Sweden-Number/tools/wrc/parser.l

672 lines
20 KiB
C

/* -*-C-*-
*
* Copyright 1998-2000 Bertho A. Stultiens (BS)
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
*
* History:
* 21-May-2000 BS - Fixed the ident requirement of resource names
* which can be keywords.
* 30-Apr-2000 BS - Reintegration into the wine-tree
* 11-Jan-2000 BS - Very drastic cleanup because we don't have a
* preprocessor in here anymore.
* 02-Jan-2000 BS - Removed the preprocessor code
* 23-Dec-1999 BS - Removed the copyright for Martin von Loewis.
* There is really nothing left of his code in
* this parser.
* 20-Jun-1998 BS - Changed the filename conversion. Filenames are
* case-sensitive inder *nix, but not under dos.
* default behaviour is to convert to lower case.
* - All backslashes are converted to forward and
* both single and double slash is recognized as
* MS/Borland does.
* - Fixed a bug in 'yywf' case that prevented
* double quoted names to be scanned properly.
*
* 19-May-1998 BS - Started to build a preprocessor.
* - Changed keyword processing completely to
* table-lookups.
*
* 20-Apr-1998 BS - Added ';' comment stripping
*
* 17-Apr-1998 BS - Made the win32 keywords optional when compiling in
* 16bit mode
*
* 15-Apr-1998 BS - Changed string handling to include escapes
* - Added unicode string handling (no codepage
* translation though).
* - 'Borrowed' the main idea of string scanning from
* the flex manual pages.
* - Added conditional handling of scanning depending
* on the state of the parser. This was mainly required
* to distinguish a file to load or raw data that
* follows. MS's definition of filenames is rather
* complex... It can be unquoted or double quoted. If
* double quoted, then the '\\' char is not automatically
* escaped according to Borland's rc compiler, but it
* accepts both "\\path\\file.rc" and "\path\file.rc".
* This makes life very hard! I go for the escaped
* version, as this seems to be the documented way...
* - Single quoted strings are now parsed and converted
* here.
* - Added comment stripping. The implementation is
* 'borrowed' from the flex manpages.
* - Rebuild string processing so that it may contain
* escaped '\0'.
*/
/* Exclusive string handling */
%x tkstr
/* Exclusive unicode string handling */
%x tklstr
/* Exclusive rcdata single quoted data handling */
%x tkrcd
/* Exclusive comment eating... */
%x comment
/* Set when stripping c-junk */
%x pp_cstrip
/* Set when scanning #line style directives */
%x pp_line
/* Set when scanning #pragma */
%x pp_pragma
%x pp_code_page
%option stack
%option noinput nounput noyy_top_state noyywrap
%option 8bit never-interactive
%option prefix="parser_"
/* Some shortcut definitions */
ws [ \f\t\r]
%{
/*#define LEX_DEBUG*/
#include "config.h"
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <ctype.h>
#include <assert.h>
#include <errno.h>
#include <limits.h>
#ifdef HAVE_UNISTD_H
#include <unistd.h>
#else
#define YY_NO_UNISTD_H
#endif
#include "../tools.h"
#include "wrc.h"
#include "utils.h"
#include "parser.h"
#include "newstruc.h"
#include "wpp_private.h"
#include "parser.tab.h"
/* Always update the current character position within a line */
#define YY_USER_ACTION char_number+=yyleng; wanted_id = want_id; want_id = 0;
#define YY_USER_INIT current_codepage = utf8_input ? CP_UTF8 : -1;
static void addcchar(char c);
static void addwchar(WCHAR s);
static string_t *get_buffered_cstring(void);
static string_t *get_buffered_wstring(void);
static string_t *make_string(char *s);
static char *cbuffer; /* Buffers for string collection */
static int cbufidx;
static int cbufalloc = 0;
static WCHAR *wbuffer;
static int wbufidx;
static int wbufalloc = 0;
static int current_codepage = -1; /* use language default */
/*
* This one is a bit tricky.
