mike@ists.ists.ca (Mike Clarkson) (06/03/90)
#! /bin/sh
# This is a shell archive. Remove anything before this line, then unpack
# it by saving it into a file and typing "sh file". To overwrite existing
# files, type "sh file -c". You can also feed this as standard input via
# unshar, or by typing "sh <file", e.g.. If this archive is complete, you
# will see the following message at the end:
# "End of archive 8 (of 9)."
# Contents: manual/texindex.c
# Wrapped by mike@sam on Sat Jun 2 18:18:15 1990
PATH=/bin:/usr/bin:/usr/ucb ; export PATH
if test -f 'manual/texindex.c' -a "${1}" != "-c" ; then
echo shar: Will not clobber existing file \"'manual/texindex.c'\"
else
echo shar: Extracting \"'manual/texindex.c'\" \(42774 characters\)
sed "s/^X//" >'manual/texindex.c' <<'END_OF_FILE'
X/* Prepare Tex index dribble output into an actual index.
X Copyright (C) 1987 Free Software Foundation, Inc.
X
X NO WARRANTY
X
X BECAUSE THIS PROGRAM IS LICENSED FREE OF CHARGE, WE PROVIDE ABSOLUTELY
XNO WARRANTY, TO THE EXTENT PERMITTED BY APPLICABLE STATE LAW. EXCEPT
XWHEN OTHERWISE STATED IN WRITING, FREE SOFTWARE FOUNDATION, INC,
XRICHARD M. STALLMAN AND/OR OTHER PARTIES PROVIDE THIS PROGRAM "AS IS"
XWITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING,
XBUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
XFITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY
XAND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE PROGRAM PROVE
XDEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR
XCORRECTION.
X
X IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW WILL RICHARD M.
XSTALLMAN, THE FREE SOFTWARE FOUNDATION, INC., AND/OR ANY OTHER PARTY
XWHO MAY MODIFY AND REDISTRIBUTE THIS PROGRAM AS PERMITTED BELOW, BE
XLIABLE TO YOU FOR DAMAGES, INCLUDING ANY LOST PROFITS, LOST MONIES, OR
XOTHER SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
XUSE OR INABILITY TO USE (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR
XDATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY THIRD PARTIES OR
XA FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS) THIS
XPROGRAM, EVEN IF YOU HAVE BEEN ADVISED OF THE POSSIBILITY OF SUCH
XDAMAGES, OR FOR ANY CLAIM BY ANY OTHER PARTY.
X
X GENERAL PUBLIC LICENSE TO COPY
X
X 1. You may copy and distribute verbatim copies of this source file
Xas you receive it, in any medium, provided that you conspicuously
Xand appropriately publish on each copy a valid copyright notice
X"Copyright (C) 1987 Free Software Foundation, Inc.", and include
Xfollowing the copyright notice a verbatim copy of the above disclaimer
Xof warranty and of this License.
X
X 2. You may modify your copy or copies of this source file or
Xany portion of it, and copy and distribute such modifications under
Xthe terms of Paragraph 1 above, provided that you also do the following:
X
X a) cause the modified files to carry prominent notices stating
X that you changed the files and the date of any change; and
X
X b) cause the whole of any work that you distribute or publish,
X that in whole or in part contains or is a derivative of this
X program or any part thereof, to be licensed at no charge to all
X third parties on terms identical to those contained in this
X License Agreement (except that you may choose to grant more extensive
X warranty protection to some or all third parties, at your option).
X
X c) You may charge a distribution fee for the physical act of
X transferring a copy, and you may at your option offer warranty
X protection in exchange for a fee.
X
XMere aggregation of another unrelated program with this program (or its
Xderivative) on a volume of a storage or distribution medium does not bring
Xthe other program under the scope of these terms.
X
X 3. You may copy and distribute this program (or a portion or derivative
Xof it, under Paragraph 2) in object code or executable form under the terms
Xof Paragraphs 1 and 2 above provided that you also do one of the following:
X
X a) accompany it with the complete corresponding machine-readable
X source code, which must be distributed under the terms of
X Paragraphs 1 and 2 above; or,
X
X b) accompany it with a written offer, valid for at least three
X years, to give any third party free (except for a nominal
X shipping charge) a complete machine-readable copy of the
X corresponding source code, to be distributed under the terms of
X Paragraphs 1 and 2 above; or,
X
X c) accompany it with the information you received as to where the
X corresponding source code may be obtained. (This alternative is
X allowed only for noncommercial distribution and only if you
X received the program in object code or executable form alone.)
X
XFor an executable file, complete source code means all the source code for
Xall modules it contains; but, as a special exception, it need not include
Xsource code for modules which are standard libraries that accompany the
Xoperating system on which the executable file runs.
X
X 4. You may not copy, sublicense, distribute or transfer this program
Xexcept as expressly provided under this License Agreement. Any attempt
Xotherwise to copy, sublicense, distribute or transfer this program is void and
Xyour rights to use the program under this License agreement shall be
Xautomatically terminated. However, parties who have received computer
Xsoftware programs from you with this License Agreement will not have
Xtheir licenses terminated so long as such parties remain in full compliance.
X
X 5. If you wish to incorporate parts of this program into other free
Xprograms whose distribution conditions are different, write to the Free
XSoftware Foundation at 675 Mass Ave, Cambridge, MA 02139. We have not yet
Xworked out a simple rule that can be stated here, but we will often permit
Xthis. We will be guided by the two goals of preserving the free status of
Xall derivatives of our free software and of promoting the sharing and reuse of
Xsoftware.
X
X In other words, you are welcome to use, share and improve this program.
