Fixing a bug which prevented proper end-of-sentence recognition

between an `unformatted' box and the following text.  As a
consequence, vertical line distances are no longer preserved in
boxes after a call to `.unformat' -- because boxes aren't
line-oriented (contrary to diversions), this doesn't make sense
anyway.

* src/roff/troff/node.cc (*node::set_unformat_flag): Add return
value.
(vertical_size_node::set_unformat_flag): New method.
* src/roff/troff/node.hh: Updated.
* src/roff/troff/input.cc (word_space_node::reread,
hmotion_node::reread): Reset `unformat' flag after usage.
(unformat_macro): Append only if `set_unformat_flag()' returns
non-zero.
* src/roff/troff/troff.man: Updated.


* src/roff/troff/troff.man, NEWS, man/groff.man: Improved
documentation of `asciify' and `unformat' requests.

7 files changed, 62 insertions, 17 deletions

diff --git a/ChangeLog b/ChangeLog
index 5e82b960..13da4b72 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,27 @@
+2001-02-16  Werner LEMBERG  <wl@gnu.org>
+
+	Fixing a bug which prevented proper end-of-sentence recognition
+	between an `unformatted' box and the following text.  As a
+	consequence, vertical line distances are no longer preserved in
+	boxes after a call to `.unformat' -- because boxes aren't
+	line-oriented (contrary to diversions), this doesn't make sense
+	anyway.
+
+	* src/roff/troff/node.cc (*node::set_unformat_flag): Add return
+	value.
+	(vertical_size_node::set_unformat_flag): New method.
+	* src/roff/troff/node.hh: Updated.
+	* src/roff/troff/input.cc (word_space_node::reread,
+	hmotion_node::reread): Reset `unformat' flag after usage.
+	(unformat_macro): Append only if `set_unformat_flag()' returns
+	non-zero.
+	* src/roff/troff/troff.man: Updated.
+
+2001-02-15  Werner LEMBERG  <wl@gnu.org>
+
+	* src/roff/troff/troff.man, NEWS, man/groff.man: Improved
+	documentation of `asciify' and `unformat' requests.
+
 2001-02-13  Werner LEMBERG  <wl@gnu.org>
 
 	Redesigned the `unformat' request.  It is no longer connected with
diff --git a/NEWS b/NEWS
index c579ae62..61f02633 100644
--- a/NEWS
+++ b/NEWS
@@ -43,13 +43,13 @@ o `box' and `boxa' are two new requests which behave similarly to `di' and
   ending the diversion).
 
 o The `asciify' request has been extended to `unformat' space characters
-  also.
+  and some other escape sequences also.
 
   `\ ' will no longer be unformatted as a space but remains an unpaddable,
   unbreakable space character.
 
-o The new `unformat' request is similar to `asciify' but only converts
-  (formatted) horizontal space back to its input form, retaining font
+o The new `unformat' request is similar to `asciify' but only handles space
+  characters and tabs specially if the diversion is reread, retaining font
   information. This makes it possible to reformat diversions; for example
   the following
 
diff --git a/man/groff.man b/man/groff.man
index 15332a5a..7b23cba4 100644
--- a/man/groff.man
+++ b/man/groff.man
@@ -873,7 +873,7 @@ to
 .argument stringvar .
 .
 .REQ .asciify diversion
-Unformat ASCII characters and spaces in
+Unformat ASCII characters, spaces, and some escape sequences in
 .argument diversion .
 .
 .REQ .backtrace
@@ -1714,7 +1714,7 @@ Underline (italicize in troff)
 input lines.
 .
 .REQ .unformat diversion
-Unformat horizontal space (no glyphs) in
+Unformat space characters and tabs, preserving font information in
 .argument diversion .
 .REQ .vpt n
 Enable vertical position traps if
diff --git a/src/roff/troff/input.cc b/src/roff/troff/input.cc
index a1673855..85bf4735 100644
--- a/src/roff/troff/input.cc
+++ b/src/roff/troff/input.cc
@@ -2136,6 +2136,7 @@ int word_space_node::reread(int *bolp)
   if (unformat) {
     for (width_list *w = orig_width; w; w = w->next)
       curenv->space(w->width, w->sentence_width);
+    unformat = 0;
     return 1;
   }
   return 0;
@@ -2150,6 +2151,7 @@ int hmotion_node::reread(int *bolp)
 {
   if (unformat && was_tab) {
     curenv->handle_tab(0);
+    unformat = 0;
     return 1;
   }
   return 0;
@@ -3748,8 +3750,8 @@ void unformat_macro()
 	if (c != 0)
 	  am.append(c);
 	else {
-	  nd->set_unformat_flag();
-	  am.append(nd);
+	  if (nd->set_unformat_flag())
+	    am.append(nd);
 	}
       }
       *m = am;
diff --git a/src/roff/troff/node.cc b/src/roff/troff/node.cc
index 02db5b56..e3e37777 100644
--- a/src/roff/troff/node.cc
+++ b/src/roff/troff/node.cc
@@ -2849,8 +2849,9 @@ vunits vmotion_node::vertical_width()
   return n;
 }
 