* We set 'want_id' in the parser to get the first
* identifier we get across in the scanner, but we
* also want it to be reset at nearly any token we
* see. Exceptions are:
* - newlines
* - comments
* - whitespace
*
* The scanner will automatically reset 'want_id'
* after *each* scanner reduction and puts is value
* into the var below. In this way we can see the
* state after the YY_RULE_SETUP (i.e. the user action;
* see above) and don't have to worry too much when
* it needs to be reset.
*/
static int wanted_id = 0;
static int save_wanted_id; /* To save across comment reductions */
struct keyword {
const char *keyword;
int token;
int isextension;
int needcase;
int alwayskw;
};
static struct keyword keywords[] = {
{ "ACCELERATORS", tACCELERATORS, 0, 0, 0},
{ "ALT", tALT, 0, 0, 0},
{ "ASCII", tASCII, 0, 0, 0},
{ "AUTO3STATE", tAUTO3STATE, 1, 0, 0},
{ "AUTOCHECKBOX", tAUTOCHECKBOX, 1, 0, 0},
{ "AUTORADIOBUTTON", tAUTORADIOBUTTON, 1, 0, 0},
{ "BEGIN", tBEGIN, 0, 0, 0},
{ "BITMAP", tBITMAP, 0, 0, 0},
{ "BLOCK", tBLOCK, 0, 0, 0},
{ "BUTTON", tBUTTON, 1, 0, 0},
{ "CAPTION", tCAPTION, 0, 0, 0},
{ "CHARACTERISTICS", tCHARACTERISTICS, 1, 0, 0},
{ "CHECKBOX", tCHECKBOX, 0, 0, 0},
{ "CHECKED", tCHECKED, 0, 0, 0},
{ "CLASS", tCLASS, 0, 0, 0},
{ "COMBOBOX", tCOMBOBOX, 0, 0, 0},
{ "CONTROL", tCONTROL, 0, 0, 0},
{ "CTEXT", tCTEXT, 0, 0, 0},
{ "CURSOR", tCURSOR, 0, 0, 0},
{ "DEFPUSHBUTTON", tDEFPUSHBUTTON, 0, 0, 0},
{ "DIALOG", tDIALOG, 0, 0, 0},
{ "DIALOGEX", tDIALOGEX, 1, 0, 0},
{ "DISCARDABLE", tDISCARDABLE, 0, 0, 0},
{ "DLGINIT", tDLGINIT, 0, 0, 0},
{ "EDITTEXT", tEDITTEXT, 0, 0, 0},
{ "END", tEND, 0, 0, 0},
{ "EXSTYLE", tEXSTYLE, 0, 0, 0},
{ "FILEFLAGS", tFILEFLAGS, 0, 0, 0},
{ "FILEFLAGSMASK", tFILEFLAGSMASK, 0, 0, 0},
{ "FILEOS", tFILEOS, 0, 0, 0},
{ "FILESUBTYPE", tFILESUBTYPE, 0, 0, 0},
{ "FILETYPE", tFILETYPE, 0, 0, 0},
{ "FILEVERSION", tFILEVERSION, 0, 0, 0},
{ "FIXED", tFIXED, 0, 0, 0},
{ "FONT", tFONT, 0, 0, 0},
{ "FONTDIR", tFONTDIR, 0, 0, 0}, /* This is a Borland BRC extension */
{ "GRAYED", tGRAYED, 0, 0, 0},
{ "GROUPBOX", tGROUPBOX, 0, 0, 0},
{ "HELP", tHELP, 0, 0, 0},
{ "HTML", tHTML, 0, 0, 0},
{ "ICON", tICON, 0, 0, 0},
{ "IMPURE", tIMPURE, 0, 0, 0},
{ "INACTIVE", tINACTIVE, 0, 0, 0},
{ "LANGUAGE", tLANGUAGE, 1, 0, 1},
{ "LISTBOX", tLISTBOX, 0, 0, 0},
{ "LOADONCALL", tLOADONCALL, 0, 0, 0},
{ "LTEXT", tLTEXT, 0, 0, 0},
{ "MENU", tMENU, 0, 0, 0},
{ "MENUBARBREAK", tMENUBARBREAK, 0, 0, 0},
{ "MENUBREAK", tMENUBREAK, 0, 0, 0},
{ "MENUEX", tMENUEX, 1, 0, 0},
{ "MENUITEM", tMENUITEM, 0, 0, 0},
{ "MESSAGETABLE", tMESSAGETABLE, 1, 0, 0},
{ "MOVEABLE", tMOVEABLE, 0, 0, 0},
{ "NOINVERT", tNOINVERT, 0, 0, 0},
{ "NOT", tNOT, 0, 0, 0},
{ "POPUP", tPOPUP, 0, 0, 0},