X You are forbidden to forbid anyone else to use, share and improve
X what you give them. Help stamp out software-hoarding! */
X
X
X#include <stdio.h>
X#include <ctype.h>
X
X#ifdef VMS
X#include <file.h>
X
X#define EXIT_SUCCESS ((1 << 28) | 1)
X#define EXIT_FATAL ((1 << 28) | 4)
X#define unlink delete
X#define tell(fd) lseek(fd, 0L, 1)
X#else
X#include <sys/file.h>
X
X#define EXIT_SUCCESS 0
X#define EXIT_FATAL 1
X#endif
X
X
X#ifndef L_XTND
X#define L_XTND 2
X#endif
X
X/* When sorting in core, this structure describes one line
X and the position and length of its first keyfield. */
X
Xstruct lineinfo
X {
X char *text; /* The actual text of the line */
X union
X { /* The start of the key (for textual comparison) */
X char *text;
X long number; /* or the numeric value (for numeric comparison) */
X } key;
X long keylen; /* Length of key field */
X };
X
X/* This structure describes a field to use as a sort key */
X
Xstruct keyfield
X {
X int startwords; /* # words to skip */
X int startchars; /* and # additional chars to skip, to start of field */
X int endwords; /* similar, from beg (or end) of line, to find end of field */
X int endchars;
X char ignore_blanks; /* Ignore spaces and tabs within the field */
X char fold_case; /* Convert upper case to lower before comparing */
X char reverse; /* Compare in reverse order */
X char numeric; /* Parse text as an integer and compare the integers */
X char positional; /* Sort according to position within the file */
X char braced; /* Count balanced-braced groupings as fields */
X };
X
X/* Vector of keyfields to use */
X
Xstruct keyfield keyfields[3];
X
X/* Number of keyfields stored in that vector. */
X
Xint num_keyfields = 3;
X
X/* Vector of input file names, terminated with a zero (null pointer) */
X
Xchar **infiles;
X
X/* Vector of corresponding output file names, or zero meaning default it */
X
Xchar **outfiles;
X
X/* Length of `infiles' */
X
Xint num_infiles;
X
X/* Pointer to the array of pointers to lines being sorted */
X
Xchar **linearray;
X
X/* The allocated length of `linearray'. */
X
Xlong lines;
X
X/* Directory to use for temporary files. On Unix, it ends with a slash. */
X
Xchar *tempdir;
X
X/* Start of filename to use for temporary files. */
X
Xchar *tempbase;
X
X/* Number of last temporary file. */
X
Xint tempcount;
X
X/* Number of last temporary file already deleted.
X Temporary files are deleted by `flush_tempfiles' in order of creation. */
X
Xint last_deleted_tempcount;
X
X/* During in-core sort, this points to the base of the data block
X which contains all the lines of data. */
X
Xchar *text_base;
X
X/* Additional command switches */
X
Xint keep_tempfiles; /* Nonzero means do not delete tempfiles -- for debugging */
X
X/* Forward declarations of functions in this file */
X
Xvoid decode_command ();
Xvoid sort_in_core ();
Xvoid sort_offline ();
Xchar **parsefile ();
Xchar *find_field ();
Xchar *find_pos ();
Xlong find_value ();
Xchar *find_braced_pos ();
Xchar *find_braced_end ();
Xvoid writelines ();
Xint compare_full ();
Xlong readline ();
Xint merge_files ();
Xint merge_direct ();
Xchar *concat ();
Xchar *maketempname ();
Xvoid flush_tempfiles ();
Xchar *tempcopy ();
X
Xextern char *mktemp ();
X
X#define MAX_IN_CORE_SORT 500000
X
Xint
Xmain (argc, argv)
X int argc;
X char **argv;
X{
X int i;
X
X tempcount = 0;
X last_deleted_tempcount = 0;
X
X /* Describe the kind of sorting to do. */
X /* The first keyfield uses the first braced field and folds case */
X keyfields[0].braced = 1;
X keyfields[0].fold_case = 1;
X keyfields[0].endwords = -1;
X keyfields[0].endchars = -1;
X /* The second keyfield uses the second braced field, numerically */
X keyfields[1].braced = 1;
X keyfields[1].numeric = 1;
X keyfields[1].startwords = 1;
X keyfields[1].endwords = -1;
X keyfields[1].endchars = -1;
X /* The third keyfield (which is ignored while discarding duplicates)
X compares the whole line */
X keyfields[2].endwords = -1;
X keyfields[2].endchars = -1;
X
X decode_command (argc, argv);
X
X tempbase = mktemp (concat ("txiXXXXXX", "", ""));
X
X /* Process input files completely, one by one. */
X
X for (i = 0; i < num_infiles; i++)
X {
X int desc;
X long ptr;
X char *outfile;
X char *p;
X
X desc = open (infiles[i], 0, 0);
X if (desc < 0) pfatal_with_name (infiles[i]);
X lseek (desc, 0, L_XTND);
X ptr = tell (desc);
X close (desc);
X
X outfile = outfiles[i];
X if (!outfile)
X {
X outfile = concat (infiles[i], "s", "");
X }
X
X if (ptr < MAX_IN_CORE_SORT)
X /* Sort a small amount of data */
X sort_in_core (infiles[i], ptr, outfile);
X else
X sort_offline (infiles[i], ptr, outfile);
X }
X
X flush_tempfiles (tempcount);
X exit (EXIT_SUCCESS);
X}
X
X/* This page decodes the command line arguments to set the parameter variables
X and set up the vector of keyfields and the vector of input files */
X
Xvoid
Xdecode_command (argc, argv)
X int argc;
X char **argv;
X{
X int i;
X char **ip;
X char **op;
X
X /* Store default values into parameter variables */
X
X#ifdef VMS
X tempdir = "sys$scratch:";
X#else
X tempdir = "/tmp/";
X#endif
X
X keep_tempfiles = 0;
X
X /* Allocate argc input files, which must be enough. */
X
X infiles = (char **) xmalloc (argc * sizeof (char *));
X outfiles = (char **) xmalloc (argc * sizeof (char *));
X ip = infiles;
X op = outfiles;
X
X /* First find all switches that control the default kind-of-sort */
X
X for (i = 1; i < argc; i++)
X {
X int tem = classify_arg (argv[i]);
X char c;
X char *p;
X
X if (tem <= 0)
X {
X *ip++ = argv[i];
X *op++ = 0;
X continue;
X }
X if (tem > 1)
X {
X if (i + 1 == argc)
X fatal ("switch %s given with no argument following it", argv[i]);
X else if (!strcmp (argv[i], "-T"))
X tempdir = argv[i + 1];
X else if (!strcmp (argv[i], "-o"))
X *(op - 1) = argv[i + 1];
X i += tem - 1;
X continue;
X }
X
X p = &argv[i][1];
X while (c = *p++)
X switch (c)
X {
X case 'k':
X keep_tempfiles = 1;
X break;
X
X default:
X fatal ("invalid command switch %c", c);
X }
X switchdone: ;
X }
X
X /* Record number of keyfields, terminate list of filenames */
X
X num_infiles = ip - infiles;
X *ip = 0;
X}
X
X/* Return 0 for an argument that is not a switch;
X for a switch, return 1 plus the number of following arguments that the switch swallows.
