Version 0.2.0: RDB_ICASE, Changelog, --version, angepasste Man-Pages

CHANGELOG.md

@@ -0,0 +1,32 @@
+# Changelog
+
+Alle wichtigen Änderungen stehen hier. Das Format folgt
+[Keep a Changelog](https://keepachangelog.com/de/1.1.0/), die Versionsnummern
+folgen [Semantic Versioning](https://semver.org/lang/de/). Die Version steht in
+`lib/rdb/version.rb`; `rdb --version` zeigt sie an.
+
+## [Unveröffentlicht]
+
+## [0.2.0] - 2026-10-11
+
+### Hinzugefügt
+- `RDB_ICASE=1` ignoriert Groß-/Kleinschreibung bei Spaltennamen. Ein exakter
+  Treffer gewinnt, mehrere Treffer sind ein Fehler; awk-Wörter und
+  Funktionsaufrufe bleiben unberührt.
+- Man-Pages von `row`, `compute`, `column`, `sorttable`, `jointable`, `rename`,
+  `validate` und `rdb` mit dem Abschnitt „Case of Column Names“.
+- `rdb man -g`: Man-Page mit groff formatiert, dazu `man/rdb.tmac` mit den
+  MWC-Makros (`.DM`, `.DE`, `.PC` ...).
+- `rdb --version`, `CHANGELOG.md`.
+
+### Geändert
+- `column` schreibt den Kopf in der Schreibweise der Tabelle, wenn die Spalte
+  existiert.
+- `demo/ledger` und `demo/chart` verwenden echte Tabs.
+
+## [0.1.0] - 2026-10-11
+
+### Hinzugefügt
+- Erster Nachbau von /rdb in Ruby: ein Programm `bin/rdb` mit Symlinks für alle
+  dokumentierten Befehle (Relational, Text, Konvertierung, Schnellzugriff,
+  Berichte, Terminal, /act), 137 Man-Pages, Demo-Daten, 76 Tests, mise-Setup.

README.md

@@ -43,11 +43,19 @@ die meisten Befehle lesen beides.
 `row`, `compute` und `validate` übergeben ihren Ausdruck an awk, nachdem sie
 Spaltennamen durch `$N` ersetzt haben. Es gilt also awk-Syntax.
 
+Spaltennamen unterscheiden Groß- und Kleinschreibung. Mit `RDB_ICASE=1` wird
+sie ignoriert (Details unter „Unterschiede zum Original“):
+
+```
+RDB_ICASE=1 compute 'wert = cost * amount' < inventory
+```
+
 ## Hilfe
 
 ```
 rdb                  alle Befehle
 rdb man search       die Original-Man-Page als Text (auch Themen: backup, dbase, dashline ...)
+rdb --version        Versionsnummer (Änderungen siehe CHANGELOG.md)
 rdb man -g search    dieselbe Seite mit groff formatiert (Beispiele sauber eingerückt)
 whatis column        Beschreibung und Syntax
 whatwill total       Befehle, die zu einem Stichwort passen
@@ -104,6 +112,14 @@ er nicht mehr passt (außer mit `-a`). Beide halten die Indizes aktuell.
   Handbuch sind als Tests enthalten (`test/rdb_test.rb`).
 - Die Indexdateien haben ein eigenes, einfaches Format (das Original ist nicht
   dokumentiert). Sie sind nicht mit alten /rdb-Dateien austauschbar.
+- Spaltennamen unterscheiden wie im Original Groß- und Kleinschreibung. Mit
+  `RDB_ICASE=1` (z. B. in `mise.toml` unter `[env]`) wird sie ignoriert:
+  ein exakter Treffer gewinnt, mehrere ähnliche sind ein Fehler. In
+  `row`/`compute`/`validate` bleiben awk-Wörter und Funktionsaufrufe unberührt
+  (eine Spalte `Length` verdeckt also nicht `length()`). Die Man-Pages von
+  `row`, `compute`, `column`, `sorttable`, `jointable`, `rename`, `validate` und
+  `rdb` haben dazu den Abschnitt „Case of Column Names“; er ist die einzige
+  Ergänzung gegenüber den Original-Seiten.
 - `jointable` braucht keine vorsortierten Tabellen.