{ "PRELOAD", tPRELOAD, 0, 0, 0},
{ "PRODUCTVERSION", tPRODUCTVERSION, 0, 0, 0},
{ "PURE", tPURE, 0, 0, 0},
{ "PUSHBUTTON", tPUSHBUTTON, 0, 0, 0},
{ "RADIOBUTTON", tRADIOBUTTON, 0, 0, 0},
{ "RCDATA", tRCDATA, 0, 0, 0},
{ "RTEXT", tRTEXT, 0, 0, 0},
{ "SCROLLBAR", tSCROLLBAR, 0, 0, 0},
{ "SEPARATOR", tSEPARATOR, 0, 0, 0},
{ "SHIFT", tSHIFT, 0, 0, 0},
{ "STATE3", tSTATE3, 1, 0, 0},
{ "STRING", tSTRING, 0, 0, 0},
{ "STRINGTABLE", tSTRINGTABLE, 0, 0, 1},
{ "STYLE", tSTYLE, 0, 0, 0},
{ "TOOLBAR", tTOOLBAR, 1, 0, 0},
{ "VALUE", tVALUE, 0, 0, 0},
{ "VERSION", tVERSION, 1, 0, 0},
{ "VERSIONINFO", tVERSIONINFO, 0, 0, 0},
{ "VIRTKEY", tVIRTKEY, 0, 0, 0}
};
#define NKEYWORDS (sizeof(keywords)/sizeof(keywords[0]))
#define KWP(p) ((const struct keyword *)(p))
static int kw_cmp_func(const void *s1, const void *s2)
{
int ret;
ret = compare_striA(KWP(s1)->keyword, KWP(s2)->keyword);
if(!ret && (KWP(s1)->needcase || KWP(s2)->needcase))
return strcmp(KWP(s1)->keyword, KWP(s2)->keyword);
else
return ret;
}
#define KW_BSEARCH
#define DO_SORT
static struct keyword *iskeyword(char *kw)
{
struct keyword *kwp;
struct keyword key;
key.keyword = kw;
key.needcase = 0;
#ifdef DO_SORT
{
/* Make sure that it is sorted for bsearsh */
static int sorted = 0;
if(!sorted)
{
qsort(keywords, NKEYWORDS, sizeof(keywords[0]), kw_cmp_func);
sorted = 1;
}
}
#endif
#ifdef KW_BSEARCH
kwp = bsearch(&key, keywords, NKEYWORDS, sizeof(keywords[0]), kw_cmp_func);
#else
{
int i;
for(i = 0; i < NKEYWORDS; i++)
{
if(!kw_cmp_func(&key, &keywords[i]))
break;
}
if(i < NKEYWORDS)
kwp = &keywords[i];
else
kwp = NULL;
}
#endif
if(kwp == NULL || (kwp->isextension && !extensions))
return NULL;
else
return kwp;
}
/* converts an integer in string form to an unsigned long and prints an error
* on overflow */
static unsigned long xstrtoul(const char *nptr, char **endptr, int base)
{
unsigned long l;
errno = 0;
l = strtoul(nptr, endptr, base);
if (l == ULONG_MAX && errno == ERANGE)
parser_error("integer constant %s is too large", nptr);
return l;
}
%}
/*
**************************************************************************
* The flexer starts here
**************************************************************************
*/
%%
/*
* Catch the GCC-style line statements here and parse them.
* This has the advantage that you can #include at any
* stage in the resource file.
* The preprocessor generates line directives in the format:
* # <linenum> "filename" <codes>
*
* Codes can be a sequence of:
* - 1 start of new file
* - 2 returning to previous
* - 3 system header
* - 4 interpret as C-code
*
* 4 is not used and 1 mutually excludes 2
* Anyhow, we are not really interested in these at all
* because we only want to know the linenumber and
* filename.
