></div>
1371 <div class="listingblock">
1372 <div class="content">
1373 <pre><code>packet:          git&gt; command=smudge
1374 packet:          git&gt; pathname=path/testfile.dat
1375 packet:          git&gt; can-delay=1
1376 packet:          git&gt; 0000
1377 packet:          git&gt; CONTENT
1378 packet:          git&gt; 0000
1379 packet:          git&lt; status=delayed
1380 packet:          git&lt; 0000</code></pre>
1381 </div></div>
1382 <div class="paragraph"><p>If the filter supports the "delay" capability then it must support the
1383 "list_available_blobs" command. If Git sends this command, then the
1384 filter is expected to return a list of pathnames representing blobs
1385 that have been delayed earlier and are now available.
1386 The list must be terminated with a flush packet followed
1387 by a "success" status that is also terminated with a flush packet. If
1388 no blobs for the delayed paths are available, yet, then the filter is
1389 expected to block the response until at least one blob becomes
1390 available. The filter can tell Git that it has no more delayed blobs
1391 by sending an empty list. As soon as the filter responds with an empty
1392 list, Git stops asking. All blobs that Git has not received at this
1393 point are considered missing and will result in an error.</p></div>
1394 <div class="listingblock">
1395 <div class="content">
1396 <pre><code>packet:          git&gt; command=list_available_blobs
1397 packet:          git&gt; 0000
1398 packet:          git&lt; pathname=path/testfile.dat
1399 packet:          git&lt; pathname=path/otherfile.dat
1400 packet:          git&lt; 0000
1401 packet:          git&lt; status=success
1402 packet:          git&lt; 0000</code></pre>
1403 </div></div>
1404 <div class="paragraph"><p>After Git received the pathnames, it will request the corresponding
1405 blobs again. These requests contain a pathname and an empty content
1406 section. The filter is expected to respond with the smudged content
1407 in the usual way as explained above.</p></div>
1408 <div class="listingblock">
1409 <div class="content">
1410 <pre><code>packet:          git&gt; command=smudge
1411 packet:          git&gt; pathname=path/testfile.dat
1412 packet:          git&gt; 0000
1413 packet:          git&gt; 0000  # empty content!
1414 packet:          git&lt; status=success
1415 packet:          git&lt; 0000
1416 packet:          git&lt; SMUDGED_CONTENT
1417 packet:          git&lt; 0000
1418 packet:          git&lt; 0000  # empty list, keep "status=success" unchanged!</code></pre>
1419 </div></div>
1420 </div>
1421 <div class="sect3">
1422 <h4 id="_example">Example</h4>
1423 <div class="paragraph"><p>A long running filter demo implementation can be found in
1424 <code>contrib/long-running-filter/example.pl</code> located in the Git
1425 core repository. If you develop your own long running filter
1426 process then the <code>GIT_TRACE_PACKET</code> environment variables can be
1427 very helpful for debugging (see <a href="git.html">git(1)</a>).</p></div>
1428 <div class="paragraph"><p>Please note that you cannot use an existing <code>filter.&lt;driver&gt;.clean</code>
1429 or <code>filter.&lt;driver&gt;.smudge</code> command with <code>filter.&lt;driver&gt;.process</code>
1430 because the former two use a different inter process communication
1431 protocol than the latter one.</p></div>
1432 </div>
1433 <div class="sect3">
1434 <h4 id="_interaction_between_checkin_checkout_attributes">Interaction between checkin/checkout attributes</h4>
1435 <div class="paragraph"><p>In the check-in codepath, the worktree file is first converted
1436 with <code>filter</code> driver (if specified and corresponding driver
1437 defined), then the result is processed with <code>ident</code> (if
1438 specified), and then finally with <code>text</code> (again, if specified
1439 and applicable).</p></div>
1440 <div class="paragraph"><p>In the check-out codepath, the blob content is first converted
1441 with <code>text</code>, and then <code>ident</code> and fed to <code>filter</code>.</p></div>
1442 </div>
1443 <div class="sect3">
1444 <h4 id="_merging_branches_with_differing_checkin_checkout_attributes">Merging branches with differing checkin/checkout attributes</h4>
1445 <div class="paragraph"><p>If you have added attributes to a file that cause the canonical
1446 repository format for that file to change, such as adding a
1447 clean/smudge filter or text/eol/ident attributes, merging anything
1448 where the attribute is not in place would normally cause merge
1449 conflicts.</p></div>
1450 <div class="paragraph"><p>To prevent these unnecessary merge conflicts, Git can be told to run a
1451 virtual check-out and check-in of all three stages of a file when
1452 resolving a three-way merge by setting the <code>merge.renormalize</code>
1453 configuration variable.  This prevents changes caused by check-in
1454 conversion from causing spurious merge conflicts when a converted file
1455 is merged with an unconverted file.</p></div>
1456 <div class="paragraph"><p>As long as a "smudge&#8594;clean" results in the same output as a "clean"
1457 even on files that are already smudged, this strategy will
1458 automatically resolve all filter-related conflicts.  Filters that do
1459 not act in this way may cause additional merge conflicts that must be
1460 resolved manually.</p></div>
1461 </div>
1462 </div>
1463 <div class="sect2">
1464 <h3 id="_generating_diff_text">Generating diff text</h3>
1465 <div class="sect3">
1466 <h4 id="_code_diff_code"><code>diff</code></h4>
1467 <div class="paragraph"><p>The attribute <code>diff</code> affects how Git generates diffs for particular
1468 files. It can tell Git whether to generate a textual patch for the path
1469 or to treat the path as a binary file.  It can also affect what line is
1470 shown on the hunk header <code>@@ -k,l +n,m @@</code> line, tell Git to use an
1471 external command to generate the diff, or ask Git to convert binary
1472 files to a text format before generating the diff.</p></div>
1473 <div class="dlist"><dl>
1474 <dt class="hdlist1">
1475 Set
1476 </dt>
1477 <dd>
1478 <p>
1479         A path to which the <code>diff</code> attribute is set is treated
1480         as text, even when they contain byte values that
1481         normally never appear in text files, such as NUL.
