etags.c

早期freebsd实现

	  if (*dbp == 0)
	    continue;

	  /* save all values for later tagging */
	  tline.size = lb.size;
	  strcpy (tline.buffer, lb.buffer);
	  save_lineno = lineno;
	  save_lcno = linecharno;

	  /* grab block name */
	  for (cp = dbp + 1; *cp && (!endtoken (*cp)); cp++)
	    continue;
	  c = cp[0];
	  cp[0] = 0;
	  strcpy (nambuf, dbp);
	  cp[0] = c;
	  dbp = cp;		/* restore dbp to e-o-token */
	  get_tagname = FALSE;
	  found_tag = TRUE;
	  continue;

	  /* and proceed to check for "extern" */
	}
      if ((!incomm1) && (!incomm2) && (!inquote) &&
	  (!found_tag) && (!get_tagname))
	{
	  /* check for proc/fn keywords */
	  switch (c | ' ')
	    {
	    case 'p':
	      if (tail ("rocedure"))	/* c = 'p', dbp has advanced */
		get_tagname = TRUE;
	      continue;
	    case 'f':
	      if (tail ("unction"))
		get_tagname = TRUE;
	      continue;
	    }
	}
    }				/* while not e-o-f */
}

/*
 * lisp tag functions
 * just look for (def or (DEF
 */

void
L_funcs (fi)
     FILE *fi;
{
  lineno = 0;
  charno = 0;
  pfcnt = 0;

  while (!feof (fi))
    {
      lineno++;
      linecharno = charno;
      charno += readline (&lb, fi);
      dbp = lb.buffer;
      if (dbp[0] == '(')
	{
	  if (L_isdef (dbp))
	    {
	      while (!isspace (*dbp))
		dbp++;
	      while (isspace (*dbp))
		dbp++;
	      L_getit ();
	    }
	  else
	    {
	      /* Check for (foo::defmumble name-defined ... */
	      while (*dbp && *dbp != ':' && !isspace (*dbp)
		     && *dbp != '(' && *dbp != ')')
		dbp++;
	      if (*dbp == ':')
		{
		  while (*dbp == ':')
		    dbp++;

		  if (L_isdef (dbp))
		    {
		      while (!isspace (*dbp))
			dbp++;
		      while (isspace (*dbp))
			dbp++;
		      L_getit ();
		    }
		}
	    }
	}
    }
}

int
L_isdef (dbp)
     char *dbp;
{
  return ((dbp[1] == 'D' || dbp[1] == 'd') &&
	  (dbp[2] == 'E' || dbp[2] == 'e') &&
	  (dbp[3] == 'F' || dbp[3] == 'f'));
}

void
L_getit ()
{
  register char *cp;
  char c;
  char nambuf[BUFSIZ];

  if (*dbp == 0)
    return;
  for (cp = dbp + 1; *cp && *cp != '(' && *cp != ' '; cp++)
    continue;
  c = cp[0];
  cp[0] = 0;
  (void) strcpy (nambuf, dbp);
  cp[0] = c;
  pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
  pfcnt++;
}

/*
 * Scheme tag functions
 * look for (def... xyzzy
 * look for (def... (xyzzy
 * look for (def ... ((...(xyzzy ....
 * look for (set! xyzzy
 */

static void get_scheme ();

void
Scheme_funcs (fi)
     FILE *fi;
{
  lineno = 0;
  charno = 0;
  pfcnt = 0;

  while (!feof (fi))
    {
      lineno++;
      linecharno = charno;
      charno += readline (&lb, fi);
      dbp = lb.buffer;
      if (dbp[0] == '(' &&
	  (dbp[1] == 'D' || dbp[1] == 'd') &&
	  (dbp[2] == 'E' || dbp[2] == 'e') &&
	  (dbp[3] == 'F' || dbp[3] == 'f'))
	{
	  while (!isspace (*dbp))
	    dbp++;
	  /* Skip over open parens and white space */
	  while (*dbp && (isspace (*dbp) || *dbp == '('))
	    dbp++;
	  get_scheme ();
	}
      if (dbp[0] == '(' &&
	  (dbp[1] == 'S' || dbp[1] == 's') &&
	  (dbp[2] == 'E' || dbp[2] == 'e') &&
	  (dbp[3] == 'T' || dbp[3] == 't') &&
	  (dbp[4] == '!' || dbp[4] == '!') &&
	  (isspace (dbp[5])))
	{
	  while (!isspace (*dbp))
	    dbp++;
	  /* Skip over white space */
	  while (isspace (*dbp))
	    dbp++;
	  get_scheme ();
	}
    }
}

static void
get_scheme ()
{
  register char *cp;
  char c;
  char nambuf[BUFSIZ];

  if (*dbp == 0)
    return;
  /* Go till you get to white space or a syntactic break */
  for (cp = dbp + 1; *cp && *cp != '(' && *cp != ')' && !isspace (*cp); cp++)
    continue;
  /* Null terminate the string there. */
  c = cp[0];
  cp[0] = 0;
  /* Copy the string */
  strcpy (nambuf, dbp);
  /* Unterminate the string */
  cp[0] = c;
  /* Announce the change */
  pfnote (nambuf, TRUE, FALSE, lb.buffer, cp - lb.buffer + 1, lineno, linecharno);
  pfcnt++;
}

