v09INF1: Posting Document v2.0 & ship.c (repost)
Chris Spell
Submitted-by: Chris Spell <sp...@seq.uncwil.edu>
Posting-number: Volume 9, Info 1
Archive-name: postguid.cgs
[ Note: Postings until May 15 will be done as I have free time.
After that time I should be able to keep up with
posting and my mail without problems. -chris ]
---------------------------------------------------------------
If you need compiled version of ship check the directory:
wuarchive.wustl.edu:/systems/hp/hp48/Posting/ship_executables
If you need the source for Info Zip zip/unzip check the directory:
wuarchive.wustl.edu:/systems/hp/hp48/Posting
If you want the entire package for DOS pick up:
wuarchive.wustl.edu:/systems/hp/hp48/Posting/posttool.zip
If you need a binary for zip/unzip other than dos see the file:
wuarchive.wustl.edu:/systems/hp/hp48/Posting/README
---------------------------------------------------------------
Comp.Sources.HP48 Posting Format
v2.0
I hate to change posting formats, but I hope the new format
will be easier for everyone. At least it will be MUCH
easier for me to put posts together and help me to get stuff out
MUCH faster.
Since we have the nice zip & unzip programs developed by the
Info ZIP people, I am going to start using these programs. These
programs should run on just about anything. The version that
I'm using is compatable with the new PKUNZIP v2.04c. The three
Info ZIP tools that I'm using are Unzip v5.0p1, Zip v1.9, and
Ship v1.0.
Each post will now consist of a Ship encoded file. This may be
a problem for some people in that you must have ship to extract
the posts. I don't really like relying on something that isn't
commonly found, but fortunately ship seems to be quite portable.
I compiled it on the pc using msc 7.0 and it worked great. I
did have to increase the stack size, but that was the only
problem. I will provide that executable in this post uuencoded.
It easily compiled on our VAX/VMS system for which I will provide
the executable in the archives. I don't have executables for
the MAC or Amiga, but I hope someone will compile the code and
send the executable to me.
I think this will work well for all of us once everyone gets ship
going. I know it will save me a lot of time and enable me to
get posts out easier. It will also give me a little more
flexability and save space in the archives. It will make
it quite simple for you and me to extract the zip file from the
posts. All one has to do is either cat all the parts in
order and pipe them to ship -u or pass the files names in order
on the command line to ship -u. It will then extract the zip
file for you without you having to remove any headers.
Another side effect of using ship is that we will not really
need the brik checksum. But since the program I use already
generates it, I will leave it in. If there is a problem with the
zip file, unzip will let you know. But I guess the checksum
will be handy in identifying bad sections of multi-part posts.
Basically, what I would like to receive for posting now is a ship
encoded zip file with the file names 8.3 characters. Remember, I
am extracting and preparing the posts on a DOS machine in most
cases. Please use the following naming chart when naming
files:
BINARIES
*. | HP 48 binary object (no extension)
*.LIB | HP 48 Library (binary)
*.RFD | HP 48 Directory containing auto-uncompressing objects.
*.DIR | HP 48 Directory (possibly binary)
DOCUMENTATION
*.DOC | HP 48 Program documentation or any documentation
*.RDM | Readme file for the HP 48 program
*.PS | Postscript document
SOURCE CODE
*.SRC | HP 48 Source Code (User RPL) in downloadable ASCII format
*.S | HP 48 Source Code (System RPL / Assembly) in HP RPLCOMP format
*.RPL | HP 48 Source COde (System RPL / Assembly) in (S)RPL.LIB format
*.DIR | HP 48 Directory
*.C | C Language source
SPECIAL (but very important time saver for me)
*.INF | Info file contains your email address, subject line, version
number and a short description. See discussion below.
I used Joe Horn's filename extensions as a basis for the above
list. I added some things that will come up for this group and
I'm sure more will be added as it comes up. The above list is
only a guideline that should be used when possible. Please make
sure when possible that you use the name of the program for the
*. part of the name. Then use the appropiate extension.
A very important time saver for me is the *.INF file I would like
you to include in the zip file. This file should have lines of
the following form:
-------
From: Chris Spell <sp...@seq.uncwil.edu>
Subject: Posting Guidelines, v2.0
This program does the following ......
