#!/usr/bin/gawk -f

# lw2xml: Conversion of Lexware band label format to xhtml
# (c) Campbell Webb <cw@camwebb.info> 2020
# License: GPLv3
# (See tail of file for documentation)

# TODO: see <https://github.com/alaskanlc/lexweb/issues>

BEGIN{

  # =====================  Initialization ==========================

  read_args()
  init()

  # Initialize Path and Prevbc (previous band class)
  Path = ""
  Prevbc = "0"
  # Start the xhtml output
  if (!XML)
    head("Dictionary")
  o("doc")

  # =====================  Line-by-line cycle ==========================
  # Read each line from the input file, until there are no more
  while ((getline < LEX) > 0) {
    Line++
    # Ignore comments, and read from start line to finish line (if given)
    if (($0) && ($0 !~ /^[ #]/) &&              \
        (Line >= STARTLINE) &&                  \
        (Line <= FINISHLINE)) {

      # Pre-process the line. See functions below
      indexmarks()
      xmlencode()
      dollar2sup()
      splitline()  # sets Bl, Bc, Word2, Rest1, Rest2
      
      # Skip some classes
      if (Bc ~ /^(cf|caus|xr|xgl|grp|smf|sc)$/)
        continue
      if (XML)
        if (Bc ~ /^(file|par|par3|com)$/)
          continue
      
      # =====================  Validation ==========================
      if (VALIDATE)
        if ((Bc != "com") && (Bc != "par") && (Bc != "par3"))
          validate()
  
      # ====================  Conversion to XHTML ==================
      # For each line, check the band label
      if (Bc == "rt") {
        # Close any open <div>s back up to the </div> inside <div class="doc">
        c("doc")
        # Open a new <div class="rt">
        o("rt")
        # These comments (#> ) following an o() or oc() can be extracted from
        #   the program to form a comprehensive outline of the XML structure
        #   with:  grep -E ' +#>' lw2xml | sed -E 's/^ +#>//g' | sort > \
        #            lw_xml_elements
        #> rt:   doc/rt
        o("rtattr")
        #> rt:   doc/rt/rtattr
        oc("rttype", substr(Bl,2))
        #> rt:   doc/rt/rtattr/rttype
        if (!XML)
          print wedge(0) "<a name=\"rt" ++Rootn "\"></a>"
        # Needed for the indexing:
        Rootlabel[Rootn] = Rest1
        oc("rtword", Rest1)
        #> rt:   doc/rt/rtattr/rtword
      }
  
      else if (Bc == "pd") {
        c("rtattr")
        oc("pd", ("/" Rest1 "/"))
        #> pd:   doc/rt/rtattr/pd
      }
  
      else if (Bc == "tag") {
        c("rtattr")
        oc("tag", Rest1)
        #> tag:  doc/rt/rtattr/tag
      }
  
      else if (Bc == "rtyp") {
        c("rtattr")
        oc("rtyp", Rest1)
        #> rtyp: doc/rt/rtattr/rtyp
      }
  
      else if (Bc == "nav") {
        c("rtattr")
        oc("nav", Rest1)
        #> nav:  doc/rt/rtattr/nav
      }
  
      else if (Bc == "df") {
        c("rtattr")
        oc("df", Rest1)
        #> df:   doc/rt/rtattr/df
      }
  
      else if (Bc == "sets") {
        c("rt")
        o("sets")
        #> sets: doc/rt/sets
      }
  
      else if (Bc == "set") {
        c("sets")
        o("set")
        #> set:  doc/rt/sets/set
        if (VALIDATE)
          if (!Aspcat[Word2])
            error(Word2 " not a known aspectual category")
        oc("setasp", Word2)
        #> set:  doc/rt/sets/set/setasp
        oc("setwords", Rest2)
        #> set:  doc/rt/sets/set/setwords
        c("sets")
      }

      else if (Bc == "th") {
        c("rt")
        o("th")
        #> th:   doc/rt/th
        o("thattr")
        #> th:   doc/rt/th/thattr
        oc("thword", Rest1)
        #> th:   doc/rt/th/thattr/thword
      }
  
      else if (Bc == "tc") {
        # tc may occur in several places. The following logic is needed to
        #   correctly balance the divs:
        if (Path ~ "thattr")
          c("thattr")
        else if (Path ~ "gc2attr")
          c("gc2attr")
        else if (Path ~ "th3")
          c("th3attr")
        
        oc("tc", Rest1)
        # These are the three places that tc can occur in the XML
        #> tc:   doc/rt/th/thattr/tc
        #> tc:   doc/rt/gc2/gc2attr/tc
        #> tc:   doc/rt/gc2/th3/th3attr/tc
      }

