lexi.c File Reference

#include "c.h"
#include <err.h>
#include <stdio.h>
#include <ctype.h>
#include <stdlib.h>
#include <string.h>
#include "indent_globs.h"
#include "indent_codes.h"
#include "indent.h"

Include dependency graph for lexi.c:

Go to the source code of this file.

Data Structures
struct	templ

Macros
#define	alphanum   1

Functions
static int	strcmp_type (const void *e1, const void *e2)
static int	is_func_definition (char *tp)
int	lexi (struct parser_state *state)
void	alloc_typenames (void)
void	add_typename (const char *key)

Variables
struct templ	specials []
const char **	typenames
int	typename_count
int	typename_top = -1
char	chartype [128]

Macro Definition Documentation

alphanum

#define alphanum   1

Definition at line 55 of file lexi.c.

Function Documentation

add_typename()

void add_typename

(

const char *

key

)

Definition at line 687 of file lexi.c.

{
    int comparison;
    const char *copy;
 
    if (typename_top + 1 >= typename_count) {
    typenames = realloc((void *)typenames,
        sizeof(typenames[0]) * (typename_count *= 2));
    if (typenames == NULL)
        err(1, NULL);
    }
    if (typename_top == -1)
    typenames[++typename_top] = copy = strdup(key);
    else if ((comparison = strcmp(key, typenames[typename_top])) >= 0) {
    /* take advantage of sorted input */
    if (comparison == 0)    /* remove duplicates */
        return;
    typenames[++typename_top] = copy = strdup(key);
    }
    else {
    int p;
 
    for (p = 0; (comparison = strcmp(key, typenames[p])) > 0; p++)
        /* find place for the new key */;
    if (comparison == 0)    /* remove duplicates */
        return;
    memmove(&typenames[p + 1], &typenames[p],
        sizeof(typenames[0]) * (++typename_top - p));
    typenames[p] = copy = strdup(key);
    }
 
    if (copy == NULL)
    err(1, NULL);
}

References err(), sort-test::key, realloc, typename_count, typename_top, and typenames.

Referenced by add_typedefs_from_file(), and set_option().

alloc_typenames()

void alloc_typenames

(

void

)

Definition at line 677 of file lexi.c.

{
 
    typenames = (const char **)malloc(sizeof(typenames[0]) *
        (typename_count = 16));
    if (typenames == NULL)
    err(1, NULL);
}

References err(), malloc, typename_count, and typenames.

Referenced by main().

is_func_definition()

static int is_func_definition ( char *  tp )

static

Definition at line 160 of file lexi.c.

{
    int     paren_depth = 0;
    int     in_comment = false;
    int     in_slash_comment = false;
    int     lastc = 0;
 
    /* We may need to look past the end of the current buffer. */
    lookahead_reset();
    for (;;) {
    int     c;
 
    /* Fetch next character. */
    if (tp < buf_end)
        c = *tp++;
    else {
        c = lookahead();
        if (c == EOF)
        break;
    }
    /* Handle comments. */
    if (in_comment) {
        if (lastc == '*' && c == '/')
        in_comment = false;
    } else if (lastc == '/' && c == '*' && !in_slash_comment)
        in_comment = true;
    else if (in_slash_comment) {
        if (c == '\n')
        in_slash_comment = false;
    } else if (lastc == '/' && c == '/')
        in_slash_comment = true;
    /* Count nested parens properly. */
    else if (c == '(')
        paren_depth++;
    else if (c == ')') {
        paren_depth--;
        /*
         * If we find unbalanced parens, we must have started inside a
         * declaration.
         */
        if (paren_depth < 0)
        return false;
    } else if (paren_depth == 0) {
        /* We are outside any parentheses or comments. */
        if (c == '{')
        return true;
        else if (c == ';' || c == ',')
        return false;
    }
    lastc = c;
    }
    /* Hit EOF --- for lack of anything better, assume "not a definition". */
    return false;
}

References buf_end, lookahead(), and lookahead_reset().

Referenced by lexi().

lexi()

int lexi

(

struct parser_state *

state

)

Definition at line 216 of file lexi.c.

