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

   1 #include "cache.h"
   2 #include "config.h"
   3 #include "userdiff.h"
   4 #include "attr.h"
   5
   6 static struct userdiff_driver *drivers;
   7 static int ndrivers;
   8 static int drivers_alloc;
   9
  10 #define PATTERNS(name, pattern, word_regex)                     \
  11         { name, NULL, -1, { pattern, REG_EXTENDED },            \
  12           word_regex "|[^[:space:]]|[\xc0-\xff][\x80-\xbf]+" }
  13 #define IPATTERN(name, pattern, word_regex)                     \
  14         { name, NULL, -1, { pattern, REG_EXTENDED | REG_ICASE }, \
  15           word_regex "|[^[:space:]]|[\xc0-\xff][\x80-\xbf]+" }
  16 static struct userdiff_driver builtin_drivers[] = {
  17 IPATTERN("ada",
  18          "!^(.*[ \t])?(is[ \t]+new|renames|is[ \t]+separate)([ \t].*)?$\n"
  19          "!^[ \t]*with[ \t].*$\n"
  20          "^[ \t]*((procedure|function)[ \t]+.*)$\n"
  21          "^[ \t]*((package|protected|task)[ \t]+.*)$",
  22          /* -- */
  23          "[a-zA-Z][a-zA-Z0-9_]*"
  24          "|[-+]?[0-9][0-9#_.aAbBcCdDeEfF]*([eE][+-]?[0-9_]+)?"
  25          "|=>|\\.\\.|\\*\\*|:=|/=|>=|<=|<<|>>|<>"),
  26 IPATTERN("fortran",
  27          "!^([C*]|[ \t]*!)\n"
  28          "!^[ \t]*MODULE[ \t]+PROCEDURE[ \t]\n"
  29          "^[ \t]*((END[ \t]+)?(PROGRAM|MODULE|BLOCK[ \t]+DATA"
  30                 "|([^'\" \t]+[ \t]+)*(SUBROUTINE|FUNCTION))[ \t]+[A-Z].*)$",
  31          /* -- */
  32          "[a-zA-Z][a-zA-Z0-9_]*"
  33          "|\\.([Ee][Qq]|[Nn][Ee]|[Gg][TtEe]|[Ll][TtEe]|[Tt][Rr][Uu][Ee]|[Ff][Aa][Ll][Ss][Ee]|[Aa][Nn][Dd]|[Oo][Rr]|[Nn]?[Ee][Qq][Vv]|[Nn][Oo][Tt])\\."
  34          /* numbers and format statements like 2E14.4, or ES12.6, 9X.
  35           * Don't worry about format statements without leading digits since
  36           * they would have been matched above as a variable anyway. */
  37          "|[-+]?[0-9.]+([AaIiDdEeFfLlTtXx][Ss]?[-+]?[0-9.]*)?(_[a-zA-Z0-9][a-zA-Z0-9_]*)?"
  38          "|//|\\*\\*|::|[/<>=]="),
  39 IPATTERN("fountain", "^((\\.[^.]|(int|ext|est|int\\.?/ext|i/e)[. ]).*)$",
  40          "[^ \t-]+"),
  41 PATTERNS("golang",
  42          /* Functions */
  43          "^[ \t]*(func[ \t]*.*(\\{[ \t]*)?)\n"
  44          /* Structs and interfaces */
  45          "^[ \t]*(type[ \t].*(struct|interface)[ \t]*(\\{[ \t]*)?)",
  46          /* -- */
  47          "[a-zA-Z_][a-zA-Z0-9_]*"
  48          "|[-+0-9.eE]+i?|0[xX]?[0-9a-fA-F]+i?"
  49          "|[-+*/<>%&^|=!:]=|--|\\+\\+|<<=?|>>=?|&\\^=?|&&|\\|\\||<-|\\.{3}"),
  50 PATTERNS("html", "^[ \t]*(<[Hh][1-6]([ \t].*)?>.*)$",
  51          "[^<>= \t]+"),
  52 PATTERNS("java",
  53          "!^[ \t]*(catch|do|for|if|instanceof|new|return|switch|throw|while)\n"
  54          "^[ \t]*(([A-Za-z_][A-Za-z_0-9]*[ \t]+)+[A-Za-z_][A-Za-z_0-9]*[ \t]*\\([^;]*)$",
  55          /* -- */
  56          "[a-zA-Z_][a-zA-Z0-9_]*"
  57          "|[-+0-9.e]+[fFlL]?|0[xXbB]?[0-9a-fA-F]+[lL]?"