X*/
X
Xint
Xclassify_arg (arg)
X char *arg;
X{
X if (!strcmp (arg, "-T") || !strcmp (arg, "-o"))
X return 2;
X if (arg[0] == '-')
X return 1;
X return 0;
X}
X
X/* Create a name for a temporary file */
X
Xchar *
Xmaketempname (count)
X int count;
X{
X char tempsuffix[10];
X sprintf (tempsuffix, "%d", count);
X return concat (tempdir, tempbase, tempsuffix);
X}
X
X/* Delete all temporary files up to the specified count */
X
Xvoid
Xflush_tempfiles (to_count)
X int to_count;
X{
X if (keep_tempfiles) return;
X while (last_deleted_tempcount < to_count)
X unlink (maketempname (++last_deleted_tempcount));
X}
X
X/* Copy an input file into a temporary file, and return the temporary file name */
X
X#define BUFSIZE 1024
X
Xchar *
Xtempcopy (idesc)
X int idesc;
X{
X char *outfile = maketempname (++tempcount);
X int odesc;
X char buffer[BUFSIZE];
X
X odesc = open (outfile, O_WRONLY | O_CREAT, 0666);
X
X if (odesc < 0) pfatal_with_name (outfile);
X
X while (1)
X {
X int nread = read (idesc, buffer, BUFSIZE);
X write (odesc, buffer, nread);
X if (!nread) break;
X }
X
X close (odesc);
X
X return outfile;
X}
X
X/* Compare two lines, provided as pointers to pointers to text,
X according to the specified set of keyfields */
X
Xint
Xcompare_full (line1, line2)
X char **line1, **line2;
X{
X int i;
X
X /* Compare using the first keyfield;
X if that does not distinguish the lines, try the second keyfield; and so on. */
X
X for (i = 0; i < num_keyfields; i++)
X {
X long length1, length2;
X char *start1 = find_field (&keyfields[i], *line1, &length1);
X char *start2 = find_field (&keyfields[i], *line2, &length2);
X int tem = compare_field (&keyfields[i], start1, length1, *line1 - text_base,
X start2, length2, *line2 - text_base);
X if (tem)
X {
X if (keyfields[i].reverse)
X return - tem;
X return tem;
X }
X }
X
X return 0; /* Lines match exactly */
X}
X
X/* Compare two lines described by structures
X in which the first keyfield is identified in advance.
X For positional sorting, assumes that the order of the lines in core
X reflects their nominal order. */
X
Xint
Xcompare_prepared (line1, line2)
X struct lineinfo *line1, *line2;
X{
X int i;
X int tem;
X char *text1, *text2;
X
X /* Compare using the first keyfield, which has been found for us already */
X if (keyfields->positional)
X {
X if (line1->text - text_base > line2->text - text_base)
X tem = 1;
X else
X tem = -1;
X }
X else if (keyfields->numeric)
X tem = line1->key.number - line2->key.number;
X else
X tem = compare_field (keyfields, line1->key.text, line1->keylen, 0, line2->key.text, line2->keylen, 0);
X if (tem)
X {
X if (keyfields->reverse)
X return - tem;
X return tem;
X }
X
X text1 = line1->text;
X text2 = line2->text;
X
X /* Compare using the second keyfield;
X if that does not distinguish the lines, try the third keyfield; and so on. */
X
X for (i = 1; i < num_keyfields; i++)
X {
X long length1, length2;
X char *start1 = find_field (&keyfields[i], text1, &length1);
X char *start2 = find_field (&keyfields[i], text2, &length2);
X int tem = compare_field (&keyfields[i], start1, length1, text1 - text_base,
X start2, length2, text2 - text_base);
X if (tem)
X {
X if (keyfields[i].reverse)
X return - tem;
X return tem;
X }
X }
X
X return 0; /* Lines match exactly */
X}
X
X/* Like compare_full but more general.
X You can pass any strings, and you can say how many keyfields to use.
X `pos1' and `pos2' should indicate the nominal positional ordering of
X the two lines in the input. */
X
Xint
Xcompare_general (str1, str2, pos1, pos2, use_keyfields)
X char *str1, *str2;
X long pos1, pos2;
X int use_keyfields;
X{
X int i;
X
X /* Compare using the first keyfield;
X if that does not distinguish the lines, try the second keyfield; and so on. */
X
X for (i = 0; i < use_keyfields; i++)
X {
X long length1, length2;
X char *start1 = find_field (&keyfields[i], str1, &length1);
X char *start2 = find_field (&keyfields[i], str2, &length2);
X int tem = compare_field (&keyfields[i], start1, length1, pos1, start2, length2, pos2);
X if (tem)
X {
X if (keyfields[i].reverse)
X return - tem;
X return tem;
X }
X }
X
X return 0; /* Lines match exactly */
X}
X
X/* Find the start and length of a field in `str' according to `keyfield'.
X A pointer to the starting character is returned, and the length
X is stored into the int that `lengthptr' points to. */
X
Xchar *
Xfind_field (keyfield, str, lengthptr)
X struct keyfield *keyfield;
X char *str;
X long *lengthptr;
X{
X char *start;
X char *end;
X char *(*fun) ();
X
X if (keyfield->braced) fun = find_braced_pos;
X else fun = find_pos;
X
X start = ( *fun )(str, keyfield->startwords, keyfield->startchars,
X keyfield->ignore_blanks);
X if (keyfield->endwords < 0)
X {
X if (keyfield->braced)
X end = find_braced_end (start);
X else
X {
X end = start;
X while (*end && *end != '\n') end++;
X }
X }
X else
X {
X end = ( *fun )(str, keyfield->endwords, keyfield->endchars, 0);
X if (end - str < start - str) end = start;
X }
X *lengthptr = end - start;
X return start;
X}
X
X/* Find a pointer to a specified place within `str',
X skipping (from the beginning) `words' words and then `chars' chars.