-------
If you will do this for me, it will save me the time of actually
trying to find your email address and version number, and
trying to extract a short description of your
program and a one line subject for the post. You understand your
post better than I and it makes sense for you to write
these things. The short description will be put at the beginning
of the post so everyone will know what is in the zip file. I
will also add a file list to the post after the description.
As usual, if a version number is not supplied I will assign
it 1.0.
I don't really care how you get the zip file to me. If you use
ship that's great, but if you uuencode it that's fine. If you
upload it to the Uploads dir on wuarchive.wustl.edu or seq, that's
fine as well, but please make sure the *.INF file is in the zip
file.
The minimum contents of each zip file should be a HP48 program
(preferbly a binary), a document file describing the binary, and
the *.INF file. I say it is preferable to get a binary because
sometimes I've gotten a user rpl program which might contain a
variable that is already in my 48. In this case,
the program generally won't download until I find and purge that
variable. This doesn't happen often, but it does happen and it
takes a lot of time to straighten out.
I think I have covered everything. If you have any suggestions
or comments, please let me know. This is new and I hope it works
well. I can't think of anything that will work better and save
me and you as much time. If there is a better way of doing this,
I would like to know. I know some people will say just use
shar, but that won't handle what needs to be done here.
Below is the source code for ship and the executables for zip and
unzip. This file is intended to be the "starter kit" and I want to
keep it small as possible.
The original files can be retreived from:
wuarchive.wustl.edu:/systems/hp48sx/Posting
seq.uncwil.edu:/hp48/Posting
I'm collecting executables for ship for various
machines/operating systems. Check the Posting directory for the
latest copies. If you compile it on a machine not listed in the
zipsrc.doc file, please upload a copy in Posting/shipUploads or
mail a copy to me.
Chris
Moderator Comp.Sources.HP48 & HP48 archive maintainer
sp...@seq.uncwil.edu
----------------------------------------------------------
/* ship.c -- Not copyrighted 1991 Mark Adler */
#define SHIPVER "ship version 1.0 September 29, 1991 Mark Adler"
/* Command for mailing (-m): the %s's are, in order, the subject prefix,
the part number (always of the form "partnnnn"), the subject suffix
(empty or " (last)" if the last part), the mailing address, and the
name of the temporary file begin mailed. The command "Mail" is for BSD
systems. You may need to change it to "mailx" for System V Unix, using
the compile option "-DMAILX". Also, on Sperry (Unisys?) SysV.3 systems,
you might try the command name "v6mail". */
#ifdef DIRENT /* If compiled with zip, DIRENT implies System V */
# define MAILX
#endif /* DIRENT */
#ifdef sun /* Except Sun's use DIRENT, but have Mail */
# ifdef MAILX
# undef MAILX
# endif /* MAILX */
#endif /* sun */
#ifdef sgi /* Silicon Graphics that way too */
# ifdef MAILX
# undef MAILX
# endif /* MAILX */
#endif /* sgi */
#ifdef VMS
# define TMPNAME "_SXXXXXX."
# define MAILCMD "mail %s /subj=\"%s %s%s\" \"%s\""
# define PATHCUT ']'
#else /* !VMS */
# define TMPNAME "_SXXXXXX"
# ifdef MAILX
# define MAILCMD "mailx -s \"%s %s%s\" \"%s\" < %s"
# else /* !MAILX */
# define MAILCMD "Mail -s \"%s %s%s\" \"%s\" < %s"
# endif /* ?MAILX */
# define PATHCUT '/'
#endif /* ?VMS */
/*
SHIP -
Ship is a program for sending binary files through email. It is designed
to supplant uuencode and uudecode. Ship encodes approximately 6.23 bits
per character mailed, compared to uuencode's 5.73 bits per character.
Ship also has these features: a 32-bit CRC check on each file; automatic
splitting of the ship output into multiple, smaller files for speedier
mailing; automatic mailing of ship's output, with subject lines for
multiple parts; and a check on the sequence of parts when unshipping.
Usage:
ship [-nnn] [-m address] [-s subject] file ...
where nnn is the maximum number of K bytes for each output file, address
is the address to send mail to, subject is a Subject" line prefix, and
file ... is a list of files to ship. If no options are given, ship
outputs to stdout. The simplest use is:
ship foo > x
where foo is converted into the mailable file, x.