1482 </p>
1483 </dd>
1484 <dt class="hdlist1">
1485 Unset
1486 </dt>
1487 <dd>
1488 <p>
1489         A path to which the <code>diff</code> attribute is unset will
1490         generate <code>Binary files differ</code> (or a binary patch, if
1491         binary patches are enabled).
1492 </p>
1493 </dd>
1494 <dt class="hdlist1">
1495 Unspecified
1496 </dt>
1497 <dd>
1498 <p>
1499         A path to which the <code>diff</code> attribute is unspecified
1500         first gets its contents inspected, and if it looks like
1501         text and is smaller than core.bigFileThreshold, it is treated
1502         as text. Otherwise it would generate <code>Binary files differ</code>.
1503 </p>
1504 </dd>
1505 <dt class="hdlist1">
1506 String
1507 </dt>
1508 <dd>
1509 <p>
1510         Diff is shown using the specified diff driver.  Each driver may
1511         specify one or more options, as described in the following
1512         section. The options for the diff driver "foo" are defined
1513         by the configuration variables in the "diff.foo" section of the
1514         Git config file.
1515 </p>
1516 </dd>
1517 </dl></div>
1518 </div>
1519 <div class="sect3">
1520 <h4 id="_defining_an_external_diff_driver">Defining an external diff driver</h4>
1521 <div class="paragraph"><p>The definition of a diff driver is done in <code>gitconfig</code>, not
1522 <code>gitattributes</code> file, so strictly speaking this manual page is a
1523 wrong place to talk about it.  However&#8230;</p></div>
1524 <div class="paragraph"><p>To define an external diff driver <code>jcdiff</code>, add a section to your
1525 <code>$GIT_DIR/config</code> file (or <code>$HOME/.gitconfig</code> file) like this:</p></div>
1526 <div class="listingblock">
1527 <div class="content">
1528 <pre><code>[diff "jcdiff"]
1529         command = j-c-diff</code></pre>
1530 </div></div>
1531 <div class="paragraph"><p>When Git needs to show you a diff for the path with <code>diff</code>
1532 attribute set to <code>jcdiff</code>, it calls the command you specified
1533 with the above configuration, i.e. <code>j-c-diff</code>, with 7
1534 parameters, just like <code>GIT_EXTERNAL_DIFF</code> program is called.
1535 See <a href="git.html">git(1)</a> for details.</p></div>
1536 </div>
1537 <div class="sect3">
1538 <h4 id="_setting_the_internal_diff_algorithm">Setting the internal diff algorithm</h4>
1539 <div class="paragraph"><p>The diff algorithm can be set through the <code>diff.algorithm</code> config key, but
1540 sometimes it may be helpful to set the diff algorithm per path. For example,
1541 one may want to use the <code>minimal</code> diff algorithm for .json files, and the
1542 <code>histogram</code> for .c files, and so on without having to pass in the algorithm
1543 through the command line each time.</p></div>
1544 <div class="paragraph"><p>First, in <code>.gitattributes</code>, assign the <code>diff</code> attribute for paths.</p></div>
1545 <div class="listingblock">
1546 <div class="content">
1547 <pre><code>*.json diff=&lt;name&gt;</code></pre>
1548 </div></div>
1549 <div class="paragraph"><p>Then, define a "diff.&lt;name&gt;.algorithm" configuration to specify the diff
1550 algorithm, choosing from <code>myers</code>, <code>patience</code>, <code>minimal</code>, or <code>histogram</code>.</p></div>
1551 <div class="listingblock">
1552 <div class="content">
1553 <pre><code>[diff "&lt;name&gt;"]
1554   algorithm = histogram</code></pre>
1555 </div></div>
1556 <div class="paragraph"><p>This diff algorithm applies to user facing diff output like git-diff(1),
1557 git-show(1) and is used for the <code>--stat</code> output as well. The merge machinery
1558 will not use the diff algorithm set through this method.</p></div>
1559 <div class="admonitionblock">
1560 <table><tr>
1561 <td class="icon">
1562 <div class="title">Note</div>
1563 </td>
1564 <td class="content">If <code>diff.&lt;name&gt;.command</code> is defined for path with the
1565 <code>diff=&lt;name&gt;</code> attribute, it is executed as an external diff driver
1566 (see above), and adding <code>diff.&lt;name&gt;.algorithm</code> has no effect, as the
1567 algorithm is not passed to the external diff driver.</td>
1568 </tr></table>
1569 </div>
1570 </div>
1571 <div class="sect3">