*/
<INITIAL,pp_cstrip>^{ws}*\#{ws}*pragma{ws}+ yy_push_state(pp_pragma);
<INITIAL,pp_cstrip>^{ws}*\#{ws}* yy_push_state(pp_line);
<pp_line>[^\n]* {
int lineno, len;
char *cptr;
char *fname;
yy_pop_state();
lineno = (int)strtol(yytext, &cptr, 10);
if(!lineno)
parser_error("Malformed '#...' line-directive; invalid linenumber");
fname = strchr(cptr, '"');
if(!fname)
parser_error("Malformed '#...' line-directive; missing filename");
fname++;
cptr = strchr(fname, '"');
if(!cptr)
parser_error("Malformed '#...' line-directive; missing terminating \"");
*cptr = '\0';
line_number = lineno - 1; /* We didn't read the newline */
input_name = xstrdup(fname);
/* ignore contents of C include files */
len = strlen(input_name);
if (len > 1 && !strcasecmp( input_name + len - 2, ".h" ))
BEGIN(pp_cstrip);
else
BEGIN(INITIAL);
}
<pp_pragma>code_page[^\n]* yyless(9); yy_pop_state(); yy_push_state(pp_code_page);
<pp_pragma>[^\n]* yy_pop_state(); if (pedantic) parser_warning("Unrecognized #pragma directive '%s'\n",yytext);
<pp_code_page>\({ws}*default{ws}*\)[^\n]* current_codepage = -1; yy_pop_state();
<pp_code_page>\({ws}*utf8{ws}*\)[^\n]* current_codepage = CP_UTF8; yy_pop_state();
<pp_code_page>\({ws}*[0-9]+{ws}*\)[^\n]* {
char *p = yytext;
yy_pop_state();
while (*p < '0' || *p > '9') p++;
current_codepage = strtol( p, NULL, 10 );
if (!is_valid_codepage( current_codepage ))
{
parser_error("Codepage %d not supported", current_codepage);
current_codepage = 0;
}
}
<pp_code_page>[^\n]* yy_pop_state(); parser_error("Malformed #pragma code_page directive");
/*
* Strip everything until a ';' taking
* into account braces {} for structures,
* classes and enums.
*/
<pp_cstrip>\n line_number++; char_number = 1;
<pp_cstrip>. ; /* ignore */
\{ return tBEGIN;
\} return tEND;
[0-9]+[lL]? { parser_lval.num = xstrtoul(yytext, 0, 10);
return (yytext[yyleng-1] == 'L' || yytext[yyleng-1] == 'l') ? tLNUMBER : tNUMBER; }
0[xX][0-9A-Fa-f]+[lL]? { parser_lval.num = xstrtoul(yytext, 0, 16);
return (yytext[yyleng-1] == 'L' || yytext[yyleng-1] == 'l') ? tLNUMBER : tNUMBER; }
0[oO][0-7]+[lL]? { parser_lval.num = xstrtoul(yytext+2, 0, 8);
return (yytext[yyleng-1] == 'L' || yytext[yyleng-1] == 'l') ? tLNUMBER : tNUMBER; }
[A-Za-z_0-9./\\][A-Za-z_0-9./\\\-]* {
struct keyword *tok = iskeyword(yytext);
if(tok)
{
if(wanted_id && !tok->alwayskw)
{
parser_lval.str = make_string(yytext);
return tIDENT;
}
else
return tok->token;
}
else
{
parser_lval.str = make_string(yytext);
return tIDENT;
}
}
/*
* Wide string scanning
*/
L\" {
yy_push_state(tklstr);
wbufidx = 0;
if(!win32)
parser_error("16bit resource contains unicode strings");
}
<tklstr>\"{ws}+ |
<tklstr>\" {
yy_pop_state();
parser_lval.str = get_buffered_wstring();
return tSTRING;
}
<tklstr>\\[0-7]{1,6} { /* octal escape sequence */
unsigned int result;
result = strtoul(yytext+1, 0, 8);
if ( result > 0xffff )
parser_error("Character constant out of range");
addwchar((WCHAR)result);
}
<tklstr>\\x[0-9a-fA-F]{4} { /* hex escape sequence */
