mksyntax.c revision 149026
11556Srgrimes/*- 21556Srgrimes * Copyright (c) 1991, 1993 31556Srgrimes * The Regents of the University of California. All rights reserved. 41556Srgrimes * 51556Srgrimes * This code is derived from software contributed to Berkeley by 61556Srgrimes * Kenneth Almquist. 71556Srgrimes * 81556Srgrimes * Redistribution and use in source and binary forms, with or without 91556Srgrimes * modification, are permitted provided that the following conditions 101556Srgrimes * are met: 111556Srgrimes * 1. Redistributions of source code must retain the above copyright 121556Srgrimes * notice, this list of conditions and the following disclaimer. 131556Srgrimes * 2. Redistributions in binary form must reproduce the above copyright 141556Srgrimes * notice, this list of conditions and the following disclaimer in the 151556Srgrimes * documentation and/or other materials provided with the distribution. 161556Srgrimes * 4. Neither the name of the University nor the names of its contributors 171556Srgrimes * may be used to endorse or promote products derived from this software 181556Srgrimes * without specific prior written permission. 191556Srgrimes * 201556Srgrimes * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 211556Srgrimes * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 221556Srgrimes * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 231556Srgrimes * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 241556Srgrimes * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 251556Srgrimes * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 261556Srgrimes * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 271556Srgrimes * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 281556Srgrimes * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 291556Srgrimes * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 301556Srgrimes * SUCH DAMAGE. 311556Srgrimes */ 321556Srgrimes 33114433Sobrien#if 0 341556Srgrimes#ifndef lint 3520425Sstevestatic char const copyright[] = 361556Srgrimes"@(#) Copyright (c) 1991, 1993\n\ 371556Srgrimes The Regents of the University of California. All rights reserved.\n"; 381556Srgrimes#endif /* not lint */ 391556Srgrimes 401556Srgrimes#ifndef lint 4136150Scharnierstatic char sccsid[] = "@(#)mksyntax.c 8.2 (Berkeley) 5/4/95"; 42114433Sobrien#endif /* not lint */ 4336150Scharnier#endif 4499110Sobrien#include <sys/cdefs.h> 4599110Sobrien__FBSDID("$FreeBSD: head/bin/sh/mksyntax.c 149026 2005-08-13 15:47:13Z stefanf $"); 461556Srgrimes 471556Srgrimes/* 481556Srgrimes * This program creates syntax.h and syntax.c. 491556Srgrimes */ 501556Srgrimes 511556Srgrimes#include <stdio.h> 5278469Sdes#include <stdlib.h> 5317987Speter#include <string.h> 541556Srgrimes#include "parser.h" 551556Srgrimes 561556Srgrimes 571556Srgrimesstruct synclass { 581556Srgrimes char *name; 591556Srgrimes char *comment; 601556Srgrimes}; 611556Srgrimes 621556Srgrimes/* Syntax classes */ 631556Srgrimesstruct synclass synclass[] = { 6417987Speter { "CWORD", "character is nothing special" }, 6517987Speter { "CNL", "newline character" }, 6617987Speter { "CBACK", "a backslash