oed.c

这是一个同样来自贝尔实验室的和UNIX有着渊源的操作系统, 其简洁的设计和实现易于我们学习和理解

	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,
/*10*/	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,
	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,
/*20*/	L' ',	L'!',	L'"',	L'#',	L'$',	L'%',	SPCS,	L'\'',
	L'₍',	L'₎',	L'*',	L'₊',	L',',	L'₋',	L'.',	L'/',
/*30*/  L'₀',	L'₁',	L'₂',	L'₃',	L'₄',	L'₅',	L'₆',	L'₇',
	L'₈',	L'₉',	L':',	L';',	TAGS,	L'₌',	TAGE,	L'?',
/*40*/  L'@',	L'A',	L'B',	L'C',	L'D',	L'E',	L'F',	L'G',
	L'H',	L'I',	L'J',	L'K',	L'L',	L'M',	L'N',	L'O',
/*50*/	L'P',	L'Q',	L'R',	L'S',	L'T',	L'U',	L'V',	L'W',
	L'X',	L'Y',	L'Z',	L'[',	L'\\',	L']',	L'^',	L'_',
/*60*/	L'`',	L'a',	L'b',	L'c',	L'd',	L'e',	L'f',	L'g',
	L'h',	L'i',	L'j',	L'k',	L'l',	L'm',	L'n',	L'o',
/*70*/	L'p',	L'q',	L'r',	L's',	L't',	L'u',	L'v',	L'w',
	L'x',	L'y',	L'z',	L'{',	L'|',	L'}',	L'~',	NONE,
};
static Rune suptab[128] = {
	/*0*/	/*1*/	/*2*/	/*3*/	/*4*/	/*5*/	/*6*/	/*7*/
/*00*/	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,
	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,
/*10*/	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,
	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,	NONE,
/*20*/	L' ',	L'!',	L'"',	L'#',	L'$',	L'%',	SPCS,	L'\'',
	L'⁽',	L'⁾',	L'*',	L'⁺',	L',',	L'⁻',	L'.',	L'/',
/*30*/  L'⁰',	L'ⁱ',	L'⁲',	L'⁳',	L'⁴',	L'⁵',	L'⁶',	L'⁷',
	L'⁸',	L'⁹',	L':',	L';',	TAGS,	L'⁼',	TAGE,	L'?',
/*40*/  L'@',	L'A',	L'B',	L'C',	L'D',	L'E',	L'F',	L'G',
	L'H',	L'I',	L'J',	L'K',	L'L',	L'M',	L'N',	L'O',
/*50*/	L'P',	L'Q',	L'R',	L'S',	L'T',	L'U',	L'V',	L'W',
	L'X',	L'Y',	L'Z',	L'[',	L'\\',	L']',	L'^',	L'_',
/*60*/	L'`',	L'a',	L'b',	L'c',	L'd',	L'e',	L'f',	L'g',
	L'h',	L'i',	L'j',	L'k',	L'l',	L'm',	L'n',	L'o',
/*70*/	L'p',	L'q',	L'r',	L's',	L't',	L'u',	L'v',	L'w',
	L'x',	L'y',	L'z',	L'{',	L'|',	L'}',	L'~',	NONE,
};

static int	tagstarts;
static char	tag[Buflen];
static int	naux;
static char	auxname[Maxaux][Buflen];
static char	auxval[Maxaux][Buflen];
static char	spec[Buflen];
static char	*auxstate[Naux];	/* vals for most recent tag */
static Entry	curentry;
#define cursize (curentry.end-curentry.start)

static char	*getspec(char *, char *);
static char	*gettag(char *, char *);
static void	dostatus(void);

