Fix and rewrite parts of the numbering code

Summary:
- Added the -U option
- Added the -D option to the manages -N/-U and -F sub-options
- Fixed and simplified the -N and -F options
- Cleaned code
- Fixed and improved the manual
- Fixed various small bugs

3 files changed, 549 insertions, 413 deletions

diff --git a/examples/lvm_menu/menu.sh b/examples/lvm_menu/menu.sh
index e3854bb..b1dd724 100755
--- a/examples/lvm_menu/menu.sh
+++ b/examples/lvm_menu/menu.sh
@@ -56,7 +56,7 @@ fi
 REP=$(../../smenu "$TITLE"                 \
             -s /Exit                       \
             -q -d -M -n 30 -c -w           \
-            -N "Exit menu" e:r             \
+            -U "Exit menu"                 \
             -a da:3,b                      \
             -e '^#' -E '/(^#)+(.*)$/ \2/v' \
             -1 'Exit' 2,r                  \
diff --git a/smenu.1 b/smenu.1
index 32eccbb..fe3e30a 100644
--- a/smenu.1
+++ b/smenu.1
@@ -9,7 +9,8 @@ and outputs the selection to stdout.
       [\fB-s\fP \fIpattern\fP] [\fB-m\fP \fImessage\fP] [\fB-w\fP] \\
       [\fB-d\fP] [\fB-M\fP] [\fB-c\fP] [\fB-l\fP] [\fB-r\fP] [\fB-b\fP] \\
       [\fB-a\fP (i|e|c|b|s|t|sf|st|da):\fIATTR\fP]... \\
-      [\fB-i\fP \fIregex\fP] [\fB-e\fP \fIregex\fP] [\fB-C\fP \
+      [\fB-i\fP \fIregex\fP] [\fB-e\fP \fIregex\fP] \\
+      [\fB-C\fP \
 [\fIa\fP|\fIs\fP|\fIi\fP|\fIr\fP|\fId\fP|\fIe\fP]<col selectors>] \\
       [\fB-R\fP \
 [\fIa\fP|\fIs\fP|\fIi\fP|\fIr\fP|\fId\fP|\fIe\fP]<row selectors>] \\
@@ -17,9 +18,8 @@ and outputs the selection to stdout.
       [\fB-I\fP \fI/regex/string/\fP[\fIg\fP][\fIv\fP][\fIs\fP][\fIi\fP]] \\
       [\fB-E\fP \fI/regex/string/\fP[\fIg\fP][\fIv\fP][\fIs\fP][\fIi\fP]] \\
       [\fB-A\fP \fIregex\fP] [\fB-Z\fP \fIregex\fP] \\
-      [\fB-N\fP \fIregex\fP \\
-      [\fB-F\fP \fItype\fP \\
-      [\fB-D\fP [\fIFargument\fP]...] \\
+      [\fB-N\fP \fIregex\fP] [\fB-U\fP \fIregex\fP] [\fB-F\fP] \
+[\fB-D\fP \fIsub-option\fP...] \\
       [\fB-1\fP \fIregex\fP [\fIATTR\fP]] \
 [\fB-2\fP \fIregex\fP [\fIATTR\fP]] ... \
 [\fB-5\fP \fIregex\fP [\fIATTR\fP]] \\
@@ -30,12 +30,14 @@ and outputs the selection to stdout.
 
       <col selectors> ::= \fIcol1\fP[-\fIcol2\fP],...|\fI<RE>\fP,...
       <row selectors> ::= \fIcol1\fP[-\fIcol2\fP],...|\fI<RE>\fP,...
-      <argument>      ::= [l|r:<char>]|[a:l|r]|[p:i|a|[e:m|r]|[w:<num>]
-                          |[o:<num>]|[s:<num>]|[d:<char>]
+      <sub-option>    ::= [l|r:<char>]|[a:left|right]|[p:included|all|
+                          [w:<num>]|[f:yes|no]|[o:<num>]|[n:<num>]|
+                          [i:<num>]
       <ATTR>          ::= [fg][/bg][,style]
-      <RE>            ::= \fB<char>\fIregex\fB<char>\fP
+      <RE>            ::= \fB<char>\fIregex\fB<char>\fR
 
       <col/row selectors> and <RE> can be freely mixed.
+      The parameters of -a and -D must be delimited by blanks.
 .fi
 .SH DESCRIPTION
 This small utility acts as a filter when no input file is given
@@ -99,6 +101,17 @@ keyboard arrow keys (\fB\(<-\fP,\fB\(da\fP,\fB\(ua\fP,\fB\(->\fP)
 or the \fIvi\fP direction keys (\fBh\fP, \fBj\fP, \fBk\fP and \fBl\fP).
 \fBHOME\fP, \fBEND\fP, \fBPgDn\fP and \fBPgUp\fP can also be used, if
 available, and have their traditional meanings.
+
+If the \fB-N\fP, \fB-U\fP or \fB-F\fP are used then it is possible to
+directly access a word by entering its number.
+The numbering created using these option is done \fBbefore\fP any words
+substitution done using \fB-S\fP, \fB-I\fP or \fB-E\fP.
+
+Using a combination of these options, it is easy to control which words
+will be numbered by adding a special symbol in it before using smenu and
+removing it (substituted by nothing) afterward using \fB-I\fP by example.
+
+\fB-E\fP gives another way to do that, see below or more.
 .SS "Searching for a word"
 The key \fB/\fP or the key combination \fB^F\fP (\fBCTRL+f\fP) can be
 used to initiate a search by prefix among the words (possibly altered
@@ -399,6 +412,7 @@ This mode is the same as column mode but without any column alignment.
 Enables \fBENTER\fP to validate the selection even in search mode.
 .IP \fB-b\fP
 Replaces all non-printable characters by a blank.
+If this results in a blank word, it will be potentially deleted.
 .IP "\fB-a \fIPREFIX:ATTR\fP [\fIPREFIX:ATTR\fP...]"
 Sets the display attributes of the elements displayed and the cursor.
 