      else if (Bc == "cnj") {
        if (Path ~ "thattr")
          c("thattr")
        else if (Path ~ "gc2attr")
          c("gc2attr")
        else if (Path ~ "th3")
          c("th3attr")

        oc("cnj", Rest1)
        #> cnj:  doc/rt/th/thattr/cnj
        #> cnj:  doc/rt/gc2/gc2attr/cnj
        #> cnj:  doc/rt/gc2/th3/th3attr/cnj
      }
      
      else if (Bc == "gl") {
        if      (Path ~ "thattr")
          c("thattr")
        else if (Path ~ "gc2attr")
          c("gc2attr")
        else if (Path ~ "th3")
          c("th3attr")
        else if (Path ~ "gc3attr")
          c("gc3attr")
        
        oc("gl", Rest1)
        #> gl:   doc/rt/th/thattr/gl
        #> gl:   doc/rt/gc2/gc2attr/gl
        #> gl:   doc/rt/gc2/th3/th3attr/gl
        #> gl:   doc/rt/gc2/gc3/gc3attr/gl
        #> gl:   doc/rt/th/gc3/gc3attr/gl
      }
      
      else if (Bc == "quo") {
        if      (Path ~ "thattr")
          c("thattr")
        else if (Path ~ "th3attr")
          c("th3attr")
        else if (Path ~ "exengs")
          c("exeng")
        
        oc("quo", Rest1)
        #> quo:  doc/rt/th/thattr/quo
        #> quo:  doc/rt/gc2/th3/th3attr/quo
        #> quo:  doc/rt/th/thexengs/exeng/quo
        #> quo:  doc/rt/gc2/gc2exengs/exeng/quo
        #> quo:  doc/rt/gc2/th3/th3exengs/exeng/quo
        #> quo:  doc/rt/gc2/gc3/gc3exengs/exeng/quo
        #> quo:  doc/rt/th/gc3/gc3exengs/exeng/quo
      }

      else if (Bc == "cit") {
        if      (Path ~ "thattr")
          c("thattr")
        else if (Path ~ "th3attr")
          c("th3attr")
        else if (Path ~ "exengs")
          c("exeng")
        
        oc("cit", Rest1)
        #> cit:  doc/rt/th/thattr/cit
        #> cit:  doc/rt/gc2/th3/th3attr/cit
        #> cit:  doc/rt/th/thexengs/exeng/cit
        #> cit:  doc/rt/gc2/gc2exengs/exeng/cit
        #> cit:  doc/rt/gc2/th3/th3exengs/exeng/cit
        #> cit:  doc/rt/gc2/gc3/gc3exengs/exeng/cit
        #> cit:  doc/rt/th/gc3/gc3exengs/exeng/cit
      }

      else if (Bc == "ex") {
        # first one
        if (Path !~ "exengs") {
          if      (Path ~ "rt/th")
            c("th")
          else if (Path ~ "gc2/th3")
            c("th3")
          else if (Path ~ "gc2/gc3")
            c("gc3")
          else if (Path ~ "gc2")
            c("gc2")

          o("exengs")
          o("exeng")
          oc("ex", Rest1)
          #> ex:   doc/rt/th/exengs/exeng/ex
          #> ex:   doc/rt/gc2/exengs/exeng/ex
          #> ex:   doc/rt/gc2/th3/exengs/exeng/ex
          #> ex:   doc/rt/gc2/gc3/gc3exengs/exeng/ex
          #> ex:   doc/rt/th/gc3/gc3exengs/exeng/ex
          # oc("excolon", ": ")
        }
        # second or later
        else {
          c("exengs")
          o("exeng")
          oc("ex", Rest1)
          # oc("excolon", ": ")
        }
      }
      
      else if (Bc == "eng") {
        c("exeng")
        oc("eng", Rest1)
        #> eng:  doc/rt/th/exengs/exeng/eng
        #> eng:  doc/rt/gc2/exengs/exeng/eng
        #> eng:  doc/rt/gc2/th3/exengs/exeng/eng
        #> eng:  doc/rt/gc2/gc3/exengs/exeng/eng
        #> eng:  doc/rt/th/gc3/exengs/exeng/eng
        if (!XML)
          oc("exsemicolon", "; ")
      }

      else if (Bc == "lit") {
        if      (Path ~ "exengs")
          c("exeng")
        else if (Path ~ "gc3attr")
          c("gc3attr")
        else if (Path ~ "gc2attr")
          c("gc2attr")

