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