⟦f114de5c8⟧

TextFile

*** ctex.ch	Tue Mar 19 15:42:34 1991
--- trie.ch	Tue Mar 19 15:43:09 1991
***************
*** 43,45 ****
  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
! % [1.2] banner
  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
--- 43,45 ----
  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
! % [1.2] banner (changed to show trie extension version)
  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
***************
*** 48,50 ****
  @y
! @d banner=='This is TeX, C Version 3.14' {printed when \TeX\ starts}
  @z
--- 48,50 ----
  @y
! @d banner=='This is TeX, C Version 3.14t3' {printed when \TeX\ starts}
  @z
***************
*** 207,212 ****
    at most |max_halfword|}
! @!trie_size=8000; {space for hyphenation patterns; should be larger for
    \.{INITEX} than it is in production versions of \TeX}
! @!trie_op_size=500; {space for ``opcodes'' in the hyphenation patterns}
! @!neg_trie_op_size=-500; {for lower trie_op_hash array bound}
  @!dvi_buf_size=16384; {size of the output buffer; must be a multiple of 8}
--- 207,214 ----
    at most |max_halfword|}
! @!trie_size=24000; {space for hyphenation patterns; should be larger for
    \.{INITEX} than it is in production versions of \TeX}
! @!trie_op_size=750; {space for ``opcodes'' in the hyphenation patterns}
! @!neg_trie_op_size=-750; {for lower trie_op_hash array bound}
! @!min_trie_op=0; {first possible trie op code for any language}
! @!max_trie_op=500; {largest possible trie op code for any language}
  @!dvi_buf_size=16384; {size of the output buffer; must be a multiple of 8}
***************
*** 1043,1044 ****
--- 1045,1100 ----
  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+ % [43.950,958,960,963] Larger tries.
+ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+ @x
+ @d trie_back(#)==trie[#].lh {backward links in |trie| holes}
+ @y
+ @d trie_back(#)==trie_tro[#] {use the opcode field now for backward links}
+ @z
+ @x
+ @<Move the data into |trie|@>=
+ h.rh:=0; h.b0:=min_quarterword; h.b1:=min_quarterword; {|trie_link:=0|,
+   |trie_op:=min_quarterword|, |trie_char:=qi(0)|}
+ if trie_root=0 then {no patterns were given}
+   begin for r:=0 to 256 do trie[r]:=h;
+ @y
+ @d clear_trie == {clear |trie[r]|}
+   begin trie_link(r):=0;
+   trie_op(r):=min_trie_op;
+   trie_char(r):=min_quarterword; {|trie_char:=qi(0)|}
+   end
+ @<Move the data into |trie|@>=
+ if trie_root=0 then {no patterns were given}
+   begin for r:=0 to 256 do clear_trie;
+ @z
+ @x
+   repeat s:=trie_link(r); trie[r]:=h; r:=s;
+ @y
+   repeat s:=trie_link(r); clear_trie; r:=s;
+ @z
+ @x
+ @!v:quarterword; {trie op code}
+ @y
+ @!v:trie_opcode; {trie op code}
+ @z
+ @x
+ if trie_o[q]<>min_quarterword then
+ @y
+ if trie_o[q]<>min_trie_op then
+ @z
+ @x
+ trie_c[p]:=si(c); trie_o[p]:=min_quarterword;
+ @y
+ trie_c[p]:=si(c); trie_o[p]:=min_trie_op;
+ @z
+ @x
+ l:=k; v:=min_quarterword;
+ @y
+ l:=k; v:=min_trie_op;
+ @z
+ @x
+ @!h:two_halves; {template used to zero out |trie|'s holes}
+ @y
+ @z
+ 
+ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  % [29.537] Use a path when calling a_open_in to do a \input; also, try
***************
*** 1167,1168 ****
--- 1223,1313 ----
  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+ % [42.920,921,923,924] Allow larger hyphenation tries.
+ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+ @x
+ Comparatively few different number sequences $n_0\ldots n_k$ actually occur,
+ since most of the |n|'s are generally zero. Therefore the number sequences
+ are encoded in such a way that |trie_op|$(z_k)$ is only one byte long.
+ If |trie_op(@t$z_k$@>)<>min_quarterword|, when $p_1\ldots p_k$ has matched