        oc("lit", "lit. " Rest1)
        #> lit:  doc/rt/th/exengs/exeng/lit
        #> lit:  doc/rt/gc2/exengs/exeng/lit
        #> lit:  doc/rt/gc2/th3/exengs/exeng/lit
        #> lit:  doc/rt/gc2/gc3/exengs/exeng/lit
        #> lit:  doc/rt/th/gc3/exengs/exeng/lit
        #> lit:  doc/rt/gc2/gc2attr/lit
        #> lit:  doc/rt/gc2/gc3/gc3attr/lit
        #> lit:  doc/rt/th/gc3/gc3attr/lit
        if (!XML)
          oc("exsemicolon", "; ")
      }

      else if (Bc == "prds") {
        c("th")
        o("prds")
        oc("prdstype", Word2)
        #> prds: doc/rt/th/prds/prdstype
        oc("prdsdef", Rest2)
        #> prds: doc/rt/th/prds/prdsdef
      }

      else if (Bc == "prd") {
        c("prds")
        o("prd")
        oc("prdpers", Word2)
        #> prd:  doc/rt/th/prds/prd/prdpers
        oc("prdwords", Rest2)
        #> prd:  doc/rt/th/prds/prd/prdwords
      }

      else if (Bc == "prdgl") {
        c("prd")
        oc("prdgl", Rest1)
        #> prdgl:doc/rt/th/prds/prd/prdgl
      }

      else if (Bc == "gc2") {
        if (Path ~ "doc/rt")
          c("rt")
        if (Path ~ "doc/lw")
          c("lw")
        o("gc2")
        o("gc2attr")
        oc("gc2word", Rest1)
        #> gc2:  doc/rt/gc2/gc2attr/gc2word
        oc("gc2type", "(" abbrev(Bl) ")")
        #> gc2:  doc/rt/gc2/gc2attr/gc2type
      }
  
      else if (Bc == "dial") {
        # first one
        if (Path !~ /dials/) {
          if (Path ~ /gc2attr/) 
            c("gc2attr")
          if (Path ~ /gc3attr/) 
            c("gc3attr")
          if (Path ~ /exeng/) 
            c("exeng")
          
          o("dials")
          oc("dial", Word2)
          #> dial: doc/rt/gc2/gc2attr/dials/dial
          #> dial: doc/rt/gc2/gc3/gc3attr/dials/dial
          #> dial: doc/rt/th/gc3/gc3attr/dials/dial
          #> dial: ... exeng/dials/dial
          if (VALIDATE)
            if (!Dial[Word2])
              error(Word2 " not a known dialect")
        }
        # not first dial
        else if (Path ~ /dials/) {
          c("dials")
          o("dialx")
          if (!XML)
            oc("dialcomma", ",")
          oc("dial", Word2)
          #> dial: doc/rt/gc2/gc2attr/dials/dialx/dial
          #> dial: doc/rt/gc2/gc3/gc3attr/dials/dialx/dial
          #> dial: doc/rt/th/gc3/gc3attr/dials/dialx/dial
          #> dial: ... exeng/dials/dialx/dial
          if (!XML)
            oc("dialcolon", ":")
          oc("dialword", Rest2)
          #> dial: doc/rt/gc2/gc2attr/dials/dialx/dialword
          #> dial: doc/rt/gc2/gc3/gc3attr/dials/dialx/dialword
          #> dial: doc/rt/th/gc3/gc3attr/dials/dialx/dialword
          #> dial: ... exeng/dials/dialx/dialword
          c("dials")
        }
      }

      else if (Bc == "asp") {
        if (Path !~ "gc3")
          c("gc2attr")
        else if (Path ~ "gc3")
          c("gc3attr")
        oc("asp", Rest1)
        #> asp:  doc/rt/gc2/gc2attr/asp
        #> asp:  doc/rt/gc2/gc3/gc3attr/asp
        #> asp:  doc/rt/th/gc3/gc3attr/asp
      }
      
      else if (Bc == "th3") {
        c("gc2")
        o("th3")
        o("th3attr")
        oc("thword", Rest1)
        #> th3:  doc/rt/gc2/th3/th3attr/thword
      }
      