/* Find tags in TeX and LaTeX input files.  */

/* TEX_toktab is a table of TeX control sequences that define tags.
   Each TEX_tabent records one such control sequence.
   CONVERT THIS TO USE THE Stab TYPE!! */

struct TEX_tabent
{
  char *name;
  int len;
};

struct TEX_tabent *TEX_toktab = NULL;	/* Table with tag tokens */

/* Default set of control sequences to put into TEX_toktab.
   The value of environment var TEXTAGS is prepended to this.  */

static char *TEX_defenv =
":chapter:section:subsection:subsubsection:eqno:label:ref:cite:bibitem:typeout";

void TEX_mode ();
struct TEX_tabent *TEX_decode_env ();
void TEX_getit ();
int TEX_Token ();

static char TEX_esc = '\\';
static char TEX_opgrp = '{';
static char TEX_clgrp = '}';

/*
 * TeX/LaTeX scanning loop.
 */

void
TEX_funcs (fi)
     FILE *fi;
{
  char *lasthit;

  lineno = 0;
  charno = 0;
  pfcnt = 0;

  /* Select either \ or ! as escape character.  */
  TEX_mode (fi);

  /* Initialize token table once from environment. */
  if (!TEX_toktab)
    TEX_toktab = TEX_decode_env ("TEXTAGS", TEX_defenv);

  while (!feof (fi))
    {
      lineno++;
      linecharno = charno;
      charno += readline (&lb, fi);
      dbp = lb.buffer;
      lasthit = dbp;

      while (!feof (fi))
	{			/* Scan each line in file */
	  lineno++;
	  linecharno = charno;
	  charno += readline (&lb, fi);
	  dbp = lb.buffer;
	  lasthit = dbp;
	  while (dbp = index (dbp, TEX_esc))	/* Look at each escape in line */
	    {
	      register int i;

	      if (!*(++dbp))
		break;
	      linecharno += dbp - lasthit;
	      lasthit = dbp;
	      i = TEX_Token (lasthit);
	      if (0 <= i)
		{
		  TEX_getit (lasthit, TEX_toktab[i].len);
		  break;	/* We only save a line once */
		}
	    }
	}
    }
}

#define TEX_LESC '\\'
#define TEX_SESC '!'
#define TEX_cmt  '%'

/* Figure out whether TeX's escapechar is '\\' or '!' and set grouping */
/* chars accordingly. */

void
TEX_mode (f)
     FILE *f;
{
  int c;

  while ((c = getc (f)) != EOF)
    {
      /* Skip to next line if we hit the TeX comment char. */
      if (c == TEX_cmt)
	while (c != '\n')
	  c = getc (f);
      else if (c == TEX_LESC || c == TEX_SESC )
	break;
    }

  if (c == TEX_LESC)
    {
      TEX_esc = TEX_LESC;
      TEX_opgrp = '{';
      TEX_clgrp = '}';
    }
  else
    {
      TEX_esc = TEX_SESC;
      TEX_opgrp = '<';
      TEX_clgrp = '>';
    }
  rewind (f);
}

/* Read environment and prepend it to the default string. */
/* Build token table. */

struct TEX_tabent *
TEX_decode_env (evarname, defenv)
     char *evarname;
     char *defenv;
{
  register char *env, *p;
  extern char *savenstr (), *index ();

  struct TEX_tabent *tab;
  int size, i;

  /* Append default string to environment. */
  env = getenv (evarname);
  if (!env)
    env = defenv;
  else
    env = concat (env, defenv, "");

  /* Allocate a token table */
  for (size = 1, p = env; p;)
    if ((p = index (p, ':')) && *(++p))
      size++;
  tab = xnew (size, struct TEX_tabent);

  /* Unpack environment string into token table. Be careful about */
  /* zero-length strings (leading ':', "::" and trailing ':') */
  for (i = 0; *env;)
    {
      p = index (env, ':');
      if (!p)			/* End of environment string. */
	p = env + strlen (env);
      if (p - env > 0)
	{			/* Only non-zero strings. */
	  tab[i].name = savenstr (env, p - env);
	  tab[i].len = strlen (tab[i].name);
	  i++;
	}
      if (*p)
	env = p + 1;
      else
	{
	  tab[i].name = NULL;	/* Mark end of table. */
	  tab[i].len = 0;
	  break;
	}
    }
  return tab;
}