unsigned int result;
result = strtoul(yytext+2, 0, 16);
addwchar((WCHAR)result);
}
<tklstr>\\x[0-9a-fA-F]{1,3} { parser_error("Invalid hex escape sequence '%s'", yytext); }
<tklstr>\\[0-9]+ parser_error("Bad escape sequence");
<tklstr>\\\n{ws}* line_number++; char_number = 1; /* backslash at EOL continues string after leading whitespace on next line */
<tklstr>\\a addwchar('\a');
<tklstr>\\b addwchar('\b');
<tklstr>\\f addwchar('\f');
<tklstr>\\n addwchar('\n');
<tklstr>\\r addwchar('\r');
<tklstr>\\t addwchar('\t');
<tklstr>\\v addwchar('\v');
<tklstr>\\. {
if (yytext[1] & 0x80)
parser_error("Invalid char %u in wide string", (unsigned char)yytext[1]);
addwchar(yytext[1]);
}
<tklstr>\\\r\n addwchar(yytext[2]); line_number++; char_number = 1;
<tklstr>\"\" addwchar('\"'); /* "bla""bla" -> "bla\"bla" */
<tklstr>\\\"\" addwchar('\"'); /* "bla\""bla" -> "bla\"bla" */
<tklstr>\"{ws}+\" ; /* "bla" "bla" -> "blabla" */
<tklstr>[^\\\n\"]+ {
char *yptr = yytext;
while(*yptr) /* FIXME: codepage translation */
{
if (*yptr & 0x80)
parser_error("Invalid char %u in wide string", (unsigned char)*yptr);
addwchar(*yptr++ & 0xff);
}
}
<tklstr>\n parser_error("Unterminated string");
/*
* Normal string scanning
*/
\" yy_push_state(tkstr); cbufidx = 0;
<tkstr>\"{ws}+ |
<tkstr>\" {
yy_pop_state();
parser_lval.str = get_buffered_cstring();
return tSTRING;
}
<tkstr>\\[0-7]{1,3} { /* octal escape sequence */
int result;
result = strtol(yytext+1, 0, 8);
if ( result > 0xff )
parser_error("Character constant out of range");
addcchar((char)result);
}
<tkstr>\\x[0-9a-fA-F]{2} { /* hex escape sequence */
int result;
result = strtol(yytext+2, 0, 16);
addcchar((char)result);
}
<tkstr>\\x[0-9a-fA-F] { parser_error("Invalid hex escape sequence '%s'", yytext); }
<tkstr>\\[0-9]+ parser_error("Bad escape sequence");
<tkstr>\\\n{ws}* line_number++; char_number = 1; /* backslash at EOL continues string after leading whitespace on next line */
<tkstr>\\a addcchar('\a');
<tkstr>\\b addcchar('\b');
<tkstr>\\f addcchar('\f');
<tkstr>\\n addcchar('\n');
<tkstr>\\r addcchar('\r');
<tkstr>\\t addcchar('\t');
<tkstr>\\v addcchar('\v');
<tkstr>\\. addcchar(yytext[1]);
<tkstr>\\\r\n addcchar(yytext[2]); line_number++; char_number = 1;
<tkstr>[^\\\n\"]+ {
char *yptr = yytext;
while(*yptr)
addcchar(*yptr++);
}
<tkstr>\"\" addcchar('\"'); /* "bla""bla" -> "bla\"bla" */
<tkstr>\\\"\" addcchar('\"'); /* "bla\""bla" -> "bla\"bla" */
<tkstr>\"{ws}+\" ; /* "bla" "bla" -> "blabla" */
<tkstr>\n parser_error("Unterminated string");
/*
* Raw data scanning
*/
\' yy_push_state(tkrcd); cbufidx = 0;
<tkrcd>\' {
yy_pop_state();
parser_lval.raw = new_raw_data();
parser_lval.raw->size = cbufidx;
parser_lval.raw->data = xmalloc(parser_lval.raw->size);
memcpy(parser_lval.raw->data, cbuffer, parser_lval.raw->size);
return tRAWDATA;
}
<tkrcd>[0-9a-fA-F]{2} {
int result;
result = strtol(yytext, 0, 16);
addcchar((char)result);
}
<tkrcd>{ws}+ ; /* Ignore space */
<tkrcd>\n line_number++; char_number = 1;