{
    int         unary_delim;    /* this is set to 1 if the current token
                 * forces a following operator to be unary */
    int         code;       /* internal code to be returned */
    char        qchar;      /* the delimiter character for a string */
 
    e_token = s_token;      /* point to start of place to save token */
    unary_delim = false;
    state->col_1 = state->last_nl;  /* tell world that this token started
                     * in column 1 iff the last thing
                     * scanned was a newline */
    state->last_nl = false;
 
    while (*buf_ptr == ' ' || *buf_ptr == '\t') {   /* get rid of blanks */
    state->col_1 = false;   /* leading blanks imply token is not in column
                 * 1 */
    if (++buf_ptr >= buf_end)
        fill_buffer();
    }
 
    /* Scan an alphanumeric token */
    if (chartype[*buf_ptr & 127] == alphanum ||
    (buf_ptr[0] == '.' && isdigit((unsigned char)buf_ptr[1]))) {
    /*
     * we have a character or number
     */
    struct templ *p;
 
    if (isdigit((unsigned char)*buf_ptr) ||
        (buf_ptr[0] == '.' && isdigit((unsigned char)buf_ptr[1]))) {
        int         seendot = 0,
                    seenexp = 0,
            seensfx = 0;
 
        /*
         * base 2, base 8, base 16:
         */
        if (buf_ptr[0] == '0' && buf_ptr[1] != '.') {
        int len;
 
        if (buf_ptr[1] == 'b' || buf_ptr[1] == 'B')
            len = strspn(buf_ptr + 2, "01") + 2;
        else if (buf_ptr[1] == 'x' || buf_ptr[1] == 'X')
            len = strspn(buf_ptr + 2, "0123456789ABCDEFabcdef") + 2;
        else
            len = strspn(buf_ptr + 1, "012345678") + 1;
        if (len > 0) {
            CHECK_SIZE_TOKEN(len);
            memcpy(e_token, buf_ptr, len);
            e_token += len;
            buf_ptr += len;
        }
        else
            diag2(1, "Unterminated literal");
        }
        else        /* base 10: */
        while (1) {
            if (*buf_ptr == '.') {
            if (seendot)
                break;
            else
                seendot++;
            }
            CHECK_SIZE_TOKEN(3);
            *e_token++ = *buf_ptr++;
            if (!isdigit((unsigned char)*buf_ptr) && *buf_ptr != '.') {
            if ((*buf_ptr != 'E' && *buf_ptr != 'e') || seenexp)
                break;
            else {
                seenexp++;
                seendot++;
                *e_token++ = *buf_ptr++;
                if (*buf_ptr == '+' || *buf_ptr == '-')
                *e_token++ = *buf_ptr++;
            }
            }
        }
 
        while (1) {
        CHECK_SIZE_TOKEN(2);
        if (!(seensfx & 1) && (*buf_ptr == 'U' || *buf_ptr == 'u')) {
            *e_token++ = *buf_ptr++;
            seensfx |= 1;
            continue;
        }
        if (!(seensfx & 2) && (strchr("fFlL", *buf_ptr) != NULL)) {
            if (buf_ptr[1] == buf_ptr[0])
                *e_token++ = *buf_ptr++;
            *e_token++ = *buf_ptr++;
            seensfx |= 2;
            continue;
        }
        break;
        }
    }
    else
        while (chartype[*buf_ptr & 127] == alphanum || *buf_ptr == BACKSLASH) {
        /* fill_buffer() terminates buffer with newline */
        if (*buf_ptr == BACKSLASH) {
            if (*(buf_ptr + 1) == '\n') {
            buf_ptr += 2;
            if (buf_ptr >= buf_end)
                fill_buffer();
            } else
                break;
        }
        CHECK_SIZE_TOKEN(1);
        /* copy it over */
        *e_token++ = *buf_ptr++;
        if (buf_ptr >= buf_end)
            fill_buffer();
        }
    *e_token = '\0';
 
    if (s_token[0] == 'L' && s_token[1] == '\0' &&
          (*buf_ptr == '"' || *buf_ptr == '\''))
        return (strpfx);
 
