mksyntax.c revision 17987
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 * 3. All advertising materials mentioning features or use of this software 171556Srgrimes * must display the following acknowledgement: 181556Srgrimes * This product includes software developed by the University of 191556Srgrimes * California, Berkeley and its contributors. 201556Srgrimes * 4. Neither the name of the University nor the names of its contributors 211556Srgrimes * may be used to endorse or promote products derived from this software 221556Srgrimes * without specific prior written permission. 231556Srgrimes * 241556Srgrimes * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 251556Srgrimes * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 261556Srgrimes * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 271556Srgrimes * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 281556Srgrimes * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 291556Srgrimes * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 301556Srgrimes * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 311556Srgrimes * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 321556Srgrimes * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 331556Srgrimes * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 341556Srgrimes * SUCH DAMAGE. 353044Sdg * 3617987Speter * $Id: mksyntax.c,v 1.5 1996/08/12 22:14:47 ache Exp $ 371556Srgrimes */ 381556Srgrimes 391556Srgrimes#ifndef lint 401556Srgrimesstatic char copyright[] = 411556Srgrimes"@(#) Copyright (c) 1991, 1993\n\ 421556Srgrimes The Regents of the University of California. All rights reserved.\n"; 431556Srgrimes#endif /* not lint */ 441556Srgrimes 451556Srgrimes#ifndef lint 4617987Speterstatic char sccsid[] = "@(#)mksyntax.c 8.2 (Berkeley) 5/4/95"; 471556Srgrimes#endif /* not lint */ 481556Srgrimes 491556Srgrimes/* 501556Srgrimes * This program creates syntax.h and syntax.c. 511556Srgrimes */ 521556Srgrimes 531556Srgrimes#include <stdio.h> 5417987Speter#include <string.h> 551556Srgrimes#include "parser.h" 561556Srgrimes 571556Srgrimes 581556Srgrimesstruct synclass { 591556Srgrimes char *name; 601556Srgrimes char *comment; 611556Srgrimes}; 621556Srgrimes 631556Srgrimes/* Syntax classes */ 641556Srgrimesstruct synclass synclass[] = { 6517987Speter { "CWORD", "character is nothing special" }, 6617987Speter { "CNL", "newline character" }, 6717987Speter { "CBACK", "a backslash character" }, 6817987Speter { "CSQUOTE", "single quote" }, 6917987Speter { "CDQUOTE", "double quote" }, 7017987Speter { "CENDQUOTE", "a terminating quote" }, 7117987Speter { "CBQUOTE", "backwards single quote" }, 7217987Speter { "CVAR", "a dollar sign" }, 7317987Speter { "CENDVAR", "a '}' character" }, 7417987Speter { "CLP", "a left paren in arithmetic" }, 7517987Speter { "CRP", "a right paren in arithmetic" }, 7617987Speter { "CEOF", "end of file" }, 7717987Speter { "CCTL", "like CWORD, except it must be escaped" }, 7817987Speter { "CSPCL", "these terminate a word" }, 7917987Speter { NULL, NULL } 801556Srgrimes}; 811556Srgrimes 821556Srgrimes 831556Srgrimes/* 841556Srgrimes * Syntax classes for is_ functions. Warning: if you add new classes 851556Srgrimes * you may have to change the definition of the is_in_name macro. 