X If `ignore_blanks' is nonzero, we skip all blanks
X after finding the specified word. */
X
Xchar *
Xfind_pos (str, words, chars, ignore_blanks)
X char *str;
X int words, chars;
X int ignore_blanks;
X{
X int i;
X char *p = str;
X
X for (i = 0; i < words; i++)
X {
X char c;
X /* Find next bunch of nonblanks and skip them. */
X while ((c = *p) == ' ' || c == '\t') p++;
X while ((c = *p) && c != '\n' && !(c == ' ' || c == '\t')) p++;
X if (!*p || *p == '\n') return p;
X }
X
X while (*p == ' ' || *p == '\t') p++;
X
X for (i = 0; i < chars; i++)
X {
X if (!*p || *p == '\n') break;
X p++;
X }
X return p;
X}
X
X/* Like find_pos but assumes that each field is surrounded by braces
X and that braces within fields are balanced. */
X
Xchar *
Xfind_braced_pos (str, words, chars, ignore_blanks)
X char *str;
X int words, chars;
X int ignore_blanks;
X{
X int i;
X int bracelevel;
X char *p = str;
X char c;
X
X for (i = 0; i < words; i++)
X {
X bracelevel = 1;
X while ((c = *p++) != '{' && c != '\n' && c);
X if (c != '{')
X return p - 1;
X while (bracelevel)
X {
X c = *p++;
X if (c == '{') bracelevel++;
X if (c == '}') bracelevel--;
X if (c == '\\') c = *p++; /* \ quotes braces and \ */
X if (c == 0 || c == '\n') return p-1;
X }
X }
X
X while ((c = *p++) != '{' && c != '\n' && c);
X
X if (c != '{')
X return p-1;
X
X if (ignore_blanks)
X while ((c = *p) == ' ' || c == '\t') p++;
X
X for (i = 0; i < chars; i++)
X {
X if (!*p || *p == '\n') break;
X p++;
X }
X return p;
X}
X
X/* Find the end of the balanced-brace field which starts at `str'.
X The position returned is just before the closing brace. */
X
Xchar *
Xfind_braced_end (str)
X char *str;
X{
X int bracelevel;
X char *p = str;
X char c;
X
X bracelevel = 1;
X while (bracelevel)
X {
X c = *p++;
X if (c == '{') bracelevel++;
X if (c == '}') bracelevel--;
X if (c == '\\') c = *p++;
X if (c == 0 || c == '\n') return p-1;
X }
X return p - 1;
X}
X
Xlong
Xfind_value (start, length)
X char *start;
X long length;
X{
X while (length != 0L) {
X if (isdigit(*start))
X return atol(start);
X length--;
X start++;
X }
X return 0l;
X}
X
X/* Vector used to translate characters for comparison.
X This is how we make all alphanumerics follow all else,
X and ignore case in the first sorting. */
Xint char_order[256];
X
Xinit_char_order ()
X{
X int i;
X for (i = 1; i < 256; i++)
X char_order[i] = i;
X
X for (i = '0'; i <= '9'; i++)
X char_order[i] += 512;
X
X for (i = 'a'; i <= 'z'; i++) {
X char_order[i] = 512 + i;
X char_order[i + 'A' - 'a'] = 512 + i;
X }
X}
X
X/* Compare two fields (each specified as a start pointer and a character count)
X according to `keyfield'. The sign of the value reports the relation between the fields */
X
Xint
Xcompare_field (keyfield, start1, length1, pos1, start2, length2, pos2)
X struct keyfield *keyfield;
X char *start1;
X long length1;
X long pos1;
X char *start2;
X long length2;
X long pos2;
X{
X if (keyfields->positional)
X {
X if (pos1 > pos2)
X return 1;
X else
X return -1;
X }
X if (keyfield->numeric)
X {
X long value = find_value (start1, length1) - find_value (start2, length2);
X if (value > 0) return 1;
X if (value < 0) return -1;
X return 0;
X }
X else
X {
X char *p1 = start1;
X char *p2 = start2;
X char *e1 = start1 + length1;
X char *e2 = start2 + length2;
X
X int fold_case = keyfield->fold_case;
X
X while (1)
X {
X int c1, c2;
X
X if (p1 == e1) c1 = 0;
X else c1 = *p1++;
X if (p2 == e2) c2 = 0;
X else c2 = *p2++;
X
X if (char_order[c1] != char_order[c2])
X return char_order[c1] - char_order[c2];
X if (!c1) break;
X }
X
X /* Strings are equal except possibly for case. */
X p1 = start1;
X p2 = start2;
X while (1)
X {
X int c1, c2;
X
X if (p1 == e1) c1 = 0;
X else c1 = *p1++;
X if (p2 == e2) c2 = 0;
X else c2 = *p2++;
X
X if (c1 != c2)
X /* Reverse sign here so upper case comes out last. */
X return c2 - c1;
X if (!c1) break;
X }
X
X return 0;
X }
X}
X
X/* A `struct linebuffer' is a structure which holds a line of text.
X `readline' reads a line from a stream into a linebuffer
X and works regardless of the length of the line. */
X
Xstruct linebuffer
X {
X long size;
X char *buffer;
X };
X
X/* Initialize a linebuffer for use */
X
Xvoid
Xinitbuffer (linebuffer)
X struct linebuffer *linebuffer;
X{
X linebuffer->size = 200;
X linebuffer->buffer = (char *) xmalloc (200);
X}
X
X/* Read a line of text from `stream' into `linebuffer'.
