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