vim-patch:8.2.0860: cannot use CTRL-A and CTRL-X on unsigned numbers

Problem:    Cannot use CTRL-A and CTRL-X on unsigned numbers.
Solution:   Add "unsigned" to 'nrformats'. (Naruhiko Nishino, closes vim/vim#6144)
aaad995f83

runtime/doc/options.txt

@@ -4208,6 +4208,15 @@ A jump table for the options with a short description can be found at |Q_op|.
 	bin	If included, numbers starting with "0b" or "0B" will be
 		considered to be binary.  Example: Using CTRL-X on
 		"0b1000" subtracts one, resulting in "0b0111".
+	unsigned    If included, numbers are recognized as unsigned. Thus a
+		leading dash or negative sign won't be considered as part of
+		the number.  Examples:
+		    Using CTRL-X on "2020" in "9-2020" results in "9-2019"
+		    (without "unsigned" it would become "9-2021").
+		    Using CTRL-A on "2020" in "9-2020" results in "9-2021"
+		    (without "unsigned" it would become "9-2019").
+		    Using CTRL-X on "0" or "18446744073709551615" (2^64) has
+		    no effect, overflow is prevented.
 	Numbers which simply begin with a digit in the range 1-9 are always
 	considered decimal.  This also happens for numbers that are not
 	recognized as octal or hex.

src/nvim/ops.c

@@ -4731,12 +4731,7 @@ int do_addsub(int op_type, pos_T *pos, int length, linenr_T Prenum1)
   char_u      *ptr;
   int c;
   int todel;
-  bool dohex;
-  bool dooct;
-  bool dobin;
-  bool doalp;
   int firstdigit;
-  bool subtract;
   bool negative = false;
   bool was_positive = true;
   bool visual = VIsual_active;
@@ -4747,10 +4742,12 @@ int do_addsub(int op_type, pos_T *pos, int length, linenr_T Prenum1)
   pos_T endpos;
   colnr_T save_coladd = 0;
 
-  dohex = (vim_strchr(curbuf->b_p_nf, 'x') != NULL);    // "heX"
+  const bool do_hex = vim_strchr(curbuf->b_p_nf, 'x') != NULL;    // "heX"
-  dooct = (vim_strchr(curbuf->b_p_nf, 'o') != NULL);    // "Octal"
+  const bool do_oct = vim_strchr(curbuf->b_p_nf, 'o') != NULL;    // "Octal"
-  dobin = (vim_strchr(curbuf->b_p_nf, 'b') != NULL);    // "Bin"
+  const bool do_bin = vim_strchr(curbuf->b_p_nf, 'b') != NULL;    // "Bin"
-  doalp = (vim_strchr(curbuf->b_p_nf, 'p') != NULL);    // "alPha"
+  const bool do_alpha = vim_strchr(curbuf->b_p_nf, 'p') != NULL;  // "alPha"
+  // "Unsigned"
+  const bool do_unsigned = vim_strchr(curbuf->b_p_nf, 'u') != NULL;
 
   if (virtual_active()) {
     save_coladd = pos->coladd;
@@ -4767,21 +4764,21 @@ int do_addsub(int op_type, pos_T *pos, int length, linenr_T Prenum1)
 
   // First check if we are on a hexadecimal number, after the "0x".
   if (!VIsual_active) {
-    if (dobin) {
+    if (do_bin) {
       while (col > 0 && ascii_isbdigit(ptr[col])) {
         col--;
         col -= utf_head_off(ptr, ptr + col);
       }
     }
 
-    if (dohex) {
+    if (do_hex) {
       while (col > 0 && ascii_isxdigit(ptr[col])) {
         col--;
         col -= utf_head_off(ptr, ptr + col);
       }
     }
-    if (dobin
+    if (do_bin
-        && dohex
+        && do_hex
         && !((col > 0
               && (ptr[col] == 'X' || ptr[col] == 'x')
               && ptr[col - 1] == '0'
@@ -4797,13 +4794,13 @@ int do_addsub(int op_type, pos_T *pos, int length, linenr_T Prenum1)
         }
     }
 
