%{ /* ** Copyright Heikki Suonsivu 1989 ** ** Kent, you owe me a beer, if you use it and meet me some day. No other ** limitations. You can use, sell, put your name in it, print it out and ** eat the printout, or use it any other way. I take no responsibility ** for the consequences. ** ** This software is Copyright (c) 1989, 1990, 1991 by Kent Landfield. ** ** Permission is hereby granted to copy, distribute or otherwise ** use any part of this package as long as you do not try to make ** money from it or pretend that you wrote it. This copyright ** notice must be maintained in any copy made. ** */ #if !defined(lint) && !defined(SABER) static char SID[] = "@(#)match.y 2.2 2/23/91"; #endif #define YYDEBUG 1 #include #include #include #include "rkive.h" #include "article.h" extern FILE *logfp; extern FILE *errfp; extern int yydebug; int parser_return_value; %} %union { int integer; char *string; char *header; /* Must be different for parser, no "" around header names */ } %token STRING %token
HEADER %left OR AND %left NOT %left INCLUDES GLOB_MATCHES %type exp /* Tries to understand some expressions. */ %% expression: /* Empty */ { return 1; } /* Empty always matches */ | exp { parser_return_value = $1; } exp: exp OR exp { $$ = $1 || $3; } | exp AND exp { $$ = $1 && $3; } | NOT exp { $$ = ! $2; } | '(' exp ')' { $$ = $2; } | HEADER INCLUDES STRING { $$ = substr($1, $3); } | STRING INCLUDES HEADER { $$ = substr($1, $3); } | HEADER GLOB_MATCHES STRING { $$ = match_str($1, $3); } %% typedef struct { char *keyword; int key; } KEYWORD; KEYWORD keys[] = { "and", AND, "or", OR, "includes", INCLUDES, "glob-matches", GLOB_MATCHES, "not", NOT, NULL, 0 }; KEYWORD headers[] = { "from", FROM, "path", PATH, "newsgroup", NEWSGROUP, "subject", SUBJECT, "message-id", MSG_ID, "reply-to", REPLY_TO, "references", REFERENCES, "date", DATE, "expire", EXPIRE, "control", CONTROL, "sender", SENDER, "followup-to", FOLLOWUP_TO, "distribution", DISTRIBUTION, "organization", ORGANIZATION, "numlines", NUMLINES, "keywords", KEYWORDS, "summary", SUMMARY, "approved", APPROVED, "supersedes", SUPERSEDES, "xref", XREF, "posting-number", POSTING_NUMBER, "submitted-by", SUBMITTED_BY, "archive-name", ARCH_NAME, "articleid", ARTICLEID, "patch-to", PATCH_TO, "environment", ENVIRONMENT, NULL, 0 }; /* Global, copy matching-expression here, then call yyparse */ char inputstring[MAXMATCHLEN]; /* Parser for expression */ yylex() { static char input_buffer[MAXMATCHLEN]; static char string_buffer[MAXMATCHLEN]; static char *p; register char *s, *string; register KEYWORD *keyword; if (*inputstring) { if (strlen(inputstring) >= MAXMATCHLEN) { (void) fprintf(logfp, "Match string too long, max %d characters", MAXMATCHLEN); exit (1); } (void) strcpy(input_buffer, inputstring); p = input_buffer; *inputstring = 0; /* We got it */ } /* Skip whitespace separating tokens */ while (*p && isspace(*p)) p++; if (!*p) { if (yydebug) (void) fprintf(logfp, "yylex: return eof\n"); return 0; /* Eof */ } if (*p == '"') { string = string_buffer; /* Collect the string, try to be intelligent with escaped '"'. ** Trailing " is not needed. */ for (p++; *p && *p != '"'; p++) { if (*p == '\\') { *string++ = *p++; if (*p) *string++ = *p; /* Skip the next char, whatever it is */ } else *string++ = *p; } p++; /* Disgard trailing '"' */ *string = 0; yylval.string = string_buffer; if (yydebug) (void) fprintf(logfp, "yylex: return string <%s>\n", string_buffer); return STRING; } else if (isalpha(*p)) { string = string_buffer; /* Collect