      # GC3
      else if (Bc == "gc3") {
        if (Path !~ "/th/") 
          c("gc2")
        if (Path ~ "/th/") { 
          c("th")
          if (VALIDATE)
            if (Bl !~ /^\.\.\.(mpn|in|pn|n)$/)
              error("Under ..th, the only gc3 allowed are n, in, pn, mpn")
        }
        o("gc3")
        o("gc3attr")
        oc("gc3word", Rest1)
        #> gc3:  doc/rt/gc2/gc3/gc3attr/gc3word
        #> gc3:  doc/rt/th/gc3/gc3attr/gc3word
        oc("gc3type", "(" abbrev(Bl) ")")
        #> gc3:  doc/rt/gc2/gc3/gc3attr/gc3type
        #> gc3:  doc/rt/th/gc3/gc3attr/gc3type
      }
  
      else if (Bc == "lw") {
        c("doc")
        o("lw")
        o("lwattr")
        oc("lwword", Rest1)
        #> lw:   doc/lw/lwattr/lwword
      }
  
      else if (Bc == "src") {
        c("lwattr")
        oc("src", Rest1 )
        #> src:  doc/lw/lwattr/src
      }

      else if (Bc == "com") {
        oc("com", Rest1 )
        #> com:  .../com
      }

      else if (Bc == "par") {
        oc("par", Rest1 )
        #> par:  .../par
      }

      else if (Bc == "par3") {
        oc("par3", Rest1 )
        #> par3:  .../par3
      }

      else if (Bc == "file") {
        c("doc")
        oc("file", Rest1 )
        #> file: doc/file
      }
      
      # Band categories and labels not known
      else if (Bc)
        error("Band category '" Bc "' not handled")
      else
        error("Band label '" Bl "' not known")
      
      # Move Bc to Prevbc, if not a com
      if ((Bc != "com") && (Bc != "par") && (Bc != "par3"))
        Prevbc = Bc
    }
  }

  c("doc")

  # Make the index
  if (MAKEINDEX)
    makeindex()
  
  # Close doc itself
  if (XML)
    print "    </doc>"
  else
    print "    </div>"
  # XHTML footer
  if (!XML)
    foot()

  # Successful program termination (exit status 0)
}



function init(     load, x, y, i, j, k, z, fol) {
  # Initialize

  # Set default sort order
  PROCINFO["sorted_in"] = "@ind_str_asc"
  
  # ** Load the syntax for validation **
  # fol[A][B] are bands allowed to follow band B, in the context of A
  #   A is needed, because whether a band can follow A depends on the context
  #   of B. E.g., nav (as a member of attr1) can be followed by df, sets
  #   th, gc2, or a new root. A is the exit Path of B.

  #   (To view the following sections, unwrap lines.
  #      M-x toggle-truncate-lines in Emacs)
  
  #   Occurrences:            ?   ?   ?    ?  ?   ?   <+  *   *  *  *  *
  fol["doc"   ]["0"   ] = "rt                                       lw file"
  fol["doc"   ]["file"] = "rt                                       lw file"
  fol["rtattr"]["rt"  ] = "   pd tag rtyp nav df sets     th gc2 rt lw file"
  fol["rtattr"]["pd"  ] = "      tag rtyp nav df sets     th gc2 rt lw file"
  fol["rtattr"]["tag" ] = "          rtyp nav df sets     th gc2 rt lw file"
  fol["rtattr"]["rtyp"] = "               nav df sets     th gc2 rt lw file"
  fol["rtattr"]["nav" ] = "                   df sets     th gc2 rt lw file"
  fol["rtattr"]["df"  ] = "                      sets     th gc2 rt lw file"
  fol["sets"  ]["sets"] = "                           set                  "
  fol["sets"  ]["set" ] = "                           set th gc2 rt lw file"

  #                              1   ?  1  ?    ?  *   *   <1   ?   ?   ?   *  *  *    <1  ?    *   *
  fol["thattr"]["th"   ] = "dial tc                                                           "
  fol["thattr"]["tc"   ] = "        cnj gl                                                    "
  fol["thattr"]["cnj"  ] = "            gl                                                    "
  fol["thattr"]["gl"   ] = "               quo cit ex                      gc3 th prds           gc2 th3 rt lw file"
  fol["thattr"]["quo"  ] = "                   cit ex                      gc3 th prds           gc2 th3 rt lw file"
  fol["thattr"]["cit"  ] = "                       ex                      gc3 th prds           gc2 th3 rt lw file"
  fol["exeng" ]["ex"   ] = "                          dial eng                                                     "
  fol["exeng" ]["dial" ] = "                               eng                                                     "
  fol["exeng" ]["eng"  ] = "                       ex          lit quo cit gc3 th prds           gc2 th3 rt lw file"
  fol["exeng" ]["lit"  ] = "                       ex              quo cit gc3 th prds           gc2 th3 rt lw file"
  fol["exeng" ]["quo"  ] = "                       ex                  cit gc3 th prds           gc2 th3 rt lw file"
  fol["exeng" ]["cit"  ] = "                       ex                      gc3 th prds           gc2 th3 rt lw file"
  fol["prds"  ]["prds" ] = "                                                           prd                         "
  fol["prd"   ]["prd"  ] = "                                               gc3 th prds prd prdgl gc2 th3 rt lw file"
  fol["prd"   ]["prdgl"] = "                                               gc3 th prds prd       gc2 th3 rt lw file"