1572 <h4 id="_defining_a_custom_hunk_header">Defining a custom hunk-header</h4>
1573 <div class="paragraph"><p>Each group of changes (called a "hunk") in the textual diff output
1574 is prefixed with a line of the form:</p></div>
1575 <div class="literalblock">
1576 <div class="content">
1577 <pre><code>@@ -k,l +n,m @@ TEXT</code></pre>
1578 </div></div>
1579 <div class="paragraph"><p>This is called a <em>hunk header</em>.  The "TEXT" portion is by default a line
1580 that begins with an alphabet, an underscore or a dollar sign; this
1581 matches what GNU <em>diff -p</em> output uses.  This default selection however
1582 is not suited for some contents, and you can use a customized pattern
1583 to make a selection.</p></div>
1584 <div class="paragraph"><p>First, in .gitattributes, you would assign the <code>diff</code> attribute
1585 for paths.</p></div>
1586 <div class="listingblock">
1587 <div class="content">
1588 <pre><code>*.tex   diff=tex</code></pre>
1589 </div></div>
1590 <div class="paragraph"><p>Then, you would define a "diff.tex.xfuncname" configuration to
1591 specify a regular expression that matches a line that you would
1592 want to appear as the hunk header "TEXT". Add a section to your
1593 <code>$GIT_DIR/config</code> file (or <code>$HOME/.gitconfig</code> file) like this:</p></div>
1594 <div class="listingblock">
1595 <div class="content">
1596 <pre><code>[diff "tex"]
1597         xfuncname = "^(\\\\(sub)*section\\{.*)$"</code></pre>
1598 </div></div>
1599 <div class="paragraph"><p>Note.  A single level of backslashes are eaten by the
1600 configuration file parser, so you would need to double the
1601 backslashes; the pattern above picks a line that begins with a
1602 backslash, and zero or more occurrences of <code>sub</code> followed by
1603 <code>section</code> followed by open brace, to the end of line.</p></div>
1604 <div class="paragraph"><p>There are a few built-in patterns to make this easier, and <code>tex</code>
1605 is one of them, so you do not have to write the above in your
1606 configuration file (you still need to enable this with the
1607 attribute mechanism, via <code>.gitattributes</code>).  The following built in
1608 patterns are available:</p></div>
1609 <div class="ulist"><ul>
1610 <li>
1611 <p>
1612 <code>ada</code> suitable for source code in the Ada language.
1613 </p>
1614 </li>
1615 <li>
1616 <p>
1617 <code>bash</code> suitable for source code in the Bourne-Again SHell language.
1618   Covers a superset of POSIX shell function definitions.
1619 </p>
1620 </li>
1621 <li>
1622 <p>
1623 <code>bibtex</code> suitable for files with BibTeX coded references.
1624 </p>
1625 </li>
1626 <li>
1627 <p>
1628 <code>cpp</code> suitable for source code in the C and C++ languages.
1629 </p>
1630 </li>
1631 <li>
1632 <p>
1633 <code>csharp</code> suitable for source code in the C# language.
1634 </p>
1635 </li>
1636 <li>
1637 <p>
1638 <code>css</code> suitable for cascading style sheets.
1639 </p>
1640 </li>
1641 <li>
1642 <p>
1643 <code>dts</code> suitable for devicetree (DTS) files.
1644 </p>
1645 </li>
1646 <li>
1647 <p>
1648 <code>elixir</code> suitable for source code in the Elixir language.
1649 </p>
1650 </li>
1651 <li>
1652 <p>
1653 <code>fortran</code> suitable for source code in the Fortran language.
1654 </p>
1655 </li>
1656 <li>
1657 <p>
1658 <code>fountain</code> suitable for Fountain documents.
1659 </p>
1660 </li>
1661 <li>
1662 <p>
1663 <code>golang</code> suitable for source code in the Go language.
1664 </p>
1665 </li>
1666 <li>
1667 <p>
1668 <code>html</code> suitable for HTML/XHTML documents.
1669 </p>
1670 </li>
1671 <li>
1672 <p>
1673 <code>java</code> suitable for source code in the Java language.
1674 </p>
1675 </li>
1676 <li>
1677 <p>
1678 <code>kotlin</code> suitable for source code in the Kotlin language.
1679 </p>
1680 </li>
1681 <li>
1682 <p>
1683 <code>markdown</code> suitable for Markdown documents.
1684 </p>
1685 </li>
1686 <li>
1687 <p>
1688 <code>matlab</code> suitable for source code in the MATLAB and Octave languages.
1689 </p>
1690 </li>
1691 <li>
1692 <p>
1693 <code>objc</code> suitable for source code in the Objective-C language.
1694 </p>
1695 </li>
1696 <li>
1697 <p>
1698 <code>pascal</code> suitable for source code in the Pascal/Delphi language.
