maint: refactor common expand(1) and unexpand(1) code

* src/expand.c, src/unexpand.c: Move global variables from here...
* src/expand-common.h, src/expand-common.c: ... to here.
* src/expand.c, src/unexpand.c: (parse_tab_stops, validate_tab_stops,
next_file): Move identical functions to new module.
(add_tab_stop): Move to new module, including additional code from
'unexpand' (keeping max_column_width) which will have no effect in when
used in 'expand'. Refactor common next-column calculation code
into a new function 'get_next_tab_column'.
* src/local.mk: (src_expand_SOURCES, src_unexpand_SOURCES): Add
'expand-common.c'; (noinst_HEADERS): Add 'expand-common.h'.
* po/POTFILES.in: Add 'expand-common.c'.
* tests/misc/expand.pl: Add more tests.
* tests/misc/unexpand.pl: Likewise.
* TODO: Move conclusions to above test after investigation.

2 files changed, 155 insertions, 1 deletions

diff --git a/tests/misc/expand.pl b/tests/misc/expand.pl
index a6b4a4d61..8a9cad144 100755
--- a/tests/misc/expand.pl
+++ b/tests/misc/expand.pl
@@ -18,7 +18,11 @@
 
 use strict;
 
+my $limits = getlimits ();
+my $UINTMAX_OFLOW = $limits->{UINTMAX_OFLOW};
+
 (my $program_name = $0) =~ s|.*/||;
+my $prog = 'expand';
 