  #                           ?       ?  ?   ?  1   ?  ?   ?  *   *  *   *   
  fol["gc2attr"]["gc2" ] = "dial     tc asp cnj gl            "
  fol["dials"  ]["dial"] = "dial eng tc asp cnj gl            "
  fol["dialx"  ]["dial"] = "dial     tc asp cnj gl            "
  fol["gc2attr"]["tc"  ] = "            asp cnj gl            "
  fol["gc2attr"]["asp" ] = "                cnj gl            "
  fol["gc2attr"]["cnj" ] = "                    gl            "
  fol["gc2attr"]["gl"  ] = "                       quo cit smf sc lit ex th3 gc3 gc2 rt lw file"
  fol["gc2attr"]["quo" ] = "                           cit smf sc lit ex th3 gc3 gc2 rt lw file"
  fol["gc2attr"]["cit" ] = "                               smf sc lit ex th3 gc3 gc2 rt lw file"
  fol["gc2attr"]["smf" ] = "                                   sc lit ex th3 gc3 gc2 rt lw file"
  fol["gc2attr"]["sc"  ] = "                                      lit ex th3 gc3 gc2 rt lw file"
  fol["gc2attr"]["lit" ] = "                                          ex th3 gc3 gc2 rt lw file"

  #                          1  ?  1  ?    ?  *   *    *   
  fol["th3attr"]["th3"  ] = "tc                                                           "
  fol["th3attr"]["tc"   ] = "   cnj gl                                                    "
  fol["th3attr"]["cnj"  ] = "       gl                                                    "
  fol["th3attr"]["gl"   ] = "          quo cit ex gc3 gc2 rt lw file"
  fol["th3attr"]["quo"  ] = "              cit ex gc3 gc2 rt lw file"
  fol["th3attr"]["cit"  ] = "                  ex gc3 gc2 rt lw file"

  #                           ?   ?  ?   ?  1   ?  ?   ?  *   *  *   *   
  fol["gc3attr"]["gc3" ] = "dial tc asp cnj gl            "
  fol["gc3attr"]["tc"  ] = "        asp cnj gl            "
  fol["gc3attr"]["asp" ] = "            cnj gl            "
  fol["gc3attr"]["cnj" ] = "                gl            "
  fol["gc3attr"]["gl"  ] = "                   quo cit smf sc lit ex th th3 gc3 gc2 rt lw file"
  fol["gc3attr"]["quo" ] = "                       cit smf sc lit ex th th3 gc3 gc2 rt lw file"
  fol["gc3attr"]["cit" ] = "                           smf sc lit ex th th3 gc3 gc2 rt lw file"
  fol["gc3attr"]["smf" ] = "                               sc lit ex th th3 gc3 gc2 rt lw file"
  fol["gc3attr"]["sc"  ] = "                                  lit ex th th3 gc3 gc2 rt lw file"
  fol["gc3attr"]["lit" ] = "                                      ex th th3 gc3 gc2 rt lw file"

  #
  fol["lwattr"]["lw" ] = "src "
  fol["lwattr"]["src"] = "gc2 rt lw file"
  
  
  # make F array from fol
  for (i in fol)
    for (j in fol[i]) {
      gsub(/^ +/,"",fol[i][j])
      gsub(/ +$/,"",fol[i][i])
      split(fol[i][j], z, / +/)
      for (k in z)
        F[i][j][z[k]] = 1
    }