861556Srgrimes */ 871556Srgrimesstruct synclass is_entry[] = { 8817987Speter { "ISDIGIT", "a digit" }, 8917987Speter { "ISUPPER", "an upper case letter" }, 9017987Speter { "ISLOWER", "a lower case letter" }, 9117987Speter { "ISUNDER", "an underscore" }, 9217987Speter { "ISSPECL", "the name of a special parameter" }, 9317987Speter { NULL, NULL } 941556Srgrimes}; 951556Srgrimes 9617987Speterstatic char writer[] = "\ 971556Srgrimes/*\n\ 981556Srgrimes * This file was generated by the mksyntax program.\n\ 991556Srgrimes */\n\ 1001556Srgrimes\n"; 1011556Srgrimes 1021556Srgrimes 10317987Speterstatic FILE *cfile; 10417987Speterstatic FILE *hfile; 10517987Speterstatic char *syntax[513]; 10617987Speterstatic int base; 10717987Speterstatic int size; /* number of values which a char variable can have */ 10817987Speterstatic int nbits; /* number of bits in a character */ 10917987Speterstatic int digit_contig;/* true if digits are contiguous */ 1101556Srgrimes 11117987Speterstatic void filltable __P((char *)); 11217987Speterstatic void init __P((void)); 11317987Speterstatic void add __P((char *, char *)); 11417987Speterstatic void print __P((char *)); 11517987Speterstatic void output_type_macros __P((void)); 11617987Speterstatic void digit_convert __P((void)); 1171556Srgrimes 11817987Speterint 11917987Spetermain(argc, argv) 12017987Speter int argc; 12117987Speter char **argv; 12217987Speter{ 1231556Srgrimes char c; 1241556Srgrimes char d; 1251556Srgrimes int sign; 1261556Srgrimes int i; 1271556Srgrimes char buf[80]; 1281556Srgrimes int pos; 1291556Srgrimes static char digit[] = "0123456789"; 1301556Srgrimes 1311556Srgrimes /* Create output files */ 1321556Srgrimes if ((cfile = fopen("syntax.c", "w")) == NULL) { 1331556Srgrimes perror("syntax.c"); 1341556Srgrimes exit(2); 1351556Srgrimes } 1361556Srgrimes if ((hfile = fopen("syntax.h", "w")) == NULL) { 1371556Srgrimes perror("syntax.h"); 1381556Srgrimes exit(2); 1391556Srgrimes } 1401556Srgrimes fputs(writer, hfile); 1411556Srgrimes fputs(writer, cfile); 1421556Srgrimes 1431556Srgrimes /* Determine the characteristics of chars. */ 1441556Srgrimes c = -1; 1451556Srgrimes if (c < 0) 1461556Srgrimes sign = 1; 1471556Srgrimes else 1481556Srgrimes sign = 0; 1491556Srgrimes for (nbits = 1 ; ; nbits++) { 1501556Srgrimes d = (1 << nbits) - 1; 1511556Srgrimes if (d == c) 1521556Srgrimes break; 1531556Srgrimes } 1541556Srgrimes printf("%s %d bit chars\n", sign? "signed" : "unsigned", nbits); 1551556Srgrimes if (nbits > 9) { 1561556Srgrimes fputs("Characters can't have more than 9 bits\n", stderr); 1571556Srgrimes exit(2); 1581556Srgrimes } 1591556Srgrimes size = (1 << nbits) + 1; 1601556Srgrimes base = 1; 1611556Srgrimes if (sign) 1621556Srgrimes base += 1 << (nbits - 1); 