-    if ((dohex
+    if ((do_hex
          && col > 0
          && (ptr[col] == 'X' || ptr[col] == 'x')
          && ptr[col - 1] == '0'
          && !utf_head_off(ptr, ptr + col - 1)
          && ascii_isxdigit(ptr[col + 1]))
-        || (dobin
+        || (do_bin
             && col > 0
             && (ptr[col] == 'B' || ptr[col] == 'b')
             && ptr[col - 1] == '0'
@@ -4818,13 +4815,13 @@ int do_addsub(int op_type, pos_T *pos, int length, linenr_T Prenum1)
 
       while (ptr[col] != NUL
              && !ascii_isdigit(ptr[col])
-             && !(doalp && ASCII_ISALPHA(ptr[col]))) {
+             && !(do_alpha && ASCII_ISALPHA(ptr[col]))) {
         col++;
       }
 
       while (col > 0
              && ascii_isdigit(ptr[col - 1])
-             && !(doalp && ASCII_ISALPHA(ptr[col]))) {
+             && !(do_alpha && ASCII_ISALPHA(ptr[col]))) {
         col--;
       }
     }
@@ -4832,7 +4829,7 @@ int do_addsub(int op_type, pos_T *pos, int length, linenr_T Prenum1)
 
   if (visual) {
     while (ptr[col] != NUL && length > 0 && !ascii_isdigit(ptr[col])
-           && !(doalp && ASCII_ISALPHA(ptr[col]))) {
+           && !(do_alpha && ASCII_ISALPHA(ptr[col]))) {
       int mb_len = utfc_ptr2len(ptr + col);
 
       col += mb_len;
@@ -4844,7 +4841,8 @@ int do_addsub(int op_type, pos_T *pos, int length, linenr_T Prenum1)
     }
 
     if (col > pos->col && ptr[col - 1] == '-'
-        && !utf_head_off(ptr, ptr + col - 1)) {
+        && !utf_head_off(ptr, ptr + col - 1)
+        && !do_unsigned) {
       negative = true;
       was_positive = false;
     }
@@ -4852,12 +4850,12 @@ int do_addsub(int op_type, pos_T *pos, int length, linenr_T Prenum1)
 
   // If a number was found, and saving for undo works, replace the number.
   firstdigit = ptr[col];
-  if (!ascii_isdigit(firstdigit) && !(doalp && ASCII_ISALPHA(firstdigit))) {
+  if (!ascii_isdigit(firstdigit) && !(do_alpha && ASCII_ISALPHA(firstdigit))) {
     beep_flush();
     goto theend;
   }
 
-  if (doalp && ASCII_ISALPHA(firstdigit)) {
+  if (do_alpha && ASCII_ISALPHA(firstdigit)) {
     // decrement or increment alphabetic character
     if (op_type == OP_NR_SUB) {
       if (CharOrd(firstdigit) < Prenum1) {
@@ -4889,7 +4887,9 @@ int do_addsub(int op_type, pos_T *pos, int length, linenr_T Prenum1)
     curwin->w_cursor.col = col;
   } else {
     if (col > 0 && ptr[col - 1] == '-'
-        && !utf_head_off(ptr, ptr + col - 1) && !visual) {
+        && !utf_head_off(ptr, ptr + col - 1)
+        && !visual
+        && !do_unsigned) {
       // negative number
       col--;
       negative = true;
@@ -4903,9 +4903,9 @@ int do_addsub(int op_type, pos_T *pos, int length, linenr_T Prenum1)
     }
 
     vim_str2nr(ptr + col, &pre, &length,
-               0 + (dobin ? STR2NR_BIN : 0)
+               0 + (do_bin ? STR2NR_BIN : 0)
-               + (dooct ? STR2NR_OCT : 0)
+               + (do_oct ? STR2NR_OCT : 0)
-               + (dohex ? STR2NR_HEX : 0),
+               + (do_hex ? STR2NR_HEX : 0),
                NULL, &n, maxlen);
 
