]>
Commit | Line | Data |
---|---|---|
6c510bee | 1 | #include "cache.h" |
35ebfd6a JH |
2 | #include "attr.h" |
3 | ||
6c510bee LT |
4 | /* |
5 | * convert.c - convert a file when checking it out and checking it in. | |
6 | * | |
7 | * This should use the pathname to decide on whether it wants to do some | |
8 | * more interesting conversions (automatic gzip/unzip, general format | |
9 | * conversions etc etc), but by default it just does automatic CRLF<->LF | |
10 | * translation when the "auto_crlf" option is set. | |
11 | */ | |
12 | ||
13 | struct text_stat { | |
14 | /* CR, LF and CRLF counts */ | |
15 | unsigned cr, lf, crlf; | |
16 | ||
17 | /* These are just approximations! */ | |
18 | unsigned printable, nonprintable; | |
19 | }; | |
20 | ||
21 | static void gather_stats(const char *buf, unsigned long size, struct text_stat *stats) | |
22 | { | |
23 | unsigned long i; | |
24 | ||
25 | memset(stats, 0, sizeof(*stats)); | |
26 | ||
27 | for (i = 0; i < size; i++) { | |
28 | unsigned char c = buf[i]; | |
29 | if (c == '\r') { | |
30 | stats->cr++; | |
31 | if (i+1 < size && buf[i+1] == '\n') | |
32 | stats->crlf++; | |
33 | continue; | |
34 | } | |
35 | if (c == '\n') { | |
36 | stats->lf++; | |
37 | continue; | |
38 | } | |
39 | if (c == 127) | |
40 | /* DEL */ | |
41 | stats->nonprintable++; | |
42 | else if (c < 32) { | |
43 | switch (c) { | |
44 | /* BS, HT, ESC and FF */ | |
45 | case '\b': case '\t': case '\033': case '\014': | |
46 | stats->printable++; | |
47 | break; | |
48 | default: | |
49 | stats->nonprintable++; | |
50 | } | |
51 | } | |
52 | else | |
53 | stats->printable++; | |
54 | } | |
55 | } | |
56 | ||
57 | /* | |
58 | * The same heuristics as diff.c::mmfile_is_binary() | |
59 | */ | |
60 | static int is_binary(unsigned long size, struct text_stat *stats) | |
61 | { | |
62 | ||
63 | if ((stats->printable >> 7) < stats->nonprintable) | |
64 | return 1; | |
65 | /* | |
66 | * Other heuristics? Average line length might be relevant, | |
67 | * as might LF vs CR vs CRLF counts.. | |
68 | * | |
69 | * NOTE! It might be normal to have a low ratio of CRLF to LF | |
70 | * (somebody starts with a LF-only file and edits it with an editor | |
71 | * that adds CRLF only to lines that are added..). But do we | |
72 | * want to support CR-only? Probably not. | |
73 | */ | |
74 | return 0; | |
75 | } | |
76 | ||
35ebfd6a | 77 | static int autocrlf_to_git(const char *path, char **bufp, unsigned long *sizep) |
6c510bee LT |
78 | { |
79 | char *buffer, *nbuf; | |
80 | unsigned long size, nsize; | |
81 | struct text_stat stats; | |
82 | ||
6c510bee LT |
83 | if (!auto_crlf) |
84 | return 0; | |
85 | ||
86 | size = *sizep; | |
87 | if (!size) | |
88 | return 0; | |
89 | buffer = *bufp; | |
90 | ||
91 | gather_stats(buffer, size, &stats); | |
92 | ||
93 | /* No CR? Nothing to convert, regardless. */ | |
94 | if (!stats.cr) | |
95 | return 0; | |
96 | ||
97 | /* | |
98 | * We're currently not going to even try to convert stuff | |
99 | * that has bare CR characters. Does anybody do that crazy | |
100 | * stuff? | |
101 | */ | |
102 | if (stats.cr != stats.crlf) | |
103 | return 0; | |
104 | ||
105 | /* | |
106 | * And add some heuristics for binary vs text, of course... | |
107 | */ | |
108 | if (is_binary(size, &stats)) | |
109 | return 0; | |
110 | ||
111 | /* | |