1631556Srgrimes digit_contig = 1; 1641556Srgrimes for (i = 0 ; i < 10 ; i++) { 1651556Srgrimes if (digit[i] != '0' + i) 1661556Srgrimes digit_contig = 0; 1671556Srgrimes } 1681556Srgrimes 16917525Sache fputs("#include <ctype.h>\n", hfile); 1701556Srgrimes fputs("#include <sys/cdefs.h>\n", hfile); 1711556Srgrimes 1721556Srgrimes /* Generate the #define statements in the header file */ 1731556Srgrimes fputs("/* Syntax classes */\n", hfile); 1741556Srgrimes for (i = 0 ; synclass[i].name ; i++) { 1751556Srgrimes sprintf(buf, "#define %s %d", synclass[i].name, i); 1761556Srgrimes fputs(buf, hfile); 17717987Speter for (pos = strlen(buf) ; pos < 32 ; pos = (pos + 8) & ~07) 1781556Srgrimes putc('\t', hfile); 1791556Srgrimes fprintf(hfile, "/* %s */\n", synclass[i].comment); 1801556Srgrimes } 1811556Srgrimes putc('\n', hfile); 1821556Srgrimes fputs("/* Syntax classes for is_ functions */\n", hfile); 1831556Srgrimes for (i = 0 ; is_entry[i].name ; i++) { 1841556Srgrimes sprintf(buf, "#define %s %#o", is_entry[i].name, 1 << i); 1851556Srgrimes fputs(buf, hfile); 18617987Speter for (pos = strlen(buf) ; pos < 32 ; pos = (pos + 8) & ~07) 1871556Srgrimes putc('\t', hfile); 1881556Srgrimes fprintf(hfile, "/* %s */\n", is_entry[i].comment); 1891556Srgrimes } 1901556Srgrimes putc('\n', hfile); 1911556Srgrimes fprintf(hfile, "#define SYNBASE %d\n", base); 1921556Srgrimes fprintf(hfile, "#define PEOF %d\n\n", -base); 1931556Srgrimes putc('\n', hfile); 1941556Srgrimes fputs("#define BASESYNTAX (basesyntax + SYNBASE)\n", hfile); 1951556Srgrimes fputs("#define DQSYNTAX (dqsyntax + SYNBASE)\n", hfile); 1961556Srgrimes fputs("#define SQSYNTAX (sqsyntax + SYNBASE)\n", hfile); 1971556Srgrimes fputs("#define ARISYNTAX (arisyntax + SYNBASE)\n", hfile); 1981556Srgrimes putc('\n', hfile); 1991556Srgrimes output_type_macros(); /* is_digit, etc. */ 2001556Srgrimes putc('\n', hfile); 2011556Srgrimes 2021556Srgrimes /* Generate the syntax tables. */ 2031556Srgrimes fputs("#include \"shell.h\"\n", cfile); 2041556Srgrimes fputs("#include \"syntax.h\"\n\n", cfile); 2051556Srgrimes init(); 2061556Srgrimes fputs("/* syntax table used when not in quotes */\n", cfile); 2071556Srgrimes add("\n", "CNL"); 2081556Srgrimes add("\\", "CBACK"); 2091556Srgrimes add("'", "CSQUOTE"); 2101556Srgrimes add("\"", "CDQUOTE"); 2111556Srgrimes add("`", "CBQUOTE"); 2121556Srgrimes add("$", "CVAR"); 2131556Srgrimes add("}", "CENDVAR"); 2141556Srgrimes add("<>();&| \t", "CSPCL"); 2151556Srgrimes print("basesyntax"); 2161556Srgrimes init(); 2171556Srgrimes fputs("\n/* syntax table used when in double quotes */\n", cfile); 2181556Srgrimes add("\n", "CNL"); 2191556Srgrimes add("\\", "CBACK"); 2201556Srgrimes add("\"", "CENDQUOTE"); 2211556Srgrimes add("`", "CBQUOTE"); 2221556Srgrimes add("$", "CVAR"); 2231556Srgrimes add("}", "CENDVAR"); 22417987Speter /* ':/' for tilde expansion, '-' for [a\-x] pattern ranges */ 22517987Speter add("!*?