X Return the length of the line. */
X
Xlong
Xreadline (linebuffer, stream)
X struct linebuffer *linebuffer;
X FILE *stream;
X{
X char *buffer = linebuffer->buffer;
X char *p = linebuffer->buffer;
X char *end = p + linebuffer->size;
X
X while (1)
X {
X int c = getc (stream);
X if (p == end)
X {
X buffer = (char *) xrealloc (buffer, linebuffer->size *= 2);
X p += buffer - linebuffer->buffer;
X end += buffer - linebuffer->buffer;
X linebuffer->buffer = buffer;
X }
X if (c < 0 || c == '\n')
X {
X *p = 0;
X break;
X }
X *p++ = c;
X }
X
X return p - buffer;
X}
X
X/* Sort an input file too big to sort in core. */
X
Xvoid
Xsort_offline (infile, nfiles, total, outfile)
X char *infile;
X long total;
X char *outfile;
X{
X int ntemps = 2 * (total + MAX_IN_CORE_SORT - 1) / MAX_IN_CORE_SORT; /* More than enough */
X char **tempfiles = (char **) xmalloc (ntemps * sizeof (char *));
X FILE *istream = fopen (infile, "r");
X int i;
X struct linebuffer lb;
X long linelength;
X int failure = 0;
X
X initbuffer (&lb);
X
X /* Read in one line of input data. */
X
X linelength = readline (&lb, istream);
X
X if (lb.buffer[0] != '\\')
X {
X error ("%s: not a texinfo index file", infile);
X return;
X }
X
X /* Split up the input into `ntemps' temporary files, or maybe fewer,
X and put the new files' names into `tempfiles' */
X
X for (i = 0; i < ntemps; i++)
X {
X char *outname = maketempname (++tempcount);
X FILE *ostream = fopen (outname, "w");
X long tempsize = 0;
X
X if (!ostream) pfatal_with_name (outname);
X tempfiles[i] = outname;
X
X /* Copy lines into this temp file as long as it does not make file "too big"
X or until there are no more lines. */
X
X while (tempsize + linelength + 1 <= MAX_IN_CORE_SORT)
X {
X tempsize += linelength + 1;
X fputs (lb.buffer, ostream);
X putc ('\n', ostream);
X
X /* Read another line of input data. */
X
X linelength = readline (&lb, istream);
X if (!linelength && feof (istream)) break;
X
X if (lb.buffer[0] != '\\')
X {
X error ("%s: not a texinfo index file", infile);
X failure = 1;
X goto fail;
X }
X }
X fclose (ostream);
X if (feof (istream)) break;
X }
X
X free (lb.buffer);
X
X fail:
X /* Record number of temp files we actually needed. */
X
X ntemps = i;
X
X /* Sort each tempfile into another tempfile.
X Delete the first set of tempfiles and put the names of the second into `tempfiles' */
X
X for (i = 0; i < ntemps; i++)
X {
X char *newtemp = maketempname (++tempcount);
X sort_in_core (&tempfiles[i], MAX_IN_CORE_SORT, newtemp);
X if (!keep_tempfiles)
X unlink (tempfiles[i]);
X tempfiles[i] = newtemp;
X }
X
X if (failure)
X return;
X
X /* Merge the tempfiles together and indexify */
X
X merge_files (tempfiles, ntemps, outfile);
X}
X
X/* Sort `infile', whose size is `total',
X assuming that is small enough to be done in-core,
X then indexify it and send the output to `outfile' (or to stdout). */
X
Xvoid
Xsort_in_core (infile, total, outfile)
X char *infile;
X long total;
X char *outfile;
X{
X char **nextline;
X char *data = (char *) xmalloc (total + 1);
X char *file_data;
X long file_size;
X int i;
X FILE *ostream = stdout;
X struct lineinfo *lineinfo;
X
X /* Read the contents of the file into the moby array `data' */
X
X int desc = open (infile, 0, 0);
X
X if (desc < 0)
X fatal ("failure reopening %s", infile);
X for (file_size = 0; ; )
X {
X if ((i = read (desc, data + file_size, total - file_size)) <= 0)
X break;
X file_size += i;
X }
X file_data = data;
X data[file_size] = 0;
X
X close (desc);
X
X if (file_size > 0 && data[0] != '\\')
X {
X error ("%s: not a texinfo index file", infile);
X return;
X }
X
X init_char_order ();
X
X /* Sort routines want to know this address */
X
X text_base = data;
X
X /* Create the array of pointers to lines, with a default size frequently enough. */
X
X lines = total / 50;
X if (!lines) lines = 2;
X linearray = (char **) xmalloc (lines * sizeof (char *));
X
X /* `nextline' points to the next free slot in this array.
X `lines' is the allocated size. */
X
X nextline = linearray;
X
X /* Parse the input file's data, and make entries for the lines. */
X
X nextline = parsefile (infile, nextline, file_data, file_size);
X if (nextline == 0)
X {
X error ("%s: not a texinfo index file", infile);
X return;
X }
X
X /* Sort the lines */
X
X /* If we have enough space, find the first keyfield of each line in advance.
X Make a `struct lineinfo' for each line, which records the keyfield
X as well as the line, and sort them. */
X
X lineinfo = (struct lineinfo *) malloc ((nextline - linearray) * sizeof (struct lineinfo));
X
X if (lineinfo)
X {
X struct lineinfo *lp;
X char **p;
X
X for (lp = lineinfo, p = linearray; p != nextline; lp++, p++)
X {
X lp->text = *p;
X lp->key.text = find_field (keyfields, *p, &lp->keylen);
X if (keyfields->numeric)
X lp->key.number = find_value (lp->key.text, lp->keylen);
X }
X
X qsort (lineinfo, nextline - linearray, sizeof (struct lineinfo), compare_prepared);
X
X for (lp = lineinfo, p = linearray; p != nextline; lp++, p++)
X *p = lp->text;
X
X free (lineinfo);
X }
X else
X qsort (linearray, nextline - linearray, sizeof (char *), compare_full);
X
X /* Open the output file */
X
X if (outfile)
X {
X ostream = fopen (outfile, "w");
X if (!ostream)
X pfatal_with_name (outfile);
X }
X
X writelines (linearray, nextline - linearray, ostream);
X if (outfile) fclose (ostream);
X
X free (linearray);
X free (data);
X}
X
X/* Parse an input string in core into lines.
X DATA is the input string, and SIZE is its length.