When -nnn is specified, but -m is not, ship writes to the files
part0001, part0002, etc., where each file has nnn or less K bytes. For
example:
ship -25 bigfoo
will write however many 25K byte or less ship files is needed to contain
bigfoo. If, say, six files are needed, then the files part0001 to part0006
will be written.
When using -m, nothing is written, either to files or to stdout; rather,
the output is mailed to the specified address. If -nnn is also specified,
then the parts are mailed separately with the subject lines part0001, etc.
If -nnn is not specified, then only one part (the whole thing) is mailed
with the subject line "part0001". For example:
ship -25 -m fred bigfoo
will mail the six parts of bigfoo to fred.
Any number of files can be shipped at once. They become part of one long
ship stream, so if, for example -25 is specified, all but the last part
will have about 25K bytes. For example:
ship -25 -m fred fee fi fo fum
will send the files fee, fi, fo, and fum to fred.
Fred will get several mail messages with the subject lines part0001, etc.
He can then save those messages as the files, say, p1, p2, p3, ...
Then he can use the command:
ship -u p?
to recreate bigfoo, or fee fi fo and fum, depending on what he was sent.
If Fred saved the wrong numbers, ship will detect this and report a
sequence error.
Note: there is enough information in the shipped parts to determine the
correct sequence. A future version of ship will prescan the files to
determine the sequence, and then process them in the correct order.
If a file being received already exists, ship -u will report an error
and exit. The -o option avoids this and allows ship to overwrite existing
files. The -o option must follow the -u option:
ship -u -o p?
In addition to the -u option, ship will unship if it sees that its name is
unship. On Unix systems, this can be done simply by linking the executable
to unship:
ln ship unship
Ship can also be used as a filter. The special file name "-" means stdin.
For example:
tar covf - foodir | compress | ship -25 -m fred -
will tar the directory foodir, compress it, and ship it to fred in 25K byte
pieces. Then, after Fred saves the files as p01, etc. at the other, end,
he can:
ship -u p? | zcat | tar xovf -
which will recreate the directory foobar and its contents. ship -u knows
to write to stdout, since the original ship put the special file name "-"
in the first part.
Ship uses a base 85 coding that needs 32-bit multiplication and division.
This can be slow on 16-bit machines, so ship provides a fast encoding
method by specifying the -f option. This method is somewhat faster even
on 32-bit machines, and has approximately a 1% penalty in the size of the
encoded result (-f gives 6.26 bits per character, on the average). The -f
option need only be used when shipping--unshipping (ship -u) automatically
detects the encoding used. For example:
ship -f -25 -m fred foo
will send foo to fred in 25K byte pieces using the fast encoding method.
You don't need to tell Fred, since ship -u will figure that out for him.
The fast encoding method is probabilistic, so it's possible for the size
penalty to be worse than 1%, and it's also possible for the fast encoding
to produce a smaller result than base 85 encoding would, all depending on
the data.
The -q option can be used with either ship or unship (ship -u) for quiet
operation--informational messages are inhibited.
You can find out the version of ship and get the command usage by using
"ship -h" or "ship -?". The version number and date and help will be
printed, and ship will exit (the rest of the command line is ignored).
Acknowledgements:
The hard-arithmetic coding algorithm was blatantly stolen from Peter
Gutmann's pgencode/pgdecode programs posted on comp.compression, with
modifications to use 86 instead of 94 characters, and to make zeros encode
better than, rather than worse than other bytes. (As Stravinsky once said:
"Mediocre composers plagiarize. Great composers steal.")