[=~:/-", "CCTL"); 2261556Srgrimes print("dqsyntax"); 2271556Srgrimes init(); 2281556Srgrimes fputs("\n/* syntax table used when in single quotes */\n", cfile); 2291556Srgrimes add("\n", "CNL"); 2301556Srgrimes add("'", "CENDQUOTE"); 23117987Speter /* ':/' for tilde expansion, '-' for [a\-x] pattern ranges */ 23217987Speter add("!*?[=~:/-", "CCTL"); 2331556Srgrimes print("sqsyntax"); 2341556Srgrimes init(); 2351556Srgrimes fputs("\n/* syntax table used when in arithmetic */\n", cfile); 2361556Srgrimes add("\n", "CNL"); 2371556Srgrimes add("\\", "CBACK"); 2381556Srgrimes add("`", "CBQUOTE"); 2391556Srgrimes add("'", "CSQUOTE"); 2401556Srgrimes add("\"", "CDQUOTE"); 2411556Srgrimes add("$", "CVAR"); 2421556Srgrimes add("}", "CENDVAR"); 2431556Srgrimes add("(", "CLP"); 2441556Srgrimes add(")", "CRP"); 2451556Srgrimes print("arisyntax"); 2461556Srgrimes filltable("0"); 2471556Srgrimes fputs("\n/* character classification table */\n", cfile); 2481556Srgrimes add("0123456789", "ISDIGIT"); 2491556Srgrimes add("abcdefghijklmnopqrstucvwxyz", "ISLOWER"); 2501556Srgrimes add("ABCDEFGHIJKLMNOPQRSTUCVWXYZ", "ISUPPER"); 2511556Srgrimes add("_", "ISUNDER"); 2521556Srgrimes add("#?$!-*@", "ISSPECL"); 2531556Srgrimes print("is_type"); 2541556Srgrimes if (! digit_contig) 2551556Srgrimes digit_convert(); 2561556Srgrimes exit(0); 2571556Srgrimes} 2581556Srgrimes 2591556Srgrimes 2601556Srgrimes 2611556Srgrimes/* 2621556Srgrimes * Clear the syntax table. 2631556Srgrimes */ 2641556Srgrimes 26517987Speterstatic void 2661556Srgrimesfilltable(dftval) 2671556Srgrimes char *dftval; 26817987Speter{ 2691556Srgrimes int i; 2701556Srgrimes 2711556Srgrimes for (i = 0 ; i < size ; i++) 2721556Srgrimes syntax[i] = dftval; 2731556Srgrimes} 2741556Srgrimes 2751556Srgrimes 2761556Srgrimes/* 2771556Srgrimes * Initialize the syntax table with default values. 2781556Srgrimes */ 2791556Srgrimes 28017987Speterstatic void 28117987Speterinit() 28217987Speter{ 2831556Srgrimes filltable("CWORD"); 2841556Srgrimes syntax[0] = "CEOF"; 2851556Srgrimes syntax[base + CTLESC] = "CCTL"; 2861556Srgrimes syntax[base + CTLVAR] = "CCTL"; 2871556Srgrimes syntax[base + CTLENDVAR] = "CCTL"; 2881556Srgrimes syntax[base + CTLBACKQ] = "CCTL"; 2891556Srgrimes syntax[base + CTLBACKQ + CTLQUOTE] = "CCTL"; 2901556Srgrimes syntax[base + CTLARI] = "CCTL"; 2911556Srgrimes syntax[base + CTLENDARI] = "CCTL"; 2921556Srgrimes} 2931556Srgrimes 2941556Srgrimes 2951556Srgrimes/* 2961556Srgrimes * Add entries to the syntax table. 2971556Srgrimes */ 2981556Srgrimes 29917987Speterstatic void 3001556Srgrimesadd(p, type) 3011556Srgrimes char *p, *type; 30217987Speter{ 3031556Srgrimes while (*p) 3041556Srgrimes syntax[*p++ + base] = type; 3051556Srgrimes} 3061556Srgrimes 3071556Srgrimes 3081556Srgrimes 3091556Srgrimes/* 3101556Srgrimes * Output the syntax table. 