@@ -590,11 +604,11 @@ The same trick with \fB-L\fP can also be used.
 .RS
 \fI\\u\fP sequences can also be used in the regexp after \fB-Z\fP.
 .RE
-.IP "\fB-N\fP [\fIparameters\fP]"
+.IP "\fB-N\fP [\fIregex\fP]"
 This option allows to number the selectable words matching a specific
 regular expression.
-These numbers are numbered from 1 and provides a direct access to
-the words.
+These numbers are numbered starting from 1 and provides a direct access
+to the words.
 
 To use this functionality, the user must enter the number which
 corresponds to the desired entry digit per digit.
@@ -605,48 +619,111 @@ will start a new number.
 If the number does not exists, then the cursor is restored to it's
 initial position.
 
-These optional parameters must respect the format \fBx\fP:\fBy\fP where
-\fBx\fP can be:
+The sub-options of the \fB-D\fP option described below can change the
+way \fB-N\fP sets and formats the numbers.
+
+This option can be used more than once with cummulative effects.
+
+\fB-N\fP, \fB-U\fP and \fB-F\fP can be mixed.
+.IP "\fB-U\fP [\fIregex\fP]"
+This option allows to un-number words.
+If placed after a previous \fB-N\fP, it can be used to remove the
+numbering of selected words.
+If placed before, the word which doesn't match its regular expression
+will be numbered by default.
+
+This mechanism is similar to to the inclusion/exclusion of words by
+\fB-i\fP and \fB-e\fP.
+
+This option can be used more than once with cummulative effects.
+
+\fB-U\fP, \fB-N\fP and \fB-F\fP can be mixed.
+.IP "\fB-F\fP"
+This option is similar to \fB-N\fP but does not generate a continuous
+flow of numbers but extracts them from the word itself.
+
+With this option you can take full control of the numbering of the
+displayed word.
+Note that the numbering does not need to be ordered.
 
-- '\f(CBl\fP'
+The resulting word after the extraction of the number must be non empty.
+
+Some sub-option are required, see the \fB-D\fP option described below.
+
+\fBNotice\fP that for this option to work correctly, all the embedded
+numbers must have the same number of digits.
+To get that, a preprocessing may be necessary on the words before using
+this program.
+
+\fB-F\fP, \fB-N\fP and \fB-U\fP can be mixed.
+.IP "\fB-D\fP [\fIparameters\fP]"
+This option allows to change the default behaviour or the \fB-N\fP,
+\fB-U\fP and \fB-F\fP options.
+
+Its optional parameters are called sub-options and must respect the
+format \fBx\fP:\fBy\fP where \fBx\fP can be:
+
+- '\f(CBl\fP' (\fB-F\fP, \fB-N\fP and \fB-U\fP options)
+.br
 where \fBy\fP is the UTF-8 character (in native or \fI\\u\fP form)
-to print before the number.
+to print before the number. The default is a single space.
 
-- '\f(CBr\fP'
+- '\f(CBr\fP' (\fB-F\fP, \fB-N\fP and \fB-U\fP options)
+.br
 where \fBy\fP is the UTF-8 character (in native or \fI\\u\fP form)
-to print after the number.
+to print after the number. The default is \f(CB)\fP.
 