<tkrcd>. parser_error("Malformed data-line");
/*
* Comment stripping
* Should never occur after preprocessing
*/
<INITIAL,pp_cstrip>"/*" {
yy_push_state(comment);
save_wanted_id = wanted_id;
if(!no_preprocess)
parser_warning("Found comments after preprocessing, please report\n");
}
<comment>[^*\n]* ;
<comment>"*"+[^*/\n]* ;
<comment>\n line_number++; char_number = 1;
<comment>"*"+"/" yy_pop_state(); want_id = save_wanted_id;
;[^\n]* want_id = wanted_id; /* not really comment, but left-over c-junk */
"//"[^\n]* want_id = wanted_id; if(!no_preprocess) parser_warning("Found comments after preprocessing, please report\n");
\n {
want_id = wanted_id;
line_number++;
char_number = 1;
if(want_nl)
{
want_nl = 0;
return tNL;
}
}
{ws}+ want_id = wanted_id; /* Eat whitespace */
<INITIAL>[ -~] return yytext[0];
<*>.|\n {
/* Catch all rule to find any unmatched text */
if(*yytext == '\n')
{
line_number++;
char_number = 1;
}
parser_error("Unmatched text '%c' (0x%02x) YY_START=%d",
isprint((unsigned char)*yytext) ? *yytext : '.', *yytext, YY_START);
}
%%
/* These dup functions copy the enclosed '\0' from
* the resource string.
*/
static void addcchar(char c)
{
if(cbufidx >= cbufalloc)
{
cbufalloc += 1024;
cbuffer = xrealloc(cbuffer, cbufalloc * sizeof(cbuffer[0]));
if(cbufalloc > 65536)
parser_warning("Reallocating string buffer larger than 64kB\n");
}
cbuffer[cbufidx++] = c;
}
static void addwchar(WCHAR s)
{
if(wbufidx >= wbufalloc)
{
wbufalloc += 1024;
wbuffer = xrealloc(wbuffer, wbufalloc * sizeof(wbuffer[0]));
if(wbufalloc > 65536)
parser_warning("Reallocating wide string buffer larger than 64kB\n");
}
wbuffer[wbufidx++] = s;
}
static string_t *get_buffered_cstring(void)
{
string_t *str = new_string();
str->size = cbufidx;
str->type = str_char;
str->str.cstr = xmalloc(cbufidx+1);
memcpy(str->str.cstr, cbuffer, cbufidx);
str->str.cstr[cbufidx] = '\0';
if (!current_codepage || current_codepage == -1 || !win32) /* store as ANSI string */
{
if (!current_codepage) parser_error("Codepage set to Unicode only, cannot use ASCII string here");
return str;
}
else /* convert to Unicode before storing */
{
string_t *str_w = convert_string_unicode( str, current_codepage );
if (check_valid_utf8( str, current_codepage ))
parser_warning( "string \"%s\" seems to be UTF-8 but codepage %u is in use, maybe use --utf8?\n",
str->str.cstr, current_codepage );
free_string( str );
return str_w;
}
}
static string_t *get_buffered_wstring(void)
{
string_t *str = new_string();
str->size = wbufidx;
str->type = str_unicode;
str->str.wstr = xmalloc((wbufidx+1)*sizeof(WCHAR));
memcpy(str->str.wstr, wbuffer, wbufidx*sizeof(WCHAR));
str->str.wstr[wbufidx] = 0;
return str;
}
static string_t *make_string(char *s)
{
string_t *ret, *str = new_string();
str->size = strlen(s);
str->type = str_char;
str->str.cstr = xmalloc(str->size+1);
memcpy(str->str.cstr, s, str->size+1);
if (current_codepage <= 0 || !win32) return str;
ret = convert_string_unicode( str, current_codepage );
free_string( str );
return ret;
}