convert.c

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