  58          "|[-+*/<>%&^|=!]="
  59          "|--|\\+\\+|<<=?|>>>?=?|&&|\\|\\|"),
  60 PATTERNS("matlab",
  61          "^[[:space:]]*((classdef|function)[[:space:]].*)$|^%%[[:space:]].*$",
  62          "[a-zA-Z_][a-zA-Z0-9_]*|[-+0-9.e]+|[=~<>]=|\\.[*/\\^']|\\|\\||&&"),
  63 PATTERNS("objc",
  64          /* Negate C statements that can look like functions */
  65          "!^[ \t]*(do|for|if|else|return|switch|while)\n"
  66          /* Objective-C methods */
  67          "^[ \t]*([-+][ \t]*\\([ \t]*[A-Za-z_][A-Za-z_0-9* \t]*\\)[ \t]*[A-Za-z_].*)$\n"
  68          /* C functions */
  69          "^[ \t]*(([A-Za-z_][A-Za-z_0-9]*[ \t]+)+[A-Za-z_][A-Za-z_0-9]*[ \t]*\\([^;]*)$\n"
  70          /* Objective-C class/protocol definitions */
  71          "^(@(implementation|interface|protocol)[ \t].*)$",
  72          /* -- */
  73          "[a-zA-Z_][a-zA-Z0-9_]*"
  74          "|[-+0-9.e]+[fFlL]?|0[xXbB]?[0-9a-fA-F]+[lL]?"
  75          "|[-+*/<>%&^|=!]=|--|\\+\\+|<<=?|>>=?|&&|\\|\\||::|->"),
  76 PATTERNS("pascal",
  77          "^(((class[ \t]+)?(procedure|function)|constructor|destructor|interface|"
  78                 "implementation|initialization|finalization)[ \t]*.*)$"
  79          "\n"
  80          "^(.*=[ \t]*(class|record).*)$",
  81          /* -- */
  82          "[a-zA-Z_][a-zA-Z0-9_]*"
  83          "|[-+0-9.e]+|0[xXbB]?[0-9a-fA-F]+"
  84          "|<>|<=|>=|:=|\\.\\."),
  85 PATTERNS("perl",
  86          "^package .*\n"
  87          "^sub [[:alnum:]_':]+[ \t]*"
  88                 "(\\([^)]*\\)[ \t]*)?" /* prototype */
  89                 /*
  90                  * Attributes.  A regex can't count nested parentheses,
  91                  * so just slurp up whatever we see, taking care not
  92                  * to accept lines like "sub foo; # defined elsewhere".
  93                  *
  94                  * An attribute could contain a semicolon, but at that
  95                  * point it seems reasonable enough to give up.
  96                  */
  97                 "(:[^;#]*)?"
  98                 "(\\{[ \t]*)?" /* brace can come here or on the next line */
  99                 "(#.*)?$\n" /* comment */
 100          "^(BEGIN|END|INIT|CHECK|UNITCHECK|AUTOLOAD|DESTROY)[ \t]*"
 101                 "(\\{[ \t]*)?" /* brace can come here or on the next line */
 102                 "(#.*)?$\n"
 103          "^=head[0-9] .*",      /* POD */
 104          /* -- */
 105          "[[:alpha:]_'][[:alnum:]_']*"
 106          "|0[xb]?[0-9a-fA-F_]*"
 107          /* taking care not to interpret 3..5 as (3.)(.5) */
 108          "|[0-9a-fA-F_]+(\\.[0-9a-fA-F_]+)?([eE][-+]?[0-9_]+)?"
 109          "|=>|-[rwxoRWXOezsfdlpSugkbctTBMAC>]|~~|::"
 110          "|&&=|\\|\\|=|//=|\\*\\*="
 111          "|&&|\\|\\||//|\\+\\+|--|\\*\\*|\\.\\.\\.?"
 112          "|[-+*/%.^&<>=!|]="
 113          "|=~|!~"
 114          "|<<|<>|<=>|>>"),
 115 PATTERNS("php",
 116          "^[\t ]*(((public|protected|private|static)[\t ]+)*function.*)$\n"
 117          "^[\t ]*((((final|abstract)[\t ]+)?class|interface|trait).*)$",
 118          /* -- */
 119          "[a-zA-Z_][a-zA-Z0-9_]*"
 120          "|[-+0-9.e]+|0[xXbB]?[0-9a-fA-F]+"
 121          "|[-+*/<>%&^|=!.]=|--|\\+\\+|<<=?|>>=?|===|&&|\\|\\||::|->"),
 122 PATTERNS("python", "^[ \t]*((class|def)[ \t].*)$",
 123          /* -- */
 124          "[a-zA-Z_][a-zA-Z0-9_]*"
 125          "|[-+0-9.e]+[jJlL]?|0[xX]?[0-9a-fA-F]+[lL]?"
