Move config line parsing function to confline.c

1 files changed, 131 insertions, 0 deletions

diff --git a/src/common/confline.c b/src/common/confline.c
index 3e3d1bd3b6..c65560bbf3 100644
--- a/src/common/confline.c
+++ b/src/common/confline.c
@@ -183,3 +183,134 @@ config_count_key(const config_line_t *a, const char *key)
   return n;
 }
 
+/** Given a string containing part of a configuration file or similar format,
+ * advance past comments and whitespace and try to parse a single line.  If we
+ * parse a line successfully, set *<b>key_out</b> to a new string holding the
+ * key portion and *<b>value_out</b> to a new string holding the value portion
+ * of the line, and return a pointer to the start of the next line.  If we run
+ * out of data, return a pointer to the end of the string.  If we encounter an
+ * error, return NULL and set *<b>err_out</b> (if provided) to an error
+ * message.
+ */
+const char *
+parse_config_line_from_str_verbose(const char *line, char **key_out,
+                                   char **value_out,
+                                   const char **err_out)
+{
+  /*
+    See torrc_format.txt for a description of the (silly) format this parses.
+   */
+  const char *key, *val, *cp;
+  int continuation = 0;
+
+  tor_assert(key_out);
+  tor_assert(value_out);
+
+  *key_out = *value_out = NULL;
+  key = val = NULL;
+  /* Skip until the first keyword. */
+  while (1) {
+    while (TOR_ISSPACE(*line))
+      ++line;
+    if (*line == '#') {
+      while (*line && *line != '\n')
+        ++line;
+    } else {
+      break;
+    }
+  }
+
+  if (!*line) { /* End of string? */
+    *key_out = *value_out = NULL;
+    return line;
+  }
+
+  /* Skip until the next space or \ followed by newline. */
+  key = line;
+  while (*line && !TOR_ISSPACE(*line) && *line != '#' &&
+         ! (line[0] == '\\' && line[1] == '\n'))
+    ++line;
+  *key_out = tor_strndup(key, line-key);
+
+  /* Skip until the value. */
+  while (*line == ' ' || *line == '\t')
+    ++line;
+
+  val = line;
+
+  /* Find the end of the line. */
+  if (*line == '\"') { // XXX No continuation handling is done here
+    if (!(line = unescape_string(line, value_out, NULL))) {
+      if (err_out)
+        *err_out = "Invalid escape sequence in quoted string";
+      return NULL;
+    }
+    while (*line == ' ' || *line == '\t')
+      ++line;
+    if (*line == '\r' && *(++line) == '\n')
+      ++line;
+    if (*line && *line != '#' && *line != '\n') {
+      if (err_out)
+        *err_out = "Excess data after quoted string";
+      return NULL;
+    }
+  } else {
+    /* Look for the end of the line. */
+    while (*line && *line != '\n' && (*line != '#' || continuation)) {
+      if (*line == '\\' && line[1] == '\n') {
+        continuation = 1;
+        line += 2;
+      } else if (*line == '#') {
+        do {
+          ++line;
+        } while (*line && *line != '\n');
+        if (*line == '\n')
+          ++line;
+      } else {
+        ++line;
+      }
+    }
+
+    if (*line == '\n') {
+      cp = line++;
+    } else {
+      cp = line;
+    }
+    /* Now back cp up to be the last nonspace character */
+    while (cp>val && TOR_ISSPACE(*(cp-1)))
+      --cp;
+
+    tor_assert(cp >= val);
+
+    /* Now copy out and decode the value. */
+    *value_out = tor_strndup(val, cp-val);
+    if (continuation) {
+      char *v_out, *v_in;
+      v_out = v_in = *value_out;
+      while (*v_in) {
+        if (*v_in == '#') {
+          do {
+            ++v_in;
+          } while (*v_in && *v_in != '\n');
+          if (*v_in == '\n')
+            ++v_in;
+        } else if (v_in[0] == '\\' && v_in[1] == '\n') {
+          v_in += 2;
+        } else {
+          *v_out++ = *v_in++;
+        }
+      }
+      *v_out = '\0';
+    }
+  }
+
+  if (*line == '#') {
+    do {
+      ++line;
+    } while (*line && *line != '\n');
+  }
+  while (TOR_ISSPACE(*line)) ++line;
+
+  return line;
+}
+