  # Band labels to band classes
  load = \
    ".file file; ..par par; ...par par3; com com; ..grp grp; "               \
    ".rt rt; .af rt; .ra rt; .lw lw; "                              \
    "pd pd; tag tag; rtyp rtyp; nav nav; df df; "                          \
    "..sets sets; set set; "                                        \
    "..th th; ...th th3; tc tc; cnj cnj; gl gl; quo quo; cit cit; caus caus; "  \
    "ex ex; eng eng; lit lit; cf cf; "                                      \
    "...prds prds; prd prd; prdgl prdgl; "                          \
    "dial dial; asp asp; smf smf; sc sc; "                           \
    ".xr xr; xgl xgl; "                                              \
    "src src; "                                                      \
                                                                        \
    "..adj  gc2  adj.;   "                                              \
    "..ads  gc2  a.d.s.; "                                              \
    "..adv  gc2  adv.;   "                                              \
    "..an   gc2  a.n.;   "                                              \
    "..c    gc2  c.;     "                                              \
    "..cnj  gc2  cnj.;   "                                              \
    "..coll gc2  coll.;  "                                              \
    "..dem  gc2  dem.;   "                                              \
    "..dir  gc2  dir.;   "                                              \
    "..enc  gc2  enc.;   "                                              \
    "..exc  gc2  exc.;   "                                              \
    "..i    gc2  i.;     "                                              \
    "..ic   gc2  i.c.;   "                                              \
    "..in   gc2  ins.n.; "                                              \
    "..int  gc2  int.;   "                                              \
    "..mpn  gc2  m.p.n.; "                                              \
    "..n    gc2  n.;     "                                              \
    "..nc   gc2  n.c.;   "                                              \
    "..nds  gc2  n.d.s;  "                                              \
    "..nenc gc2  n.enc.; "                                              \
    "..nfaf gc2  n.f.af.;"                                              \
    "..nfsf gc2  n.f.sf.;"                                              \
    "..ni   gc2  n.i.;   "                                              \
    "..nic  gc2  n.i.c.; "                                              \
    "..nsf  gc2  n.sf.;  "                                              \
    "..padj gc2  p.adj.; "                                              \
    "..pf   gc2  pf.;    "                                              \
    "..pn   gc2  p.n.;   "                                              \
    "..pp   gc2  pp.;    "                                              \
    "..pro  gc2  pro.;   "                                              \
    "..psn  gc2  ps.n.;  "                                              \
    "..scnj gc2  s.cnj.; "                                              \
    "..sf   gc2  sf.;    "                                              \
    "..tfs  gc2  t.f.s.; "                                              \
    "..venc gc2  v.enc.; "                                              \
    "..vfaf gc2  v.f.af.;"                                              \
    "..vfsf gc2  v.f.sf.;"                                              \
    "..voc  gc2  voc.;   "                                              \
    "..vpf  gc2  v.pf.;  "                                              \
    "..vsf  gc2  v.sf.;  "                                              \
    "..vsf1 gc2  v.sf.1; "                                              \
                                                                        \
    "...adj  gc3  adj.;   "                                             \
    "...ads  gc3  a.d.s.; "                                             \
    "...adv  gc3  adv.;   "                                             \
    "...an   gc3  a.n.;   "                                             \
    "...c    gc3  c.;     "                                             \
    "...cnj  gc3  cnj.;   "                                             \
    "...coll gc3  coll.;  "                                             \
    "...dem  gc3  dem.;   "                                             \
    "...dir  gc3  dir.;   "                                             \
    "...enc  gc3  enc.;   "                                              \
    "...exc  gc3  exc.;   "                                              \
    "...i    gc3  i.;     "                                              \
    "...ic   gc3  i.c.;   "                                              \
    "...ifs  gc3  i.f.s.; "                                             \
    "...in   gc3  ins.n.; "                                              \
    "...int  gc3  int.;   "                                              \
    "...mpn  gc3  m.p.n.; "                                              \
    "...n    gc3  n.;     "                                              \
    "...nc   gc3  n.c.;   "                                              \
    "...nds  gc3  n.d.s;  "                                              \
    "...nenc gc3  n.enc.; "                                              \
    "...ni   gc3  n.i.;   "                                              \
    "...nic  gc3  n.i.c.; "                                              \
    "...padj gc3  p.adj.; "                                              \
    "...pf   gc3  pf.;    "                                              \
    "...pn   gc3  p.n.;   "                                              \
    "...pp   gc3  pp.;    "                                              \
    "...pro  gc3  pro.;   "                                              \
    "...psn  gc3  ps.n.;  "                                              \
    "...scnj gc3  s.cnj.; "                                              \
    "...sf   gc3  sf.;    "                                              \
    "...venc gc3  v.enc.; "                                              \
    "...voc  gc3  voc."                                                 \
  
  split(load, x, ";")
  for (i in x) {
    gsub(/^ +/,"",x[i])
    split(x[i], y)
    Bl2bc[y[1]] = y[2]
    if (y[3])
      Abbrev[y[1]] = y[3]
  }

  # Allowed aspectual cats
  load = "bisec|conc|cns|cona|cont|cust|dist|dur|durØ|durD|mom|mult|neu|per|"\
    "pers|prog|rep|rev|sem|tran"
  split(load, x, "|")
  for (i in x)
    Aspcat[x[i]] = 1