     // ignore leading '-' for hex, octal and bin numbers
@@ -4916,7 +4916,7 @@ int do_addsub(int op_type, pos_T *pos, int length, linenr_T Prenum1)
     }
 
     // add or subtract
-    subtract = false;
+    bool subtract = false;
     if (op_type == OP_NR_SUB) {
       subtract ^= true;
     }
@@ -4948,6 +4948,17 @@ int do_addsub(int op_type, pos_T *pos, int length, linenr_T Prenum1)
       }
     }
 
+    if (do_unsigned && negative) {
+      if (subtract) {
+        // sticking at zero.
+        n = (uvarnumber_T)0;
+      } else {
+        // sticking at 2^64 - 1.
+        n = (uvarnumber_T)(-1);
+      }
+      negative = false;
+    }
+
     if (visual && !was_positive && !negative && col > 0) {
       // need to remove the '-'
       col--;
@@ -5029,7 +5040,7 @@ int do_addsub(int op_type, pos_T *pos, int length, linenr_T Prenum1)
     // total length of the number remains the same.
     // Don't do this when
     // the result may look like an octal number.
-    if (firstdigit == '0' && !(dooct && pre == 0)) {
+    if (firstdigit == '0' && !(do_oct && pre == 0)) {
       while (length-- > 0) {
         *ptr++ = '0';
       }

src/nvim/option.c

@@ -292,7 +292,8 @@ typedef struct vimoption {
 
 static char *(p_ambw_values[]) =      { "single", "double", NULL };
 static char *(p_bg_values[]) =        { "light", "dark", NULL };
-static char *(p_nf_values[]) =        { "bin", "octal", "hex", "alpha", NULL };
+static char *(p_nf_values[]) =        { "bin", "octal", "hex", "alpha",
+                                        "unsigned", NULL };
 static char *(p_ff_values[]) =        { FF_UNIX, FF_DOS, FF_MAC, NULL };
 static char *(p_wak_values[]) =       { "yes", "menu", "no", NULL };
 static char *(p_mousem_values[]) =    { "extend", "popup", "popup_setpos",

src/nvim/testdir/test_increment.vim

@@ -1,4 +1,4 @@
-" Tests for using Ctrl-A/Ctrl-X on visual selections
+" Tests for using Ctrl-A/Ctrl-X
 
 func SetUp()
   new dummy
@@ -779,6 +779,50 @@ func Test_increment_empty_line()
   bwipe!
 endfunc
 
+" Try incrementing/decrementing a number when nrformats contains unsigned
+func Test_increment_unsigned()
+  set nrformats+=unsigned
+
+  call setline(1, '0')
+  exec "norm! gg0\<C-X>"
+  call assert_equal('0', getline(1))
+
+  call setline(1, '3')
+  exec "norm! gg010\<C-X>"
+  call assert_equal('0', getline(1))
+
+  call setline(1, '-0')
+  exec "norm! gg0\<C-X>"
+  call assert_equal("-0", getline(1))
+
+  call setline(1, '-11')
+  exec "norm! gg08\<C-X>"
+  call assert_equal('-3', getline(1))
+
+  " NOTE: 18446744073709551615 == 2^64 - 1
+  call setline(1, '18446744073709551615')
+  exec "norm! gg0\<C-A>"
+  call assert_equal('18446744073709551615', getline(1))
+
+  call setline(1, '-18446744073709551615')
+  exec "norm! gg0\<C-A>"
+  call assert_equal('-18446744073709551615', getline(1))
+
+  call setline(1, '-18446744073709551614')
+  exec "norm! gg08\<C-A>"
+  call assert_equal('-18446744073709551615', getline(1))
+
+  call setline(1, '-1')
+  exec "norm! gg0\<C-A>"
+  call assert_equal('-2', getline(1))
+
+  call setline(1, '-3')
+  exec "norm! gg08\<C-A>"
+  call assert_equal('-11', getline(1))
+
+  set nrformats-=unsigned
+endfunc
+
 func Test_normal_increment_with_virtualedit()
   set virtualedit=all