 126          "|[-+*/<>%&^|=!]=|//=?|<<=?|>>=?|\\*\\*=?"),
 127          /* -- */
 128 PATTERNS("ruby", "^[ \t]*((class|module|def)[ \t].*)$",
 129          /* -- */
 130          "(@|@@|\\$)?[a-zA-Z_][a-zA-Z0-9_]*"
 131          "|[-+0-9.e]+|0[xXbB]?[0-9a-fA-F]+|\\?(\\\\C-)?(\\\\M-)?."
 132          "|//=?|[-+*/<>%&^|=!]=|<<=?|>>=?|===|\\.{1,3}|::|[!=]~"),
 133 PATTERNS("rust",
 134          "^[\t ]*((pub(\\([^\\)]+\\))?[\t ]+)?((async|const|unsafe|extern([\t ]+\"[^\"]+\"))[\t ]+)?(struct|enum|union|mod|trait|fn|impl(<.+>)?)[ \t]+[^;]*)$",
 135          /* -- */
 136          "[a-zA-Z_][a-zA-Z0-9_]*"
 137          "|[-+_0-9.eE]+(f32|f64|u8|u16|u32|u64|u128|usize|i8|i16|i32|i64|i128|isize)?"
 138          "|0[box]?[0-9a-fA-F_]+(u8|u16|u32|u64|u128|usize|i8|i16|i32|i64|i128|isize)?"
 139          "|[-+*\\/<>%&^|=!:]=|<<=?|>>=?|&&|\\|\\||->|=>|\\.{2}=|\\.{3}|::"),
 140 PATTERNS("bibtex", "(@[a-zA-Z]{1,}[ \t]*\\{{0,1}[ \t]*[^ \t\"@',\\#}{~%]*).*$",
 141          "[={}\"]|[^={}\" \t]+"),
 142 PATTERNS("tex", "^(\\\\((sub)*section|chapter|part)\\*{0,1}\\{.*)$",
 143          "\\\\[a-zA-Z@]+|\\\\.|[a-zA-Z0-9\x80-\xff]+"),
 144 PATTERNS("cpp",
 145          /* Jump targets or access declarations */
 146          "!^[ \t]*[A-Za-z_][A-Za-z_0-9]*:[[:space:]]*($|/[/*])\n"
 147          /* functions/methods, variables, and compounds at top level */
 148          "^((::[[:space:]]*)?[A-Za-z_].*)$",
 149          /* -- */
 150          "[a-zA-Z_][a-zA-Z0-9_]*"
 151          "|[-+0-9.e]+[fFlL]?|0[xXbB]?[0-9a-fA-F]+[lLuU]*"
 152          "|[-+*/<>%&^|=!]=|--|\\+\\+|<<=?|>>=?|&&|\\|\\||::|->\\*?|\\.\\*"),
 153 PATTERNS("csharp",
 154          /* Keywords */
 155          "!^[ \t]*(do|while|for|if|else|instanceof|new|return|switch|case|throw|catch|using)\n"
 156          /* Methods and constructors */
 157          "^[ \t]*(((static|public|internal|private|protected|new|virtual|sealed|override|unsafe|async)[ \t]+)*[][<>@.~_[:alnum:]]+[ \t]+[<>@._[:alnum:]]+[ \t]*\\(.*\\))[ \t]*$\n"
 158          /* Properties */
 159          "^[ \t]*(((static|public|internal|private|protected|new|virtual|sealed|override|unsafe)[ \t]+)*[][<>@.~_[:alnum:]]+[ \t]+[@._[:alnum:]]+)[ \t]*$\n"
 160          /* Type definitions */
 161          "^[ \t]*(((static|public|internal|private|protected|new|unsafe|sealed|abstract|partial)[ \t]+)*(class|enum|interface|struct)[ \t]+.*)$\n"
 162          /* Namespace */
 163          "^[ \t]*(namespace[ \t]+.*)$",
 164          /* -- */
 165          "[a-zA-Z_][a-zA-Z0-9_]*"
 166          "|[-+0-9.e]+[fFlL]?|0[xXbB]?[0-9a-fA-F]+[lL]?"
 167          "|[-+*/<>%&^|=!]=|--|\\+\\+|<<=?|>>=?|&&|\\|\\||::|->"),
 168 IPATTERN("css",
 169          "![:;][[:space:]]*$\n"
 170          "^[_a-z0-9].*$",
 171          /* -- */
 172          /*
 173           * This regex comes from W3C CSS specs. Should theoretically also
 174           * allow ISO 10646 characters U+00A0 and higher,
 175           * but they are not handled in this regex.