X Data goes in LINEARRAY starting at NEXTLINE.
X The value returned is the first entry in LINEARRAY still unused.
X Value 0 means input file contents are invalid. */
X
Xchar **
Xparsefile (filename, nextline, data, size)
X char *filename;
X char **nextline;
X char *data;
X long size;
X{
X char *p, *end;
X char **line = nextline;
X
X p = data;
X end = p + size;
X *end = 0;
X
X while (p != end)
X {
X if (p[0] != '\\')
X return 0;
X
X *line = p;
X while (*p && *p != '\n') p++;
X if (p != end) p++;
X
X line++;
X if (line == linearray + lines)
X {
X char **old = linearray;
X linearray = (char **) xrealloc (linearray, sizeof (char *) * (lines *= 4));
X line += linearray - old;
X }
X }
X
X return line;
X}
X
X/* Indexification is a filter applied to the sorted lines
X as they are being written to the output file.
X Multiple entries for the same name, with different page numbers,
X get combined into a single entry with multiple page numbers.
X The first braced field, which is used for sorting, is discarded.
X However, its first character is examined, folded to lower case,
X and if it is different from that in the previous line fed to us
X a \initial line is written with one argument, the new initial.
X
X If an entry has four braced fields, then the second and third
X constitute primary and secondary names.
X In this case, each change of primary name
X generates a \primary line which contains only the primary name,
X and in between these are \secondary lines which contain
X just a secondary name and page numbers.
X*/
X
X/* The last primary name we wrote a \primary entry for.
X If only one level of indexing is being done, this is the last name seen */
Xchar *lastprimary;
Xint lastprimarylength; /* Length of storage allocated for lastprimary */
X
X/* Similar, for the secondary name. */
Xchar *lastsecondary;
Xint lastsecondarylength;
X
X/* Zero if we are not in the middle of writing an entry.
X One if we have written the beginning of an entry but have not
X yet written any page numbers into it.
X Greater than one if we have written the beginning of an entry
X plus at least one page number. */
Xint pending;
X
X/* The initial (for sorting purposes) of the last primary entry written.
X When this changes, a \initial {c} line is written */
X
Xchar * lastinitial;
X
Xint lastinitiallength;
X
X/* When we need a string of length 1 for the value of lastinitial,
X store it here. */
X
Xchar lastinitial1[2];
X
X/* Initialize static storage for writing an index */
X
Xvoid
Xinit_index ()
X{
X pending = 0;
X lastinitial = lastinitial1;
X lastinitial1[0] = 0;
X lastinitial1[1] = 0;
X lastinitiallength = 0;
X lastprimarylength = 100;
X lastprimary = (char *) xmalloc (lastprimarylength + 1);
X bzero (lastprimary, lastprimarylength + 1);
X lastsecondarylength = 100;
X lastsecondary = (char *) xmalloc (lastsecondarylength + 1);
X bzero (lastsecondary, lastsecondarylength + 1);
X}
X
X/* Indexify. Merge entries for the same name,
X insert headers for each initial character, etc. */
X
Xindexify (line, ostream)
X char *line;
X FILE *ostream;
X{
X char *primary, *secondary, *pagenumber;
X int primarylength, secondarylength, pagelength;
X int len = strlen (line);
X int nosecondary;
X int initiallength;
X char *initial;
X char initial1[2];
X register char *p;
X
X /* First, analyze the parts of the entry fed to us this time */
X
X p = find_braced_pos (line, 0, 0, 0);
X if (*p == '{')
X {
X initial = p;
X /* Get length of inner pair of braces starting at p,
X including that inner pair of braces. */
X initiallength = find_braced_end (p + 1) + 1 - p;
X }
X else
X {
X initial = initial1;
X initial1[0] = *p;
X initial1[1] = 0;
X initiallength = 1;
X
X if (initial1[0] >= 'a' && initial1[0] <= 'z')
X initial1[0] -= 040;
X }
X
X pagenumber = find_braced_pos (line, 1, 0, 0);
X pagelength = find_braced_end (pagenumber) - pagenumber;
X if (pagelength == 0)
X abort ();
X
X primary = find_braced_pos (line, 2, 0, 0);
X primarylength = find_braced_end (primary) - primary;
X
X secondary = find_braced_pos (line, 3, 0, 0);
X nosecondary = !*secondary;
X if (!nosecondary)
X secondarylength = find_braced_end (secondary) - secondary;
X
X /* If the primary is different from before, make a new primary entry */
X if (strncmp (primary, lastprimary, primarylength))
X {
X /* Close off current secondary entry first, if one is open */
X if (pending)
X {
X fputs ("}\n", ostream);
X pending = 0;
X }
X
X /* If this primary has a different initial, include an entry for the initial */
X if (initiallength != lastinitiallength ||
X strncmp (initial, lastinitial, initiallength))
X {
X fprintf (ostream, "\\initial {");
X fwrite (initial, 1, initiallength, ostream);
X fprintf (ostream, "}\n", initial);
X if (initial == initial1)
X {
X lastinitial = lastinitial1;
X *lastinitial1 = *initial1;
X }
X else
X {
X lastinitial = initial;
X }
X lastinitiallength = initiallength;
X }
X
X /* Make the entry for the primary. */
X if (nosecondary)
X fputs ("\\entry {", ostream);
X else
X fputs ("\\primary {", ostream);
X fwrite (primary, primarylength, 1, ostream);
X if (nosecondary)
X {
X fputs ("}{", ostream);
X pending = 1;
X }
X else
X fputs ("}\n", ostream);
X
X /* Record name of most recent primary */
X if (lastprimarylength < primarylength)
X {
X lastprimarylength = primarylength + 100;
X lastprimary = (char *) xrealloc (lastprimary,
X 1 + lastprimarylength);
X }
X strncpy (lastprimary, primary, primarylength);
X lastprimary[primarylength] = 0;
X
X /* There is no current secondary within this primary, now */
X lastsecondary[0] = 0;
X }
X
X /* Should not have an entry with no subtopic following one with a subtopic */
X
X if (nosecondary && *lastsecondary)
X error ("entry %s follows an entry with a secondary name", line);
X
X /* Start a new secondary entry if necessary */
X if (!nosecondary && strncmp (secondary, lastsecondary, secondarylength))
X {
X if (pending)
X {
X fputs ("}\n", ostream);
X pending = 0;
X }
X
X /* Write the entry for the secondary. */
X fputs ("\\secondary {", ostream);
X fwrite (secondary, secondarylength, 1, ostream);
X fputs ("}{", ostream);
X pending = 1;
X
X /* Record name of most recent secondary */
X if (lastsecondarylength < secondarylength)
X {
X lastsecondarylength = secondarylength + 100;
X lastsecondary = (char *) xrealloc (lastsecondary,
X 1 + lastsecondarylength);
X }
X strncpy (lastsecondary, secondary, secondarylength);
X lastsecondary[secondarylength] = 0;
X }
X
X /* Here to add one more page number to the current entry */
X if (pending++ != 1)
X fputs (", ", ostream); /* Punctuate first, if this is not the first */
X fwrite (pagenumber, pagelength, 1, ostream);
X}
X
X/* Close out any unfinished output entry */
X
Xvoid
Xfinish_index (ostream)
X FILE *ostream;
X{
X if (pending)
X fputs ("}\n", ostream);
X free (lastprimary);
X free (lastsecondary);
X}
X
X/* Copy the lines in the sorted order.