*/
/* tailor.h -- Not copyrighted 1991 Mark Adler */
/* const's are inconsistently used across ANSI libraries--kill for all
header files. */
#define const
/* Use prototypes and ANSI libraries if __STDC__ */
#ifdef __STDC__
# ifndef PROTO
# define PROTO
# endif /* !PROTO */
# define MODERN
#endif /* __STDC__ */
/* Use prototypes and ANSI libraries if Silicon Graphics */
#ifdef sgi
# ifndef PROTO
# define PROTO
# endif /* !PROTO */
# define MODERN
#endif /* sgi */
/* Define MSDOS for Turbo C as well as Microsoft C */
#ifdef __POWERC /* For Power C too */
# define __TURBOC__
#endif /* __POWERC */
#ifdef __TURBOC__
# ifndef MSDOS
# define MSDOS
# endif /* !MSDOS */
#endif /* __TURBOC__ */
/* Use prototypes and ANSI libraries if Microsoft or Borland C */
#ifdef MSDOS
# ifndef PROTO
# define PROTO
# endif /* !PROTO */
# define MODERN
#endif /* MSDOS */
/* Turn off prototypes if requested */
#ifdef NOPROTO
# ifdef PROTO
# undef PROTO
# endif /* PROTO */
#endif /* NOPROT */
/* Used to remove arguments in function prototypes for non-ANSI C */
#ifdef PROTO
# define OF(a) a
#else /* !PROTO */
# define OF(a) ()
#endif /* ?PROTO */
/* Allow far and huge allocation for small model (Microsoft C or Turbo C) */
#ifdef MSDOS
# ifdef __TURBOC__
# include <alloc.h>
# else /* !__TURBOC__ */
# include <malloc.h>
# define farmalloc _fmalloc
# define farfree _ffree
# endif /* ?__TURBOC__ */
#else /* !MSDOS */
# define huge
# define far
# define near
# define farmalloc malloc
# define farfree free
#endif /* ?MSDOS */
/* Define MSVMS if either MSDOS or VMS defined */
#ifdef MSDOS
# define MSVMS
#else /* !MSDOS */
# ifdef VMS
# define MSVMS
# endif /* VMS */
#endif /* ?MSDOS */
/* Define void, voidp, and extent (size_t) */
#include <stdio.h>
#ifdef MODERN
# ifndef M_XENIX
# include <stddef.h>
# endif /* !M_XENIX */
# include <stdlib.h>
typedef size_t extent;
typedef void voidp;
#else /* !MODERN */
typedef unsigned int extent;
# define void int
typedef char voidp;
#endif /* ?MODERN */
/* Get types and stat */
#ifdef VMS
# include <types.h>
# include <stat.h>
#else /* !VMS */
# include <sys/types.h>
# include <sys/stat.h>
#endif /* ?VMS */
/* Cheap fix for unlink on VMS */
#ifdef VMS
# define unlink delete
#endif /* VMS */
/* For Pyramid */
#ifdef pyr
# define strrchr rindex
# define ZMEM
#endif /* pyr */
/* File operations--use "b" for binary if allowed */
#ifdef MODERN
# define FOPR "rb"
# define FOPM "r+b"
# define FOPW "w+b"
#else /* !MODERN */
# define FOPR "r"
# define FOPM "r+"
# define FOPW "w+"
#endif /* ?MODERN */
/* Fine tuning */
#ifndef MSDOS
# define BSZ 8192 /* Buffer size for files */
#else /* !MSDOS */
# define BSZ 4096 /* Keep precious NEAR space */
/* BSZ can't be 8192 even for compact model because of 64K limitation
* in im_lmat.c. If you run out of memory when processing a large number
* files, use the compact model and reduce BSZ to 2048 here and in
* im_lm.asm.