+ the letters in |hc[(l-k+1)..l@,]| of language |t|,
+ we perform all of the required operations
+ for this pattern by carrying out the following little program: Set
+ |v:=trie_op(@t$z_k$@>)|. Then set |v:=v+op_start[t]|,
+ |hyf[l-hyf_distance[v]]:=@tmax@>(hyf[l-hyf_distance[v]], hyf_num[v])|,
+ and |v:=hyf_next[v]|; repeat, if necessary, until |v=min_quarterword|.
+ @y
+ The theory that comparatively few different number sequences $n_0\ldots n_k$
+ actually occur, since most of the |n|'s are generally zero, seems to fail
+ at least for the large German hyphenation patterns.
+ Therefore the number sequences cannot any longer be encoded in such a way
+ that |trie_op|$(z_k)$ is only one byte long.
+ We have introduced a new constant |max_trie_op| for the maximum allowable
+ hyphenation operation code value; |max_trie_op| might be different for
+ \TeX\ and \.{INITEX} and must not exceed |max_halfword|.
+ An opcode will occupy a halfword if |max_trie_op| exceeds |max_quarterword|
+ or a quarterword otherwise.
+ @^system dependencies@>
+ If |trie_op(@t$z_k$@>)<>min_trie_op|, when $p_1\ldots p_k$ has matched
+ the letters in |hc[(l-k+1)..l@,]| of language |t|,
+ we perform all of the required operations
+ for this pattern by carrying out the following little program: Set
+ |v:=trie_op(@t$z_k$@>)|. Then set |v:=v+op_start[t]|,
+ |hyf[l-hyf_distance[v]]:=@tmax@>(hyf[l-hyf_distance[v]], hyf_num[v])|,
+ and |v:=hyf_next[v]|; repeat, if necessary, until |v=min_trie_op|.
+ @z
+ @x
+ @!trie_pointer=0..trie_size; {an index into |trie|}
+ @y
+ @!trie_opcode=min_trie_op..max_trie_op;  {a trie opcode}
+ @!trie_pointer=0..trie_size; {an index into |trie|}
+ @z
+ @x
+ @ @d trie_link(#)==trie[#].rh {``downward'' link in a trie}
+ @d trie_char(#)==trie[#].b1 {character matched at this trie location}
+ @d trie_op(#)==trie[#].b0 {program for hyphenation at this trie location}
+ @y
+ @ For more than 255 trie op codes, the three fields |trie_link|, |trie_char|,
+ and |trie_op| will no longer fit into one memory word; thus using web2c
+ we define |trie| as three array instead of an array of records.
+ The variant will be implented by reusing the opcode field later on with
+ another macro.
+ @d trie_link(#)==trie_trl[#] {``downward'' link in a trie}
+ @d trie_char(#)==trie_trc[#] {character matched at this trie location}
+ @d trie_op(#)==trie_tro[#] {program for hyphenation at this trie location}
+ @z
+ @x
+ @!trie:array[trie_pointer] of two_halves; {|trie_link|, |trie_char|, |trie_op|
+ @y
+ @!trie_trl:array[trie_pointer] of halfword; {|trie_link|}
+ @!trie_tro:array[trie_pointer] of halfword; {|trie_op| and |trie_link|}
+ @!trie_trc:array[trie_pointer] of quarterword; {|trie_char|}
+ @z
+ @x
+ @!hyf_next:array[1..trie_op_size] of quarterword; {continuation code}
+ @y
+ @!hyf_next:array[1..trie_op_size] of trie_opcode; {continuation code}
+ @z
+ @x
+     begin if trie_op(z)<>min_quarterword then
+ @y
+     begin if trie_op(z)<>min_trie_op then
+ @z
+ @x
+ until v=min_quarterword;
+ @y
+ until v=min_trie_op;
+ @z
+ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+ % [43.943] Larger tries, also in documentation parts.
+ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+ @x
+ |hyf_next[@t$v^\prime$@>]=min_quarterword|.
+ @y
+ |hyf_next[@t$v^\prime$@>]=min_trie_op|.
+ @z
+ @x
+ $$\hbox{|@t$v^\prime$@>:=new_trie_op(0,1,min_quarterword)|,\qquad
+ @y
+ $$\hbox{|@t$v^\prime$@>:=new_trie_op(0,1,min_trie_op)|,\qquad
+ @z
+ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  % [43.?] Hack around the negative lower bound here.
***************
*** 1183,1188 ****
--- 1328,1364 ----
  
+ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+ % [43.943,944] Larger hyphenation tries.
+ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+ @x
+ @!trie_used:array[ASCII_code] of quarterword;
+ @y
+ @!trie_used:array[ASCII_code] of trie_opcode;
+ @z
+ @x
+ @!trie_op_val:array[1..trie_op_size] of quarterword;
+ @y
+ @!trie_op_val:array[1..trie_op_size] of trie_opcode;
+ @z
+ @x
+ tini
+ @y
+ tini@;
+ @!max_op_used:trie_opcode; {largest opcode used for any language}
+ @!small_op:boolean; {flag used while dumping or undumping}
+ @z
+ @x
+ |new_trie_op| could return |min_quarterword| (thereby simply ignoring
+ @y
+ |new_trie_op| could return |min_trie_op| (thereby simply ignoring
+ @z
  @x
+ function new_trie_op(@!d,@!n:small_number;@!v:quarterword):quarterword;
+ label exit;
  var h:-trie_op_size..trie_op_size; {trial hash location}
+ @!u:quarterword; {trial op code}
  @y
+ function new_trie_op(@!d,@!n:small_number;@!v:trie_opcode):trie_opcode;
+ label exit;
  var h:neg_trie_op_size..trie_op_size; {trial hash location}
+ @!u:trie_opcode; {trial op code}
  @z
***************
*** 1200,1201 ****
--- 1376,1413 ----
  %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+ % [43.944,945,946] And larger tries again.
+ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+ @x
+     if u=max_quarterword then
+       overflow("pattern memory ops per language",
+         max_quarterword-min_quarterword);
+     incr(trie_op_ptr); incr(u); trie_used[cur_lang]:=u;
+ @y
+     if u=max_trie_op then
+       overflow("pattern memory ops per language",
+       max_trie_op-min_trie_op);
+     incr(trie_op_ptr); incr(u); trie_used[cur_lang]:=u;
+     if u>max_op_used then max_op_used:=u;
+ @z
+ @x
+ op_start[0]:=-min_quarterword;
+ @y
+ op_start[0]:=-min_trie_op;
+ @z
+ @x
+ for k:=0 to 255 do trie_used[k]:=min_quarterword;
+ @y
+ for k:=0 to 255 do trie_used[k]:=min_trie_op;
+ @z
+ @x
+ trie_op_ptr:=0;
+ @y
+ max_op_used:=min_trie_op;
+ trie_op_ptr:=0;
+ @z
+ @x
+ @t\hskip10pt@>@!trie_o:packed array[trie_pointer] of quarterword;
+ @y
+ @t\hskip10pt@>@!trie_o:packed array[trie_pointer] of trie_opcode;
+ @z
+ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  % [43.947] And another.
***************
*** 1555,1557 ****
  @y
! dump_things(trie[0], trie_max+1);
  dump_int(trie_op_ptr);
--- 1767,1771 ----
  @y
! dump_things(trie_trl[0], trie_max+1);
! dump_things(trie_tro[0], trie_max+1);
! dump_things(trie_trc[0], trie_max+1);
  dump_int(trie_op_ptr);
***************
*** 1571,1573 ****
  @y
! undump_things(trie[0], j+1);
  undump_size(0)(trie_op_size)('trie op size')(j); @+init trie_op_ptr:=j;@+tini
--- 1785,1789 ----
  @y
! undump_things(trie_trl[0], j+1);
! undump_things(trie_tro[0], j+1);
! undump_things(trie_trc[0], j+1);
  undump_size(0)(trie_op_size)('trie op size')(j); @+init trie_op_ptr:=j;@+tini
DataMuseum.dk

DKUUG/EUUG Conference tapes

⟦f114de5c8⟧ TextFile

Derivation

TextFile