 # Turn off localization of executable's output.
 @ENV{qw(LANGUAGE LANG LC_ALL)} = ('C') x 3;
@@ -27,13 +31,127 @@ my @Tests =
   (
    ['t1', '--tabs=3',     {IN=>"a\tb"}, {OUT=>"a  b"}],
    ['t2', '--tabs=3,6,9', {IN=>"a\tb\tc\td\te"}, {OUT=>"a  b  c  d e"}],
+   ['t3', '--tabs="3 6 9"',   {IN=>"a\tb\tc\td\te"}, {OUT=>"a  b  c  d e"}],
+   # Leading space/commas are silently ignored; Mixing space/commas is allowed.
+   # (a side-effect of allowing direct "-3,9" parameter).
+   ['t4', '--tabs=", 3,6 9"', {IN=>"a\tb\tc\td\te"}, {OUT=>"a  b  c  d e"}],
+   # tab stops parameter without values
+   ['t5', '--tabs=""',        {IN=>"a\tb\tc"}, {OUT=>"a       b       c"}],
+   ['t6', '--tabs=","',       {IN=>"a\tb\tc"}, {OUT=>"a       b       c"}],
+   ['t7', '--tabs=" "',       {IN=>"a\tb\tc"}, {OUT=>"a       b       c"}],
+
+   # Input field wider than the specified tab list
+   ['t8', '--tabs=6,9', {IN=>"a\tbbbbbbbbbbbbb\tc"},
+    {OUT=>"a     bbbbbbbbbbbbb c"}],
+
    ['i1', '--tabs=3 -i', {IN=>"\ta\tb"}, {OUT=>"   a\tb"}],
    ['i2', '--tabs=3 -i', {IN=>" \ta\tb"}, {OUT=>"   a\tb"}],
+
+   # Undocumented feature:
+   #   treat "expand -7"  as "expand --tabs 7" ,
+   #   and   "expand -90" as "expand --tabs 90",
+   ['u1', '-3',    {IN=>"a\tb\tc"}, {OUT=>"a  b  c"}],
+   ['u2', '-4 -9', {IN=>"a\tb\tc"}, {OUT=>"a   b    c"}],
+   ['u3', '-11',   {IN=>"a\tb\tc"}, {OUT=>"a          b          c"}],
+   # Test all digits (for full code coverage)
+   ['u4', '-2 -6', {IN=>"a\tb\tc"}, {OUT=>"a b   c"}],
+   ['u5', '-7',    {IN=>"a\tb"},    {OUT=>"a      b"}],
+   ['u6', '-8',    {IN=>"a\tb"},    {OUT=>"a       b"}],
+   # This syntax is handled internally as "-3, -9"
+   ['u7', '-3,9',  {IN=>"a\tb\tc"}, {OUT=>"a  b     c"}],
+
+   # Multiple non-empty files
+   ['f1', '--tabs=4',
+    {IN=>{"in1" => "a\tb\n"}}, {IN=>{"in2" => "c\td\n"}},
+    {OUT=>"a   b\nc   d\n"}],
+   # Multiple files, first file is empty
+   ['f2', '--tabs=4',
+    {IN=>{"in1" => ""}}, {IN=>{"in2" => "c\td\n"}},
+    {OUT=>"c   d\n"}],
+   # Multiple files, second file is empty
+   ['f3', '--tabs=4',
+    {IN=>{"in1" => "a\tb\n"}}, {IN=>{"in2" => ""}},
+    {OUT=>"a   b\n"}],
+
+
+   # Test '\b' (back-space) - subtract one column.
+   #
+   # Note:
+   # In a terminal window, 'expand' will appear to erase the 'a' characters
+   # due to overwriting them with spaces:
+   #
+   #    $ printf 'aaa\b\b\bc\td\n'
+   #    caa     d
+   #    $ printf 'aaa\b\b\bc\td\n' | expand
+   #    c       d
+   #
+   # However the characters are all printed:
+   #
+   #    $ printf 'aaa\b\b\bc\td\n' | expand | od -An -ta
+   #      a   a   a  bs  bs  bs   c  sp  sp  sp  sp  sp  sp  sp   d  nl
+   #
+   # If users ever report a problem with these tests and just
+   # copy&paste from the terminal, their report will be confusing
+   # (the 'a' will not appear).
+   #
+   # To see an example, enable the 'b-confusing' test, and examine the
+   # reported log:
+   #
+   #     expand.pl: test b-confusing: stdout mismatch
+   #     *** b-confusing.2       Fri Jun 24 15:43:21 2016
+   #     --- b-confusing.O       Fri Jun 24 15:43:21 2016
+   #     ***************
+   #     *** 1 ****
+   #     ! c       d
+   #     --- 1 ----
+   #     ! c       d
+   #
+   # ['b-confusing','', {IN=>"aaa\b\b\bc\td\n"}, {OUT=>"c       d\n"}],
+
+   ['b1','', {IN=>"aaa\b\b\bc\td\n"}, {OUT=>"aaa\b\b\bc       d\n"}],
+
+   # \b as first character, when column is zero
+   ['b2','', {IN=>"\bc\td"}, {OUT=>"\bc       d"}],
+
+   # Testing tab list adjusted due to backspaces
+   # ('b3' is the baseline without backspaces).
+   ['b3','--tabs 2,4,6,10',
+    {IN=>"1\t2\t3\t4\t5\n" .
+         "a\tb\tc\td\te\n"},
+    {OUT=>"1 2 3 4   5\n" .
+          "a b c d   e\n"}],
+
+   # On screen this will appear the same as 'b3'
+   ['b4','--tabs 2,4,6,10',
+    {IN=>"1\t2\t3\t4\t5\n" .
+         "a\tbHELLO\b\b\b\b\b\tc\td\te\n"},
+    {OUT=>"1 2 3 4   5\n" .
+          "a bHELLO\b\b\b\b\b c d   e\n"}],
+
+   # On screen on 'bHE' will appear (LLO overwritten by spaces),
+   # 'c' should align with 4, 'd' with 5:
+   #   1 2 3 4   5
+   #   a bHE c   d e
+   ['b5','--tabs 2,4,6,10',
+    {IN=>"1\t2\t3\t4\t5\n" .
+         "a\tbHELLO\b\b\b\tc\td\te\n"},
+    {OUT=>"1 2 3 4   5\n" .
+          "a bHELLO\b\b\b c   d e\n"}],
+
+
+   # Test errors
+   ['e1', '--tabs="a"', {IN=>''}, {OUT=>''}, {EXIT=>1},
+    {ERR => "$prog: tab size contains invalid character(s): 'a'\n"}],
+   ['e2', "-t $UINTMAX_OFLOW", {IN=>''}, {OUT=>''}, {EXIT=>1},
+    {ERR => "$prog: tab stop is too large '$UINTMAX_OFLOW'\n"}],
+   ['e3', '--tabs=0',   {IN=>''}, {OUT=>''}, {EXIT=>1},
+    {ERR => "$prog: tab size cannot be 0\n"}],
+   ['e4', '--tabs=3,3', {IN=>''}, {OUT=>''}, {EXIT=>1},
+    {ERR => "$prog: tab sizes must be ascending\n"}],
   );
 
 my $save_temps = $ENV{DEBUG};
 my $verbose = $ENV{VERBOSE};
 