-- '\f(CBa\fP'
+- '\f(CBa\fP' (\fB-F\fP, \fB-N\fP and \fB-U\fP options)
+.br
 where \fBy\fP is '\f(CBleft\fP' (or one of its prefixes) if the number
 must be \fIleft\fP aligned, or '\f(CBright\fP' (or one of its prefixes)
-if it must be \fIright\fP aligned.
-
-- '\f(CBp\fP'
-where \fBy\fP is '\f(CBincluded\fP' (or one of its prefixes)
-or '\f(CBall\fP' (or one of its prefixes) for the initial padding of
-the non numbered words. '\f(CBincluded\fP' means that only
-\fIincluded\fP word will be padded while '\f(CBall\fP' means pad
-\fIall\fP words.
-
-- '\f(CBe\fP'
-where \fBy\fP is '\f(CBmatch\fP' (or one of its prefixes) if the words
-must \fImatch\fP the regular expression or '\f(CBinvert\fP' (or one of
-its prefixes) if they must \fBNOT\fP match it.
-
-- '\f(CBw\fP'
-where \fBy\fP is the width of the number between 1 and 5 included.
+if it must be \fIright\fP aligned. The default is \f(CBright\fP.
+
+- '\f(CBp\fP' (\fB-F\fP, \fB-N\fP and \fB-U\fP options)
+.br
+where \fBy\fP is '\f(CBincluded\fP' (or one of its prefixes) 
+or '\f(CBall\fP' (or one of its prefixes) for the initial \fIp\fPadding of
+the non numbered words. '\f(CBincluded\fP' means that only \fIincluded\fP
+word will be padded while '\f(CBall\fP' means pad \fIall\fP words. The
+default is \f(CBall\fP.
+
+- '\f(CBw\fP' (\fB-F\fP, \fB-N\fP and \fB-U\fP options)
+.br
+where \fBy\fP is the \fIw\fPidth of the number between 1 and 5 included.
+
+- '\f(CBf\fP' (\fB-F\fP, \fB-N\fP and \fB-U\fP options)
+.br
+where \fBy\fP controls if the numbering must \fIf\fPollow the last
+extracted number (defaults to \f(CByes\fP) or if it must remain
+independant.
+
+- '\f(CBo\fP' (\fB-F\fP option)
+.br
+where \fBy\fP is the \fIo\fPffset of the first multibyte character of
+the number to extract from the word (defaults to \f(CB0\fP).
+
+- '\f(CBn\fP' (\fB-F\fP option)
+.br
+where \fBy\fP is the \fIn\fPumber of multibyte characters to extract
+from the word starting at the offset given by the \f(CBo\fP sub-option.
+
+- '\f(CBi\fP' (\fB-F\fP option)
+.br
+where \fBy\fP is number of multibyte characters to \fIi\fPgnore after
+the extracted number
 
 Example: \f(CWr:\\> l:\\< a:l\fP
 
-To number all words with the default parameters, use the syntax:
-"\f(CW-N .\fP"
-which is a shortcut for: "\f(CW-N . l:' ' r:')' a:r p:a e:m\fP"
+To number all words with the default parameters, use the
+syntax: "\f(CW-N .\fP" which is a shortcut for:
+"\f(CW-N . l:' ' r:')' a:r p:a\fP"
 
-The \fIpadding\fP directive specifies whether spaces must also be added
-in front of excluded words or not to improve compactness.
+The \fIpadding\fP sub-option specifies whether spaces must also be
+added in front of excluded words or not to improve compactness.
 
-When the \f(CBw\fP optional parameter is not given the width of the
-numbers is determined automatically.
+When the \f(CBw\fP sub-option is not given the width of the numbers is
+determined automatically but if \fB-F\fP is set and the value of the
+\f(CBn\fP sub-option is given then this value is used.
 .IP "\fB-1\fP ... \fB-5\fP \fIregex\fP [\fIATTR\fP]"
 Allows to give up to 5 classes of words specified by regular expressions
 a special display color.
@@ -858,11 +935,11 @@ Imagine a file named \fBsample.mnu\fP with the following content:
 
 .nf
 \f(CR--8<---------------------------------
-[1] "First Entry" [3] "Third entry"
-[2] "Second entry" [4] "Fourth entry"
-@@@ @@@ [5] "Fifth entry"
+"1 First Entry" "3 Third entry"
+"2 Second entry" "4 Fourth entry"
+@@@ "5 Fifth entry"
 @@@
-[Quit] "Exit menu"
+"0 Exit menu"
 --8<---------------------------------
 \fP
 .fi
@@ -870,25 +947,26 @@ Imagine a file named \fBsample.mnu\fP with the following content:
 Then this quite esoteric command will render it (centered on the screen) as:
 
 .nf
-\f(CR+--------------------------------+
-|Test menu                       |
-|1    First Entry  3 Third entry |
-|2    Second entry 4 Fourth entry|
-|                  5 Fifth entry |
-|                                |
-|Quit Exit menu                  |
-+--------------------------------+
+\f(CR+----------------------------------+
+|            Test menu             |
+|                                  |
+| 1) First Entry   3) Third entry  |
+| 2) Second entry  4) Fourth entry |
+|                  5) Fifth entry  |
+|                                  |
+| 0) Exit menu                     |
++----------------------------------+
 \fP
 .fi
 
 with the cursor on \fIQuit\fP and only the numbers and "Quit" selectable.
 
-\f(CBR=$(smenu -q -d -s/Q -M -n 30 -c               \\
-          -e "@+" -E '/@+//'                   \\
-          -i '\\[ *[^ ]+ *\\]' -I '/[][ ]//g'    \\
-          -m "Test menu" < sample.mnu)\fP
+\f(CBR=$(smenu R=$(./smenu -q -d -s/Exit -M -n 30 -c      \\
+                      -e "@+" -E '/@+/ /'            \\
+                      -F -D n:1 i:1                  \\
+                      -m "Test menu"$'\n' < sample.mnu)
 
-The selected number or string will be available in \f(CBR\fP
+The selected entry will be available in \f(CBR\fP
 
 Try to understand it as an exercise.
 .SH ENVIRONMENT
diff --git a/smenu.c b/smenu.c
index e1392c1..b41e42b 100644
--- a/smenu.c
+++ b/smenu.c
@@ -349,12 +349,13 @@ enum filter_types
   EXCLUDE_FILTER
 };
 