  # Minto-Nenana|Chena|M-N&Chena|Toklat|UpperS|Salcha|Goodpaster||Gwichin
  load = "MN|C|MC|T|Us|S|G|TG|OC|GG|all|all+Sl|E|ETy|I|IIl|IIlO|IIlSl|IIlU-O|I-IO|IL|Il|Il-I|IlI|IlI-OU|IlN|IlO|IlO-UI|IlOU-NL|IlT|IlU|Il-UIO|IlU-O|IO|I-O|I-OU|IO-U|IOU|IOUSl|ISl|IT|I-U|IU|IUIl|IUO|I-UO|IU-O|L|L-ki|ki|LN|L-N|L-NO|L-NOU|L-NUO|LO|L-O|L-OIl|L-ONU|LOU|L-OUI|L-UO|L-UOIl|L-UON|N|NIl|NIlO|NL|NL-IlOU|NLU|NL-UO|NLU-O|NO|NOIl|N-OU|N-OULIl|O|OE|OEIl|OETy|OI|O-I|OII|OIIl|OIIl-U|OIL|OIl|O-IlI|OIlSl|OIl-U|OIlU|OIU|O-IU|OI-U|OIUSl|OL|O-N|ON|ONIl|ON-L|ONL|O-NU|OSl|OSlIl|OSl-IU|OU|O-U|OU-I|O-UI|OUI|OUIl|OUIlN|OUL|OUN|O-UN|OUNIl-L|O-UO|S|Sl|Sl-IlN|SlO|SlOTyIl|SlTy|Su|T|T-UIO|Ty|U|UE|UI|U-I|UIIl|UIIl-O|UIl|UIl-IO|U-IO|UI-O|UIO|U-L|UL|UN|UNIl|UNL|UO|U-O|U-OI|UO-I|U-OIl|UOIl|UONIl-L|USl|UU"
  split(load, x, "|")
  for (i in x)
    Dial[x[i]] = 1
}

function o (class) {
  # open a div
  if (!XML) print wedge(0) "<div class=\"" class "\">"
  else      print wedge(0) "<" class ">"
  if (Path) 
    Path = Path "/" class
  else Path = class
}

function wedge (reduce,  p) {
  # print "[ " reduce " " (2*split(Path,p,"/") - 2*reduce +4) " " Path " ]"
  return sprintf("%*s", (2*split(Path,p,"/") - 2*reduce +4) ,"")
}

function oc (class, text) {
  # open and close a div
  if (!XML)  print wedge(0) "<div class=\"" class "\">" text "</div>"
  else       print wedge(0) "<" class ">" text "</" class ">"
}

function c(p,     oldlevel, newlevel, z, i) {
  # "close the hierarchy n levels to the new path indicated in p"
  oldlevel = split(Path, z, "/")
  Path = ""
  for (i = 1; i <= oldlevel ; i++) {
    # Rebuild the path to the new level
    Path = Path "/" z[i]
    # stop at the new terminal level
    #   note: this only works if all the possible path elements are unique
    if (p == z[i]) {
      newlevel = i
      break
    }
  }
  # check
  if (!newlevel) {
    error("c(" p ") not found")
    exit 1
  }
  # fix "/doc/..." to "doc/..."
  gsub(/^\//,"",Path)

  # Close the divs
  if (!XML)
    for (i = 1; i <= (oldlevel - newlevel); i++)
      print wedge(i-(oldlevel-newlevel)) "</div>" # <!-- " i " -->"
  else
    for (i = 1; i <= (oldlevel - newlevel); i++)
      print wedge(i-(oldlevel-newlevel)) "</" z[oldlevel+1-i] ">"
}

function splitline (     n, z, i) {
  Bl = Bc = Word2 = Rest1 = Rest2
  gsub(/ +$/,"",$0)
  n = split($0, z)
  Bl = z[1]
  Bc = Bl2bc[Bl]
  Word2 = z[2]
  Rest1 = z[2]
  for (i = 3; i <= n; i++) Rest1 = Rest1 " " z[i]
  Rest2 = z[3]
  for (i = 4; i <= n; i++) Rest2 = Rest2 " " z[i]
}

function head(title) {
  
  # Use html5
  print "<!DOCTYPE html>"
  print "<html xmlns=\"http://www.w3.org/1999/xhtml\">"
  print "  <head>\n    <title>" title "</title>"
  print "    <meta http-equiv=\"Content-Type\" content=\"text/html; " \
           "charset=utf-8\" />"
  print "    <link rel=\"stylesheet\" href=\"lw.css\"/>"
  # print "<link href=\"https://fonts.googleapis.com/css?family=Montserrat\" " \
  #         "rel=\"stylesheet\"/>"
  # print "<link href=\"../img/akflora.ico\" rel=\"shortcut icon\"  \
  #         type=\"image/x-icon\"/>"
  print "  </head>\n  <body>"
}