-void node::set_unformat_flag()
+int node::set_unformat_flag()
 {
+  return 1;
 }
 
 int node::character_type()
@@ -3733,9 +3734,10 @@ node *word_space_node::copy()
   return new word_space_node(n, set, w_new, unformat);
 }
 
-void word_space_node::set_unformat_flag()
+int word_space_node::set_unformat_flag()
 {
   unformat = 1;
+  return 1;
 }
 
 void word_space_node::tprint(troff_output_file *out)
@@ -4396,6 +4398,11 @@ const char *vertical_size_node::type()
   return "vertical_size_node";
 }
 
+int vertical_size_node::set_unformat_flag()
+{
+  return 0;
+}
+
 int vertical_size_node::force_tprint()
 {
   return 0;
@@ -4411,9 +4418,10 @@ const char *hmotion_node::type()
   return "hmotion_node";
 }
 
-void hmotion_node::set_unformat_flag()
+int hmotion_node::set_unformat_flag()
 {
   unformat = 1;
+  return 1;
 }
 
 int hmotion_node::force_tprint()
diff --git a/src/roff/troff/node.h b/src/roff/troff/node.h
index ac80b967..77076e7b 100644
--- a/src/roff/troff/node.h
+++ b/src/roff/troff/node.h
@@ -57,7 +57,7 @@ struct node {
 
   virtual ~node();
   virtual node *copy() = 0;
-  virtual void set_unformat_flag();
+  virtual int set_unformat_flag();
   virtual int force_tprint() = 0;
   virtual hunits width();
   virtual hunits subscript_correction();
@@ -197,7 +197,7 @@ public:
   ~word_space_node();
   node *copy();
   int reread(int *);
-  void set_unformat_flag();
+  int set_unformat_flag();
   void tprint(troff_output_file *);
   int same(node *);
   void asciify(macro *);
@@ -267,6 +267,7 @@ public:
   void set_vertical_size(vertical_size *);
   void asciify(macro *);
   node *copy();
+  int set_unformat_flag();
   int same(node *);
   const char *type();
   int force_tprint();
@@ -284,7 +285,7 @@ public:
   : node(next), n(i), was_tab(flag1), unformat(flag2) {}
   node *copy();
   int reread(int *);
-  void set_unformat_flag();
+  int set_unformat_flag();
   void asciify(macro *);
   void tprint(troff_output_file *);
   hunits width();
diff --git a/src/roff/troff/troff.man b/src/roff/troff/troff.man
index 548533d1..7b57ee08 100644
--- a/src/roff/troff/troff.man
+++ b/src/roff/troff/troff.man
@@ -744,7 +744,8 @@ This request `unformats' the diversion
 .I xx
 in such a way that
 .SM ASCII
-and space characters that were formatted and diverted into
+and space characters (and some escape sequences) that were formatted and
+diverted into
 .I xx
 will be treated like ordinary input characters when
 .I xx
@@ -1553,9 +1554,18 @@ request.
 .TP
 .BI .unformat\  xx
 This request `unformats' the diversion
-.IR xx ,
-converting all horizontal space back into input characters.
-Glyph information (font, font size, etc.) is retained.
+.IR xx .
+Contrary to the
+.B .asciify
+request, which tries to convert formatted elements of the diversion back
+to input tokens as much as possible,
+.B .unformat
+will only handle tabs and spaces between words (usually caused by spaces
+or newlines in the input) specially.
+The former are treated as if they were input tokens, and the latter are
+stretchable again.
+Note that the vertical size of lines is not preserved.
+Glyph information (font, font size, space width, etc.) is retained.
 Useful in conjunction with the
 .B .box
 and