X Each line is copied out of the input file it was found in. */
X
Xvoid
Xwritelines (linearray, nlines, ostream)
X char **linearray;
X int nlines;
X FILE *ostream;
X{
X char **stop_line = linearray + nlines;
X char **next_line;
X
X init_index ();
X
X /* Output the text of the lines, and free the buffer space */
X
X for (next_line = linearray; next_line != stop_line; next_line++)
X {
X /* If -u was specified, output the line only if distinct from previous one. */
X if (next_line == linearray
X /* Compare previous line with this one, using only the explicitly specd keyfields */
X || compare_general (*(next_line - 1), *next_line, 0L, 0L, num_keyfields - 1))
X {
X char *p = *next_line;
X char c;
X while ((c = *p++) && c != '\n');
X *(p-1) = 0;
X indexify (*next_line, ostream);
X }
X }
X
X finish_index (ostream);
X}
X
X/* Assume (and optionally verify) that each input file is sorted;
X merge them and output the result.
X Returns nonzero if any input file fails to be sorted.
X
X This is the high-level interface that can handle an unlimited number of files. */
X
X#define MAX_DIRECT_MERGE 10
X
Xint
Xmerge_files (infiles, nfiles, outfile)
X char **infiles;
X int nfiles;
X char *outfile;
X{
X char **tempfiles;
X int ntemps;
X int i;
X int value = 0;
X int start_tempcount = tempcount;
X
X if (nfiles <= MAX_DIRECT_MERGE)
X return merge_direct (infiles, nfiles, outfile);
X
X /* Merge groups of MAX_DIRECT_MERGE input files at a time,
X making a temporary file to hold each group's result. */
X
X ntemps = (nfiles + MAX_DIRECT_MERGE - 1) / MAX_DIRECT_MERGE;
X tempfiles = (char **) xmalloc (ntemps * sizeof (char *));
X for (i = 0; i < ntemps; i++)
X {
X int nf = MAX_DIRECT_MERGE;
X if (i + 1 == ntemps)
X nf = nfiles - i * MAX_DIRECT_MERGE;
X tempfiles[i] = maketempname (++tempcount);
X value |= merge_direct (&infiles[i * MAX_DIRECT_MERGE], nf, tempfiles[i]);
X }
X
X /* All temporary files that existed before are no longer needed
X since their contents have been merged into our new tempfiles.
X So delete them. */
X flush_tempfiles (start_tempcount);
X
X /* Now merge the temporary files we created. */
X
X merge_files (tempfiles, ntemps, outfile);
X
X free (tempfiles);
X
X return value;
X}
X
X/* Assume (and optionally verify) that each input file is sorted;
X merge them and output the result.
X Returns nonzero if any input file fails to be sorted.
X
X This version of merging will not work if the number of
X input files gets too high. Higher level functions
X use it only with a bounded number of input files. */
X
Xint
Xmerge_direct (infiles, nfiles, outfile)
X char **infiles;
X int nfiles;
X char *outfile;
X{
X char **ip = infiles;
X struct linebuffer *lb1, *lb2;
X struct linebuffer **thisline, **prevline;
X FILE **streams;
X int i;
X int nleft;
X int lossage = 0;
X int *file_lossage;
X struct linebuffer *prev_out = 0;
X FILE *ostream = stdout;
X
X if (outfile)
X {
X ostream = fopen (outfile, "w");
X }
X if (!ostream) pfatal_with_name (outfile);
X
X init_index ();
X
X if (nfiles == 0)
X {
X if (outfile)
X fclose (ostream);
X return 0;
X }
X
X /* For each file, make two line buffers.
X Also, for each file, there is an element of `thisline'
X which points at any time to one of the file's two buffers,
X and an element of `prevline' which points to the other buffer.
X `thisline' is supposed to point to the next available line from the file,
X while `prevline' holds the last file line used,
X which is remembered so that we can verify that the file is properly sorted. */
X
X /* lb1 and lb2 contain one buffer each per file */
X lb1 = (struct linebuffer *) xmalloc (nfiles * sizeof (struct linebuffer));
X lb2 = (struct linebuffer *) xmalloc (nfiles * sizeof (struct linebuffer));
X
X /* thisline[i] points to the linebuffer holding the next available line in file i,
X or is zero if there are no lines left in that file. */
X thisline = (struct linebuffer **) xmalloc (nfiles * sizeof (struct linebuffer *));
X /* prevline[i] points to the linebuffer holding the last used line from file i.