    while (*buf_ptr == ' ' || *buf_ptr == '\t') {   /* get rid of blanks */
        if (++buf_ptr >= buf_end)
        fill_buffer();
    }
    state->keyword = 0;
    if (state->last_token == structure && !state->p_l_follow) {
                /* if last token was 'struct' and we're not
                 * in parentheses, then this token
                 * should be treated as a declaration */
        state->last_u_d = true;
        return (decl);
    }
    /*
     * Operator after identifier is binary unless last token was 'struct'
     */
    state->last_u_d = (state->last_token == structure);
 
    p = bsearch(s_token,
        specials,
        sizeof(specials) / sizeof(specials[0]),
        sizeof(specials[0]),
        strcmp_type);
    if (p == NULL) {    /* not a special keyword... */
        char *u;
 
        /* ... so maybe a type_t or a typedef */
        if ((auto_typedefs && ((u = strrchr(s_token, '_')) != NULL) &&
            strcmp(u, "_t") == 0) || (typename_top >= 0 &&
          bsearch(s_token, typenames, typename_top + 1,
            sizeof(typenames[0]), strcmp_type))) {
        state->keyword = 4; /* a type name */
        state->last_u_d = true;
            goto found_typename;
        }
    } else {            /* we have a keyword */
        state->keyword = p->rwcode;
        state->last_u_d = true;
        switch (p->rwcode) {
        case 7:     /* it is a switch */
        return (swstmt);
        case 8:     /* a case or default */
        return (casestmt);
 
        case 3:     /* a "struct" */
        /* FALLTHROUGH */
        case 4:     /* one of the declaration keywords */
        found_typename:
        if (state->p_l_follow) {
            /* inside parens: cast, param list, offsetof or sizeof */
            state->cast_mask |= (1 << state->p_l_follow) & ~state->not_cast_mask;
        }
        if (state->last_token == period || state->last_token == unary_op) {
            state->keyword = 0;
            break;
        }
        if (p != NULL && p->rwcode == 3)
            return (structure);
        if (state->p_l_follow)
            break;
        return (decl);
 
        case 5:     /* if, while, for */
        return (sp_paren);
 
        case 6:     /* do, else */
        return (sp_nparen);
 
        case 10:        /* storage class specifier */
        return (storage);
 
        case 11:        /* typedef */
        return (type_def);
 
        default:        /* all others are treated like any other
                 * identifier */
        return (ident);
        }           /* end of switch */
    }           /* end of if (found_it) */
    if (*buf_ptr == '(' && state->tos <= 1 && state->ind_level == 0 &&
        state->in_parameter_declaration == 0 && state->block_init == 0) {
        if (is_func_definition(buf_ptr)) {
        strncpy(state->procname, token, sizeof state->procname - 1);
        if (state->in_decl)
            state->in_parameter_declaration = 1;
        return (funcname);
        }
    }
    /*
     * The following hack attempts to guess whether or not the current
     * token is in fact a declaration keyword -- one that has been
     * typedefd
     */
    else if (!state->p_l_follow && !state->block_init &&
        !state->in_stmt &&
        ((*buf_ptr == '*' && buf_ptr[1] != '=') ||
        isalpha((unsigned char)*buf_ptr)) &&
        (state->last_token == semicolon || state->last_token == lbrace ||
        state->last_token == rbrace)) {
        state->keyword = 4; /* a type name */
        state->last_u_d = true;
        return decl;
    }
    if (state->last_token == decl)  /* if this is a declared variable,
                     * then following sign is unary */
        state->last_u_d = true; /* will make "int a -1" work */
    return (ident);     /* the ident is not in the list */
    }               /* end of processing for alphanum character */
 
    /* Scan a non-alphanumeric token */
 
    CHECK_SIZE_TOKEN(3);        /* things like "<<=" */
    *e_token++ = *buf_ptr;      /* if it is only a one-character token, it is
                 * moved here */
    *e_token = '\0';
    if (++buf_ptr >= buf_end)
    fill_buffer();
 
    switch (*token) {
    case '\n':
    unary_delim = state->last_u_d;
    state->last_nl = true;  /* remember that we just had a newline */
    code = (had_eof ? 0 : newline);
 