function foot() {
  print "  </body>\n</html>";
}


function abbrev(bl) {
  if (!Abbrev[bl]) {
    printf "%5d: '%s' has no abbreviation\n", Line, bl > "/dev/stderr"
    return bl
  }
  else
    return Abbrev[bl]
}

function validate(   z, n) {
  
  # find the context from the tail element of Path
  n = split(Path, z, "/")
  # if the current band class cannot follow the previous band class in the
  #   current context, write an error
  if (!F[z[n]][Prevbc][Bc])
    error(sprintf("%-5s (in context %-9s) cannot follow %-7s", ("'" Bc "'"), \
                  ("'" z[n] "'"), ("'" Prevbc "'")))
}

function error(msg) {
  printf "ERR %5d : %s\n", Line, msg > "/dev/stderr"

  if (DEBUG) {
    printf "LINE      : %-65s\n", substr($0,1,65)   > "/dev/stderr"
    # printf "PATH      : %-65s\n", substr(Path,1,65) > "/dev/stderr"
  }
}

function xmlencode() {
  gsub(/&/,"\\&amp;",$0)
  gsub(/</,"\\&lt;", $0)
  gsub(/>/,"\\&gt;", $0)
}

function   indexmarks(     i, j, ind) {
  if ($0 ~ /\*/) {
    # (can't use a single regex, so char by char)
    ind = 0; j = ""
    for (i = 1; i <= length($0); i++) {
      if (ind) {
        if ((substr($0, i, 1) !~ /[[:alpha:]]/) || \
            (i == length($0))) {
          if (i == length($0))
            j = j substr($0, i, 1)
          Index[j][Rootn]++
          ind = 0
          j = ""
        }
        else j = j substr($0, i, 1)
      }
      else if (substr($0, i, 1) == "*")
        ind = 1
    }
  }
  gsub(/\*/,"",$0)
  gsub(/\]/,"", $0)
}

function dollar2sup() {
  $0 = gensub(/\$(.)/,"<sup>\\1</sup>","G",$0)
}

function case_fold_compare(i1, v1, i2, v2,    l, r) {
    l = tolower(i1)
    r = tolower(i2)

    if (l < r)
      return -1
    else if (l == r)
      return 0
    else
      return 1
}



function makeindex(    si) {
  PROCINFO["sorted_in"] = "@ind_num_asc"
  IGNORECASE = 1
  # note IGNORECASE does not affect array subscripting, so asorti() needed
  asorti(Index, si)
  
  o("index")
  oc("indexlabel", "Index")
  o("indexitems")
  for (i in si) {
    o("indexitem")
    oc("indexword", (si[i] ":"))
    o("indexrefs")
    # numerical index. Ordered the same as the occurence of the entry in
    # the lexware file. Not necessarily in alphabetical order
    for (j in Index[si[i]])
      oc("indexref", ("<a href=\"#rt" j "\">" Rootlabel[j] "</a>, "))
    c("indexitems")
  }
  c("doc")

}


function read_args(    arg) {
  # Defaults
  DEBUG = 0
  VALIDATE = 1
  STARTLINE = 1
  FINISHLINE = 1000000000

  for (arg = 1; arg < ARGC; arg++) {
    if      (ARGV[arg] == "--s")     STARTLINE  = ARGV[arg+1]
    else if (ARGV[arg] == "--f")     FINISHLINE = ARGV[arg+1]
    else if (ARGV[arg] == "--d")     DEBUG      = 1
    else if (ARGV[arg] == "--nv")    VALIDATE   = 0
    else if (ARGV[arg] == "--xml")   XML        = 1
    else if (ARGV[arg] == "--index") MAKEINDEX  = 1
  }
  LEX = ARGV[ARGC-1]
  if (XML)
    MAKEINDEX = 0

  # removed to make it portable to Windows
  # ("test -f " LEX " ; echo $?") | getline exitcode
  # if (exitcode) {
  if (ARGC == 1) {
    print "Usage: lw2xml [ --index --s <start line> --f <finish line> ] " \
      "<lexware file>"
    exit 1
  }
}

# =================== DOCUMENTATION ======================================

# Variable naming convention: Global variables begin with a Capital letter,
#   local variables with a lowercase letter

# XHTML Document structure
# ------------------------
# 
# * XML schema choices made to reflect blocking choice in html document.
# * If a block may need to be formatted independently, it needs a unique id
#   e.g., prdsexengs, but if it will always be the same, the id can be
#   reused in different contexts (e.g., exeng)
# * Regenerate with: grep -E ' +#>' lw2xml | sed -E 's/^ +#>/# /g'
#