character" }, 6717987Speter { "CSQUOTE", "single quote" }, 6817987Speter { "CDQUOTE", "double quote" }, 6917987Speter { "CENDQUOTE", "a terminating quote" }, 7017987Speter { "CBQUOTE", "backwards single quote" }, 7117987Speter { "CVAR", "a dollar sign" }, 7217987Speter { "CENDVAR", "a '}' character" }, 7317987Speter { "CLP", "a left paren in arithmetic" }, 7417987Speter { "CRP", "a right paren in arithmetic" }, 7517987Speter { "CEOF", "end of file" }, 7617987Speter { "CCTL", "like CWORD, except it must be escaped" }, 7717987Speter { "CSPCL", "these terminate a word" }, 7817987Speter { NULL, NULL } 791556Srgrimes}; 801556Srgrimes 811556Srgrimes 821556Srgrimes/* 831556Srgrimes * Syntax classes for is_ functions. Warning: if you add new classes 841556Srgrimes * you may have to change the definition of the is_in_name macro. 851556Srgrimes */ 861556Srgrimesstruct synclass is_entry[] = { 8717987Speter { "ISDIGIT", "a digit" }, 8817987Speter { "ISUPPER", "an upper case letter" }, 8917987Speter { "ISLOWER", "a lower case letter" }, 9017987Speter { "ISUNDER", "an underscore" }, 9117987Speter { "ISSPECL", "the name of a special parameter" }, 9217987Speter { NULL, NULL } 931556Srgrimes}; 941556Srgrimes 9517987Speterstatic char writer[] = "\ 961556Srgrimes/*\n\ 971556Srgrimes * This file was generated by the mksyntax program.\n\ 981556Srgrimes */\n\ 991556Srgrimes\n"; 1001556Srgrimes 1011556Srgrimes 10217987Speterstatic FILE *cfile; 10317987Speterstatic FILE *hfile; 10417987Speterstatic char *syntax[513]; 10517987Speterstatic int base; 10617987Speterstatic int size; /* number of values which a char variable can have */ 10717987Speterstatic int nbits; /* number of bits in a character */ 10817987Speterstatic int digit_contig;/* true if digits are contiguous */ 1091556Srgrimes 11090111Simpstatic void filltable(char *); 11190111Simpstatic void init(void); 11290111Simpstatic void add(char *, char *); 11390111Simpstatic void print(char *); 11490111Simpstatic void output_type_macros(void); 11590111Simpstatic void digit_convert(void); 1161556Srgrimes 11717987Speterint 11890111Simpmain(int argc __unused, char **argv __unused) 11917987Speter{ 1201556Srgrimes char c; 1211556Srgrimes char d; 1221556Srgrimes int sign; 1231556Srgrimes int i; 1241556Srgrimes char buf[80]; 1251556Srgrimes int pos; 1261556Srgrimes static char digit[] = "0123456789"; 1271556Srgrimes 1281556Srgrimes /* Create output files */ 1291556Srgrimes if ((cfile = fopen("syntax.c", "w")) == NULL) { 1301556Srgrimes perror("syntax.c"); 1311556Srgrimes exit(2); 1321556Srgrimes } 1331556Srgrimes if ((hfile = fopen("syntax.h", "w")) == NULL) { 1341556Srgrimes perror("syntax.h"); 1351556Srgrimes exit(2); 1361556Srgrimes } 1371556Srgrimes fputs(writer, hfile); 1381556Srgrimes fputs(writer, cfile); 1391556Srgrimes 1401556Srgrimes /* Determine the characteristics of chars. */ 1411556Srgrimes c = -1; 1421556Srgrimes if (c < 0) 1431556Srgrimes sign = 1; 1441556Srgrimes else 1451556Srgrimes sign = 0; 1461556Srgrimes for (nbits = 1 ; ; nbits++) { 1471556Srgrimes d = (1 << nbits) - 1; 1481556Srgrimes if (d == c) 1491556Srgrimes break; 1501556Srgrimes } 15120425Ssteve#if 0 