 - `ve` ist ein zeilenorientierter Editor, kein Bildschirmeditor.
 - `soundex` nutzt den Standard-Algorithmus. Die Beispielwerte im Handbuch

bin/rdb

@@ -12,6 +12,7 @@ Signal.trap('PIPE', 'SYSTEM_DEFAULT')
 if name == 'rdb'
   case args.first
   when nil, '-h', '--help' then name = 'rdb.list'
+  when '-v', '--version' then puts "rdb #{Rdb::VERSION}"; exit
   when 'install' then name = 'rdb.install'; args.shift
   else name = args.shift
   end

lib/rdb.rb

@@ -7,6 +7,7 @@ require 'open3'
 require 'date'
 require 'tmpdir'
 require 'fileutils'
+require_relative 'rdb/version'
 
 module Rdb
   ROOT = File.expand_path('..', __dir__)
@@ -16,6 +17,21 @@ module Rdb
   COMMANDS = {}
 
   class << self
+    # RDB_ICASE=1 makes column names match without regard to case.  An exact
+    # match always wins; several case-insensitive matches are an error.
+    def icase?
+      !%w[0 no false].include?(ENV['RDB_ICASE'].to_s.downcase) && !ENV['RDB_ICASE'].to_s.empty?
+    end
+
+    def column_index(head, name)
+      exact = head.index(name)
+      return exact if exact || !icase?
+
+      hits = head.each_index.select { |i| head[i].casecmp?(name) }
+      die "column name '#{name}' is ambiguous: #{hits.map { |i| head[i] }.join(', ')}" if hits.size > 1
+      hits.first
+    end
+
     def die(msg)
       abort "#{File.basename($PROGRAM_NAME)}: #{msg}"
     end

lib/rdb/cmd/relational.rb

@@ -15,8 +15,9 @@ Rdb.register('column', 'Display columns of a table in any order', 'column [ Colu
   t = table_in
   next t.write if args.empty?
 
-  idx = args.map { |a| t.head.index(a) }
+  idx = args.map { |a| Rdb.column_index(t.head, a) }
-  t.with(head: args, rows: t.rows.map { |r| idx.map { |i| i ? r[i] : '' } }).write
+  head = args.each_with_index.map { |a, n| idx[n] ? t.head[idx[n]] : a }
+  t.with(head: head, rows: t.rows.map { |r| idx.map { |i| i ? r[i] : '' } }).write
 end
 Rdb.alias_for('project', 'column')
 

lib/rdb/env.rb

@@ -40,6 +40,24 @@ module Rdb
 
     # Replace column names in an awk expression with $N, leaving strings,
     # regexes and other identifiers alone.
+    AWK_WORDS = %w[BEGIN END NR NF FS OFS ORS RS FNR FILENAME SUBSEP RSTART RLENGTH ENVIRON ARGC ARGV
+                   if else while for do in break continue next exit return function delete getline
+                   print printf length substr index split sub gsub match sprintf sin cos atan2 exp log
+                   sqrt int rand srand tolower toupper system].freeze
+
+    # With RDB_ICASE, a word that is not an awk keyword or function call may
+    # name a column in any case, provided that exactly one column fits.
+    def icase_column(names, expr, i)
+      return unless Rdb.icase?