    /*
     * if data has been exhausted, the newline is a dummy, and we should
     * return code to stop
     */
    break;
 
    case '\'':          /* start of quoted character */
    case '"':           /* start of string */
    qchar = *token;
    do {            /* copy the string */
        while (1) {     /* move one character or [/<char>]<char> */
        if (*buf_ptr == '\n') {
            diag2(1, "Unterminated literal");
            goto stop_lit;
        }
        CHECK_SIZE_TOKEN(2);
        *e_token = *buf_ptr++;
        if (buf_ptr >= buf_end)
            fill_buffer();
        if (*e_token == BACKSLASH) {    /* if escape, copy extra char */
            if (*buf_ptr == '\n')   /* check for escaped newline */
            ++line_no;
            *++e_token = *buf_ptr++;
            ++e_token;  /* we must increment this again because we
                 * copied two chars */
            if (buf_ptr >= buf_end)
            fill_buffer();
        }
        else
            break;  /* we copied one character */
        }           /* end of while (1) */
    } while (*e_token++ != qchar);
stop_lit:
    code = ident;
    break;
 
    case ('('):
    case ('['):
    unary_delim = true;
    code = lparen;
    break;
 
    case (')'):
    case (']'):
    code = rparen;
    break;
 
    case '#':
    unary_delim = state->last_u_d;
    code = preesc;
    break;
 
    case '?':
    unary_delim = true;
    code = question;
    break;
 
    case (':'):
    code = colon;
    unary_delim = true;
    break;
 
    case (';'):
    unary_delim = true;
    code = semicolon;
    break;
 
    case ('{'):
    unary_delim = true;
 
    /*
     * if (state->in_or_st) state->block_init = 1;
     */
    /* ?    code = state->block_init ? lparen : lbrace; */
    code = lbrace;
    break;
 
    case ('}'):
    unary_delim = true;
    /* ?    code = state->block_init ? rparen : rbrace; */
    code = rbrace;
    break;
 
    case 014:           /* a form feed */
    unary_delim = state->last_u_d;
    state->last_nl = true;  /* remember this so we can set 'state->col_1'
                 * right */
    code = form_feed;
    break;
 
    case (','):
    unary_delim = true;
    code = comma;
    break;
 
    case '.':
    unary_delim = false;
    code = period;
    break;
 
    case '-':
    case '+':           /* check for -, +, --, ++ */
    code = (state->last_u_d ? unary_op : binary_op);
    unary_delim = true;
 
    if (*buf_ptr == token[0]) {
        /* check for doubled character */
        *e_token++ = *buf_ptr++;
        /* buffer overflow will be checked at end of loop */
        if (state->last_token == ident || state->last_token == rparen) {
        code = (state->last_u_d ? unary_op : postop);
        /* check for following ++ or -- */
        unary_delim = false;
        }
    }
    else if (*buf_ptr == '=')
        /* check for operator += */
        *e_token++ = *buf_ptr++;
    else if (*buf_ptr == '>') {
        /* check for operator -> */
        *e_token++ = *buf_ptr++;
        unary_delim = false;
        code = unary_op;
        state->want_blank = false;
    }
    break;          /* buffer overflow will be checked at end of
                 * switch */
 
    case '=':
    if (state->in_or_st)
        state->block_init = 1;
#ifdef undef
    if (chartype[*buf_ptr & 127] == opchar) {   /* we have two char assignment */
        e_token[-1] = *buf_ptr++;
        if ((e_token[-1] == '<' || e_token[-1] == '>') && e_token[-1] == *buf_ptr)
        *e_token++ = *buf_ptr++;
        *e_token++ = '=';   /* Flip =+ to += */
        *e_token = 0;
    }
#else
    if (*buf_ptr == '=') {/* == */
        *e_token++ = '=';   /* Flip =+ to += */
        buf_ptr++;
        *e_token = 0;
    }
#endif
    code = binary_op;
    unary_delim = true;
    break;
    /* can drop thru!!! */
 