1521556Srgrimes printf("%s %d bit chars\n", sign? "signed" : "unsigned", nbits); 15320425Ssteve#endif 1541556Srgrimes if (nbits > 9) { 1551556Srgrimes fputs("Characters can't have more than 9 bits\n", stderr); 1561556Srgrimes exit(2); 1571556Srgrimes } 1581556Srgrimes size = (1 << nbits) + 1; 1591556Srgrimes base = 1; 1601556Srgrimes if (sign) 1611556Srgrimes base += 1 << (nbits - 1); 1621556Srgrimes digit_contig = 1; 1631556Srgrimes for (i = 0 ; i < 10 ; i++) { 1641556Srgrimes if (digit[i] != '0' + i) 1651556Srgrimes digit_contig = 0; 1661556Srgrimes } 1671556Srgrimes 16820425Ssteve fputs("#include <sys/cdefs.h>\n", hfile); 16917525Sache fputs("#include <ctype.h>\n", hfile); 1701556Srgrimes 1711556Srgrimes /* Generate the #define statements in the header file */ 1721556Srgrimes fputs("/* Syntax classes */\n", hfile); 1731556Srgrimes for (i = 0 ; synclass[i].name ; i++) { 1741556Srgrimes sprintf(buf, "#define %s %d", synclass[i].name, i); 1751556Srgrimes fputs(buf, hfile); 17617987Speter for (pos = strlen(buf) ; pos < 32 ; pos = (pos + 8) & ~07) 1771556Srgrimes putc('\t', hfile); 1781556Srgrimes fprintf(hfile, "/* %s */\n", synclass[i].comment); 1791556Srgrimes } 1801556Srgrimes putc('\n', hfile); 1811556Srgrimes fputs("/* Syntax classes for is_ functions */\n", hfile); 1821556Srgrimes for (i = 0 ; is_entry[i].name ; i++) { 1831556Srgrimes sprintf(buf, "#define %s %#o", is_entry[i].name, 1 << i); 1841556Srgrimes fputs(buf, hfile); 18517987Speter for (pos = strlen(buf) ; pos < 32 ; pos = (pos + 8) & ~07) 1861556Srgrimes putc('\t', hfile); 1871556Srgrimes fprintf(hfile, "/* %s */\n", is_entry[i].comment); 1881556Srgrimes } 1891556Srgrimes putc('\n', hfile); 1901556Srgrimes fprintf(hfile, "#define SYNBASE %d\n", base); 1911556Srgrimes fprintf(hfile, "#define PEOF %d\n\n", -base); 1921556Srgrimes putc('\n', hfile); 1931556Srgrimes fputs("#define BASESYNTAX (basesyntax + SYNBASE)\n", hfile); 1941556Srgrimes fputs("#define DQSYNTAX (dqsyntax + SYNBASE)\n", hfile); 1951556Srgrimes fputs("#define SQSYNTAX (sqsyntax + SYNBASE)\n", hfile); 1961556Srgrimes fputs("#define ARISYNTAX (arisyntax + SYNBASE)\n", hfile); 1971556Srgrimes putc('\n', hfile); 1981556Srgrimes output_type_macros(); /* is_digit, etc. */ 1991556Srgrimes putc('\n', hfile); 2001556Srgrimes 2011556Srgrimes /* Generate the syntax tables. */ 2021556Srgrimes fputs("#include \"shell.h\"\n", cfile); 2031556Srgrimes fputs("#include \"syntax.h\"\n\n", cfile); 2041556Srgrimes init(); 2051556Srgrimes fputs("/* syntax table used when not in quotes */\n", cfile); 2061556Srgrimes add("\n", "CNL"); 2071556Srgrimes add("\\", "CBACK"); 2081556Srgrimes add("'", "CSQUOTE"); 2091556Srgrimes add("\"", "CDQUOTE"); 2101556Srgrimes add("`", "CBQUOTE"); 2111556Srgrimes add("$", "CVAR"); 2121556Srgrimes add("}", "CENDVAR"); 2131556Srgrimes add("<>();&| \t", "CSPCL"); 2141556Srgrimes print("basesyntax"); 2151556Srgrimes init(); 2161556Srgrimes fputs("\n/* syntax table used when in double quotes */\n", cfile); 2171556Srgrimes add("\n", "CNL"); 2181556Srgrimes add("\\", "CBACK"); 2191556Srgrimes add("\"", "CENDQUOTE"); 2201556Srgrimes add("`", "CBQUOTE"); 2211556Srgrimes add("$", "CVAR"); 2221556Srgrimes add("}", "CENDVAR"); 22317987Speter /* ':/' for tilde expansion, '-' for [a\-x] pattern ranges */ 22417987Speter add("!