+/* Used by the -N -F and -D options */
+/* """""""""""""""""""""""""""""""" */
 enum daccess_types
 {
-  DACCESS_NONE  = 0,
-  DACCESS_AUTO  = 1,
-  DACCESS_POS   = 2,
-  DACCESS_DELIM = 4
+  DA_TYPE_NONE = 0, /* must be 0 (boolean value) */
+  DA_TYPE_AUTO = 1,
+  DA_TYPE_POS  = 2
 };
 
 /* Used when managing the -R option */
@@ -619,16 +620,15 @@ struct daccess_s
   int    type;       /* 1: auto, 2: delimiter, 4: positional               */
   char * left;       /* character to put before the direct access selector */
   char * right;      /* character to put after the direct access selector  */
-  char * delim;      /* field delimiter                                    */
   char   alignment;  /* l: left; r: right                                  */
   char   padding;    /* a: all; i: only included words are padded          */
-  char   expression; /* m: match; r: reverse match the regular expression  */
-  char   remove;     /* y: remove from word, n: don't                      */
   int    length;     /* selector size (5 max)                              */
-  int    start_off;  /* offset to the start of the selector                */
-  int    end_off;    /* offset to the end of the selector                  */
-  int    width;      /* max width of the selector                          */
-  int    skip;       /* Number of multibytes to skip after the number      */
+  int    flength;    /* 0 or length + 3 (full prefix lengh                 */
+  int    offset;     /* offset to the start of the selector                */
+  int    size;       /* size in bytes of the selector to extract           */
+  int    ignore;     /* number of multibytes to ignore after the number    */
+  int    follow;     /* y: the numbering follows the last nuber set        */
+  int    def_number; /* 1: the numbering is on by default 0: it is not     */
 };
 
 /* **************** */
@@ -779,7 +779,9 @@ short_usage(void)
   fprintf(stderr, "       [-S /regex/repl/[g][v][s][i]] ");
   fprintf(stderr, "[-I /regex/repl/[g][v][s][i]]             \\\n");
   fprintf(stderr, "       [-E /regex/repl/[g][v][s][i]] ");
-  fprintf(stderr, "[-A regex] [-Z regex] [-N regex [arg]...] \\\n");
+  fprintf(stderr, "[-A regex] [-Z regex]                     \\\n");
+  fprintf(stderr, "       [-N regex] [-U regex] [-F] [-D arg...] ");
+  fprintf(stderr, "                                 \\\n");
   fprintf(stderr, "       [-1 regex [attr]] [-2 regex [attr]]... ");
   fprintf(stderr, "[-5 regex [attr]] [-g] [-q]      \\\n");
   fprintf(stderr, "       [-W bytes] [-L bytes] [-T [separator]] ");
@@ -789,13 +791,15 @@ short_usage(void)
   fprintf(stderr, "       <col selectors> ::= col1[-col2]...|<RE>...\n");
   fprintf(stderr, "       <row selectors> ::= row1[-row2]...|<RE>...\n");
   fprintf(stderr, "       <prefix>        ::= i|e|c|b|s|t|sf|st|da\n");
-  fprintf(stderr, "       <arg>           ::= [l|r:<char>]|[a:l|r]|[p:i|a|");
-  fprintf(stderr, "[e:m|i]|[w:<size>]\n");
-  fprintf(stderr, "         (ex: l:'(' e:r)\n");
+  fprintf(stderr, "       <arg>           ::= [l|r:<char>]|[a:l|r]|[p:i|a]|");
+  fprintf(stderr, "[w:<size>]|\n");
+  fprintf(stderr, "                           [f:y|n]|[o:<num>]|[n:<num>]|");
+  fprintf(stderr, "[i:<num>]\n");
+  fprintf(stderr, "         Ex: l:'(' a:l\n");
   fprintf(stderr, "       <attr>          ::= [fg][/bg][,style] \n");
-  fprintf(stderr, "         (ex: 7/4,bu)\n");
+  fprintf(stderr, "         Ex: 7/4,bu\n");
   fprintf(stderr, "       <RE>            ::= <char>regex<char> \n");
-  fprintf(stderr, "         (ex: /regex/ or :regex:)\n\n");
+  fprintf(stderr, "         Ex: /regex/ or :regex:\n\n");
   fprintf(stderr, "       <col/row selectors> and <RE> can be freely mixed ");
   fprintf(stderr, "when used\n");
   fprintf(stderr, "       with -C and -R (ex: 2,/x/,4-5).\n");
@@ -866,8 +870,15 @@ usage(void)
           "-Z forces a class of words to be the latest of the line they "
           "appear in.\n");
   fprintf(stderr,
-          "-N numbers and provides a direct access to words matching (or not) "
-          "a specific regex.\n");
+          "-N/-U numbers/un-numbers and provides a direct access to words "
+          "matching\n");
+  fprintf(stderr, "   (or not) a specific regex.\n");
+  fprintf(stderr,
+          "-F numbers and provides a direct access to words by extracting the "
+          "number\n");
+  fprintf(stderr, "   from the words.\n");
+  fprintf(stderr,
+          "-D sets sub-options to modify the behaviour of -N, -U and -F.\n");
   fprintf(stderr,
           "-1,-2,...,-5 gives specific colors to up to 5 classes of "
           "selectable words.\n");
@@ -887,7 +898,7 @@ usage(void)
   fprintf(stderr, "\nNavigation keys are:\n");
   fprintf(stderr, "  - Left/Down/Up/Right arrows or h/j/k/l.\n");
   fprintf(stderr, "  - Home/End.\n");
-  fprintf(stderr, "  - Numbers if some words are numbered (-N).\n");
+  fprintf(stderr, "  - Numbers if some words are numbered (-N/-U/-F).\n");
   fprintf(stderr, "  - SPACE to search for the next match of a previously\n");
   fprintf(stderr, "          entered search prefix if any, see below.\n\n");
   fprintf(stderr, "Other useful keys are:\n");
@@ -4450,7 +4461,7 @@ disp_word(word_t * word_a, int pos, int search_mode, char * buffer,
         /* prints the leading spaces */
         /* """"""""""""""""""""""""" */
         tputs(TPARM1(exit_attribute_mode), 1, outch);
-        printf("%.*s", daccess.length + 3, tmp_word);
+        printf("%.*s", daccess.flength, tmp_word);
       }
 