*/
#endif /* ?MSDOS */
/* end of tailor.h */
#ifdef MODERN
# include <string.h>
#else /* !MODERN */
voidp *malloc();
long atol();
char *strcpy();
char *strrchr();
#endif /* ?MODERN */
/* Library functions not in (most) header files */
char *mktemp OF((char *));
int unlink OF((char *));
#ifdef MSDOS /* Use binary mode for binary files */
# include <io.h>
# include <fcntl.h>
#endif /* MSDOS */
#define LNSZ 1025 /* size of line buffer */
typedef unsigned long ulg; /* 32-bit unsigned integer */
typedef struct { /* associates a CRC with a file */
FILE *f; /* pointer to associated file stream */
ulg c; /* CRC register */
ulg b; /* four byte buffer */
int n; /* buffer count */
} cfile;
/* Function prototypes */
#ifdef MODERN
void err(int, char *);
cfile *chook(FILE *);
char *nopath(char *);
void newship(void);
void endship(int);
void newline(char *);
void ship(char *, FILE *);
void mkinv(void);
void decode(unsigned char *, cfile *);
void unship(char **, int, int);
void help(void);
void main(int, char **);
#endif /* MODERN */
/* Globals for ship() */
char sname[9]; /* current ship file name */
FILE *sfile; /* current ship file */
ulg slns; /* number of lines written to ship file */
ulg slmax; /* maximum number of lines per ship file */
int fast; /* true for arithmetic coding, else base 85 */
int mail; /* true if mailing */
char mpspc[9]; /* prealloced space for prefix */
char *mprefix = mpspc; /* identification for this mailing */
char *mdest; /* mail destination */
char mname[10]; /* temporary file name if mailing */
ulg ccnt; /* count of bytes read or written */
int noisy = 1; /* false to inhibit informational messages */
/* Errors */
#define SE_ARG 1
#define SE_FIND 2
#define SE_NONE 3
#define SE_PART 4
#define SE_FORM 5
#define SE_CONT 6
#define SE_CRC 7
#define SE_MAIL 8
#define SE_OVER 9
#define SE_FULL 10
#define SE_MANY 11
#define SE_MEM 12
char *errors[] = {
/* 1 */ "invalid argument ",
/* 2 */ "could not find ",
/* 3 */ "no files received",
/* 4 */ "unfinished file ",
/* 5 */ "invalid ship format in ",
/* 6 */ "wrong sequence for ",
/* 7 */ "CRC check failed on ",
/* 8 */ "mail command failed: ",
/* 9 */ "attempted to overwrite ",
/* 10 */ "could not write to ",
/* 11 */ "too many output files!",
/* 12 */ "out of memory"
};
/* Set of 86 characters used for the base 85 digits (last one not used), and
the 86 character arithmetic coding. Selected to be part of both the ASCII
printable characters, and the common EBCDIC printable characters whose
ASCII translations are universal. */
unsigned char safe[] = {
'{','"','#','$','%','&','\'','(',')','*','+',',','-','.','/',
'0','1','2','3','4','5','6','7','8','9',':',';','<','=','>','?','@',
'A','B','C','D','E','F','G','H','I','J','K','L','M',
'N','O','P','Q','R','S','T','U','V','W','X','Y','Z','_',
'a','b','c','d','e','f','g','h','i','j','k','l','m',
'n','o','p','q','r','s','t','u','v','w','x','y','z','}'};
#define LOWSZ (sizeof(safe)-64) /* low set size for fast coding */
/* Special replacement pairs--if first of each pair is received, it is