/*
 * cmd is one of:
 *    'p': normal print
 *    'h': just print headwords
 *    'P': print raw
 */
void
oedprintentry(Entry e, int cmd)
{
	char *p, *pe;
	int t, a, i;
	long r, rprev, rlig;
	Rune *transtab;

	p = e.start;
	pe = e.end;
	transtab = normtab;
	rprev = NONE;
	changett(0, 0, 0);
	curentry = e;
	if(cmd == 'h')
		outinhibit = 1;
	while(p < pe) {
		if(cmd == 'r') {
			outchar(*p++);
			continue;
		}
		r = transtab[(*p++)&0x7F];
		if(r < NONE) {
			/* Emit the rune, but buffer in case of ligature */
			if(rprev != NONE)
				outrune(rprev);
			rprev = r;
		} else if(r == SPCS) {
			/* Start of special character name */
			p = getspec(p, pe);
			r = lookassoc(spectab, asize(spectab), spec);
			if(r == -1) {
				if(debug)
					err("spec %ld %d %s",
						e.doff, cursize, spec);
				r = L'�';
			}
			if(r >= LIGS && r < LIGE) {
				/* handle possible ligature */
				rlig = liglookup(r, rprev);
				if(rlig != NONE)
					rprev = rlig;	/* overwrite rprev */
				else {
					/* could print accent, but let's not */
					if(rprev != NONE) outrune(rprev);
					rprev = NONE;
				}
			} else if(r >= MULTI && r < MULTIE) {
				if(rprev != NONE) {
					outrune(rprev);
					rprev = NONE;
				}
				outrunes(multitab[r-MULTI]);
			} else if(r == PAR) {
				if(rprev != NONE) {
					outrune(rprev);
					rprev = NONE;
				}
				outnl(1);
			} else {
				if(rprev != NONE) outrune(rprev);
				rprev = r;
			}
		} else if(r == TAGS) {
			/* Start of tag name */
			if(rprev != NONE) {
				outrune(rprev);
				rprev = NONE;
			}
			p = gettag(p, pe);
			t = lookassoc(tagtab, asize(tagtab), tag);
			if(t == -1) {
				if(debug)
					err("tag %ld %d %s",
						e.doff, cursize, tag);
				continue;
			}
			for(i = 0; i < Naux; i++)
				auxstate[i] = 0;
			for(i = 0; i < naux; i++) {
				a = lookassoc(auxtab, asize(auxtab), auxname[i]);
				if(a == -1) {
					if(debug)
						err("aux %ld %d %s",
							e.doff, cursize, auxname[i]);
				} else
					auxstate[a] = auxval[i];
			}
			switch(t){
			case E:
			case Ve:
				outnl(0);
				if(tagstarts)
					dostatus();
				break;
			case Ed:
			case Etym:
				outchar(tagstarts? '[' : ']');
				break;
			case Pr:
				outchar(tagstarts? '(' : ')');
				break;
			case In:
				transtab = changett(transtab, subtab, tagstarts);
				break;
			case Hm:
			case Su:
			case Fq:
				transtab = changett(transtab, suptab, tagstarts);
				break;
			case Gk:
				transtab = changett(transtab, grtab, tagstarts);
				break;
			case Ph:
				transtab = changett(transtab, phtab, tagstarts);
				break;
			case Hw:
				if(cmd == 'h') {
					if(!tagstarts)
						outchar(' ');
					outinhibit = !tagstarts;
				}
				break;
			case S0:
			case S1:
			case S2:
			case S3:
			case S4:
			case S5:
			case S6:
			case S7a:
			case S7n:
			case Sn:
			case Sgk:
				if(tagstarts) {
					outnl(2);
					dostatus();
					if(auxstate[Num]) {
						if(t == S3 || t == S5) {
							i = atoi(auxstate[Num]);
							while(i--)
								outchar('*');
							outchars("  ");
						} else if(t == S7a || t == S7n || t == Sn) {
							outchar('(');
							outchars(auxstate[Num]);
							outchars(") ");
						} else if(t == Sgk) {
							i = grtab[auxstate[Num][0]];
							if(i != NONE)
								outrune(i);
							outchars(".  ");
						} else {
							outchars(auxstate[Num]);
							outchars(".  ");
						}
					}
				}
				break;
			case Cb:
			case Db:
			case Qp:
			case P:
				if(tagstarts)
					outnl(1);
				break;
			case Table:
				/*
				 * Todo: gather columns, justify them, etc.
				 * For now, just let colums come out as rows
				 */
				if(!tagstarts)
					outnl(0);
				break;
			case Col:
				if(tagstarts)
					outnl(0);
				break;
			case Dn:
				if(tagstarts)
					outchar('/');
				break;
			}
		}
	}
	if(cmd == 'h') {
		outinhibit = 0;
		outnl(0);
	}
}