/* Record a tag defined by a TeX command of length LEN and starting at NAME.
   The name being defined actually starts at (NAME + LEN + 1).
   But we seem to include the TeX command in the tag name.  */

void
TEX_getit (name, len)
     char *name;
     int len;
{
  char *p = name + len;
  char nambuf[BUFSIZ];

  if (*name == 0)
    return;

  /* Let tag name extend to next group close (or end of line) */
  while (*p && *p != TEX_clgrp)
    p++;
  (void) strncpy (nambuf, name, p - name);
  nambuf[p - name] = 0;

  pfnote (nambuf, TRUE, FALSE, lb.buffer, strlen (lb.buffer), lineno, linecharno);
  pfcnt++;
}

/* If the text at CP matches one of the tag-defining TeX command names,
   return the index of that command in TEX_toktab.
   Otherwise return -1.  */

/* Keep the capital `T' in `Token' for dumb truncating compilers
   (this distinguishes it from `TEX_toktab' */
int
TEX_Token (cp)
     char *cp;
{
  int i;

  for (i = 0; TEX_toktab[i].len > 0; i++)
    if (strncmp (TEX_toktab[i].name, cp, TEX_toktab[i].len) == 0)
      return i;
  return -1;
}

/* Support for Prolog.  */

/* whole head (not only functor, but also arguments)
   is gotten in compound term. */

void
prolog_getit (s, lineno, linecharno)
     char *s;
     int lineno;
     long linecharno;
{
  char nambuf[BUFSIZ], *save_s, tmpc;
  int insquote, npar;

  save_s = s;
  insquote = FALSE;
  npar = 0;
  while (1)
    {
      if (*s == '\0')		/* syntax error. */
	return;
      else if (insquote && *s == '\'' && *(s + 1) == '\'')
	s += 2;
      else if (*s == '\'')
	{
	  insquote = !insquote;
	  s++;
	}
      else if (!insquote && *s == '(')
	{
	  npar++;
	  s++;
	}
      else if (!insquote && *s == ')')
	{
	  npar--;
	  s++;
	  if (npar == 0)
	    break;
	  else if (npar < 0)	/* syntax error. */
	    return;
	}
      else if (!insquote && *s == '.' && (isspace (*(s + 1)) || *(s + 1) == '\0'))
	{			/* fullstop. */
	  if (npar != 0)	/* syntax error. */
	    return;
	  s++;
	  break;
	}
      else
	s++;
    }
  tmpc = *s;
  *s = '\0';
  strcpy (nambuf, save_s);
  *s = tmpc;
  pfnote (nambuf, TRUE, save_s, strlen (nambuf), lineno, linecharno);
}

/* It is assumed that prolog predicate starts from column 0. */

void
prolog_funcs (fi)
     FILE *fi;
{
  void skip_comment (), prolog_getit ();

  lineno = linecharno = charno = 0;
  while (!feof (fi))
    {
      lineno++;
      linecharno += charno;
      charno = readline (&lb, fi) + 1;	/* 1 for newline. */
      dbp = lb.buffer;
      if (isspace (dbp[0]))	/* not predicate header. */
	continue;
      else if (dbp[0] == '%')	/* comment. */
	continue;
      else if (dbp[0] == '/' && dbp[1] == '*')	/* comment. */
	skip_comment (&lb, fi, &lineno, &linecharno);
      else			/* found. */
	prolog_getit (dbp, lineno, linecharno);
    }
}

void
skip_comment (plb, fi, plineno, plinecharno)
     struct linebuffer *plb;
     FILE *fi;
     int *plineno;		/* result */
     long *plinecharno;		/* result */
{
  while (!substr ("*/", plb->buffer))
    {
      (*plineno)++;
      *plinecharno += readline (plb, fi) + 1;
    }				/* 1 for newline. */
}

/* Return TRUE if 'sub' exists somewhere in 's'. */

int
substr (sub, s)
     char *sub;
     char *s;
{
  while (*s && (s = index (s, *sub)))
    if (prestr (sub, s))
      return (TRUE);
    else
      s++;
  return (FALSE);
}

/* Return TRUE if 'pre' is prefix of string 's'. */

int
prestr (pre, s)
     char *pre;
     char *s;
{
  if (*pre == '\0')
    return (TRUE);
  else if (*pre == *s)
    return (prestr (pre + 1, s + 1));
  else
    return (FALSE);
}

/* Initialize a linebuffer for use */

void
initbuffer (linebuffer)
     struct linebuffer *linebuffer;
{
  linebuffer->size = 200;
  linebuffer->buffer = xnew (200, char);
}

/*
 * Read a line of text from `stream' into `linebuffer'.
 * Return the number of characters read from `stream',
 * which is the length of the line including the newline, if any.
 */
long
readline (linebuffer, stream)
     struct linebuffer *linebuffer;
     register FILE *stream;
{
  char *buffer = linebuffer->buffer;
  register char *p = linebuffer->buffer;
  register char *pend;
  int newline;			/* 1 if ended with newline, 0 if ended with EOF */