*?[=~:/-", "CCTL"); 2251556Srgrimes print("dqsyntax"); 2261556Srgrimes init(); 2271556Srgrimes fputs("\n/* syntax table used when in single quotes */\n", cfile); 2281556Srgrimes add("\n", "CNL"); 2291556Srgrimes add("'", "CENDQUOTE"); 23017987Speter /* ':/' for tilde expansion, '-' for [a\-x] pattern ranges */ 23117987Speter add("!*?[=~:/-", "CCTL"); 2321556Srgrimes print("sqsyntax"); 2331556Srgrimes init(); 2341556Srgrimes fputs("\n/* syntax table used when in arithmetic */\n", cfile); 2351556Srgrimes add("\n", "CNL"); 2361556Srgrimes add("\\", "CBACK"); 2371556Srgrimes add("`", "CBQUOTE"); 2381556Srgrimes add("'", "CSQUOTE"); 2391556Srgrimes add("\"", "CDQUOTE"); 2401556Srgrimes add("$", "CVAR"); 2411556Srgrimes add("}", "CENDVAR"); 2421556Srgrimes add("(", "CLP"); 2431556Srgrimes add(")", "CRP"); 2441556Srgrimes print("arisyntax"); 2451556Srgrimes filltable("0"); 2461556Srgrimes fputs("\n/* character classification table */\n", cfile); 2471556Srgrimes add("0123456789", "ISDIGIT"); 2481556Srgrimes add("abcdefghijklmnopqrstucvwxyz", "ISLOWER"); 2491556Srgrimes add("ABCDEFGHIJKLMNOPQRSTUCVWXYZ", "ISUPPER"); 2501556Srgrimes add("_", "ISUNDER"); 2511556Srgrimes add("#?$!-*@", "ISSPECL"); 2521556Srgrimes print("is_type"); 2531556Srgrimes if (! digit_contig) 2541556Srgrimes digit_convert(); 2551556Srgrimes exit(0); 2561556Srgrimes} 2571556Srgrimes 2581556Srgrimes 2591556Srgrimes 2601556Srgrimes/* 2611556Srgrimes * Clear the syntax table. 2621556Srgrimes */ 2631556Srgrimes 26417987Speterstatic void 26590111Simpfilltable(char *dftval) 26617987Speter{ 2671556Srgrimes int i; 2681556Srgrimes 2691556Srgrimes for (i = 0 ; i < size ; i++) 2701556Srgrimes syntax[i] = dftval; 2711556Srgrimes} 2721556Srgrimes 2731556Srgrimes 2741556Srgrimes/* 2751556Srgrimes * Initialize the syntax table with default values. 2761556Srgrimes */ 2771556Srgrimes 27817987Speterstatic void 27990111Simpinit(void) 28017987Speter{ 2811556Srgrimes filltable("CWORD"); 2821556Srgrimes syntax[0] = "CEOF"; 2831556Srgrimes syntax[base + CTLESC] = "CCTL"; 2841556Srgrimes syntax[base + CTLVAR] = "CCTL"; 2851556Srgrimes syntax[base + CTLENDVAR] = "CCTL"; 2861556Srgrimes syntax[base + CTLBACKQ] = "CCTL"; 2871556Srgrimes syntax[base + CTLBACKQ + CTLQUOTE] = "CCTL"; 2881556Srgrimes syntax[base + CTLARI] = "CCTL"; 2891556Srgrimes syntax[base + CTLENDARI] = "CCTL"; 29038887Stegge syntax[base + CTLQUOTEMARK] = "CCTL"; 2911556Srgrimes} 2921556Srgrimes 2931556Srgrimes 2941556Srgrimes/* 2951556Srgrimes * Add entries to the syntax table. 2961556Srgrimes */ 2971556Srgrimes 29817987Speterstatic void 29990111Simpadd(char *p, char *type) 30017987Speter{ 3011556Srgrimes while (*p) 3021556Srgrimes syntax[*p++ + base] = type; 3031556Srgrimes} 3041556Srgrimes 3051556Srgrimes 3061556Srgrimes 3071556Srgrimes/* 3081556Srgrimes * Output the syntax table. 