treated like the second member of the pair. You're probably
wondering why. The first pair is for compatibility with an
earlier version of ship that used ! for the base 85 zero digit.
However, there exist ASCII-EBCDIC translation tables that don't
know about exclamation marks. The second set has mysterious
historical origins that are best left unspoken ... */
unsigned char aliases[] = {'!','{','|','+',0};
/* Inverse of safe[], filled in by mkinv() */
unsigned char invsafe[256];
/* Table of CRC-32's of all single byte values (made by makecrc.c) */
ulg crctab[] = {
0x00000000L, 0x77073096L, 0xee0e612cL, 0x990951baL, 0x076dc419L,
0x706af48fL, 0xe963a535L, 0x9e6495a3L, 0x0edb8832L, 0x79dcb8a4L,
0xe0d5e91eL, 0x97d2d988L, 0x09b64c2bL, 0x7eb17cbdL, 0xe7b82d07L,
0x90bf1d91L, 0x1db71064L, 0x6ab020f2L, 0xf3b97148L, 0x84be41deL,
0x1adad47dL, 0x6ddde4ebL, 0xf4d4b551L, 0x83d385c7L, 0x136c9856L,
0x646ba8c0L, 0xfd62f97aL, 0x8a65c9ecL, 0x14015c4fL, 0x63066cd9L,
0xfa0f3d63L, 0x8d080df5L, 0x3b6e20c8L, 0x4c69105eL, 0xd56041e4L,
0xa2677172L, 0x3c03e4d1L, 0x4b04d447L, 0xd20d85fdL, 0xa50ab56bL,
0x35b5a8faL, 0x42b2986cL, 0xdbbbc9d6L, 0xacbcf940L, 0x32d86ce3L,
0x45df5c75L, 0xdcd60dcfL, 0xabd13d59L, 0x26d930acL, 0x51de003aL,
0xc8d75180L, 0xbfd06116L, 0x21b4f4b5L, 0x56b3c423L, 0xcfba9599L,
0xb8bda50fL, 0x2802b89eL, 0x5f058808L, 0xc60cd9b2L, 0xb10be924L,
0x2f6f7c87L, 0x58684c11L, 0xc1611dabL, 0xb6662d3dL, 0x76dc4190L,
0x01db7106L, 0x98d220bcL, 0xefd5102aL, 0x71b18589L, 0x06b6b51fL,
0x9fbfe4a5L, 0xe8b8d433L, 0x7807c9a2L, 0x0f00f934L, 0x9609a88eL,
0xe10e9818L, 0x7f6a0dbbL, 0x086d3d2dL, 0x91646c97L, 0xe6635c01L,
0x6b6b51f4L, 0x1c6c6162L, 0x856530d8L, 0xf262004eL, 0x6c0695edL,
0x1b01a57bL, 0x8208f4c1L, 0xf50fc457L, 0x65b0d9c6L, 0x12b7e950L,
0x8bbeb8eaL, 0xfcb9887cL, 0x62dd1ddfL, 0x15da2d49L, 0x8cd37cf3L,
0xfbd44c65L, 0x4db26158L, 0x3ab551ceL, 0xa3bc0074L, 0xd4bb30e2L,
0x4adfa541L, 0x3dd895d7L, 0xa4d1c46dL, 0xd3d6f4fbL, 0x4369e96aL,
0x346ed9fcL, 0xad678846L, 0xda60b8d0L, 0x44042d73L, 0x33031de5L,
0xaa0a4c5fL, 0xdd0d7cc9L, 0x5005713cL, 0x270241aaL, 0xbe0b1010L,
0xc90c2086L, 0x5768b525L, 0x206f85b3L, 0xb966d409L, 0xce61e49fL,
0x5edef90eL, 0x29d9c998L, 0xb0d09822L, 0xc7d7a8b4L, 0x59b33d17L,
0x2eb40d81L, 0xb7bd5c3bL, 0xc0ba6cadL, 0xedb88320L, 0x9abfb3b6L,
0x03b6e20cL, 0x74b1d29aL, 0xead54739L, 0x9dd277afL, 0x04db2615L,
0x73dc1683L, 0xe3630b12L, 0x94643b84L, 0x0d6d6a3eL, 0x7a6a5aa8L,
0xe40ecf0bL, 0x9309ff9dL, 0x0a00ae27L, 0x7d079eb1L, 0xf00f9344L,
0x8708a3d2L, 0x1e01f268L, 0x6906c2feL, 0xf762575dL, 0x806567cbL,
0x196c3671L, 0x6e6b06e7L, 0xfed41b76L, 0x89d32be0L, 0x10da7a5aL,
0x67dd4accL, 0xf9b9df6fL, 0x8ebeeff9L, 0x17b7be43L, 0x60b08ed5L,
0xd6d6a3e8L, 0xa1d1937eL, 0x38d8c2c4L, 0x4fdff252L, 0xd1bb67f1L,
0xa6bc5767L, 0x3fb506ddL, 0x48b2364bL, 0xd80d2bdaL, 0xaf0a1b4cL,
0x36034af6L, 0x41047a60L, 0xdf60efc3L, 0xa867df55L, 0x316e8eefL,
0x4669be79L, 0xcb61b38cL, 0xbc66831aL, 0x256fd2a0L, 0x5268e236L,
0xcc0c7795L, 0xbb0b4703L, 0x220216b9L, 0x5505262fL, 0xc5ba3bbeL,
0xb2bd0b28L, 0x2bb45a92L, 0x5cb36a04L, 0xc2d7ffa7L, 0xb5d0cf31L,
0x2cd99e8bL, 0x5bdeae1dL, 0x9b64c2b0L, 0xec63f226L, 0x756aa39cL,