1699 </p>
1700 </li>
1701 <li>
1702 <p>
1703 <code>perl</code> suitable for source code in the Perl language.
1704 </p>
1705 </li>
1706 <li>
1707 <p>
1708 <code>php</code> suitable for source code in the PHP language.
1709 </p>
1710 </li>
1711 <li>
1712 <p>
1713 <code>python</code> suitable for source code in the Python language.
1714 </p>
1715 </li>
1716 <li>
1717 <p>
1718 <code>ruby</code> suitable for source code in the Ruby language.
1719 </p>
1720 </li>
1721 <li>
1722 <p>
1723 <code>rust</code> suitable for source code in the Rust language.
1724 </p>
1725 </li>
1726 <li>
1727 <p>
1728 <code>scheme</code> suitable for source code in the Scheme language.
1729 </p>
1730 </li>
1731 <li>
1732 <p>
1733 <code>tex</code> suitable for source code for LaTeX documents.
1734 </p>
1735 </li>
1736 </ul></div>
1737 </div>
1738 <div class="sect3">
1739 <h4 id="_customizing_word_diff">Customizing word diff</h4>
1740 <div class="paragraph"><p>You can customize the rules that <code>git diff --word-diff</code> uses to
1741 split words in a line, by specifying an appropriate regular expression
1742 in the "diff.*.wordRegex" configuration variable.  For example, in TeX
1743 a backslash followed by a sequence of letters forms a command, but
1744 several such commands can be run together without intervening
1745 whitespace.  To separate them, use a regular expression in your
1746 <code>$GIT_DIR/config</code> file (or <code>$HOME/.gitconfig</code> file) like this:</p></div>
1747 <div class="listingblock">
1748 <div class="content">
1749 <pre><code>[diff "tex"]
1750         wordRegex = "\\\\[a-zA-Z]+|[{}]|\\\\.|[^\\{}[:space:]]+"</code></pre>
1751 </div></div>
1752 <div class="paragraph"><p>A built-in pattern is provided for all languages listed in the
1753 previous section.</p></div>
1754 </div>
1755 <div class="sect3">
1756 <h4 id="_performing_text_diffs_of_binary_files">Performing text diffs of binary files</h4>
1757 <div class="paragraph"><p>Sometimes it is desirable to see the diff of a text-converted
1758 version of some binary files. For example, a word processor
1759 document can be converted to an ASCII text representation, and
1760 the diff of the text shown. Even though this conversion loses
1761 some information, the resulting diff is useful for human
1762 viewing (but cannot be applied directly).</p></div>
1763 <div class="paragraph"><p>The <code>textconv</code> config option is used to define a program for
1764 performing such a conversion. The program should take a single
1765 argument, the name of a file to convert, and produce the
1766 resulting text on stdout.</p></div>
1767 <div class="paragraph"><p>For example, to show the diff of the exif information of a
1768 file instead of the binary information (assuming you have the
1769 exif tool installed), add the following section to your
1770 <code>$GIT_DIR/config</code> file (or <code>$HOME/.gitconfig</code> file):</p></div>
1771 <div class="listingblock">
1772 <div class="content">
1773 <pre><code>[diff "jpg"]
1774         textconv = exif</code></pre>
1775 </div></div>
1776 <div class="admonitionblock">
1777 <table><tr>
1778 <td class="icon">
1779 <div class="title">Note</div>
1780 </td>
1781 <td class="content">The text conversion is generally a one-way conversion;
1782 in this example, we lose the actual image contents and focus
1783 just on the text data. This means that diffs generated by
1784 textconv are <em>not</em> suitable for applying. For this reason,
1785 only <code>git diff</code> and the <code>git log</code> family of commands (i.e.,
1786 log, whatchanged, show) will perform text conversion. <code>git
1787 format-patch</code> will never generate this output. If you want to
1788 send somebody a text-converted diff of a binary file (e.g.,
1789 because it quickly conveys the changes you have made), you
1790 should generate it separately and send it as a comment <em>in
1791 addition to</em> the usual binary diff that you might send.</td>
1792 </tr></table>
1793 </div>
1794 <div class="paragraph"><p>Because text conversion can be slow, especially when doing a
1795 large number of them with <code>git log -p</code>, Git provides a mechanism
1796 to cache the output and use it in future diffs.  To enable
1797 caching, set the "cachetextconv" variable in your diff driver&#8217;s
1798 config. For example:</p></div>
1799 <div class="listingblock">
1800 <div class="content">
1801 <pre><code>[diff "jpg"]
1802         textconv = exif
1803         cachetextconv = true</code></pre>
1804 </div></div>
1805 <div class="paragraph"><p>This will cache the result of running "exif" on each blob
1806 indefinitely. If you change the textconv config variable for a
1807 diff driver, Git will automatically invalidate the cache entries
1808 and re-run the textconv filter. If you want to invalidate the
1809 cache manually (e.g., because your version of "exif" was updated
1810 and now produces better output), you can remove the cache
1811 manually with <code>git update-ref -d refs/notes/textconv/jpg</code> (where
1812 "jpg" is the name of the diff driver, as in the example above).</p></div>
1813 </div>
1814 <div class="sect3">
1815 <h4 id="_choosing_textconv_versus_external_diff">Choosing textconv versus external diff</h4>
1816 <div class="paragraph"><p>If you want to show differences between binary or specially-formatted
1817 blobs in your repository, you can choose to use either an external diff
1818 command, or to use textconv to convert them to a diff-able text format.
