1: #ifndef lint 2: static char *sccsid = "@(#)sortbib.c 4.1 (Berkeley) 5/6/83"; 3: #endif 4: 5: #include <stdio.h> 6: #include <signal.h> 7: #define BUF BUFSIZ 8: #define MXFILES 16 9: 10: char *tempfile; /* temporary file for sorting keys */ 11: char *keystr = "AD"; /* default sorting on author and date */ 12: int multauth = 0; /* by default sort on senior author only */ 13: int oneauth; /* has there been author in the record? */ 14: 15: main(argc, argv) /* sortbib: sort bibliographic database in place */ 16: int argc; 17: char *argv[]; 18: { 19: FILE *fp[MXFILES], *tfp, *fopen(); 20: int i, onintr(); 21: char *mktemp(); 22: 23: if (argc == 1) /* can't use stdin for seeking anyway */ 24: { 25: puts("Usage: sortbib [-sKEYS] database [...]"); 26: puts("\t-s: sort by fields in KEYS (default is AD)"); 27: exit(1); 28: } 29: if (argc > 2 && argv[1][0] == '-' && argv[1][1] == 's') 30: { 31: keystr = argv[1]+2; 32: eval(keystr); /* evaluate A+ for multiple authors */ 33: argv++; argc--; 34: } 35: if (argc > MXFILES+1) /* too many open file streams */ 36: { 37: fprintf(stderr, 38: "sortbib: More than %d databases specified\n", MXFILES); 39: exit(1); 40: } 41: for (i = 1; i < argc; i++) /* open files in arg list */ 42: if ((fp[i-1] = fopen(argv[i], "r")) == NULL) 43: error(argv[i]); 44: tempfile = "/tmp/SbibXXXXX"; /* tempfile for sorting keys */ 45: mktemp(tempfile); 46: if (signal(SIGINT,SIG_IGN) != SIG_IGN) /* remove if interrupted */ 47: signal(SIGINT, onintr); 48: if ((tfp = fopen(tempfile, "w")) == NULL) 49: error(tempfile); 50: for (i = 0; i < argc-1; i++) /* read keys from bib files */ 51: sortbib(fp[i], tfp, i); 52: fclose(tfp); 53: deliver(fp, tfp); /* do disk seeks and read from biblio files */ 54: unlink(tempfile); 55: exit(0); 56: } 57: 58: int rsmode = 0; /* record separator: 1 = null line, 2 = bracket */ 59: 60: sortbib(fp, tfp, i) /* read records, prepare list for sorting */ 61: FILE *fp, *tfp; 62: int i; 63: { 64: long offset, lastoffset = 0, ftell(); /* byte offsets in file */ 65: int length, newrec, recno = 0; /* reclen, new rec'd?, number */ 66: char line[BUF], fld[4][BUF]; /* one line, the sort fields */ 67: 68: /* measure byte offset, then get new line */ 69: while (offset = ftell(fp), fgets(line, BUF, fp)) 70: { 71: if (recno == 0) /* accept record w/o initial newline */ 72: newrec = 1; 73: if (line[0] == '\n') /* accept null line record separator */ 74: { 75: if (!rsmode) 76: rsmode = 1; /* null line mode */ 77: if (rsmode == 1) 78: newrec = 1; 79: } 80: if (line[0] == '.' && line[1] == '[') /* also accept .[ .] */ 81: { 82: if (!rsmode) 83: rsmode = 2; /* bracket pair mode */ 84: if (rsmode == 2) 85: newrec = 1; 86: } 87: if (newrec) /* by whatever means above */ 88: { 89: newrec = 0; 90: length = offset - lastoffset; /* measure rec len */ 91: if (length > BUF*8) 92: { 93: fprintf(stderr, 94: "sortbib: record %d longer than %d (%d)\n", 95: recno, BUF*8, length); 96: exit(1); 97: } 98: if (recno++) /* info for sorting */ 99: { 100: fprintf(tfp, "%d %D %d : %s %s %s %s\n", 101: i, lastoffset, length, 102: fld[0], fld[1], fld[2], fld[3]); 103: if (ferror(tfp)) 104: error(tempfile); 105: } 106: *fld[0] = *fld[1] = *fld[2] = *fld[3] = NULL; 107: oneauth = 0; /* reset number of authors */ 108: lastoffset = offset; /* save for next time */ 109: } 110: if (line[0] == '%') /* parse out fields to be sorted */ 111: parse(line, fld); 112: } 113: offset = ftell(fp); /* measure byte offset at EOF */ 114: length = offset - lastoffset; /* measure final record length */ 115: if (length > BUF*8) 116: { 117: fprintf(stderr, "sortbib: record %d longer than %d (%d)\n", 118: recno, BUF*8, length); 119: exit(1); 120: } 121: if (line[0] != '\n') /* ignore null line just before EOF */ 122: { 123: fprintf(tfp, "%d %D %d : %s %s %s %s\n", 124: i, lastoffset, length, 125: fld[0], fld[1], fld[2], fld[3]); 126: if (ferror(tfp)) 127: error(tempfile); /* disk error in /tmp */ 128: } 129: } 130: 131: deliver(fp, tfp) /* deliver sorted entries out of database(s) */ 132: FILE *fp[], *tfp; 133: { 134: char str[BUF], buff[BUF*8]; /* for tempfile & databases */ 135: char cmd[80], *sprintf(); /* for using system sort command */ 136: long int offset; 137: int i, length; 138: 139: /* when sorting, ignore case distinctions; tab char is ':' */ 140: sprintf(cmd, "sort -ft: +1 %s -o %s", tempfile, tempfile); 141: if (system(cmd) == 127) 142: error("sortbib"); 143: tfp = fopen(tempfile, "r"); 144: while (fgets(str, sizeof(str), tfp)) 145: { 146: /* get file pointer, record offset, and length */ 147: if (sscanf(str, "%d %D %d :", &i, &offset, &length) != 3) 148: error("sortbib: sorting error"); 149: /* seek to proper disk location in proper file */ 150: if (fseek(fp[i], offset, 0) == -1) 151: error("sortbib"); 152: /* read exactly one record from bibliography */ 153: if (fread(buff, sizeof(*buff), length, fp[i]) == 0) 154: error("sortbib"); 155: /* add newline between unseparated records */ 156: if (buff[0] != '\n' && rsmode == 1) 157: putchar('\n'); 158: /* write record buffer to standard output */ 159: if (fwrite(buff, sizeof(*buff), length, stdout) == 0) 160: error("sortbib"); 161: } 162: } 163: 164: parse(line, fld) /* get fields out of line, prepare for sorting */ 165: char line[]; 166: char fld[][BUF]; 167: { 168: char wd[8][BUF/4], *strcat(); 169: int n, i, j; 170: 171: for (i = 0; i < 8; i++) /* zap out old strings */ 172: *wd[i] = NULL; 173: n = sscanf(line, "%s %s %s %s %s %s %s %s", 174: wd[0], wd[1], wd[2], wd[3], wd[4], wd[5], wd[6], wd[7]); 175: for (i = 0; i < 4; i++) 176: { 177: if (wd[0][1] == keystr[i]) 178: { 179: if (wd[0][1] == 'A') 180: { 181: if (oneauth && !multauth) /* no repeat */ 182: break; 183: else if (oneauth) /* mult auths */ 184: strcat(fld[i], "~~"); 185: if (!endcomma(wd[n-2])) /* surname */ 186: strcat(fld[i], wd[n-1]); 187: else { /* jr. or ed. */ 188: strcat(fld[i], wd[n-2]); 189: n--; 190: } 191: strcat(fld[i], " "); 192: for (j = 1; j < n-1; j++) 193: strcat(fld[i], wd[j]); 194: oneauth = 1; 195: } 196: else if (wd[0][1] == 'D') 197: { 198: strcat(fld[i], wd[n-1]); /* year */ 199: if (n > 2) 200: strcat(fld[i], wd[1]); /* month */ 201: } 202: else if (wd[0][1] == 'T' || wd[0][1] == 'J') 203: { 204: j = 1; 205: if (article(wd[1])) /* skip article */ 206: j++; 207: for (; j < n; j++) 208: strcat(fld[i], wd[j]); 209: } 210: else /* any other field */ 211: for (j = 1; j < n; j++) 212: strcat(fld[i], wd[j]); 213: } 214: /* %Q quorporate or queer author - unreversed %A */ 215: else if (wd[0][1] == 'Q' && keystr[i] == 'A') 216: for (j = 1; j < n; j++) 217: strcat(fld[i], wd[j]); 218: } 219: } 220: 221: article(str) /* see if string contains an article */ 222: char *str; 223: { 224: if (strcmp("The", str) == 0) /* English */ 225: return(1); 226: if (strcmp("A", str) == 0) 227: return(1); 228: if (strcmp("An", str) == 0) 229: return(1); 230: if (strcmp("Le", str) == 0) /* French */ 231: return(1); 232: if (strcmp("La", str) == 0) 233: return(1); 234: if (strcmp("Der", str) == 0) /* German */ 235: return(1); 236: if (strcmp("Die", str) == 0) 237: return(1); 238: if (strcmp("Das", str) == 0) 239: return(1); 240: if (strcmp("El", str) == 0) /* Spanish */ 241: return(1); 242: if (strcmp("Den", str) == 0) /* Scandinavian */ 243: return(1); 244: return(0); 245: } 246: 247: eval(keystr) /* evaluate key string for A+ marking */ 248: char keystr[]; 249: { 250: int i, j; 251: 252: for (i = 0, j = 0; keystr[i]; i++, j++) 253: { 254: if (keystr[i] == '+') 255: { 256: multauth = 1; 257: i++; 258: } 259: keystr[j] = keystr[i]; 260: } 261: keystr[j] = NULL; 262: } 263: 264: error(s) /* exit in case of various system errors */ 265: char *s; 266: { 267: perror(s); 268: exit(1); 269: } 270: 271: onintr() /* remove tempfile in case of interrupt */ 272: { 273: fprintf(stderr, "\nInterrupt\n"); 274: unlink(tempfile); 275: exit(1); 276: } 277: 278: endcomma(str) 279: char *str; 280: { 281: int n; 282: 283: n = strlen(str) - 1; 284: if (str[n] == ',') 285: { 286: str[n] = NULL; 287: return(1); 288: } 289: return(0); 290: }