t7408: merge short tests, factor out testing method
[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 IPATTERN("css",
152          "![:;][[:space:]]*$\n"
153          "^[_a-z0-9].*$",
154          /* -- */
155          /*
156           * This regex comes from W3C CSS specs. Should theoretically also
157           * allow ISO 10646 characters U+00A0 and higher,
158           * but they are not handled in this regex.
159           */
160          "-?[_a-zA-Z][-_a-zA-Z0-9]*" /* identifiers */
161          "|-?[0-9]+|\\#[0-9a-fA-F]+" /* numbers */
162 ),
163 { "default", NULL, -1, { NULL, 0 } },
164 };
165 #undef PATTERNS
166 #undef IPATTERN
167
168 static struct userdiff_driver driver_true = {
169         "diff=true",
170         NULL,
171         0,
172         { NULL, 0 }
173 };
174
175 static struct userdiff_driver driver_false = {
176         "!diff",
177         NULL,
178         1,
179         { NULL, 0 }
180 };
181
182 static struct userdiff_driver *userdiff_find_by_namelen(const char *k, int len)
183 {
184         int i;
185         for (i = 0; i < ndrivers; i++) {
186                 struct userdiff_driver *drv = drivers + i;
187                 if (!strncmp(drv->name, k, len) && !drv->name[len])
188                         return drv;
189         }
190         for (i = 0; i < ARRAY_SIZE(builtin_drivers); i++) {
191                 struct userdiff_driver *drv = builtin_drivers + i;
192                 if (!strncmp(drv->name, k, len) && !drv->name[len])
193                         return drv;
194         }
195         return NULL;
196 }
197
198 static int parse_funcname(struct userdiff_funcname *f, const char *k,
199                 const char *v, int cflags)
200 {
201         if (git_config_string(&f->pattern, k, v) < 0)
202                 return -1;
203         f->cflags = cflags;
204         return 0;
205 }
206
207 static int parse_tristate(int *b, const char *k, const char *v)
208 {
209         if (v && !strcasecmp(v, "auto"))
210                 *b = -1;
211         else
212                 *b = git_config_bool(k, v);
213         return 0;
214 }
215
216 static int parse_bool(int *b, const char *k, const char *v)
217 {
218         *b = git_config_bool(k, v);
219         return 0;
220 }
221
222 int userdiff_config(const char *k, const char *v)
223 {
224         struct userdiff_driver *drv;
225         const char *name, *type;
226         int namelen;
227
228         if (parse_config_key(k, "diff", &name, &namelen, &type) || !name)
229                 return 0;
230
231         drv = userdiff_find_by_namelen(name, namelen);
232         if (!drv) {
233                 ALLOC_GROW(drivers, ndrivers+1, drivers_alloc);
234                 drv = &drivers[ndrivers++];
235                 memset(drv, 0, sizeof(*drv));
236                 drv->name = xmemdupz(name, namelen);
237                 drv->binary = -1;
238         }
239
240         if (!strcmp(type, "funcname"))
241                 return parse_funcname(&drv->funcname, k, v, 0);
242         if (!strcmp(type, "xfuncname"))
243                 return parse_funcname(&drv->funcname, k, v, REG_EXTENDED);
244         if (!strcmp(type, "binary"))
245                 return parse_tristate(&drv->binary, k, v);
246         if (!strcmp(type, "command"))
247                 return git_config_string(&drv->external, k, v);
248         if (!strcmp(type, "textconv"))
249                 return git_config_string(&drv->textconv, k, v);
250         if (!strcmp(type, "cachetextconv"))
251                 return parse_bool(&drv->textconv_want_cache, k, v);
252         if (!strcmp(type, "wordregex"))
253                 return git_config_string(&drv->word_regex, k, v);
254
255         return 0;
256 }
257
258 struct userdiff_driver *userdiff_find_by_name(const char *name) {
259         int len = strlen(name);
260         return userdiff_find_by_namelen(name, len);
261 }
262
263 struct userdiff_driver *userdiff_find_by_path(const char *path)
264 {
265         static struct git_attr *attr;
266         struct git_attr_check check;
267
268         if (!attr)
269                 attr = git_attr("diff");
270         check.attr = attr;
271
272         if (!path)
273                 return NULL;
274         if (git_check_attr(path, 1, &check))
275                 return NULL;
276
277         if (ATTR_TRUE(check.value))
278                 return &driver_true;
279         if (ATTR_FALSE(check.value))
280                 return &driver_false;
281         if (ATTR_UNSET(check.value))
282                 return NULL;
283         return userdiff_find_by_name(check.value);
284 }
285
286 struct userdiff_driver *userdiff_get_textconv(struct userdiff_driver *driver)
287 {
288         if (!driver->textconv)
289                 return NULL;
290
291         if (driver->textconv_want_cache && !driver->textconv_cache) {
292                 struct notes_cache *c = xmalloc(sizeof(*c));
293                 struct strbuf name = STRBUF_INIT;
294
295                 strbuf_addf(&name, "textconv/%s", driver->name);
296                 notes_cache_init(c, name.buf, driver->textconv);
297                 driver->textconv_cache = c;
298         }
299
300         return driver;
301 }