1819 Which method you choose depends on your exact situation.</p></div>
1820 <div class="paragraph"><p>The advantage of using an external diff command is flexibility. You are
1821 not bound to find line-oriented changes, nor is it necessary for the
1822 output to resemble unified diff. You are free to locate and report
1823 changes in the most appropriate way for your data format.</p></div>
1824 <div class="paragraph"><p>A textconv, by comparison, is much more limiting. You provide a
1825 transformation of the data into a line-oriented text format, and Git
1826 uses its regular diff tools to generate the output. There are several
1827 advantages to choosing this method:</p></div>
1828 <div class="olist arabic"><ol class="arabic">
1829 <li>
1830 <p>
1831 Ease of use. It is often much simpler to write a binary to text
1832    transformation than it is to perform your own diff. In many cases,
1833    existing programs can be used as textconv filters (e.g., exif,
1834    odt2txt).
1835 </p>
1836 </li>
1837 <li>
1838 <p>
1839 Git diff features. By performing only the transformation step
1840    yourself, you can still utilize many of Git&#8217;s diff features,
1841    including colorization, word-diff, and combined diffs for merges.
1842 </p>
1843 </li>
1844 <li>
1845 <p>
1846 Caching. Textconv caching can speed up repeated diffs, such as those
1847    you might trigger by running <code>git log -p</code>.
1848 </p>
1849 </li>
1850 </ol></div>
1851 </div>
1852 <div class="sect3">
1853 <h4 id="_marking_files_as_binary">Marking files as binary</h4>
1854 <div class="paragraph"><p>Git usually guesses correctly whether a blob contains text or binary
1855 data by examining the beginning of the contents. However, sometimes you
1856 may want to override its decision, either because a blob contains binary
1857 data later in the file, or because the content, while technically
1858 composed of text characters, is opaque to a human reader. For example,
1859 many postscript files contain only ASCII characters, but produce noisy
1860 and meaningless diffs.</p></div>
1861 <div class="paragraph"><p>The simplest way to mark a file as binary is to unset the diff
1862 attribute in the <code>.gitattributes</code> file:</p></div>
1863 <div class="listingblock">
1864 <div class="content">
1865 <pre><code>*.ps -diff</code></pre>
1866 </div></div>
1867 <div class="paragraph"><p>This will cause Git to generate <code>Binary files differ</code> (or a binary
1868 patch, if binary patches are enabled) instead of a regular diff.</p></div>
1869 <div class="paragraph"><p>However, one may also want to specify other diff driver attributes. For
1870 example, you might want to use <code>textconv</code> to convert postscript files to
1871 an ASCII representation for human viewing, but otherwise treat them as
1872 binary files. You cannot specify both <code>-diff</code> and <code>diff=ps</code> attributes.
1873 The solution is to use the <code>diff.*.binary</code> config option:</p></div>
1874 <div class="listingblock">
1875 <div class="content">
1876 <pre><code>[diff "ps"]
1877   textconv = ps2ascii
1878   binary = true</code></pre>
1879 </div></div>
1880 </div>
1881 </div>
1882 <div class="sect2">
1883 <h3 id="_performing_a_three_way_merge">Performing a three-way merge</h3>
1884 <div class="sect3">
1885 <h4 id="_code_merge_code"><code>merge</code></h4>
1886 <div class="paragraph"><p>The attribute <code>merge</code> affects how three versions of a file are
1887 merged when a file-level merge is necessary during <code>git merge</code>,
1888 and other commands such as <code>git revert</code> and <code>git cherry-pick</code>.</p></div>
1889 <div class="dlist"><dl>
1890 <dt class="hdlist1">
1891 Set
1892 </dt>
1893 <dd>
1894 <p>
1895         Built-in 3-way merge driver is used to merge the
1896         contents in a way similar to <em>merge</em> command of <code>RCS</code>
1897         suite.  This is suitable for ordinary text files.
1898 </p>
1899 </dd>
1900 <dt class="hdlist1">
1901 Unset
1902 </dt>
1903 <dd>
1904 <p>
1905         Take the version from the current branch as the
1906         tentative merge result, and declare that the merge has
1907         conflicts.  This is suitable for binary files that do
1908         not have a well-defined merge semantics.
1909 </p>
1910 </dd>
1911 <dt class="hdlist1">
1912 Unspecified
1913 </dt>
1914 <dd>
1915 <p>
1916         By default, this uses the same built-in 3-way merge
1917         driver as is the case when the <code>merge</code> attribute is set.
1918         However, the <code>merge.default</code> configuration variable can name
1919         different merge driver to be used with paths for which the
1920         <code>merge</code> attribute is unspecified.