/*
 * Return offset into bdict where next oed entry after fromoff starts.
 * Oed entries start with <e>, <ve>, <e st=...>, or <ve st=...>
 */
long
oednextoff(long fromoff)
{
	long a, n;
	int c;

	a = Bseek(bdict, fromoff, 0);
	if(a < 0)
		return -1;
	n = 0;
	for(;;) {
		c = Bgetc(bdict);
		if(c < 0)
			break;
		if(c == '<') {
			c = Bgetc(bdict);
			if(c == 'e') {
				c = Bgetc(bdict);
				if(c == '>' || c == ' ')
					n = 3;
			} else if(c == 'v' && Bgetc(bdict) == 'e') {
				c = Bgetc(bdict);
				if(c == '>' || c == ' ')
					n = 4;
			}
			if(n)
				break;
		}
	}
	return (Boffset(bdict)-n);
}

static char *prkey =
"KEY TO THE PRONUNCIATION\n"
"\n"
"I. CONSONANTS\n"
"b, d, f, k, l, m, n, p, t, v, z: usual English values\n"
"\n"
"g as in go (gəʊ)\n"
"h  ...  ho! (həʊ)\n"
"r  ...  run (rʌn), terrier (ˈtɛriə(r))\n"
"(r)...  her (hɜː(r))\n"
"s  ...  see (siː), success (səkˈsɜs)\n"
"w  ...  wear (wɛə(r))\n"
"hw ...  when (hwɛn)\n"
"j  ...  yes (jɛs)\n"
"θ  ...  thin (θin), bath (bɑːθ)\n"
"ð  ...  then (ðɛn), bathe (beɪð)\n"
"ʃ  ...  shop (ʃɒp), dish (dɪʃ)\n"
"tʃ ...  chop (tʃɒp), ditch (dɪtʃ)\n"
"ʒ  ...  vision (ˈvɪʒən), déjeuner (deʒøne)\n"
"dʒ ...  judge (dʒʌdʒ)\n"
"ŋ  ...  singing (ˈsɪŋɪŋ), think (θiŋk)\n"
"ŋg ...  finger (ˈfiŋgə(r))\n"
"\n"
"Foreign\n"
"ʎ as in It. seraglio (serˈraʎo)\n"
"ɲ  ...  Fr. cognac (kɔɲak)\n"
"x  ...  Ger. ach (ax), Sc. loch (lɒx)\n"
"ç  ...  Ger. ich (ɪç), Sc. nicht (nɪçt)\n"
"ɣ  ...  North Ger. sagen (ˈzaːɣən)\n"
"c  ...  Afrikaans baardmannetjie (ˈbaːrtmanəci)\n"
"ɥ  ...  Fr. cuisine (kɥizin)\n"
"\n"
"II. VOWELS AND DIPTHONGS\n"
"\n"
"Short\n"
"ɪ as in pit (pɪt), -ness (-nɪs)\n"
"ɛ  ...  pet (pɛt), Fr. sept (sɛt)\n"
"æ  ...  pat (pæt)\n"
"ʌ  ...  putt (pʌt)\n"
"ɒ  ...  pot (pɒt)\n"
"ʊ  ...  put (pʊt)\n"
"ə  ...  another (əˈnʌðə(r))\n"
"(ə)...  beaten (ˈbiːt(ə)n)\n"
"i  ...  Fr. si (si)\n"
"e  ...  Fr. bébé (bebe)\n"
"a  ...  Fr. mari (mari)\n"
"ɑ  ...  Fr. bâtiment (bɑtimã)\n"
"ɔ  ...  Fr. homme (ɔm)\n"
"o  ...  Fr. eau (o)\n"
"ø  ...  Fr. peu (pø)\n"
"œ  ...  Fr. boeuf (bœf), coeur (kœr)\n"
"u  ...  Fr. douce (dus)\n"
"ʏ  ...  Ger. Müller (ˈmʏlər)\n"
"y  ...  Fr. du (dy)\n"
"\n"
"Long\n"
"iː as in bean (biːn)\n"
"ɑː ...  barn (bɑːn)\n"
"ɔː ...  born (bɔːn)\n"
"uː ...  boon (buːn)\n"
"ɜː ...  burn (bɜːn)\n"
"eː ...  Ger. Schnee (ʃneː)\n"
"ɛː ...  Ger. Fähre (ˈfɛːrə)\n"
"aː ...  Ger. Tag (taːk)\n"
"oː ...  Ger. Sohn (zoːn)\n"
"øː ...  Ger. Goethe (gøːtə)\n"
"yː ...  Ger. grün (gryːn)\n"
"\n"
"Nasal\n"
"ɛ˜, æ˜ as in Fr. fin (fɛ˜, fæ˜)\n"
"ã  ...  Fr. franc (frã)\n"
"ɔ˜ ...  Fr. bon (bɔ˜n)\n"
"œ˜ ...  Fr. un (œ˜)\n"
"\n"
"Dipthongs, etc.\n"
"eɪ as in bay (beɪ)\n"
"aɪ ...  buy (baɪ)\n"
"ɔɪ ...  boy (bɔɪ)\n"
"əʊ ...  no (nəʊ)\n"
"aʊ ...  now (naʊ)\n"
"ɪə ...  peer (pɪə(r))\n"
"ɛə ...  pair (pɛə(r))\n"
"ʊə ...  tour (tʊə(r))\n"
"ɔə ...  boar (bɔə(r))\n"
"\n"
"III. STRESS\n"
"\n"
"Main stress: ˈ preceding stressed syllable\n"
"Secondary stress: ˌ preceding stressed syllable\n"
"\n"
"E.g.: pronunciation (prəˌnʌnsɪˈeɪʃ(ə)n)\n";
/* TODO: find transcriptions of foreign consonents, œ, ʏ, nasals */

