git.oblomov.eu Git - git/blob - userdiff.c

   1 #include "cache.h"
   2 #include "userdiff.h"
   3 #include "attr.h"
   4
   5 static struct userdiff_driver *drivers;
   6 static int ndrivers;
   7 static int drivers_alloc;
   8
   9 #define PATTERNS(name, pattern, word_regex)                     \
  10         { name, NULL, -1, { pattern, REG_EXTENDED },            \
  11           word_regex "|[^[:space:]]|[\xc0-\xff][\x80-\xbf]+" }
  12 #define IPATTERN(name, pattern, word_regex)                     \
  13         { name, NULL, -1, { pattern, REG_EXTENDED | REG_ICASE }, \
  14           word_regex "|[^[:space:]]|[\xc0-\xff][\x80-\xbf]+" }
  15 static struct userdiff_driver builtin_drivers[] = {
  16 IPATTERN("ada",
  17          "!^(.*[ \t])?(is[ \t]+new|renames|is[ \t]+separate)([ \t].*)?$\n"
  18          "!^[ \t]*with[ \t].*$\n"
  19          "^[ \t]*((procedure|function)[ \t]+.*)$\n"
  20          "^[ \t]*((package|protected|task)[ \t]+.*)$",
  21          /* -- */
  22          "[a-zA-Z][a-zA-Z0-9_]*"
  23          "|[-+]?[0-9][0-9#_.aAbBcCdDeEfF]*([eE][+-]?[0-9_]+)?"
  24          "|=>|\\.\\.|\\*\\*|:=|/=|>=|<=|<<|>>|<>"),
  25 IPATTERN("fortran",
  26          "!^([C*]|[ \t]*!)\n"
  27          "!^[ \t]*MODULE[ \t]+PROCEDURE[ \t]\n"
  28          "^[ \t]*((END[ \t]+)?(PROGRAM|MODULE|BLOCK[ \t]+DATA"
  29                 "|([^'\" \t]+[ \t]+)*(SUBROUTINE|FUNCTION))[ \t]+[A-Z].*)$",
  30          /* -- */
  31          "[a-zA-Z][a-zA-Z0-9_]*"
  32          "|\\.([Ee][Qq]|[Nn][Ee]|[Gg][TtEe]|[Ll][TtEe]|[Tt][Rr][Uu][Ee]|[Ff][Aa][Ll][Ss][Ee]|[Aa][Nn][Dd]|[Oo][Rr]|[Nn]?[Ee][Qq][Vv]|[Nn][Oo][Tt])\\."
  33          /* numbers and format statements like 2E14.4, or ES12.6, 9X.
  34           * Don't worry about format statements without leading digits since
  35           * they would have been matched above as a variable anyway. */
  36          "|[-+]?[0-9.]+([AaIiDdEeFfLlTtXx][Ss]?[-+]?[0-9.]*)?(_[a-zA-Z0-9][a-zA-Z0-9_]*)?"
  37          "|//|\\*\\*|::|[/<>=]="),
  38 IPATTERN("fountain", "^((\\.[^.]|(int|ext|est|int\\.?/ext|i/e)[. ]).*)$",
  39          "[^ \t-]+"),
  40 PATTERNS("html", "^[ \t]*(<[Hh][1-6][ \t].*>.*)$",
  41          "[^<>= \t]+"),
  42 PATTERNS("java",
  43          "!^[ \t]*(catch|do|for|if|instanceof|new|return|switch|throw|while)\n"
  44          "^[ \t]*(([A-Za-z_][A-Za-z_0-9]*[ \t]+)+[A-Za-z_][A-Za-z_0-9]*[ \t]*\\([^;]*)$",
  45          /* -- */
  46          "[a-zA-Z_][a-zA-Z0-9_]*"
  47          "|[-+0-9.e]+[fFlL]?|0[xXbB]?[0-9a-fA-F]+[lL]?"