X This is just for verifying that file i is properly sorted. */
X prevline = (struct linebuffer **) xmalloc (nfiles * sizeof (struct linebuffer *));
X /* streams[i] holds the input stream for file i. */
X streams = (FILE **) xmalloc (nfiles * sizeof (FILE *));
X /* file_lossage[i] is nonzero if we already know file i is not properly sorted. */
X file_lossage = (int *) xmalloc (nfiles * sizeof (int));
X
X /* Allocate and initialize all that storage */
X
X for (i = 0; i < nfiles; i++)
X {
X initbuffer (&lb1[i]);
X initbuffer (&lb2[i]);
X thisline[i] = &lb1[i];
X prevline[i] = &lb2[i];
X file_lossage[i] = 0;
X streams[i] = fopen (infiles[i], "r");
X if (!streams[i])
X pfatal_with_name (infiles[i]);
X
X readline (thisline[i], streams[i]);
X }
X
X /* Keep count of number of files not at eof */
X nleft = nfiles;
X
X while (nleft)
X {
X struct linebuffer *best = 0;
X struct linebuffer *exch;
X int bestfile = -1;
X int i;
X
X /* Look at the next avail line of each file; choose the least one. */
X
X for (i = 0; i < nfiles; i++)
X {
X if (thisline[i] &&
X (!best ||
X 0 < compare_general (best->buffer, thisline[i]->buffer,
X (long) bestfile, (long) i, num_keyfields)))
X {
X best = thisline[i];
X bestfile = i;
X }
X }
X
X /* Output that line, unless it matches the previous one and we don't want duplicates */
X
X if (!(prev_out &&
X !compare_general (prev_out->buffer, best->buffer, 0L, 1L, num_keyfields - 1)))
X indexify (best->buffer, ostream);
X prev_out = best;
X
X /* Now make the line the previous of its file, and fetch a new line from that file */
X
X exch = prevline[bestfile];
X prevline[bestfile] = thisline[bestfile];
X thisline[bestfile] = exch;
X
X while (1)
X {
X /* If the file has no more, mark it empty */
X
X if (feof (streams[bestfile]))
X {
X thisline[bestfile] = 0;
X nleft--; /* Update the number of files still not empty */
X break;
X }
X readline (thisline[bestfile], streams[bestfile]);
X if (thisline[bestfile]->buffer[0] || !feof (streams[bestfile])) break;
X }
X }
X
X finish_index (ostream);
X
X /* Free all storage and close all input streams */
X
X for (i = 0; i < nfiles; i++)
X {
X fclose (streams[i]);
X free (lb1[i].buffer);
X free (lb2[i].buffer);
X }
X free (file_lossage);
X free (lb1);
X free (lb2);
X free (thisline);
X free (prevline);
X free (streams);
X
X if (outfile)
X fclose (ostream);
X
X return lossage;
X}
X
X/* Print error message and exit. */
X
Xfatal (s1, s2)
X char *s1, *s2;
X{
X error (s1, s2);
X exit (EXIT_FATAL);
X}
X
X/* Print error message. `s1' is printf control string, `s2' is arg for it. */
X
Xerror (s1, s2)
X char *s1, *s2;
X{
X printf ("texindex: ");
X printf (s1, s2);
X printf ("\n");
X}
X
Xperror_with_name (name)
X char *name;
X{
X#ifdef VMS
X#include <errno.h>
X extern noshare int sys_nerr;
X extern noshare char *sys_errlist[];
X#else
X extern int errno, sys_nerr;
X extern char *sys_errlist[];
X#endif
X char *s;
X
X if (errno < sys_nerr)
X s = concat ("", sys_errlist[errno], " for %s");
X else
X s = "cannot open %s";
X error (s, name);
X}
X
Xpfatal_with_name (name)
X char *name;
X{
X extern int errno, sys_nerr;
X extern char *sys_errlist[];
X char *s;
X
X if (errno < sys_nerr)
X s = concat ("", sys_errlist[errno], " for %s");
X else
X s = "cannot open %s";
X fatal (s, name);
X}
X
X/* Return a newly-allocated string whose contents concatenate those of s1, s2, s3. */
X
Xchar *
Xconcat (s1, s2, s3)
X char *s1, *s2, *s3;
X{
X int len1 = strlen (s1), len2 = strlen (s2), len3 = strlen (s3);
X char *result = (char *) xmalloc (len1 + len2 + len3 + 1);
X
X strcpy (result, s1);
X strcpy (result + len1, s2);
X strcpy (result + len1 + len2, s3);
X *(result + len1 + len2 + len3) = 0;
X
X return result;
X}
X
X/* Like malloc but get fatal error if memory is exhausted. */
X
Xint
Xxmalloc (size)
X int size;
X{
X int result = malloc (size);
X if (!result)
X fatal ("virtual memory exhausted", 0);
X return result;
X}
X
X
Xint
Xxrealloc (ptr, size)
X char *ptr;
X int size;
X{
X int result = realloc (ptr, size);
X if (!result)
X fatal ("virtual memory exhausted");
X return result;
X}
X
Xbzero (b, length)
X register char *b;
X register int length;
X{
X#ifdef VMS
X short zero = 0;
X long max_str = 65535;
X
X while (length > max_str) {
X (void) LIB$MOVC5 (&zero, &zero, &zero, &max_str, b);
X length -= max_str;
X b += max_str;
X }
X (void) LIB$MOVC5 (&zero, &zero, &zero, &length, b);
X#else
X while (length-- > 0)
X *b++ = 0;
X#endif /* not VMS */
X}
END_OF_FILE
if test 42774 -ne `wc -c <'manual/texindex.c'`; then
echo shar: \"'manual/texindex.c'\" unpacked with wrong size!
fi
# end of 'manual/texindex.c'
fi
echo shar: End of archive 8 \(of 9\).
cp /dev/null ark8isdone
MISSING=""
for I in 1 2 3 4 5 6 7 8 9 ; do
if test ! -f ark${I}isdone ; then
MISSING="${MISSING} ${I}"
fi
done
if test "${MISSING}" = "" ; then
echo You have unpacked all 9 archives.
rm -f ark[1-9]isdone ark[1-9][0-9]isdone
else
echo You still need to unpack the following archives:
echo " " ${MISSING}
fi
## End of shell archive.
exit 0
--
Mike Clarkson mike@ists.ists.ca
Institute for Space and Terrestrial Science uunet!attcan!ists!mike
York University, North York, Ontario, FORTRAN - just say no.
CANADA M3J 1P3 +1 (416) 736-5611