+
+      word = expr[i..][/\A\w+/].to_s
+      return if word.empty? || AWK_WORDS.include?(word) || expr[i + word.size..].to_s.match?(/\A\s*\(/)
+
+      hits = names.select { |n, _| n.casecmp?(word) }
+      Rdb.die "column name '#{word}' is ambiguous: #{hits.map(&:first).join(', ')}" if hits.size > 1
+      hits.first
+    end
+
     def to_awk(expr, head)
       names = head.each_with_index.sort_by { |n, _| -n.size }
       out = +''
@@ -59,6 +77,7 @@ module Rdb
           i = j + 1
         elsif c.match?(/[A-Za-z_#]/) && (i.zero? || !expr[i - 1].match?(/[\w$]/))
           hit = names.find { |n, _| !n.empty? && expr[i, n.size] == n && !expr[i + n.size].to_s.match?(/\w/) }
+          hit ||= icase_column(names, expr, i)
           if hit
             out << "$#{hit[1] + 1}"
             i += hit[0].size

lib/rdb/fast.rb

@@ -53,7 +53,7 @@ module Rdb
       end
 
       def column(name)
-        @head.index(name) || Rdb.die("#{path}: no column named '#{name}'")
+        Rdb.column_index(@head, name) || Rdb.die("#{path}: no column named '#{name}'")
       end
 
       def row_at(offset)

lib/rdb/table.rb

@@ -60,7 +60,7 @@ module Rdb
     end
 
     def index_of(name)
-      @head.index(name) || Rdb.die("no column named '#{name}'")
+      Rdb.column_index(@head, name) || Rdb.die("no column named '#{name}'")
     end
 
     def to_s

lib/rdb/version.rb

@@ -0,0 +1,5 @@
+# frozen_string_literal: true
+
+module Rdb
+  VERSION = '0.2.0'
+end

man/column.man

@@ -113,6 +113,22 @@ Phone   (111) 222-3333
 Name    Chiquita  Banana
 Phone   1234
 .DE
+.SH "Case of Column Names"
+.I "(Ruby re-creation only.)"
+Column names are case sensitive, as in the original
+.BR /rdb .
+If the environment variable
+.B RDB_ICASE
+is set to 1, upper and lower case are ignored when column names are looked up.
+An exact match always wins;
+if several columns differ only in case, the command stops with an
+.I ambiguous
+error.
+Names that are also
+.B awk
+reserved words or are followed by
+.B (
+are never taken for columns.
 .SH "See Also"
 .B
 project,

man/compute.man

@@ -355,6 +355,22 @@ compute 'Value = sprintf ("%7.2f", ( Cost * Amount ))' < inventory
 .DE
 .PP
 These commands will produce the first table under Examples, above.
+.SH "Case of Column Names"
+.I "(Ruby re-creation only.)"
+Column names are case sensitive, as in the original
+.BR /rdb .
+If the environment variable
+.B RDB_ICASE
+is set to 1, upper and lower case are ignored when column names are looked up.
+An exact match always wins;
+if several columns differ only in case, the command stops with an
+.I ambiguous
+error.
+Names that are also
+.B awk
+reserved words or are followed by
+.B (
+are never taken for columns.
 .SH "See Also"
 COHERENT Lexicon:
 .B

man/jointable.man

@@ -393,6 +393,22 @@ before
 .BR chart .
 It says that the standard input file
 will be file 1 or on the left side of the table.
+.SH "Case of Column Names"
+.I "(Ruby re-creation only.)"
+Column names are case sensitive, as in the original
+.BR /rdb .
+If the environment variable
+.B RDB_ICASE
+is set to 1, upper and lower case are ignored when column names are looked up.
+An exact match always wins;
+if several columns differ only in case, the command stops with an
+.I ambiguous
+error.
+Names that are also
+.B awk
+reserved words or are followed by
+.B (
+are never taken for columns.
 .SH "See Also"
 .B
 column,

man/rdb.man

@@ -10,6 +10,22 @@ prints on the standard output all commands in directory
 .BR $RDB/bin .
 This command is a shell script, which you should edit if its output
 is not to your liking.
+.SH "Case of Column Names"
+.I "(Ruby re-creation only.)"
+Column names are case sensitive, as in the original
+.BR /rdb .
+If the environment variable
+.B RDB_ICASE
+is set to 1, upper and lower case are ignored when column names are looked up.
+An exact match always wins;
+if several columns differ only in case, the command stops with an
+.I ambiguous
+error.
+Names that are also
+.B awk
+reserved words or are followed by
+.B (
+are never taken for columns.
 .SH "See Also"
 .B
 act,

man/rename.man

@@ -112,6 +112,22 @@ State   Costa Rica
 Zip     123456789
 Phone   1234
 .DE
+.SH "Case of Column Names"
+.I "(Ruby re-creation only.)"
+Column names are case sensitive, as in the original
+.BR /rdb .
+If the environment variable
+.B RDB_ICASE
+is set to 1, upper and lower case are ignored when column names are looked up.
+An exact match always wins;
+if several columns differ only in case, the command stops with an
+.I ambiguous
+error.
+Names that are also
+.B awk
+reserved words or are followed by
+.B (
+are never taken for columns.