1921 </p>
1922 </dd>
1923 <dt class="hdlist1">
1924 String
1925 </dt>
1926 <dd>
1927 <p>
1928         3-way merge is performed using the specified custom
1929         merge driver.  The built-in 3-way merge driver can be
1930         explicitly specified by asking for "text" driver; the
1931         built-in "take the current branch" driver can be
1932         requested with "binary".
1933 </p>
1934 </dd>
1935 </dl></div>
1936 </div>
1937 <div class="sect3">
1938 <h4 id="_built_in_merge_drivers">Built-in merge drivers</h4>
1939 <div class="paragraph"><p>There are a few built-in low-level merge drivers defined that
1940 can be asked for via the <code>merge</code> attribute.</p></div>
1941 <div class="dlist"><dl>
1942 <dt class="hdlist1">
1943 text
1944 </dt>
1945 <dd>
1946 <p>
1947         Usual 3-way file level merge for text files.  Conflicted
1948         regions are marked with conflict markers <code>&lt;&lt;&lt;&lt;&lt;&lt;&lt;</code>,
1949         <code>=======</code> and <code>&gt;&gt;&gt;&gt;&gt;&gt;&gt;</code>.  The version from your branch
1950         appears before the <code>=======</code> marker, and the version
1951         from the merged branch appears after the <code>=======</code>
1952         marker.
1953 </p>
1954 </dd>
1955 <dt class="hdlist1">
1956 binary
1957 </dt>
1958 <dd>
1959 <p>
1960         Keep the version from your branch in the work tree, but
1961         leave the path in the conflicted state for the user to
1962         sort out.
1963 </p>
1964 </dd>
1965 <dt class="hdlist1">
1966 union
1967 </dt>
1968 <dd>
1969 <p>
1970         Run 3-way file level merge for text files, but take
1971         lines from both versions, instead of leaving conflict
1972         markers.  This tends to leave the added lines in the
1973         resulting file in random order and the user should
1974         verify the result. Do not use this if you do not
1975         understand the implications.
1976 </p>
1977 </dd>
1978 </dl></div>
1979 </div>
1980 <div class="sect3">
1981 <h4 id="_defining_a_custom_merge_driver">Defining a custom merge driver</h4>
1982 <div class="paragraph"><p>The definition of a merge driver is done in the <code>.git/config</code>
1983 file, not in the <code>gitattributes</code> file, so strictly speaking this
1984 manual page is a wrong place to talk about it.  However&#8230;</p></div>
1985 <div class="paragraph"><p>To define a custom merge driver <code>filfre</code>, add a section to your
1986 <code>$GIT_DIR/config</code> file (or <code>$HOME/.gitconfig</code> file) like this:</p></div>
1987 <div class="listingblock">
1988 <div class="content">
1989 <pre><code>[merge "filfre"]
1990         name = feel-free merge driver
1991         driver = filfre %O %A %B %L %P
1992         recursive = binary</code></pre>
1993 </div></div>
1994 <div class="paragraph"><p>The <code>merge.*.name</code> variable gives the driver a human-readable
1995 name.</p></div>
1996 <div class="paragraph"><p>The &#8216;merge.*.driver` variable&#8217;s value is used to construct a
1997 command to run to merge ancestor&#8217;s version (<code>%O</code>), current
1998 version (<code>%A</code>) and the other branches&#8217; version (<code>%B</code>).  These
1999 three tokens are replaced with the names of temporary files that
2000 hold the contents of these versions when the command line is
2001 built. Additionally, %L will be replaced with the conflict marker
2002 size (see below).</p></div>
2003 <div class="paragraph"><p>The merge driver is expected to leave the result of the merge in
2004 the file named with <code>%A</code> by overwriting it, and exit with zero
2005 status if it managed to merge them cleanly, or non-zero if there
2006 were conflicts.  When the driver crashes (e.g. killed by SEGV),
2007 it is expected to exit with non-zero status that are higher than
2008 128, and in such a case, the merge results in a failure (which is
2009 different from producing a conflict).</p></div>
2010 <div class="paragraph"><p>The <code>merge.*.recursive</code> variable specifies what other merge
2011 driver to use when the merge driver is called for an internal
2012 merge between common ancestors, when there are more than one.
2013 When left unspecified, the driver itself is used for both
2014 internal merge and the final merge.</p></div>
2015 <div class="paragraph"><p>The merge driver can learn the pathname in which the merged result
2016 will be stored via placeholder <code>%P</code>.</p></div>
2017 </div>
2018 <div class="sect3">
2019 <h4 id="_code_conflict_marker_size_code"><code>conflict-marker-size</code></h4>
2020 <div class="paragraph"><p>This attribute controls the length of conflict markers left in
2021 the work tree file during a conflicted merge.  Only setting to
2022 the value to a positive integer has any meaningful effect.</p></div>
2023 <div class="paragraph"><p>For example, this line in <code>.gitattributes</code> can be used to tell the merge
2024 machinery to leave much longer (instead of the usual 7-character-long)
2025 conflict markers when merging the file <code>Documentation/git-merge.txt</code>
2026 results in a conflict.</p></div>
2027 <div class="listingblock">
2028 <div class="content">
2029 <pre><code>Documentation/git-merge.txt     conflict-marker-size=32</code></pre>
2030 </div></div>
2031 </div>
2032 </div>
2033 <div class="sect2">
2034 <h3 id="_checking_whitespace_errors">Checking whitespace errors</h3>
2035 <div class="sect3">
2036 <h4 id="_code_whitespace_code"><code>whitespace</code></h4>
2037 <div class="paragraph"><p>The <code>core.whitespace</code> configuration variable allows you to define what
2038 <em>diff</em> and <em>apply</em> should consider whitespace errors for all paths in
2039 the project (See <a href="git-config.html">git-config(1)</a>).  This attribute gives you finer
