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