       /* Set the search cursor attribute */
@@ -4470,17 +4481,13 @@ disp_word(word_t * word_a, int pos, int search_mode, char * buffer,
       /* Print and overwrite the beginning of the word with the search */
       /* buffer content if it is not empty                             */
       /* """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""" */
-      if (daccess.length > 0)
-        tputs(tmp_word + daccess.length + 3, 1, outch);
-      else
-        tputs(tmp_word, 1, outch);
+      tputs(tmp_word + daccess.flength, 1, outch);
 
       if (buffer[0] != '\0')
       {
         int i = 0;
 
         int buf_width;
-        int total_daccess_sel_len;
 
         /* Calculate the space taken by the buffer on screen */
         /* """"""""""""""""""""""""""""""""""""""""""""""""" */
@@ -4488,16 +4495,9 @@ disp_word(word_t * word_a, int pos, int search_mode, char * buffer,
                              mbstowcs(NULL, buffer, 0));
         free(w);
 
-        /* Size of the direct access selector to skip to reach the word */
-        /* """""""""""""""""""""""""""""""""""""""""""""""""""""""""""" */
-        if (daccess.length > 0)
-          total_daccess_sel_len = daccess.length + 3;
-        else
-          total_daccess_sel_len = 0;
-
         /* Put the cursor at the beginning of the word */
         /* """"""""""""""""""""""""""""""""""""""""""" */
-        for (i = 0; i < e - s + 1 - total_daccess_sel_len; i++)
+        for (i = 0; i < e - s + 1 - daccess.flength; i++)
           tputs(TPARM1(cursor_left), 1, outch);
 
         /* Set the buffer display attribute */
@@ -4512,17 +4512,14 @@ disp_word(word_t * word_a, int pos, int search_mode, char * buffer,
 
         /* Put back the cursor after the word */
         /* """""""""""""""""""""""""""""""""" */
-        for (i = 0; i < e - s - buf_width + 1 - total_daccess_sel_len; i++)
+        for (i = 0; i < e - s - buf_width + 1 - daccess.flength; i++)
           tputs(TPARM1(cursor_right), 1, outch);
       }
     }
     else
     {
-      int offset = 0;
       if (daccess.length > 0)
       {
-        offset = daccess.length + 3;
-
         /* If this word is not numbered, reset the display */
         /* attributes before printing the leading spaces.  */
         /* """"""""""""""""""""""""""""""""""""""""""""""" */
@@ -4531,7 +4528,7 @@ disp_word(word_t * word_a, int pos, int search_mode, char * buffer,
           /* Print the non significant part of the word */
           /* """""""""""""""""""""""""""""""""""""""""" */
           tputs(TPARM1(exit_attribute_mode), 1, outch);
-          printf("%.*s", offset - 1, word_a[pos].str);
+          printf("%.*s", daccess.flength - 1, word_a[pos].str);
           tputs(TPARM1(exit_attribute_mode), 1, outch);
           fputc(' ', stdout);
         }
@@ -4601,7 +4598,7 @@ disp_word(word_t * word_a, int pos, int search_mode, char * buffer,
       }
 
       mb_strprefix(tmp_word, word_a[pos].str, (int)word_a[pos].mb - 1, &p);
-      tputs(tmp_word + offset, 1, outch);
+      tputs(tmp_word + daccess.flength, 1, outch);
     }
     tputs(TPARM1(exit_attribute_mode), 1, outch);
   }
@@ -4641,7 +4638,7 @@ disp_word(word_t * word_a, int pos, int search_mode, char * buffer,
       /* """"""""""""""""""""""""""""""""""""""""""""""""""""" */
       tputs(TPARM1(exit_attribute_mode), 1, outch);
       if (daccess.padding == 'a')
-        for (i = 0; i < daccess.length + 3; i++)
+        for (i = 0; i < daccess.flength; i++)
           fputc(' ', stdout);
     }
 
@@ -4694,7 +4691,7 @@ disp_word(word_t * word_a, int pos, int search_mode, char * buffer,
 
     if ((daccess.length > 0 && daccess.padding == 'a')
         || word_a[pos].is_numbered)
-      tputs(tmp_word + daccess.length + 3, 1, outch);
+      tputs(tmp_word + daccess.flength, 1, outch);
     else
       tputs(tmp_word, 1, outch);
 
