# -*-tcl-*- nowrap

This is one of four example files for the Dictionary Package for Alpha.
The other three files are
   man
   Country-codes
   Whisky-distillers

A dictionary file ideally has three sectors: The first sector is a header
with a description for human beings --- this is what you are reading now. 
The third sector is the main one: it is the raw data in alphabetical order. 
The second sector is a description for the Dictionary Programme, referred
to as the boot sector.  Here the programme learns how the file is
structured: how the alphabetical order is defined, and how the result 
should be presented on the screen.  As illustrated by the example 
'Whisky-distillers', the two first sectors are not mandatory, if just 
the file is in standard alphabetical order.  (In this case fallback 
definitions stand in.  These are defined in the file dictionary.tcl)

This example (which constitutes a dictionary of Tcl commands) illustrates
the use of an index file.  The main data file is 'Tcl 8.4 Commands.txt'
located in Help folder.  The file you are reading now is the index file:
every entry in the main sector has as its value a pair of integers, the
first being the position of the Tcl command in the data file, the second
being the number of bytes this entry occupies.  In this way the Dictionary
programme uses alphabetical order in this file to find the offset, and then
the format proc (which usually is meant just for formatting the return of
the lookup programme) takes a 'second round' reading the actual data from
the main file.  The advantage of using index file is that it can be faster
and more economical (no need to create a whole data file if the data
already exists elsewhere).  The disadvantage is that each time the main
data file changes, the index file must be recompiled.  (For this particular
case there is actually a mechanism which rebuilds the index if the data
file is newer.  The new index is placed in the Cache folder inside your
AlphaPrefs folder.

Another feature of this particular dictionary is that it specifies
proprietary window geometry for the dictionary console.  (Every dictionary
has its own consoleGeometry variable: you can save the current geometry in
the menu...)


<!-- BEGIN TCL

# This proc takes the first word of the input string
# strips all accents,
# and then removes all non-alpha letters
proc normalForm { chunk } {
  #just take the first word:
  regexp {[^ \t]+} $chunk chunk
  
  regsub -all {[ˉ̌]} $chunk {a} chunk
  regsub -all {[]} $chunk {c} chunk
  regsub -all {[]} $chunk {e} chunk
  regsub -all {[]} $chunk {i} chunk
  regsub -all {[]} $chunk {n} chunk
  regsub -all {[Ϳ]} $chunk {o} chunk
  regsub -all {[]} $chunk {u} chunk
  regsub -all {[]} $chunk {y} chunk
  regsub -all {[]} $chunk {ae} chunk
  regsub -all {[]} $chunk {oe} chunk

  regsub -all {[^A-Za-z_]} $chunk {} chunk
  return [string tolower $chunk]
}

# The raw output from the lookup proc consists in a command name,
# an offset and a size.  So go to the tcl-commands file and get this
# chunk:
proc formatOutput { linje } {
  global HOME
  if { ![regexp {([a-z_]+) (\d+) (\d+)$} $linje dummy headword offset size] } {
      return ""
  }
  variable currentHeadword $headword
  set ordstrom [open [file join $HOME Help "Tcl 8.4 Commands.txt"] r]
  seek $ordstrom $offset
  set res [read $ordstrom $size]
  close $ordstrom
  regsub -all "\n" $res "\r" res
  return $res
}


# Since the output format (given in "Tcl 8.4 Commands.txt") is a long 
# and narrow column, this dictionary has chosen to specify a matching
# window format.  It is computed from the screen dimensions.
# 
# Note: the name of the dictionary (which is chosen by the user, not by
# this file!) is always equal to the namespace tail of the dictionary,
# and when the boot sector is sourced we are in that namespace.  Hence
# we know:
set nameOfThisDictionary [namespace tail [namespace current]]
set consoleWidth 428
set consoleHeight [expr {$::screenHeight - 44}]
set leftOffset [expr {$::screenWidth - $consoleWidth}]
global tcl_platform
if { $tcl_platform(platform) == "macintosh" } {
    # Correct for window borders:
    incr consoleHeight -6
    incr leftOffset -7
}  
set topOffset 59
# (Here we assume statusbar at bottom.  When we actually open the
# window we check where the statusbar is and adjust correspondingly...)

global DicomodeVars
::console::create "* $nameOfThisDictionary *" -mode Text -minormode "dictionary" \
  -font Monaco -fontsize $DicomodeVars(fontSize) \
  -g $leftOffset $topOffset $consoleWidth $consoleHeight

unset consoleWidth
unset consoleHeight
unset leftOffset
unset topOffset
unset nameOfThisDictionary

END TCL -->
  
after 112 4635
append 4815 621
array 5504 6936
bgerror 12508 3464
binary 16040 25723
break 41831 761
catch 42660 2025
cd 44753 385
clock 45206 12419
close 57693 2115
concat 59876 843
continue 60787 780
encoding 61635 3089
eof 64792 666
error 65526 1977
eval 67571 726
exec 68367 17760
exit 86195 341
expr 86604 17461
fblocked 104133 977
fconfigure 105178 24643
fcopy 129889 5536
file 135493 22106
fileevent 157667 5511
filename 163246 11125
flush 174439 1037
for 175544 2005
foreach 177617 2753
format 180438 9871
gets 190377 1909
glob 192354 9256
global 201678 704
history 202450 4431
http 206949 25014
if 232031 1341
incr 233440 637
info 234145 10234
interp 244447 35451
join 279966 527
lappend 280561 918
lindex 281547 3109
linsert 284724 868
list 285660 1117
llength 286845 366
load 287279 5835
lrange 293182 1060
lreplace 294310 1553
lsearch 295931 4608
lset 300607 3772
lsort 304448 6757
msgcat 311273 12702
namespace 324043 25850
open 349961 12097
package 362126 10356
pid 372550 779
pkg_mkIndex 373398 12119
proc 385585 3276
puts 388929 2817
pwd 391814 256
re_syntax 392138 27921
read 420127 3008
regexp 423203 7173
registry 430447 6950
regsub 437468 5122
rename 442658 675
resource 443401 8015
return 451484 3125
Safe 454677 17583
scan 472328 10215
seek 482611 2156
set 484835 1608
socket 486511 6422
source 493001 1911
split 494980 1061
string 496109 16538
subst 512715 5089
switch 517872 3901
Tcl 521841 32591
tell 554500 932
time 555500 580
trace 556148 22524
unknown 578740 3228
unset 582036 1391
update 583495 1702
uplevel 585265 3014
upvar 588347 4891
variable 593306 2204
vwait 595578 1237
while 596883 1757