    case '>':
    case '<':
    case '!':           /* ops like <, <<, <=, !=, etc */
    if (*buf_ptr == '>' || *buf_ptr == '<' || *buf_ptr == '=') {
        *e_token++ = *buf_ptr;
        if (++buf_ptr >= buf_end)
        fill_buffer();
    }
    if (*buf_ptr == '=')
        *e_token++ = *buf_ptr++;
    code = (state->last_u_d ? unary_op : binary_op);
    unary_delim = true;
    break;
 
    case '*':
    unary_delim = true;
    if (!state->last_u_d) {
        if (*buf_ptr == '=')
        *e_token++ = *buf_ptr++;
        code = binary_op;
        break;
    }
    while (*buf_ptr == '*' || isspace((unsigned char)*buf_ptr)) {
        if (*buf_ptr == '*') {
        CHECK_SIZE_TOKEN(1);
        *e_token++ = *buf_ptr;
        }
        if (++buf_ptr >= buf_end)
        fill_buffer();
    }
    code = unary_op;
    break;
 
    default:
    if (token[0] == '/' && *buf_ptr == '*') {
        /* it is start of comment */
        *e_token++ = '*';
 
        if (++buf_ptr >= buf_end)
        fill_buffer();
 
        code = comment;
        unary_delim = state->last_u_d;
        break;
    }
    while (*(e_token - 1) == *buf_ptr || *buf_ptr == '=') {
        /*
         * handle ||, &&, etc, and also things as in int *****i
         */
        CHECK_SIZE_TOKEN(1);
        *e_token++ = *buf_ptr;
        if (++buf_ptr >= buf_end)
        fill_buffer();
    }
    code = (state->last_u_d ? unary_op : binary_op);
    unary_delim = true;
 
 
    }               /* end of switch */
    if (buf_ptr >= buf_end) /* check for input buffer empty */
    fill_buffer();
    state->last_u_d = unary_delim;
    CHECK_SIZE_TOKEN(1);
    *e_token = '\0';        /* null terminate the token */
    return (code);
}

References alphanum, auto_typedefs, BACKSLASH, binary_op, buf_end, buf_ptr, casestmt, chartype, CHECK_SIZE_TOKEN, colon, comma, comment, decl, diag2(), e_token, fill_buffer(), form_feed, funcname, had_eof, ident, is_func_definition(), lbrace, len, line_no, lparen, newline, period, postop, preesc, question, rbrace, rparen, templ::rwcode, s_token, semicolon, sp_nparen, sp_paren, specials, storage, strcmp_type(), strpfx, structure, swstmt, token, type_def, typename_top, typenames, and unary_op.

Referenced by main().

strcmp_type()

static int strcmp_type ( const void *  e1, const void *  e2  )

static

Definition at line 142 of file lexi.c.

{
    return (strcmp(e1, *(const char * const *)e2));
}

Referenced by lexi().

Variable Documentation

chartype

char chartype[128]

Initial value:

=
{               
    0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0,
    0, 3, 0, 0, 1, 3, 3, 0,
    0, 0, 3, 3, 0, 3, 0, 3,
    1, 1, 1, 1, 1, 1, 1, 1,
    1, 1, 0, 0, 3, 3, 3, 3,
    0, 1, 1, 1, 1, 1, 1, 1,
    1, 1, 1, 1, 1, 1, 1, 1,
    1, 1, 1, 1, 1, 1, 1, 1,
    1, 1, 1, 0, 0, 0, 3, 1,
    0, 1, 1, 1, 1, 1, 1, 1,
    1, 1, 1, 1, 1, 1, 1, 1,
    1, 1, 1, 1, 1, 1, 1, 1,
    1, 1, 1, 0, 3, 0, 3, 0
}

Definition at line 119 of file lexi.c.

Referenced by lexi().

specials

struct templ specials[]

Definition at line 1 of file lexi.c.

Referenced by lexi().

typename_count

int typename_count

Definition at line 116 of file lexi.c.

Referenced by add_typename(), and alloc_typenames().

typename_top

int typename_top = -1

Definition at line 117 of file lexi.c.

Referenced by add_typename(), and lexi().

typenames

const char** typenames

Definition at line 115 of file lexi.c.

Referenced by add_typename(), alloc_typenames(), get_object_address_opf_member(), lexi(), type_in_list_does_not_exist_skipping(), and TypeNameListToString().