@@ -5133,19 +5130,11 @@ search_next(tst_node_t * tst, word_t * word_a, char * search_buf,
 {
   wchar_t * w;
   int       found = 0;
-  int       total_daccess_sel_len;
-
-  /* Size of the direct access selector to skip to reach the word */
-  /* """""""""""""""""""""""""""""""""""""""""""""""""""""""""""" */
-  if (daccess.length < 0)
-    total_daccess_sel_len = 0;
-  else
-    total_daccess_sel_len = daccess.length + 3;
 
   /* Consider a word under the cursor found if it matches the search prefix. */
   /* """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""" */
   if (!after_only)
-    if (memcmp(word_a[current].str + total_daccess_sel_len, search_buf,
+    if (memcmp(word_a[current].str + daccess.flength, search_buf,
                strlen(search_buf))
         == 0)
       return 1;
@@ -5597,11 +5586,16 @@ main(int argc, char * argv[])
 
   int index; /* generic counter */
 
-  int     daccess_index        = 1;
-  char *  daccess_locator_type = NULL;
-  char *  daccess_pattern      = NULL;
-  regex_t daccess_pattern_re; /* variable to store the compiled *
-                               * direct access pattern (-N) RE  */
+  int    daccess_index        = 1;
+  char * daccess_locator_type = NULL;
+
+  char *  daccess_np = NULL;
+  regex_t daccess_np_re; /* variable to store the compiled *
+                          * direct access pattern (-N) RE  */
+
+  char *  daccess_up = NULL;
+  regex_t daccess_up_re; /* variable to store the compiled *
+                          * direct access pattern (-U) RE  */
 
   char * include_pattern     = NULL;
   char * exclude_pattern     = NULL;
@@ -5896,21 +5890,24 @@ main(int argc, char * argv[])
   timeout.remain        = 0;
   timeout.reached       = 0;
 
-  daccess.type       = DACCESS_NONE;
+  daccess.type       = DA_TYPE_NONE;
   daccess.left       = " ";
   daccess.right      = ")";
   daccess.alignment  = 'r';
   daccess.padding    = 'a';
-  daccess.expression = 'm';
   daccess.length     = -2;
-  daccess.width      = 0;
-  daccess.skip       = 0;
+  daccess.flength    = 0;
+  daccess.offset     = 0;
+  daccess.size       = 0;
+  daccess.ignore     = 0;
+  daccess.follow     = 'y';
+  daccess.def_number = -1;
 
   /* Command line options analysis */
   /* """"""""""""""""""""""""""""" */
   while ((opt = egetopt(argc, argv,
                         "Vf:h?X:x:qdMba:i:e:S:I:E:A:Z:1:2:3:4:5:C:R:"
-                        "kclwrgn:t%m:s:W:L:T%P%pN:F:"))
+                        "kclwrgn:t%m:s:W:L:T%P%pN:U:FD:"))
          != -1)
   {
     switch (opt)
@@ -6227,6 +6224,9 @@ main(int argc, char * argv[])
 
           optind--;
 
+          if (*argv[optind] == '-')
+            TELL("A blank is required before the first sub-option -- ");
+
           /* Parse the arguments arguments */
           /* """"""""""""""""""""""""""""" */
           while (argv[optind] && *argv[optind] != '-')
@@ -6382,183 +6382,165 @@ main(int argc, char * argv[])
         toggle.autotag = 1;
         break;
 
-      case 'N':
-      {
-        int       pos;
-        wchar_t * w;
-        int       n;
-
-        if (daccess_pattern == NULL)
-          daccess_pattern = xstrdup(optarg);
-        else
-          TELL("Option already given -- ");
+      case 'D':
+        if (optarg && *optarg != '-')
+        {
+          int       pos;
+          wchar_t * w;
+          int       n;
 
-        daccess.type |= DACCESS_AUTO; /* auto */
+          /* Parse optional additional arguments */
+          /* """"""""""""""""""""""""""""""""""" */
+          optind--;
 
-        /* Parse optional additional arguments */
-        /* """"""""""""""""""""""""""""""""""" */
-        while (argv[optind] && *argv[optind] != '-')
-        {
-          if (argv[optind][1] != ':')
-            TELL("Bad format -- ");
+          if (*argv[optind] == '-')
+            TELL("A blank is required before the first sub-option -- ");
 
-          switch (*(argv[optind]))
+          while (argv[optind] && *argv[optind] != '-')
           {
-            case 'l': /* left char */
-              daccess.left = strdup(argv[optind] + 2);
-              mb_interpret(daccess.left, &langinfo);
+            if (argv[optind][1] != ':')
+              TELL("Bad format -- ");
 
-              if (mb_strlen(daccess.left) != 1)
-                TELL("Too many characters after l: -- ");
+            switch (*(argv[optind]))
+            {
+              case 'l': /* left char */
+                daccess.left = strdup(argv[optind] + 2);
+                mb_interpret(daccess.left, &langinfo);
 
-              n = wcswidth((w = mb_strtowcs(daccess.left)), 1);
-              free(w);
+                if (mb_strlen(daccess.left) != 1)
+                  TELL("Too many characters after l: -- ");
 
-              if (n > 1)
-                TELL("A multi columns character is not allowed after l: -- ");
-              break;
+                n = wcswidth((w = mb_strtowcs(daccess.left)), 1);
+                free(w);
 