-my $prog = 'expand';
 my $fail = run_tests ($program_name, $prog, \@Tests, $save_temps, $verbose);
 exit $fail;
diff --git a/tests/misc/unexpand.pl b/tests/misc/unexpand.pl
index c592c5a52..6ba6d405c 100755
--- a/tests/misc/unexpand.pl
+++ b/tests/misc/unexpand.pl
@@ -90,6 +90,42 @@ my @Tests =
      # setting of e.g., _POSIX2_VERSION=1.
      ['obs-ovflo', "-$limits->{UINTMAX_OFLOW}", {IN=>''}, {OUT=>''},
       {EXIT => 1}, {ERR => "$prog: tab stop value is too large\n"}],
+
+
+     # Test input with backspaces '\b' ('bs1' is the baseline, without \b)
+     # Note: If users report errors in these tests, copy&pasting results from
+     # their terminate output might be confusing due to '\b' overriding
+     # characters. For details see '\b' tests in 'expand.pl'.
+     ['bs1', '-a -t4', {IN=>"aa  c\n"},    {OUT=>"aa\tc\n"}],
+     ['bs2', '-a -t4', {IN=>"aa\b  c\n"},  {OUT=>"aa\b  c\n"}],
+     ['bs3', '-a -t4', {IN=>"aa\b   c\n"}, {OUT=>"aa\b\tc\n"}],
+     ['bs4', '-a -t3', {IN=>"aa\b  c\n"},  {OUT=>"aa\b\tc\n"}],
+
+     # Undocumented feature:
+     #   treat "unexpand -7"  as "unexpand --first-only --tabs 7" ,
+     #   and   "unexpand -90" as "unexpand --first-only --tabs 90",
+     ['u1', '-a -3',    {IN=>"a  b  c"}, {OUT=>"a\tb\tc"}],
+     ['u2', '-a -4,9',  {IN=>"a   b    c"}, {OUT=>"a\tb\tc"}],
+     ['u3', '-a -11',   {IN=>"a          b"}, {OUT=>"a\tb"}],
+     # Test all digits (for full code coverage)
+     ['u4', '-a -2,6',  {IN=>"a b   c"}, {OUT=>"a b\tc"}],
+     ['u5', '-a -7',    {IN=>"a      b"},    {OUT=>"a\tb"}],
+     ['u6', '-a -8',    {IN=>"a       b"},    {OUT=>"a\tb"}],
+     # This syntax is handled internally as "-3, -9"
+     ['u7', '-a -3,9',  {IN=>"a  b     c"}, {OUT=>"a\tb\tc"}],
+     # Default (without -a) is --first-only:
+     ['u8', '-3',  {IN=>"   a   b"}, {OUT=>"\ta   b"}],
+
+     # Arguably this should minimize translation as is done on Solaris.
+     # I.e., not modify the input.  But since the result is equivalent,
+     # and to be consistent in output with older versions, we output
+     # a '\t' rather than a space for the second tab position.
+     # For more detailed comparison with other implementations see:
+     # http://lists.gnu.org/archive/html/coreutils/2016-06/msg00015.html
+     # http://lists.gnu.org/archive/html/coreutils/2016-07/msg00011.html
+     ['ts1', '-t8,9', {IN=>"x\t \t y\n"},    {OUT=>"x\t\t\t y\n"}],
+     # There is no ambiguity here. This should always be the output.
+     ['ts2', '-t5,8', {IN=>"x\t \t y\n"},    {OUT=>"x\t\t y\n"}],
     );
 
 my $save_temps = $ENV{DEBUG};