 176           */
 177          "-?[_a-zA-Z][-_a-zA-Z0-9]*" /* identifiers */
 178          "|-?[0-9]+|\\#[0-9a-fA-F]+" /* numbers */
 179 ),
 180 { "default", NULL, -1, { NULL, 0 } },
 181 };
 182 #undef PATTERNS
 183 #undef IPATTERN
 184
 185 static struct userdiff_driver driver_true = {
 186         "diff=true",
 187         NULL,
 188         0,
 189         { NULL, 0 }
 190 };
 191
 192 static struct userdiff_driver driver_false = {
 193         "!diff",
 194         NULL,
 195         1,
 196         { NULL, 0 }
 197 };
 198
 199 static struct userdiff_driver *userdiff_find_by_namelen(const char *k, int len)
 200 {
 201         int i;
 202         for (i = 0; i < ndrivers; i++) {
 203                 struct userdiff_driver *drv = drivers + i;
 204                 if (!strncmp(drv->name, k, len) && !drv->name[len])
 205                         return drv;
 206         }
 207         for (i = 0; i < ARRAY_SIZE(builtin_drivers); i++) {
 208                 struct userdiff_driver *drv = builtin_drivers + i;
 209                 if (!strncmp(drv->name, k, len) && !drv->name[len])
 210                         return drv;
 211         }
 212         return NULL;
 213 }
 214
 215 static int parse_funcname(struct userdiff_funcname *f, const char *k,
 216                 const char *v, int cflags)
 217 {
 218         if (git_config_string(&f->pattern, k, v) < 0)
 219                 return -1;
 220         f->cflags = cflags;
 221         return 0;
 222 }
 223
 224 static int parse_tristate(int *b, const char *k, const char *v)
 225 {
 226         if (v && !strcasecmp(v, "auto"))
 227                 *b = -1;
 228         else
 229                 *b = git_config_bool(k, v);
 230         return 0;
 231 }
 232
 233 static int parse_bool(int *b, const char *k, const char *v)
 234 {
 235         *b = git_config_bool(k, v);
 236         return 0;
 237 }
 238
 239 int userdiff_config(const char *k, const char *v)
 240 {
 241         struct userdiff_driver *drv;
 242         const char *name, *type;
 243         int namelen;
 244
 245         if (parse_config_key(k, "diff", &name, &namelen, &type) || !name)
 246                 return 0;
 247
 248         drv = userdiff_find_by_namelen(name, namelen);
 249         if (!drv) {
 250                 ALLOC_GROW(drivers, ndrivers+1, drivers_alloc);
 251                 drv = &drivers[ndrivers++];
 252                 memset(drv, 0, sizeof(*drv));
 253                 drv->name = xmemdupz(name, namelen);
 254                 drv->binary = -1;
 255         }
 256
 257         if (!strcmp(type, "funcname"))
 258                 return parse_funcname(&drv->funcname, k, v, 0);
 259         if (!strcmp(type, "xfuncname"))
 260                 return parse_funcname(&drv->funcname, k, v, REG_EXTENDED);
 261         if (!strcmp(type, "binary"))
 262                 return parse_tristate(&drv->binary, k, v);
 263         if (!strcmp(type, "command"))
 264                 return git_config_string(&drv->external, k, v);
 265         if (!strcmp(type, "textconv"))
 266                 return git_config_string(&drv->textconv, k, v);
 267         if (!strcmp(type, "cachetextconv"))
 268                 return parse_bool(&drv->textconv_want_cache, k, v);
 269         if (!strcmp(type, "wordregex"))
 270                 return git_config_string(&drv->word_regex, k, v);
 271
 272         return 0;
 273 }
 274
 275 struct userdiff_driver *userdiff_find_by_name(const char *name)
 276 {
 277         int len = strlen(name);
 278         return userdiff_find_by_namelen(name, len);
 279 }
 280
 281 struct userdiff_driver *userdiff_find_by_path(struct index_state *istate,
 282                                               const char *path)
 283 {
 284         static struct attr_check *check;
 285
 286         if (!check)
 287                 check = attr_check_initl("diff", NULL);
 288         if (!path)
 289                 return NULL;
 290         git_check_attr(istate, path, check);
 291
 292         if (ATTR_TRUE(check->items[0].value))
 293                 return &driver_true;
 294         if (ATTR_FALSE(check->items[0].value))
 295                 return &driver_false;
 296         if (ATTR_UNSET(check->items[0].value))
 297                 return NULL;
 298         return userdiff_find_by_name(check->items[0].value);
 299 }
 300
 301 struct userdiff_driver *userdiff_get_textconv(struct repository *r,
 302                                               struct userdiff_driver *driver)
 303 {
 304         if (!driver->textconv)
 305                 return NULL;
 306
 307         if (driver->textconv_want_cache && !driver->textconv_cache) {
 308                 struct notes_cache *c = xmalloc(sizeof(*c));
 309                 struct strbuf name = STRBUF_INIT;
 310
 311                 strbuf_addf(&name, "textconv/%s", driver->name);
 312                 notes_cache_init(r, c, name.buf, driver->textconv);
 313                 driver->textconv_cache = c;
 314                 strbuf_release(&name);
 315         }
 316
 317         return driver;
 318 }