Make `-w' mode work much better. This is INCREDIBLY poorly specified in

any other deroff manual, and as I don't think anybody actually uses
deroff, I don't feel compelled to research its behaviour too much and
can just do what's logical.

diff --git a/demandoc.1 b/demandoc.1
index 30f6d2e7b19cb502eef7c82d51ecc8eaa1a32908..9415fe7016a1ea07928a88fbcd061dce287deeab 100644 (file)
--- a/demandoc.1
+++ b/demandoc.1
@@ -1,4 +1,4 @@
-.\"    $Id: demandoc.1,v 1.2 2011/09/01 20:55:50 kristaps Exp $
+.\"    $Id: demandoc.1,v 1.3 2011/09/01 22:09:50 kristaps Exp $
 .\"
 .\" Copyright (c) 2011 Kristaps Dzonsons <kristaps@bsd.lv>
 .\"
@@ -39,9 +39,10 @@ Its arguments are as follows:
 Output a word list.
 This outputs each word of text on its own line.
 A
-.Qq word
-starts with at least two letters and consists of at least three letters
-total.
+.Qq word ,
+in this case, refers to whitespace-delimited terms beginning with at
+least two letters after opening punctuation and not consisting of any
+escape sequences.
 .It Ar
 The input files.
 .El
@@ -51,12 +52,13 @@ If
 is not provided,
 .Nm
 accepts standard input.
+If a document is not well-formed, it is skipped.
 .Pp
 By default,
 .Nm
 parses its input and outputs only text nodes, preserving line column
 position.
-If a document is not well-formed, it is skipped.
+Escape sequences are omitted from the output.
 .Pp
 The
 .Fl i ,

diff --git a/demandoc.c b/demandoc.c
index 63c7b7bbb11f078733ef3495d9bd1ce65a91f8f6..76548de0e59d1b53728148a8d66d3441ec5f19b4 100644 (file)
--- a/demandoc.c
+++ b/demandoc.c
@@ -1,4 +1,4 @@
-/*     $Id: demandoc.c,v 1.4 2011/09/01 20:55:50 kristaps Exp $ */
+/*     $Id: demandoc.c,v 1.5 2011/09/01 22:09:50 kristaps Exp $ */
 /*
  * Copyright (c) 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -121,7 +121,8 @@ pmandoc(struct mparse *mp, int fd, const char *fn, int list)
        else
                return;
 
-       putchar('\n');
+       if ( ! list)
+               putchar('\n');
 }
 
 /*
@@ -131,12 +132,58 @@ static void
 pstring(const char *p, int col, int *colp, int list)
 {
        enum mandoc_esc  esc;
+       const char      *start;
+       int              emit;
+
+       /*
+        * Print as many column spaces til we achieve parity with the
+        * input document.
+        */
+
+again:
+       if (list && '\0' != *p) {
+               while (isspace((unsigned char)*p))
+                       p++;
+
+               while ('\'' == *p || '(' == *p || '"' == *p)
+                       p++;
+
+               emit = isalpha((unsigned char)p[0]) &&
+                       isalpha((unsigned char)p[1]);
+
+               for (start = p; '\0' != *p; p++)
+                       if ('\\' == *p) {
+                               p++;
+                               esc = mandoc_escape(&p, NULL, NULL);
+                               if (ESCAPE_ERROR == esc)
+                                       return;
+                               emit = 0;
+                       } else if (isspace((unsigned char)*p))
+                               break;
+
+               if (emit && p - start >= 2) {
+                       for ( ; start != p; start++)
+                               if (ASCII_HYPH == *start)
+                                       putchar('-');
+                               else
+                                       putchar((unsigned char)*start);
+                       putchar('\n');
+               }
+
+               if (isspace((unsigned char)*p))
+                       goto again;
+
+               return;
+       }
 
        while (*colp < col) {
                putchar(' ');
                (*colp)++;
        }
 
+       /*
+        * Print the input word, skipping any special characters.
+        */
        while ('\0' != *p) 
                if ('\\' == *p) {
                        p++;
@@ -153,6 +200,14 @@ static void
 pline(int line, int *linep, int *col, int list)
 {
 
+       if (list)
+               return;
+
+       /*
+        * Print out as many lines as needed to reach parity with the
+        * original input. 
+        */
+
        while (*linep < line) {
                putchar('\n');
                (*linep)++;