2040 control per path.</p></div>
2041 <div class="dlist"><dl>
2042 <dt class="hdlist1">
2043 Set
2044 </dt>
2045 <dd>
2046 <p>
2047         Notice all types of potential whitespace errors known to Git.
2048         The tab width is taken from the value of the <code>core.whitespace</code>
2049         configuration variable.
2050 </p>
2051 </dd>
2052 <dt class="hdlist1">
2053 Unset
2054 </dt>
2055 <dd>
2056 <p>
2057         Do not notice anything as error.
2058 </p>
2059 </dd>
2060 <dt class="hdlist1">
2061 Unspecified
2062 </dt>
2063 <dd>
2064 <p>
2065         Use the value of the <code>core.whitespace</code> configuration variable to
2066         decide what to notice as error.
2067 </p>
2068 </dd>
2069 <dt class="hdlist1">
2070 String
2071 </dt>
2072 <dd>
2073 <p>
2074         Specify a comma separated list of common whitespace problems to
2075         notice in the same format as the <code>core.whitespace</code> configuration
2076         variable.
2077 </p>
2078 </dd>
2079 </dl></div>
2080 </div>
2081 </div>
2082 <div class="sect2">
2083 <h3 id="_creating_an_archive">Creating an archive</h3>
2084 <div class="sect3">
2085 <h4 id="_code_export_ignore_code"><code>export-ignore</code></h4>
2086 <div class="paragraph"><p>Files and directories with the attribute <code>export-ignore</code> won&#8217;t be added to
2087 archive files.</p></div>
2088 </div>
2089 <div class="sect3">
2090 <h4 id="_code_export_subst_code"><code>export-subst</code></h4>
2091 <div class="paragraph"><p>If the attribute <code>export-subst</code> is set for a file then Git will expand
2092 several placeholders when adding this file to an archive.  The
2093 expansion depends on the availability of a commit ID, i.e., if
2094 <a href="git-archive.html">git-archive(1)</a> has been given a tree instead of a commit or a
2095 tag then no replacement will be done.  The placeholders are the same
2096 as those for the option <code>--pretty=format:</code> of <a href="git-log.html">git-log(1)</a>,
2097 except that they need to be wrapped like this: <code>$Format:PLACEHOLDERS$</code>
2098 in the file.  E.g. the string <code>$Format:%H$</code> will be replaced by the
2099 commit hash.  However, only one <code>%(describe)</code> placeholder is expanded
2100 per archive to avoid denial-of-service attacks.</p></div>
2101 </div>
2102 </div>
2103 <div class="sect2">
2104 <h3 id="_packing_objects">Packing objects</h3>
2105 <div class="sect3">
2106 <h4 id="_code_delta_code"><code>delta</code></h4>
2107 <div class="paragraph"><p>Delta compression will not be attempted for blobs for paths with the
2108 attribute <code>delta</code> set to false.</p></div>
2109 </div>
2110 </div>
2111 <div class="sect2">
2112 <h3 id="_viewing_files_in_gui_tools">Viewing files in GUI tools</h3>
2113 <div class="sect3">
2114 <h4 id="_code_encoding_code"><code>encoding</code></h4>
2115 <div class="paragraph"><p>The value of this attribute specifies the character encoding that should
2116 be used by GUI tools (e.g. <a href="gitk.html">gitk(1)</a> and <a href="git-gui.html">git-gui(1)</a>) to
2117 display the contents of the relevant file. Note that due to performance
2118 considerations <a href="gitk.html">gitk(1)</a> does not use this attribute unless you
2119 manually enable per-file encodings in its options.</p></div>
2120 <div class="paragraph"><p>If this attribute is not set or has an invalid value, the value of the
2121 <code>gui.encoding</code> configuration variable is used instead
2122 (See <a href="git-config.html">git-config(1)</a>).</p></div>
2123 </div>
2124 </div>
2125 </div>
2126 </div>
2127 <div class="sect1">
2128 <h2 id="_using_macro_attributes">USING MACRO ATTRIBUTES</h2>
2129 <div class="sectionbody">
2130 <div class="paragraph"><p>You do not want any end-of-line conversions applied to, nor textual diffs
2131 produced for, any binary file you track.  You would need to specify e.g.</p></div>
2132 <div class="listingblock">
2133 <div class="content">
2134 <pre><code>*.jpg -text -diff</code></pre>
2135 </div></div>
2136 <div class="paragraph"><p>but that may become cumbersome, when you have many attributes.  Using
2137 macro attributes, you can define an attribute that, when set, also
2138 sets or unsets a number of other attributes at the same time.  The
2139 system knows a built-in macro attribute, <code>binary</code>:</p></div>
2140 <div class="listingblock">