void
oedprintkey(void)
{
	Bprint(bout, "%s", prkey);
}

/*
 * f points just after a '&', fe points at end of entry.
 * Accumulate the special name, starting after the &
 * and continuing until the next '.', in spec[].
 * Return pointer to char after '.'.
 */
static char *
getspec(char *f, char *fe)
{
	char *t;
	int c, i;

	t = spec;
	i = sizeof spec;
	while(--i > 0) {
		c = *f++;
		if(c == '.' || f == fe)
			break;
		*t++ = c;
	}
	*t = 0;
	return f;
}

/*
 * f points just after '<'; fe points at end of entry.
 * Expect next characters from bin to match:
 *  [/][^ >]+( [^>=]+=[^ >]+)*>
 *      tag   auxname auxval
 * Accumulate the tag and its auxilliary information in
 * tag[], auxname[][] and auxval[][].
 * Set tagstarts=1 if the tag is 'starting' (has no '/'), else 0.
 * Set naux to the number of aux pairs found.
 * Return pointer to after final '>'.
 */
static char *
gettag(char *f, char *fe)
{
	char *t;
	int c, i;

	t = tag;
	c = *f++;
	if(c == '/')
		tagstarts = 0;
	else {
		tagstarts = 1;
		*t++ = c;
	}
	i = Buflen;
	naux = 0;
	while(--i > 0) {
		c = *f++;
		if(c == '>' || f == fe)
			break;
		if(c == ' ') {
			*t = 0;
			t = auxname[naux];
			i = Buflen;
			if(naux < Maxaux-1)
				naux++;
		} else if(naux && c == '=') {
			*t = 0;
			t = auxval[naux-1];
			i = Buflen;
		} else
			*t++ = c;
	}
	*t = 0;
	return f;
}

static void
dostatus(void)
{
	char *s;

	s = auxstate[St];
	if(s) {
		if(strcmp(s, "obs") == 0)
			outrune(L'†');
		else if(strcmp(s, "ali") == 0)
			outrune(L'‖');
		else if(strcmp(s, "err") == 0 || strcmp(s, "spu") == 0)
			outrune(L'¶');
		else if(strcmp(s, "xref") == 0)
			{/* nothing */}
		else if(debug)
			err("status %ld %d %s", curentry.doff, cursize, s);
	}
}