 .SH "See Also"
 .B
 column

man/row.man

@@ -175,6 +175,22 @@ and
 .BR compute ,
 read the COHERENT manual's tutorial on
 .BR awk .
+.SH "Case of Column Names"
+.I "(Ruby re-creation only.)"
+Column names are case sensitive, as in the original
+.BR /rdb .
+If the environment variable
+.B RDB_ICASE
+is set to 1, upper and lower case are ignored when column names are looked up.
+An exact match always wins;
+if several columns differ only in case, the command stops with an
+.I ambiguous
+error.
+Names that are also
+.B awk
+reserved words or are followed by
+.B (
+are never taken for columns.
 .SH "See Also"
 .B
 compute,

man/sorttable.man

@@ -165,6 +165,22 @@ People are often confused by
 .B sort
 because it does not behave the way one would expect.
 Study its entry in the COHERENT Lexicon.
+.SH "Case of Column Names"
+.I "(Ruby re-creation only.)"
+Column names are case sensitive, as in the original
+.BR /rdb .
+If the environment variable
+.B RDB_ICASE
+is set to 1, upper and lower case are ignored when column names are looked up.
+An exact match always wins;
+if several columns differ only in case, the command stops with an
+.I ambiguous
+error.
+Names that are also
+.B awk
+reserved words or are followed by
+.B (
+are never taken for columns.
 .SH "See Also"
 COHERENT Lexicon:
 .B

man/validate.man

@@ -133,6 +133,22 @@ that is easy to find, such as
 .PP
 If you only want to see the error lines, use the command
 .BR row .
+.SH "Case of Column Names"
+.I "(Ruby re-creation only.)"
+Column names are case sensitive, as in the original
+.BR /rdb .
+If the environment variable
+.B RDB_ICASE
+is set to 1, upper and lower case are ignored when column names are looked up.
+An exact match always wins;
+if several columns differ only in case, the command stops with an
+.I ambiguous
+error.
+Names that are also
+.B awk
+reserved words or are followed by
+.B (
+are never taken for columns.
 .SH "See Also"
 .SH "Logical Conditions"
 .B

test/rdb_test.rb

@@ -478,4 +478,41 @@ class ActTest < RdbCase
     assert_includes out, 'Duplicate Short names in chart'
     assert st.success?, err
   end
+
+  ICASE = { 'RDB_ICASE' => '1' }.freeze
+
+  def test_column_names_are_case_sensitive_by_default
+    _o, err, st = run3('column', 'cost', stdin: File.read(File.join(@dir, 'inventory')))
+    assert st.success?, err
+    assert_equal %w[cost], _o.lines.first.chomp.split("\t")
+  end
+
+  def test_icase_column_and_sorttable
+    out, err, st = run3('column', 'COST', 'item', stdin: File.read(File.join(@dir, 'inventory')), env: ICASE)
+    assert st.success?, err
+    assert_equal %w[Cost Item], out.lines.first.chomp.split("\t")
+    out, = run3('sorttable', '-nr', 'cost', stdin: File.read(File.join(@dir, 'inventory')), env: ICASE)
+    assert_equal 'scales', data(out).first.last
+  end
+
+  def test_icase_compute_and_awk_words
+    src = File.read(File.join(@dir, 'inventory'))
+    out, err, st = run3('compute', 'value = sprintf("%5.2f", cost * AMOUNT)', stdin: src, env: ICASE)
+    assert st.success?, err
+    assert_equal '1312.75', data(out)[5][3].strip
+    # a column called Length must not swallow the length() function
+    t = "Length\tName\n------\t----\n3\tab\n"
+    out, err, st = run3('compute', 'Length = length(name)', stdin: t, env: ICASE)
+    assert st.success?, err
+    assert_equal [%w[2 ab]], data(out)
+  end
+
+  def test_icase_ambiguous_column_is_an_error
+    t = "Name\tname\n----\t----\na\tb\n"
+    _o, err, st = run3('column', 'NAME', stdin: t, env: ICASE)
+    refute st.success?
+    assert_match(/ambiguous/, err)
+    out, = run3('column', 'name', stdin: t, env: ICASE)
+    assert_equal [%w[b]], data(out)
+  end
 end