3091556Srgrimes */ 3101556Srgrimes 31117987Speterstatic void 31290111Simpprint(char *name) 31317987Speter{ 3141556Srgrimes int i; 3151556Srgrimes int col; 3161556Srgrimes 3171556Srgrimes fprintf(hfile, "extern const char %s[];\n", name); 3181556Srgrimes fprintf(cfile, "const char %s[%d] = {\n", name, size); 3191556Srgrimes col = 0; 3201556Srgrimes for (i = 0 ; i < size ; i++) { 3211556Srgrimes if (i == 0) { 3221556Srgrimes fputs(" ", cfile); 3231556Srgrimes } else if ((i & 03) == 0) { 3241556Srgrimes fputs(",\n ", cfile); 3251556Srgrimes col = 0; 3261556Srgrimes } else { 3271556Srgrimes putc(',', cfile); 3281556Srgrimes while (++col < 9 * (i & 03)) 3291556Srgrimes putc(' ', cfile); 3301556Srgrimes } 3311556Srgrimes fputs(syntax[i], cfile); 3321556Srgrimes col += strlen(syntax[i]); 3331556Srgrimes } 3341556Srgrimes fputs("\n};\n", cfile); 3351556Srgrimes} 3361556Srgrimes 3371556Srgrimes 3381556Srgrimes 3391556Srgrimes/* 3401556Srgrimes * Output character classification macros (e.g. is_digit). If digits are 3411556Srgrimes * contiguous, we can test for them quickly. 3421556Srgrimes */ 3431556Srgrimes 34417987Speterstatic char *macro[] = { 34584936Stegge "#define is_digit(c)\t((is_type+SYNBASE)[c] & ISDIGIT)", 346149026Sstefanf "#define is_eof(c)\t((c) == PEOF)", 347149026Sstefanf "#define is_alpha(c)\t(((c) < CTLESC || (c) > CTLQUOTEMARK) && isalpha((unsigned char) (c)))", 348149026Sstefanf "#define is_name(c)\t(((c) < CTLESC || (c) > CTLQUOTEMARK) && ((c) == '_' || isalpha((unsigned char) (c))))", 349149026Sstefanf "#define is_in_name(c)\t(((c) < CTLESC || (c) > CTLQUOTEMARK) && ((c) == '_' || isalnum((unsigned char) (c))))", 3501556Srgrimes "#define is_special(c)\t((is_type+SYNBASE)[c] & (ISSPECL|ISDIGIT))", 3511556Srgrimes NULL 3521556Srgrimes}; 3531556Srgrimes 35417987Speterstatic void 35590111Simpoutput_type_macros(void) 35617987Speter{ 3571556Srgrimes char **pp; 3581556Srgrimes 3591556Srgrimes if (digit_contig) 3601556Srgrimes macro[0] = "#define is_digit(c)\t((unsigned)((c) - '0') <= 9)"; 3611556Srgrimes for (pp = macro ; *pp ; pp++) 3621556Srgrimes fprintf(hfile, "%s\n", *pp); 3631556Srgrimes if (digit_contig) 3641556Srgrimes fputs("#define digit_val(c)\t((c) - '0')\n", hfile); 3651556Srgrimes else 3661556Srgrimes fputs("#define digit_val(c)\t(digit_value[c])\n", hfile); 3671556Srgrimes} 3681556Srgrimes 3691556Srgrimes 3701556Srgrimes 3711556Srgrimes/* 3721556Srgrimes * Output digit conversion table (if digits are not contiguous). 3731556Srgrimes */ 3741556Srgrimes 37517987Speterstatic void 37690111Simpdigit_convert(void) 37717987Speter{ 3781556Srgrimes int maxdigit; 3791556Srgrimes static char digit[] = "0123456789"; 3801556Srgrimes char *p; 3811556Srgrimes int i; 3821556Srgrimes 3831556Srgrimes maxdigit = 0; 3841556Srgrimes for (p = digit ; *p ; p++) 3851556Srgrimes if (*p > maxdigit) 3861556Srgrimes maxdigit = *p; 3871556Srgrimes fputs("extern const char digit_value[];\n", hfile); 3881556Srgrimes fputs("\n\nconst char digit_value[] = {\n", cfile); 3891556Srgrimes for (i = 0 ; i <= maxdigit ; i++) { 3901556Srgrimes for (p = digit ; *p && *p != i ; p++); 3911556Srgrimes if (*p == '\0') 3921556Srgrimes p = digit; 393104367Stjr fprintf(cfile, " %d,\n", (int)(p - digit)); 3941556Srgrimes } 3951556Srgrimes fputs("};\n", cfile); 3961556Srgrimes} 397