Merge branch 'rn/glossary-typofix'
[git] / 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("bibtex", "(@[a-zA-Z]{1,}[ \t]*\\{{0,1}[ \t]*[^ \t\"@',\\#}{~%]*).*$",
124          "[={}\"]|[^={}\" \t]+"),
125 PATTERNS("tex", "^(\\\\((sub)*section|chapter|part)\\*{0,1}\\{.*)$",
126          "\\\\[a-zA-Z@]+|\\\\.|[a-zA-Z0-9\x80-\xff]+"),
127 PATTERNS("cpp",
128          /* Jump targets or access declarations */
129          "!^[ \t]*[A-Za-z_][A-Za-z_0-9]*:[[:space:]]*($|/[/*])\n"
130          /* functions/methods, variables, and compounds at top level */
131          "^((::[[:space:]]*)?[A-Za-z_].*)$",
132          /* -- */
133          "[a-zA-Z_][a-zA-Z0-9_]*"
134          "|[-+0-9.e]+[fFlL]?|0[xXbB]?[0-9a-fA-F]+[lLuU]*"
135          "|[-+*/<>%&^|=!]=|--|\\+\\+|<<=?|>>=?|&&|\\|\\||::|->\\*?|\\.\\*"),
136 PATTERNS("csharp",
137          /* Keywords */
138          "!^[ \t]*(do|while|for|if|else|instanceof|new|return|switch|case|throw|catch|using)\n"
139          /* Methods and constructors */
140          "^[ \t]*(((static|public|internal|private|protected|new|virtual|sealed|override|unsafe)[ \t]+)*[][<>@.~_[:alnum:]]+[ \t]+[<>@._[:alnum:]]+[ \t]*\\(.*\\))[ \t]*$\n"
141          /* Properties */
142          "^[ \t]*(((static|public|internal|private|protected|new|virtual|sealed|override|unsafe)[ \t]+)*[][<>@.~_[:alnum:]]+[ \t]+[@._[:alnum:]]+)[ \t]*$\n"
143          /* Type definitions */
144          "^[ \t]*(((static|public|internal|private|protected|new|unsafe|sealed|abstract|partial)[ \t]+)*(class|enum|interface|struct)[ \t]+.*)$\n"
145          /* Namespace */
146          "^[ \t]*(namespace[ \t]+.*)$",
147          /* -- */
148          "[a-zA-Z_][a-zA-Z0-9_]*"
149          "|[-+0-9.e]+[fFlL]?|0[xXbB]?[0-9a-fA-F]+[lL]?"
150          "|[-+*/<>%&^|=!]=|--|\\+\\+|<<=?|>>=?|&&|\\|\\||::|->"),
151 { "default", NULL, -1, { NULL, 0 } },
152 };
153 #undef PATTERNS
154 #undef IPATTERN
155
156 static struct userdiff_driver driver_true = {
157         "diff=true",
158         NULL,
159         0,
160         { NULL, 0 }
161 };
162
163 static struct userdiff_driver driver_false = {
164         "!diff",
165         NULL,
166         1,
167         { NULL, 0 }
168 };
169
170 static struct userdiff_driver *userdiff_find_by_namelen(const char *k, int len)
171 {
172         int i;
173         for (i = 0; i < ndrivers; i++) {
174                 struct userdiff_driver *drv = drivers + i;
175                 if (!strncmp(drv->name, k, len) && !drv->name[len])
176                         return drv;
177         }
178         for (i = 0; i < ARRAY_SIZE(builtin_drivers); i++) {
179                 struct userdiff_driver *drv = builtin_drivers + i;
180                 if (!strncmp(drv->name, k, len) && !drv->name[len])
181                         return drv;
182         }
183         return NULL;
184 }
185
186 static int parse_funcname(struct userdiff_funcname *f, const char *k,
187                 const char *v, int cflags)
188 {
189         if (git_config_string(&f->pattern, k, v) < 0)
190                 return -1;
191         f->cflags = cflags;
192         return 0;
193 }
194
195 static int parse_tristate(int *b, const char *k, const char *v)
196 {
197         if (v && !strcasecmp(v, "auto"))
198                 *b = -1;
199         else
200                 *b = git_config_bool(k, v);
201         return 0;
202 }
203
204 static int parse_bool(int *b, const char *k, const char *v)
205 {
206         *b = git_config_bool(k, v);
207         return 0;
208 }
209
210 int userdiff_config(const char *k, const char *v)
211 {
212         struct userdiff_driver *drv;
213         const char *name, *type;
214         int namelen;
215
216         if (parse_config_key(k, "diff", &name, &namelen, &type) || !name)
217                 return 0;
218
219         drv = userdiff_find_by_namelen(name, namelen);
220         if (!drv) {
221                 ALLOC_GROW(drivers, ndrivers+1, drivers_alloc);
222                 drv = &drivers[ndrivers++];
223                 memset(drv, 0, sizeof(*drv));
224                 drv->name = xmemdupz(name, namelen);
225                 drv->binary = -1;
226         }
227
228         if (!strcmp(type, "funcname"))
229                 return parse_funcname(&drv->funcname, k, v, 0);
230         if (!strcmp(type, "xfuncname"))
231                 return parse_funcname(&drv->funcname, k, v, REG_EXTENDED);
232         if (!strcmp(type, "binary"))
233                 return parse_tristate(&drv->binary, k, v);
234         if (!strcmp(type, "command"))
235                 return git_config_string(&drv->external, k, v);
236         if (!strcmp(type, "textconv"))
237                 return git_config_string(&drv->textconv, k, v);
238         if (!strcmp(type, "cachetextconv"))
239                 return parse_bool(&drv->textconv_want_cache, k, v);
240         if (!strcmp(type, "wordregex"))
241                 return git_config_string(&drv->word_regex, k, v);
242
243         return 0;
244 }
245
246 struct userdiff_driver *userdiff_find_by_name(const char *name) {
247         int len = strlen(name);
248         return userdiff_find_by_namelen(name, len);
249 }
250
251 struct userdiff_driver *userdiff_find_by_path(const char *path)
252 {
253         static struct git_attr *attr;
254         struct git_attr_check check;
255
256         if (!attr)
257                 attr = git_attr("diff");
258         check.attr = attr;
259
260         if (!path)
261                 return NULL;
262         if (git_check_attr(path, 1, &check))
263                 return NULL;
264
265         if (ATTR_TRUE(check.value))
266                 return &driver_true;
267         if (ATTR_FALSE(check.value))
268                 return &driver_false;
269         if (ATTR_UNSET(check.value))
270                 return NULL;
271         return userdiff_find_by_name(check.value);
272 }
273
274 struct userdiff_driver *userdiff_get_textconv(struct userdiff_driver *driver)
275 {
276         if (!driver->textconv)
277                 return NULL;
278
279         if (driver->textconv_want_cache && !driver->textconv_cache) {
280                 struct notes_cache *c = xmalloc(sizeof(*c));
281                 struct strbuf name = STRBUF_INIT;
282
283                 strbuf_addf(&name, "textconv/%s", driver->name);
284                 notes_cache_init(c, name.buf, driver->textconv);
285                 driver->textconv_cache = c;
286         }
287
288         return driver;
289 }