dedup-{format,sift}.c: Find ways to save space by making hardlinks.

[rsync-backup] / dedup-format.c

diff --git a/dedup-format.c b/dedup-format.c
new file mode 100644 (file)
index 0000000..c3d0bc3
--- /dev/null
+++ b/dedup-format.c
@@ -0,0 +1,70 @@
+#include <errno.h>
+#include <stdio.h>
+#include <string.h>
+
+#include <mLib/dstr.h>
+#include <mLib/macros.h>
+#include <mLib/quis.h>
+#include <mLib/report.h>
+
+struct field {
+  const char *p;
+  size_t n;
+};
+
+int main(int argc, char *argv[])
+{
+  int i, j;
+  FILE *fp;
+  dstr d = DSTR_INIT, dd = DSTR_INIT;
+  const char *fn, *p, *q;
+  size_t tagsz;
+  struct field f[7];
+
+  ego(argv[0]);
+
+  for (i = 1; i < argc; i++) {
+    fn = argv[i];
+    if ((fp = fopen(fn, "r")) == 0)
+      die(1, "fopen(%s): %s", fn, strerror(errno));
+    q = strrchr(fn, '.'); tagsz = q ? q - fn : strlen(fn);
+
+    for (;;) {
+      DRESET(&d); DRESET(&dd);
+      if (dstr_putline(&d, fp) == EOF) break;
+
+      if (!d.len || d.buf[0] == '*' || d.buf[0] == '[') continue;
+
+      for (p = d.buf, j = 0; j < N(f); j++) {
+       while (*p == ' ') p++;
+       if (!*p) break;
+       f[j].p = p;
+       if (j == N(f) - 1) {
+         f[j++].n = d.buf + d.len - p;
+         break;
+       }
+       while (*p && *p != ' ') p++;
+       if (!*p) break;
+       f[j].n = p - f[j].p;
+      }
+      if (j < N(f)) die(1, "failed to parse input");
+
+      DPUTM(&dd, f[0].p, f[0].n); DPUTC(&dd, ' ');
+      DPUTM(&dd, f[2].p, f[2].n); DPUTC(&dd, ' ');
+      DPUTM(&dd, f[3].p, f[3].n); DPUTC(&dd, ' ');
+      DPUTM(&dd, f[4].p, f[4].n); DPUTC(&dd, ' ');
+      DPUTM(&dd, f[5].p, f[5].n); DPUTC(&dd, '|');
+      DPUTM(&dd, fn, tagsz); DPUTC(&dd, '|');
+      DPUTM(&dd, f[6].p, f[6].n); DPUTC(&dd, '\n');
+      fwrite(dd.buf, 1, dd.len, stdout);
+    }
+
+    if (ferror(fp)) die(1, "ferror(%s)", fn);
+    fclose(fp);
+  }
+
+  if (ferror(stdout) || fflush(stdout) || fclose(stdout))
+    die(1, "stdout: %s", strerror(errno));
+
+  return (0);
+}