  48          "|[-+*/<>%&^|=!]="
  49          "|--|\\+\\+|<<=?|>>>?=?|&&|\\|\\|"),
  50 PATTERNS("matlab",
  51          "^[[:space:]]*((classdef|function)[[:space:]].*)$|^%%[[:space:]].*$",
  52          "[a-zA-Z_][a-zA-Z0-9_]*|[-+0-9.e]+|[=~<>]=|\\.[*/\\^']|\\|\\||&&"),
  53 PATTERNS("objc",
  54          /* Negate C statements that can look like functions */
  55          "!^[ \t]*(do|for|if|else|return|switch|while)\n"
  56          /* Objective-C methods */
  57          "^[ \t]*([-+][ \t]*\\([ \t]*[A-Za-z_][A-Za-z_0-9* \t]*\\)[ \t]*[A-Za-z_].*)$\n"
  58          /* C functions */
  59          "^[ \t]*(([A-Za-z_][A-Za-z_0-9]*[ \t]+)+[A-Za-z_][A-Za-z_0-9]*[ \t]*\\([^;]*)$\n"
  60          /* Objective-C class/protocol definitions */
  61          "^(@(implementation|interface|protocol)[ \t].*)$",
  62          /* -- */
  63          "[a-zA-Z_][a-zA-Z0-9_]*"
  64          "|[-+0-9.e]+[fFlL]?|0[xXbB]?[0-9a-fA-F]+[lL]?"
  65          "|[-+*/<>%&^|=!]=|--|\\+\\+|<<=?|>>=?|&&|\\|\\||::|->"),
  66 PATTERNS("pascal",
  67          "^(((class[ \t]+)?(procedure|function)|constructor|destructor|interface|"
  68                 "implementation|initialization|finalization)[ \t]*.*)$"
  69          "\n"
  70          "^(.*=[ \t]*(class|record).*)$",
  71          /* -- */
  72          "[a-zA-Z_][a-zA-Z0-9_]*"
  73          "|[-+0-9.e]+|0[xXbB]?[0-9a-fA-F]+"
  74          "|<>|<=|>=|:=|\\.\\."),
  75 PATTERNS("perl",
  76          "^package .*\n"
  77          "^sub [[:alnum:]_':]+[ \t]*"
  78                 "(\\([^)]*\\)[ \t]*)?" /* prototype */
  79                 /*
  80                  * Attributes.  A regex can't count nested parentheses,
  81                  * so just slurp up whatever we see, taking care not
  82                  * to accept lines like "sub foo; # defined elsewhere".
  83                  *
  84                  * An attribute could contain a semicolon, but at that
  85                  * point it seems reasonable enough to give up.
  86                  */
  87                 "(:[^;#]*)?"
  88                 "(\\{[ \t]*)?" /* brace can come here or on the next line */
  89                 "(#.*)?$\n" /* comment */
  90          "^(BEGIN|END|INIT|CHECK|UNITCHECK|AUTOLOAD|DESTROY)[ \t]*"
  91                 "(\\{[ \t]*)?" /* brace can come here or on the next line */
  92                 "(#.*)?$\n"
  93          "^=head[0-9] .*",      /* POD */
  94          /* -- */
  95          "[[:alpha:]_'][[:alnum:]_']*"
  96          "|0[xb]?[0-9a-fA-F_]*"
  97          /* taking care not to interpret 3..5 as (3.)(.5) */
  98          "|[0-9a-fA-F_]+(\\.[0-9a-fA-F_]+)?([eE][-+]?[0-9_]+)?"
  99          "|=>|-[rwxoRWXOezsfdlpSugkbctTBMAC>]|~~|::"
 100          "|&&=|\\|\\|=|//=|\\*\\*="
 101          "|&&|\\|\\||//|\\+\\+|--|\\*\\*|\\.\\.\\.?"
 102          "|[-+*/%.^&<>=!|]="
 103          "|=~|!~"
 104          "|<<|<>|<=>|>>"),
 105 PATTERNS("php",
 106          "^[\t ]*(((public|protected|private|static)[\t ]+)*function.*)$\n"
 107          "^[\t ]*(class.*)$",
 108          /* -- */
 109          "[a-zA-Z_][a-zA-Z0-9_]*"
 110          "|[-+0-9.e]+|0[xXbB]?[0-9a-fA-F]+"
 111          "|[-+*/<>%&^|=!.]=|--|\\+\\+|<<=?|>>=?|===|&&|\\|\\||::|->"),
 112 PATTERNS("python", "^[ \t]*((class|def)[ \t].*)$",
 113          /* -- */
 114          "[a-zA-Z_][a-zA-Z0-9_]*"
 115          "|[-+0-9.e]+[jJlL]?|0[xX]?[0-9a-fA-F]+[lL]?"