3111556Srgrimes */ 3121556Srgrimes 31317987Speterstatic void 3141556Srgrimesprint(name) 3151556Srgrimes char *name; 31617987Speter{ 3171556Srgrimes int i; 3181556Srgrimes int col; 3191556Srgrimes 3201556Srgrimes fprintf(hfile, "extern const char %s[];\n", name); 3211556Srgrimes fprintf(cfile, "const char %s[%d] = {\n", name, size); 3221556Srgrimes col = 0; 3231556Srgrimes for (i = 0 ; i < size ; i++) { 3241556Srgrimes if (i == 0) { 3251556Srgrimes fputs(" ", cfile); 3261556Srgrimes } else if ((i & 03) == 0) { 3271556Srgrimes fputs(",\n ", cfile); 3281556Srgrimes col = 0; 3291556Srgrimes } else { 3301556Srgrimes putc(',', cfile); 3311556Srgrimes while (++col < 9 * (i & 03)) 3321556Srgrimes putc(' ', cfile); 3331556Srgrimes } 3341556Srgrimes fputs(syntax[i], cfile); 3351556Srgrimes col += strlen(syntax[i]); 3361556Srgrimes } 3371556Srgrimes fputs("\n};\n", cfile); 3381556Srgrimes} 3391556Srgrimes 3401556Srgrimes 3411556Srgrimes 3421556Srgrimes/* 3431556Srgrimes * Output character classification macros (e.g. is_digit). If digits are 3441556Srgrimes * contiguous, we can test for them quickly. 3451556Srgrimes */ 3461556Srgrimes 34717987Speterstatic char *macro[] = { 3481556Srgrimes "#define is_digit(c)\t((is_type+SYNBASE)[c] & ISDIGIT)", 34917562Sache "#define is_alpha(c)\t((c) != PEOF && ((c) < CTLESC || (c) > CTLENDARI) && isalpha((unsigned char) (c)))", 35017562Sache "#define is_name(c)\t((c) != PEOF && ((c) < CTLESC || (c) > CTLENDARI) && ((c) == '_' || isalpha((unsigned char) (c))))", 35117562Sache "#define is_in_name(c)\t((c) != PEOF && ((c) < CTLESC || (c) > CTLENDARI) && ((c) == '_' || isalnum((unsigned char) (c))))", 3521556Srgrimes "#define is_special(c)\t((is_type+SYNBASE)[c] & (ISSPECL|ISDIGIT))", 3531556Srgrimes NULL 3541556Srgrimes}; 3551556Srgrimes 35617987Speterstatic void 35717987Speteroutput_type_macros() 35817987Speter{ 3591556Srgrimes char **pp; 3601556Srgrimes 3611556Srgrimes if (digit_contig) 3621556Srgrimes macro[0] = "#define is_digit(c)\t((unsigned)((c) - '0') <= 9)"; 3631556Srgrimes for (pp = macro ; *pp ; pp++) 3641556Srgrimes fprintf(hfile, "%s\n", *pp); 3651556Srgrimes if (digit_contig) 3661556Srgrimes fputs("#define digit_val(c)\t((c) - '0')\n", hfile); 3671556Srgrimes else 3681556Srgrimes fputs("#define digit_val(c)\t(digit_value[c])\n", hfile); 3691556Srgrimes} 3701556Srgrimes 3711556Srgrimes 3721556Srgrimes 3731556Srgrimes/* 3741556Srgrimes * Output digit conversion table (if digits are not contiguous). 3751556Srgrimes */ 3761556Srgrimes 37717987Speterstatic void 37817987Speterdigit_convert() 37917987Speter{ 3801556Srgrimes int maxdigit; 3811556Srgrimes static char digit[] = "0123456789"; 3821556Srgrimes char *p; 3831556Srgrimes int i; 3841556Srgrimes 3851556Srgrimes maxdigit = 0; 3861556Srgrimes for (p = digit ; *p ; p++) 3871556Srgrimes if (*p > maxdigit) 3881556Srgrimes maxdigit = *p; 3891556Srgrimes fputs("extern const char digit_value[];\n", hfile); 3901556Srgrimes fputs("\n\nconst char digit_value[] = {\n", cfile); 3911556Srgrimes for (i = 0 ; i <= maxdigit ; i++) { 3921556Srgrimes for (p = digit ; *p && *p != i ; p++); 3931556Srgrimes if (*p == '\0') 3941556Srgrimes p = digit; 3951556Srgrimes fprintf(cfile, " %d,\n", p - digit); 3961556Srgrimes } 3971556Srgrimes fputs("};\n", cfile); 3981556Srgrimes} 399