-            case 'r': /* right char */
-              daccess.right = strdup(argv[optind] + 2);
-              mb_interpret(daccess.right, &langinfo);
+                if (n > 1)
+                  TELL("A multi columns character is not allowed after l: -- ");
+                break;
 
-              if (mb_strlen(daccess.right) != 1)
-                TELL("Too many characters after r: -- ");
+              case 'r': /* right char */
+                daccess.right = strdup(argv[optind] + 2);
+                mb_interpret(daccess.right, &langinfo);
 
-              n = wcswidth((w = mb_strtowcs(daccess.right)), 1);
-              free(w);
+                if (mb_strlen(daccess.right) != 1)
+                  TELL("Too many characters after r: -- ");
 
-              if (n > 1)
-                TELL("A multi columns character is not allowed after r: -- ");
-              break;
+                n = wcswidth((w = mb_strtowcs(daccess.right)), 1);
+                free(w);
 
-            case 'a': /* alignment */
-              if (strprefix("left", argv[optind] + 2))
-                daccess.alignment = 'l';
-              else if (strprefix("right", argv[optind] + 2))
-                daccess.alignment = 'r';
-              else
-                TELL("Bad format -- ");
-              break;
+                if (n > 1)
+                  TELL("A multi columns character is not allowed after r: -- ");
+                break;
 
-            case 'p': /* padding */
-              if (strprefix("all", argv[optind] + 2))
-                daccess.padding = 'a';
-              else if (strprefix("included", argv[optind] + 2))
-                daccess.padding = 'i';
-              else
-                TELL("Bad format -- ");
-              break;
+              case 'a': /* alignment */
+                if (strprefix("left", argv[optind] + 2))
+                  daccess.alignment = 'l';
+                else if (strprefix("right", argv[optind] + 2))
+                  daccess.alignment = 'r';
+                else
+                  TELL("Bad format -- ");
+                break;
 
-            case 'e': /* expression */
-              if (strprefix("match", argv[optind] + 2))
-                daccess.expression = 'm';
-              else if (strprefix("invert", argv[optind] + 2))
-                daccess.expression = 'i';
-              else
-                TELL("Bad format -- ");
-              break;
+              case 'p': /* padding */
+                if (strprefix("all", argv[optind] + 2))
+                  daccess.padding = 'a';
+                else if (strprefix("included", argv[optind] + 2))
+                  daccess.padding = 'i';
+                else
+                  TELL("Bad format -- ");
+                break;
 
-            case 'w': /* width */
-              if (sscanf(argv[optind] + 2, "%d%n", &daccess.length, &pos) != 1)
-                TELL("Bad format -- ");
-              if (argv[optind][pos + 2] != '\0')
-                TELL("Bad format -- ");
-              break;
+              case 'w': /* width */
+                if (sscanf(argv[optind] + 2, "%d%n", &daccess.length, &pos)
+                    != 1)
+                  TELL("Bad format -- ");
+                if (argv[optind][pos + 2] != '\0')
+                  TELL("Bad format -- ");
+                if (daccess.length <= 0 || daccess.length > 5)
+                  TELL("w suboption must be between 1 and 5 -- ");
+                break;
 
-            default:
-              TELL("Bad format -- ");
-          }
+              case 'o': /* start offset */
+                if (sscanf(argv[optind] + 2, "%d%n", &daccess.offset, &pos)
+                    != 1)
+                  TELL("Bad format -- ");
+                if (argv[optind][pos + 2] != '\0')
+                  TELL("Bad format -- ");
+                break;
 
-          if (daccess.length <= 0 || daccess.length > 5)
-            daccess.length = -2; /* special value -> auto */
+              case 'n': /* numbor of digits to extract */
+                if (sscanf(argv[optind] + 2, "%d%n", &daccess.size, &pos) != 1)
+                  TELL("Bad format -- ");
+                if (argv[optind][pos + 2] != '\0')
+                  TELL("Bad format -- ");
+                if (daccess.size <= 0 || daccess.size > 5)
+                  TELL("n suboption must be between 1 and 5 -- ");
+                break;
 
-          optind++;
-        }
-      }
+              case 'i': /* Number of multibytes to ignore after the selector to
+                           extract */
+                if (sscanf(argv[optind] + 2, "%d%n", &daccess.ignore, &pos)
+                    != 1)
+                  TELL("Bad format -- ");
+                if (argv[optind][pos + 2] != '\0')
+                  TELL("Bad format -- ");
+                break;
 
-      break;
+              case 'f': /* follow */
+                if (strprefix("yes", argv[optind] + 2))
+                  daccess.follow = 'y';
+                else if (strprefix("no", argv[optind] + 2))
+                  daccess.follow = 'n';
+                else
+                  TELL("Bad format -- ");
+                break;
 
-      case 'F':
-      {
-        int       pos;
-        wchar_t * w;
-        int       n;
+              default:
+                TELL("Bad format -- ");
+            }
 
-        if (daccess_locator_type == NULL)
-          daccess_locator_type = xstrdup(optarg);
-        else
-          TELL("Option already given -- ");
+            if (daccess.length <= 0 || daccess.length > 5)
+              daccess.length = -2; /* special value -> auto */
 