112 | * Ok, allocate a new buffer, fill it in, and return true | |
113 | * to let the caller know that we switched buffers on it. | |
114 | */ | |
115 | nsize = size - stats.crlf; | |
116 | nbuf = xmalloc(nsize); | |
117 | *bufp = nbuf; | |
118 | *sizep = nsize; | |
119 | do { | |
120 | unsigned char c = *buffer++; | |
121 | if (c != '\r') | |
122 | *nbuf++ = c; | |
123 | } while (--size); | |
124 | ||
125 | return 1; | |
126 | } | |
127 | ||
35ebfd6a | 128 | static int autocrlf_to_working_tree(const char *path, char **bufp, unsigned long *sizep) |
6c510bee LT |
129 | { |
130 | char *buffer, *nbuf; | |
131 | unsigned long size, nsize; | |
132 | struct text_stat stats; | |
133 | unsigned char last; | |
134 | ||
135 | /* | |
136 | * FIXME! Other pluggable conversions should go here, | |
137 | * based on filename patterns. Right now we just do the | |
138 | * stupid auto-CRLF one. | |
139 | */ | |
d7f46334 | 140 | if (auto_crlf <= 0) |
6c510bee LT |
141 | return 0; |
142 | ||
143 | size = *sizep; | |
144 | if (!size) | |
145 | return 0; | |
146 | buffer = *bufp; | |
147 | ||
148 | gather_stats(buffer, size, &stats); | |
149 | ||
150 | /* No LF? Nothing to convert, regardless. */ | |
151 | if (!stats.lf) | |
152 | return 0; | |
153 | ||
154 | /* Was it already in CRLF format? */ | |
155 | if (stats.lf == stats.crlf) | |
156 | return 0; | |
157 | ||
158 | /* If we have any bare CR characters, we're not going to touch it */ | |
159 | if (stats.cr != stats.crlf) | |
160 | return 0; | |
161 | ||
162 | if (is_binary(size, &stats)) | |
163 | return 0; | |
164 | ||
165 | /* | |
166 | * Ok, allocate a new buffer, fill it in, and return true | |
167 | * to let the caller know that we switched buffers on it. | |
168 | */ | |
169 | nsize = size + stats.lf - stats.crlf; | |
170 | nbuf = xmalloc(nsize); | |
171 | *bufp = nbuf; | |
172 | *sizep = nsize; | |
173 | last = 0; | |
174 | do { | |
175 | unsigned char c = *buffer++; | |
176 | if (c == '\n' && last != '\r') | |
177 | *nbuf++ = '\r'; | |
178 | *nbuf++ = c; | |
179 | last = c; | |
180 | } while (--size); | |
181 | ||
182 | return 1; | |
183 | } | |
35ebfd6a JH |
184 | |
185 | static void setup_crlf_check(struct git_attr_check *check) | |
186 | { | |
187 | static struct git_attr *attr_crlf; | |
188 | ||
189 | if (!attr_crlf) | |
190 | attr_crlf = git_attr("crlf", 4); | |
191 | check->attr = attr_crlf; | |
192 | } | |
193 | ||
194 | static int git_path_is_binary(const char *path) | |
195 | { | |
196 | struct git_attr_check attr_crlf_check; | |
197 | ||
198 | setup_crlf_check(&attr_crlf_check); | |
199 | ||
200 | /* | |
201 | * If crlf is not mentioned, default to autocrlf; | |
202 | * disable autocrlf only when crlf attribute is explicitly | |
203 | * unset. | |
204 | */ | |
205 | return (!git_checkattr(path, 1, &attr_crlf_check) && | |
206 | (0 == attr_crlf_check.isset)); | |
207 | } | |
208 | ||
209 | int convert_to_git(const char *path, char **bufp, unsigned long *sizep) | |
210 | { | |
211 | if (git_path_is_binary(path)) | |
212 | return 0; | |
213 | return autocrlf_to_git(path, bufp, sizep); | |
214 | } | |
215 | ||
216 | int convert_to_working_tree(const char *path, char **bufp, unsigned long *sizep) | |
217 | { | |
218 | if (git_path_is_binary(path)) | |
219 | return 0; | |
220 | return autocrlf_to_working_tree(path, bufp, sizep); | |
221 | } |