-void tokenize(char* raw, long tokens[4])
-{
- // set each token to 0
- for (int i = 0; i < 4; ++i)
- tokens[i] = 0;
-
- // grab the first two words
- char* words[2];
- words[0] = (char*) xmalloc(strlen(raw) + 1);
- words[1] = (char*) xmalloc(strlen(raw) + 1);
- int word_count = sscanf(raw, "%s%s", words[0], words[1]);
-
- // make space for substrings and zero it out
- char chunk_data[][6] = {
- {"\0\0\0\0\0"},
- {"\0\0\0\0\0"},
- {"\0\0\0\0\0"},
- {"\0\0\0\0\0"},
- };
-
- // break the words into up to 4 5-char substrings
- sscanf(words[0], "%5s%5s", chunk_data[0], chunk_data[1]);
- if (word_count == 2)
- sscanf(words[1], "%5s%5s", chunk_data[2], chunk_data[3]);
- free(words[0]);
- free(words[1]);
-
- // uppercase all the substrings
- for (int i = 0; i < 4; ++i)
- for (unsigned int j = 0; j < strlen(chunk_data[i]); ++j)
- chunk_data[i][j] = (char) toupper(chunk_data[i][j]);
-
- // pack the substrings
- for (int i = 0; i < 4; ++i)
- tokens[i] = token_to_packed(chunk_data[i]);
+void tokenize(char* raw, struct command_t *cmd)
+{
+ memset(cmd, '\0', sizeof(struct command_t));
+
+ /* Bound prefix on the %s would be needed to prevent buffer
+ * overflow. but we shortstop this more simply by making each
+ * raw-input buffer as long as the enrire inout buffer. */
+ sscanf(raw, "%s%s", cmd->raw1, cmd->raw2);
+
+ /* (ESR) In oldstyle mode, simulate the uppercasing and truncating
+ * effect on raw tokens of packing them into sixbit characters, 5
+ * to a 32-bit word. This is something the FORTRAN version did
+ * becuse archaic FORTRAN had no string types. Don Wood's
+ * mechanical translation of 2.5 to C retained the packing and
+ * thus this misfeature.
+ *
+ * It's philosophically questionable whether this is the right
+ * thing to do even in oldstyle mode. On one hand, the text
+ * mangling was not authorial intent, but a result of limitations
+ * in their tools. On the other, not simulating this misbehavior
+ * goes against the goal of making oldstyle as accurate as
+ * possible an emulation of the original UI.
+ */
+ if (settings.oldstyle) {
+ cmd->raw1[TOKLEN + TOKLEN] = cmd->raw2[TOKLEN + TOKLEN] = '\0';
+ for (size_t i = 0; i < strlen(cmd->raw1); i++)
+ cmd->raw1[i] = toupper(cmd->raw1[i]);
+ for (size_t i = 0; i < strlen(cmd->raw2); i++)
+ cmd->raw2[i] = toupper(cmd->raw2[i]);
+ }