2141 <div class="content">
2142 <pre><code>*.jpg binary</code></pre>
2143 </div></div>
2144 <div class="paragraph"><p>Setting the "binary" attribute also unsets the "text" and "diff"
2145 attributes as above.  Note that macro attributes can only be "Set",
2146 though setting one might have the effect of setting or unsetting other
2147 attributes or even returning other attributes to the "Unspecified"
2148 state.</p></div>
2149 </div>
2150 </div>
2151 <div class="sect1">
2152 <h2 id="_defining_macro_attributes">DEFINING MACRO ATTRIBUTES</h2>
2153 <div class="sectionbody">
2154 <div class="paragraph"><p>Custom macro attributes can be defined only in top-level gitattributes
2155 files (<code>$GIT_DIR/info/attributes</code>, the <code>.gitattributes</code> file at the
2156 top level of the working tree, or the global or system-wide
2157 gitattributes files), not in <code>.gitattributes</code> files in working tree
2158 subdirectories.  The built-in macro attribute "binary" is equivalent
2159 to:</p></div>
2160 <div class="listingblock">
2161 <div class="content">
2162 <pre><code>[attr]binary -diff -merge -text</code></pre>
2163 </div></div>
2164 </div>
2165 </div>
2166 <div class="sect1">
2167 <h2 id="_notes">NOTES</h2>
2168 <div class="sectionbody">
2169 <div class="paragraph"><p>Git does not follow symbolic links when accessing a <code>.gitattributes</code>
2170 file in the working tree. This keeps behavior consistent when the file
2171 is accessed from the index or a tree versus from the filesystem.</p></div>
2172 </div>
2173 </div>
2174 <div class="sect1">
2175 <h2 id="_examples">EXAMPLES</h2>
2176 <div class="sectionbody">
2177 <div class="paragraph"><p>If you have these three <code>gitattributes</code> file:</p></div>
2178 <div class="listingblock">
2179 <div class="content">
2180 <pre><code>(in $GIT_DIR/info/attributes)
2181
2182 a*      foo !bar -baz
2183
2184 (in .gitattributes)
2185 abc     foo bar baz
2186
2187 (in t/.gitattributes)
2188 ab*     merge=filfre
2189 abc     -foo -bar
2190 *.c     frotz</code></pre>
2191 </div></div>
2192 <div class="paragraph"><p>the attributes given to path <code>t/abc</code> are computed as follows:</p></div>
2193 <div class="olist arabic"><ol class="arabic">
2194 <li>
2195 <p>
2196 By examining <code>t/.gitattributes</code> (which is in the same
2197    directory as the path in question), Git finds that the first
2198    line matches.  <code>merge</code> attribute is set.  It also finds that
2199    the second line matches, and attributes <code>foo</code> and <code>bar</code>
2200    are unset.
2201 </p>
2202 </li>
2203 <li>
2204 <p>
2205 Then it examines <code>.gitattributes</code> (which is in the parent
2206    directory), and finds that the first line matches, but
2207    <code>t/.gitattributes</code> file already decided how <code>merge</code>, <code>foo</code>
2208    and <code>bar</code> attributes should be given to this path, so it
2209    leaves <code>foo</code> and <code>bar</code> unset.  Attribute <code>baz</code> is set.
2210 </p>
2211 </li>
2212 <li>
2213 <p>
2214 Finally it examines <code>$GIT_DIR/info/attributes</code>.  This file
2215    is used to override the in-tree settings.  The first line is
2216    a match, and <code>foo</code> is set, <code>bar</code> is reverted to unspecified
2217    state, and <code>baz</code> is unset.
2218 </p>
2219 </li>
2220 </ol></div>
2221 <div class="paragraph"><p>As the result, the attributes assignment to <code>t/abc</code> becomes:</p></div>
2222 <div class="listingblock">
2223 <div class="content">
2224 <pre><code>foo     set to true
2225 bar     unspecified
2226 baz     set to false
2227 merge   set to string value "filfre"
2228 frotz   unspecified</code></pre>
2229 </div></div>
2230 </div>
2231 </div>
2232 <div class="sect1">
2233 <h2 id="_see_also">SEE ALSO</h2>
2234 <div class="sectionbody">
2235 <div class="paragraph"><p><a href="git-check-attr.html">git-check-attr(1)</a>.</p></div>
2236 </div>
2237 </div>
2238 <div class="sect1">
2239 <h2 id="_git">GIT</h2>
2240 <div class="sectionbody">
2241 <div class="paragraph"><p>Part of the <a href="git.html">git(1)</a> suite</p></div>
2242 </div>
2243 </div>
2244 </div>
2245 <div id="footnotes"><hr /></div>
2246 <div id="footer">
2247 <div id="footer-text">
2248 Last updated
2249  2023-06-30 12:48:29 PDT
2250 </div>
2251 </div>
2252 </body>
2253 </html>