/* texuni, a program to convert TeX's accented characters to Unicode.
Version 2015-10-17.
Copyright 2015 Dmitri Pavlov.
Distributed under the terms of GNU General Public License version 3. */
#include <ctype.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>

struct {
  char c;
  const char *s;
} accents[] = {
  {'`', " ̀"},
  {'\'', " ́"},
  {'^', " ̂"},
  {'"', " ̈"},
  {'~', " ̃"},
  {'=', " ̄"},
  {'.', " ̇"},
  {'u', " ̆"},
  {'v', " ̌"},
  {'H', " ̋"},
  {'t', " ͡"},
  {'c', " ̧"},
  {'d', " ̣"},
  {'b', " ̱"},
};

struct {
  const char *c, *s;
} sletters[] = {
  {"oe", "œ"},
  {"OE", "Œ"},
  {"ae", "æ"},
  {"AE", "Æ"},
  {"aa", "å"},
  {"AA", "Å"},
  {"o", "ø"},
  {"O", "Ø"},
  {"l", "ł"},
  {"L", "Ł"},
  {"ss", "ß"},
  {"i", "i"},
  {"j", "j"},
/* The following does not work with accents:
  {"i", "ı"},
  {"j", "ȷ"},
*/
};

int line = 1, column = 0;

void die(const char *s) {
  fprintf(stderr, "line %d, column %d: %s\n", line, column, s);
  exit(1);
}

int get(void) {
  int c = getchar();
  if (ferror(stdin)) {
    perror("getchar");
    exit(1);
  }
  if (c == '\n') {
    column = 0;
    line++;
  } else if (c != EOF)
    column++;
  return c;
}

int sget(void) {
  int c = get();
  if (c == EOF)
    die("unexpected end of file");
  return c;
}

void put(int c) {
  if (putchar(c) == EOF) {
    perror("putchar");
    exit(1);
  }
  fflush(stdout);
}

void putstr(const char *s) {
  if (fputs(s, stdout) == EOF) {
    perror("fputs");
    exit(1);
  }
}

int main(void) {
  int c;
  while ((c = get()) != EOF)
    if (c != '\\' && c != '{') {
      put(c);
    } else {
      int accent = -1, curly = 0;
      if (c == '{') {
        curly++;
        c = sget();
        if (c != '\\') {
          put('{');
          put(c);
          continue;
        }
      }
      c = sget();
      int k;
      for (k = 0; k < sizeof(accents) / sizeof(accents[0]); k++)
        if (accents[k].c == c) {
          int d = get();
          if (isalpha(c) && isalpha(d)) {
            if (curly)
              put('{');
            put('\\');
            put(c);
            put(d);
            goto next;
          }
          if (d == ' ')
            while ((d = sget()) == ' ')
              ;
          if (d == '{') {
            d = sget();
            if (d == '\\') {
              accent = k;
              c = sget();
              curly++;
              break;
            }
            put(d);
            if (sget() != '}')
              die("expected }");
          } else if (d == '\\') {
            accent = k;
            c = sget();
            break;
          } else
            put(d);
          putstr(accents[k].s + 1);
          if (curly) {
            c = sget();
            if (c != '}')
              die("expected }");
          }
          goto next;
        }
      char seq[3] = {c, 0, 0};
      int d = sget();
      if (isalpha(d)) {
        seq[1] = d;
        d = sget();
      }
      if (isalpha(d)) {
        if (curly)
          put('{');
        put('\\');
        put(c);
        if (seq[1] != 0)
          put(seq[1]);
        put(d);
        goto next;
      }

      if (d == '{') {
        d = sget();
        curly++;
      }
      for (k = 0; k < sizeof(sletters) / sizeof(sletters[0]); k++)
        if (strcmp(sletters[k].c, seq) == 0) {
          if (curly && d != '}')
            die("expected }");
          putstr(sletters[k].s);
          if (accent >= 0)
            putstr(accents[accent].s + 1);
          if (d == '}' && curly == 2)
            d = sget();
          if (!isspace(d) && !curly)
            put(d);
          goto next;
        }
      if (curly)
        put('{');
      put('\\');
      put(c);
      if (seq[1] != 0)
        put(seq[1]);
      put(d);
    next:
      ;
    }
  return 0;
}