 116          "|[-+*/<>%&^|=!]=|//=?|<<=?|>>=?|\\*\\*=?"),
 117          /* -- */
 118 PATTERNS("ruby", "^[ \t]*((class|module|def)[ \t].*)$",
 119          /* -- */
 120          "(@|@@|\\$)?[a-zA-Z_][a-zA-Z0-9_]*"
 121          "|[-+0-9.e]+|0[xXbB]?[0-9a-fA-F]+|\\?(\\\\C-)?(\\\\M-)?."
 122          "|//=?|[-+*/<>%&^|=!]=|<<=?|>>=?|===|\\.{1,3}|::|[!=]~"),
 123 PATTERNS("shell", "^[ \t]*([a-zA-Z_0-9]+[ \t]*\\(\\)).*",
 124          /* -- */
 125          "\\$?[a-zA-Z_0-9]+|&&|\\|\\|"),
 126 PATTERNS("bibtex", "(@[a-zA-Z]{1,}[ \t]*\\{{0,1}[ \t]*[^ \t\"@',\\#}{~%]*).*$",
 127          "[={}\"]|[^={}\" \t]+"),
 128 PATTERNS("tex", "^(\\\\((sub)*section|chapter|part)\\*{0,1}\\{.*)$",
 129          "\\\\[a-zA-Z@]+|\\\\.|[a-zA-Z0-9\x80-\xff]+"),
 130 PATTERNS("cpp",
 131          /* Jump targets or access declarations */
 132          "!^[ \t]*[A-Za-z_][A-Za-z_0-9]*:[[:space:]]*($|/[/*])\n"
 133          /* functions/methods, variables, and compounds at top level */
 134          "^((::[[:space:]]*)?[A-Za-z_].*)$",
 135          /* -- */
 136          "[a-zA-Z_][a-zA-Z0-9_]*"
 137          "|[-+0-9.e]+[fFlL]?|0[xXbB]?[0-9a-fA-F]+[lLuU]*"
 138          "|[-+*/<>%&^|=!]=|--|\\+\\+|<<=?|>>=?|&&|\\|\\||::|->\\*?|\\.\\*"),
 139 PATTERNS("csharp",
 140          /* Keywords */
 141          "!^[ \t]*(do|while|for|if|else|instanceof|new|return|switch|case|throw|catch|using)\n"
 142          /* Methods and constructors */
 143          "^[ \t]*(((static|public|internal|private|protected|new|virtual|sealed|override|unsafe)[ \t]+)*[][<>@.~_[:alnum:]]+[ \t]+[<>@._[:alnum:]]+[ \t]*\\(.*\\))[ \t]*$\n"
 144          /* Properties */
 145          "^[ \t]*(((static|public|internal|private|protected|new|virtual|sealed|override|unsafe)[ \t]+)*[][<>@.~_[:alnum:]]+[ \t]+[@._[:alnum:]]+)[ \t]*$\n"
 146          /* Type definitions */
 147          "^[ \t]*(((static|public|internal|private|protected|new|unsafe|sealed|abstract|partial)[ \t]+)*(class|enum|interface|struct)[ \t]+.*)$\n"
 148          /* Namespace */
 149          "^[ \t]*(namespace[ \t]+.*)$",
 150          /* -- */
 151          "[a-zA-Z_][a-zA-Z0-9_]*"
 152          "|[-+0-9.e]+[fFlL]?|0[xXbB]?[0-9a-fA-F]+[lL]?"
 153          "|[-+*/<>%&^|=!]=|--|\\+\\+|<<=?|>>=?|&&|\\|\\||::|->"),
 154 IPATTERN("css",
 155          "![:;][[:space:]]*$\n"
 156          "^[_a-z0-9].*$",
 157          /* -- */
 158          /*
 159           * This regex comes from W3C CSS specs. Should theoretically also
 160           * allow ISO 10646 characters U+00A0 and higher,
 161           * but they are not handled in this regex.