-        if (strprefix("delimiter", daccess_locator_type))
-          daccess.type |= DACCESS_DELIM;
-        else if (strprefix("positional", daccess_locator_type))
-          daccess.type |= DACCESS_POS;
+            optind++;
+          }
+        }
         else
-          TELL("Bad locator type -- ");
-
-        /* Parse optional additional arguments */
-        /* """"""""""""""""""""""""""""""""""" */
-        while (argv[optind] && *argv[optind] != '-')
-        {
-          if (argv[optind][1] != ':')
-            TELL("Bad format -- ");
-
-          switch (*(argv[optind]))
-          {
-            case 'd': /* delimiter */
-              daccess.delim = strdup(argv[optind] + 2);
-              mb_interpret(daccess.delim, &langinfo);
+          TELL("Option requires an argument -- ");
 
-              if (mb_strlen(daccess.delim) != 1)
-                TELL("Too many characters after d: -- ");
+        break;
 
-              n = wcswidth((w = mb_strtowcs(daccess.delim)), 1);
-              free(w);
-              break;
+      case 'N':
+        if (daccess_np == NULL)
+        {
+          daccess_np = concat("(", optarg, ")", NULL);
+          daccess.type |= DA_TYPE_AUTO; /* auto */
+        }
+        else
+          daccess_np = concat(daccess_np, "|(", optarg, ")", NULL);
 
-            case 'o': /* start offset */
-              if (sscanf(argv[optind] + 2, "%d%n", &daccess.start_off, &pos)
-                  != 1)
-                TELL("Bad format -- ");
-              if (argv[optind][pos + 2] != '\0')
-                TELL("Bad format -- ");
-              break;
+        if (daccess.def_number < 0)
+          daccess.def_number = 0;
 
-            case 'r': /* remove from word */
-              if (strprefix("yes", argv[optind] + 2))
-                daccess.remove = 'y';
-              else if (strprefix("no", argv[optind] + 2))
-                daccess.remove = 'n';
-              else
-                TELL("Bad format -- ");
-              break;
+        break;
 
-            case 'w': /* max width */
-              if (sscanf(argv[optind] + 2, "%d%n", &daccess.width, &pos) != 1)
-                TELL("Bad format -- ");
-              if (argv[optind][pos + 2] != '\0')
-                TELL("Bad format -- ");
-              break;
+      case 'U':
+        if (daccess_up == NULL)
+        {
+          daccess_up = concat("(", optarg, ")", NULL);
+          daccess.type |= DA_TYPE_AUTO; /* auto */
+        }
+        else
+          daccess_up = concat(daccess_up, "|(", optarg, ")", NULL);
 
-            case 's': /* skip */
-              if (sscanf(argv[optind] + 2, "%d%n", &daccess.skip, &pos) != 1)
-                TELL("Bad format -- ");
-              if (argv[optind][pos + 2] != '\0')
-                TELL("Bad format -- ");
-              break;
-          }
+        if (daccess.def_number < 0)
+          daccess.def_number = 1;
 
-          optind++;
-        }
+        break;
 
-        /* The width is mandatory */
-        /* """""""""""""""""""""" */
-        if (daccess.width <= 0)
-          TELL("The required 'width' suboption must me greater than 0 -- ");
-      }
+      case 'F':
+        daccess.type |= DA_TYPE_POS;
 
-      break;
+        break;
 
       case '?':
         fputc('\n', stderr);
@@ -6735,7 +6717,7 @@ main(int argc, char * argv[])
     free(local_ini_file);
   }
 
-  word_buffer = xcalloc(1, limits.word_length);
+  word_buffer = xcalloc(1, daccess.flength + limits.word_length + 1);
 
   /* If some attributes were not set, set their default values */
   /* """"""""""""""""""""""""""""""""""""""""""""""""""""""""" */
@@ -6935,11 +6917,18 @@ main(int argc, char * argv[])
     col_index = cols_number = 0;
   }
 
-  if (daccess_pattern
-      && regcomp(&daccess_pattern_re, daccess_pattern, REG_EXTENDED | REG_NOSUB)
-           != 0)
+  if (daccess_np
+      && regcomp(&daccess_np_re, daccess_np, REG_EXTENDED | REG_NOSUB) != 0)
+  {
+    fprintf(stderr, "Bad regular expression %s\n", daccess_np);
+
+    exit(EXIT_FAILURE);
+  }
+
+  if (daccess_up
+      && regcomp(&daccess_up_re, daccess_up, REG_EXTENDED | REG_NOSUB) != 0)
   {
-    fprintf(stderr, "Bad regular expression %s\n", daccess_pattern);
+    fprintf(stderr, "Bad regular expression %s\n", daccess_up);
 
     exit(EXIT_FAILURE);
   }
@@ -7702,85 +7691,24 @@ main(int argc, char * argv[])
       }
     }
 
-    /* Save the original word */
-    /* """""""""""""""""""""" */
-    word           = &word_a[wi];
-    unaltered_word = xstrdup(word->str);
+    word = &word_a[wi];
 
-    /* Possibly modify the word according to -S/-I/-E arguments */
-    /* """""""""""""""""""""""""""""""""""""""""""""""""""""""" */
+    /* Make sure that daccess.length >= daccess.size */
+    /* with DA_TYPE_POS.                             */
+    /* """"""""""""""""""""