keyword/header */ for (; *p && (isalpha(*p) || !isspace(*p)); p++) *string++ = isupper(*p) ? tolower(*p) : *p; *string = 0; /* Any of the operands? */ for (keyword = keys; keyword->keyword; keyword++) { if (!strcmp(keyword->keyword, string_buffer)) { if (yydebug) (void) fprintf(logfp, "yylex: return keyword %d (%s)\n", keyword->key, keyword->keyword); return keyword->key; } } /* A header? */ for (keyword = headers; keyword->keyword; keyword++) if (!strcmp(keyword->keyword, string_buffer)) { #ifdef NNTP if (! *header.header[keyword->key]) { if (nntp_retrieve_article() == ERROR_ENCOUNTERED) { fprintf(logfp, "Error retrieving NNTP article.\n"); return 0; } } #endif yylval.header = header.header[keyword->key]; if (yydebug) (void) fprintf(logfp, "yylex: return header %s <%s>\n", keyword->keyword, yylval.header); return HEADER; } /* Nope, maybe its a string without "". Note, it converts the string to lower-case. */ yylval.string = string_buffer; if (yydebug) (void) fprintf(logfp,"yylex: return unquoted string <%s>\n",string_buffer); return STRING; } else if (strchr("()", *p)) return *p++; else { if (yydebug) (void) fprintf(logfp, "yylex: Bad character '%c'\n", *p); return 0; /* Eof, could be something else, but I'm too tired now */ } /* Never here... (?) */ } yyerror(s) char *s; { (void) fprintf(errfp,"Cannot parse match pattern: %s\n",s); exit (1); } /* ** int match_str( char *string, char *format ) { return(0); } ** ** Search the "string" to see if it matches the format specified. ** ** The characters `[', `]', `*', `?', `^', `-', and `\' are considered ** special metacharacters and have the following meanings... ** `*' matches any set of characters, ** `?' matches any one character, ** [..] matches any character specified in the brackets, ** [^..] matches any character that is *not* specified in the brackets. ** The use of the brackets is suppose to work in much the same ** manner as the shell in that you can use [a-z] notation as well. ** ** If there is a need to match one of the special metacharacters characters ** the character must be prefixed with a backslash (`\') to negate its ** special meaning. ** ** This function returns 1 if there is a match and 0 otherwise. ** */ int match_str(string, format) char *string; char *format; { register char *s = string; register char *f = format; register char ch; register char nextch; register int negate; while ((ch = *f++)) { switch (ch) { case '*': /* ** Now let's match against what is left... */ while ((ch = *f++) == '?' || ch == '*') { if (ch == '?' && *s++ == 0) return 0; } if (ch == 0) /* at the end of the format ? */ return 1; if (ch == '\\') nextch = *f; else nextch = ch; for (;;) { if ((ch == '[' || *s == nextch)) { if (match_str (s, f - 1)) return 1; } if (*s++ == 0) return 0; } case '?': if (!*s) return 0; else ++s; break; case '\\': if (*f++ != *s++) return 0; break; case '[': nextch = *s++; if (*f == '^') { negate = 1; f++; } else negate = 0; for(ch = *f++;;) { register char sp = ch, rp = ch; sp = ch; rp = ch; if (ch == '\\') { sp = *f++; rp = sp; } if (!ch) return (0); ch = *f++; if (ch == '-') { rp = *f++; if (rp == '\\') rp = *f++; if (!rp) return (0); ch = *f++; } if (nextch >= sp && nextch <= rp) { /* ** Ignore the rest of [...] previously matched. */ while (ch != ']') { if (!ch || !(ch = *f++)) return (0); if (ch == '\\') f++; } if (negate) return 0; } if (ch == ']') break; } if (!negate) return 0; break; default: if (ch != *s++) return 0; break; } } return(*s ? 0 : 1); }