 162           */
 163          "-?[_a-zA-Z][-_a-zA-Z0-9]*" /* identifiers */
 164          "|-?[0-9]+|\\#[0-9a-fA-F]+" /* numbers */
 165 ),
 166 { "default", NULL, -1, { NULL, 0 } },
 167 };
 168 #undef PATTERNS
 169 #undef IPATTERN
 170
 171 static struct userdiff_driver driver_true = {
 172         "diff=true",
 173         NULL,
 174         0,
 175         { NULL, 0 }
 176 };
 177
 178 static struct userdiff_driver driver_false = {
 179         "!diff",
 180         NULL,
 181         1,
 182         { NULL, 0 }
 183 };
 184
 185 static struct userdiff_driver *userdiff_find_by_namelen(const char *k, int len)
 186 {
 187         int i;
 188         for (i = 0; i < ndrivers; i++) {
 189                 struct userdiff_driver *drv = drivers + i;
 190                 if (!strncmp(drv->name, k, len) && !drv->name[len])
 191                         return drv;
 192         }
 193         for (i = 0; i < ARRAY_SIZE(builtin_drivers); i++) {
 194                 struct userdiff_driver *drv = builtin_drivers + i;
 195                 if (!strncmp(drv->name, k, len) && !drv->name[len])
 196                         return drv;
 197         }
 198         return NULL;
 199 }
 200
 201 static int parse_funcname(struct userdiff_funcname *f, const char *k,
 202                 const char *v, int cflags)
 203 {
 204         if (git_config_string(&f->pattern, k, v) < 0)
 205                 return -1;
 206         f->cflags = cflags;
 207         return 0;
 208 }
 209
 210 static int parse_tristate(int *b, const char *k, const char *v)
 211 {
 212         if (v && !strcasecmp(v, "auto"))
 213                 *b = -1;
 214         else
 215                 *b = git_config_bool(k, v);
 216         return 0;
 217 }
 218
 219 static int parse_bool(int *b, const char *k, const char *v)
 220 {
 221         *b = git_config_bool(k, v);
 222         return 0;
 223 }
 224
 225 int userdiff_config(const char *k, const char *v)
 226 {
 227         struct userdiff_driver *drv;
 228         const char *name, *type;
 229         int namelen;
 230
 231         if (parse_config_key(k, "diff", &name, &namelen, &type) || !name)
 232                 return 0;
 233
 234         drv = userdiff_find_by_namelen(name, namelen);
 235         if (!drv) {
 236                 ALLOC_GROW(drivers, ndrivers+1, drivers_alloc);
 237                 drv = &drivers[ndrivers++];
 238                 memset(drv, 0, sizeof(*drv));
 239                 drv->name = xmemdupz(name, namelen);
 240                 drv->binary = -1;
 241         }
 242
 243         if (!strcmp(type, "funcname"))
 244                 return parse_funcname(&drv->funcname, k, v, 0);
 245         if (!strcmp(type, "xfuncname"))
 246                 return parse_funcname(&drv->funcname, k, v, REG_EXTENDED);
 247         if (!strcmp(type, "binary"))
 248                 return parse_tristate(&drv->binary, k, v);
 249         if (!strcmp(type, "command"))
 250                 return git_config_string(&drv->external, k, v);
 251         if (!strcmp(type, "textconv"))
 252                 return git_config_string(&drv->textconv, k, v);
 253         if (!strcmp(type, "cachetextconv"))
 254                 return parse_bool(&drv->textconv_want_cache, k, v);
 255         if (!strcmp(type, "wordregex"))
 256                 return git_config_string(&drv->word_regex, k, v);
 257
 258         return 0;
 259 }
 260
 261 struct userdiff_driver *userdiff_find_by_name(const char *name) {
 262         int len = strlen(name);
 263         return userdiff_find_by_namelen(name, len);
 264 }
 265
 266 struct userdiff_driver *userdiff_find_by_path(const char *path)
 267 {
 268         static struct git_attr *attr;
 269         struct git_attr_check check;
 270
 271         if (!attr)
 272                 attr = git_attr("diff");
 273         check.attr = attr;
 274
 275         if (!path)
 276                 return NULL;
 277         if (git_check_attr(path, 1, &check))
 278                 return NULL;
 279
 280         if (ATTR_TRUE(check.value))
 281                 return &driver_true;
 282         if (ATTR_FALSE(check.value))
 283                 return &driver_false;
 284         if (ATTR_UNSET(check.value))
 285                 return NULL;
 286         return userdiff_find_by_name(check.value);
 287 }
 288
 289 struct userdiff_driver *userdiff_get_textconv(struct userdiff_driver *driver)
 290 {
 291         if (!driver->textconv)
 292                 return NULL;
 293
 294         if (driver->textconv_want_cache && !driver->textconv_cache) {
 295                 struct notes_cache *c = xmalloc(sizeof(*c));
 296                 struct strbuf name = STRBUF_INIT;
 297
 298                 strbuf_addf(&name, "textconv/%s", driver->name);
 299                 notes_cache_init(c, name.buf, driver->textconv);
 300                 driver->textconv_cache = c;
 301         }
 302
 303         return driver;
 304 }