0x026d930aL, 0x9c0906a9L, 0xeb0e363fL, 0x72076785L, 0x05005713L,
0x95bf4a82L, 0xe2b87a14L, 0x7bb12baeL, 0x0cb61b38L, 0x92d28e9bL,
0xe5d5be0dL, 0x7cdcefb7L, 0x0bdbdf21L, 0x86d3d2d4L, 0xf1d4e242L,
0x68ddb3f8L, 0x1fda836eL, 0x81be16cdL, 0xf6b9265bL, 0x6fb077e1L,
0x18b74777L, 0x88085ae6L, 0xff0f6a70L, 0x66063bcaL, 0x11010b5cL,
0x8f659effL, 0xf862ae69L, 0x616bffd3L, 0x166ccf45L, 0xa00ae278L,
0xd70dd2eeL, 0x4e048354L, 0x3903b3c2L, 0xa7672661L, 0xd06016f7L,
0x4969474dL, 0x3e6e77dbL, 0xaed16a4aL, 0xd9d65adcL, 0x40df0b66L,
0x37d83bf0L, 0xa9bcae53L, 0xdebb9ec5L, 0x47b2cf7fL, 0x30b5ffe9L,
0xbdbdf21cL, 0xcabac28aL, 0x53b39330L, 0x24b4a3a6L, 0xbad03605L,
0xcdd70693L, 0x54de5729L, 0x23d967bfL, 0xb3667a2eL, 0xc4614ab8L,
0x5d681b02L, 0x2a6f2b94L, 0xb40bbe37L, 0xc30c8ea1L, 0x5a05df1bL,
0x2d02ef8dL
};
/* Macro to update the CRC shift register one byte at a time */
#define CRC(c,b) (crctab[((int)(c)^(int)(b))&0xff]^((c)>>8))
char *errname = "ship error";
char *warname = "ship warning";
void err(n, m)
int n; /* error number */
char *m; /* additional error information */
{
if (n == SE_FIND || n == SE_FULL)
perror(errname);
fputs(errname, stderr);
fputs(": ", stderr);
fputs(errors[n - 1], stderr);
fputs(m, stderr);
putc('\n', stderr);
if (*mname)
unlink(mname);
#ifdef VMS
exit(0);
#else /* !VMS */
exit(n);
#endif /* ?VMS */
}
cfile *chook(f)
FILE *f; /* file stream */
/* Inherit the file stream structure and add a CRC and buffer for appending
a CRC on reads and checking the CRC on writes. Return a pointer to the
cfile structure, or NULL if the malloc() failed. Also, if MSDOS, set the
file mode to binary to avoid LF<->CRLF conversions. */
{
cfile *c; /* allocated cfile structure */
#ifdef MSDOS
/* Set file mode to binary for MSDOS systems */
setmode(fileno(f), O_BINARY);
#endif /* MSDOS */
/* Allocate and fill structure */
if ((c = (cfile *)malloc(sizeof(cfile))) != NULL)
{
c->f = f; /* file stream */
c->b = 0; /* empty fifo (for output) */
c->c = 0xffffffffL; /* preload CRC register */
c->n = 0; /* fifo is empty (output) or */
} /* no CRC bytes given (input) */
return c;
}
/* cgetc(x)--like getc(f), but appends a 32-bit CRC to the end of the stream.
Return the byte read (the last four of which will be the CRC) or EOF. */
#define cgete(x) (x->n==4?EOF:(x->c=x->n++?x->c>>8:~x->c,(int)x->c&0xff))
#define cgetc(x) (x->n==0&&(b=getc(x->f))!=EOF?(ccnt++,x->c=CRC(x->c,b),b):cgete(c))
/* cputc(d,x)--like putc(d,f), but delays four bytes and computes a CRC.
x is a cfile *, and d is expected to be an ulg. */
#define cputf(x) (int)(x->c=CRC(x->c,x->b),putc((int)x->b&0xff,x->f),ccnt++)
#define cputc(d,x) (x->n!=4?x->n++:cputf(x),x->b=(x->b>>8)+((ulg)(d)<<24))
char *nopath(p)
char *p; /* full file name */
/* Extract just the name of file--remove and subdirectories or devices */
{
#ifdef MSDOS
char *q = "/\\:"; /* MSDOS delimiters */
#else /* !MSDOS */
#ifdef VMS
char *q = "]:"; /* VMS delimiters */
#else /* !VMS */
char *q = "/"; /* Unix delimiter */
#endif /* ?VMS */
#endif /* ?MSDOS */
char *r; /* result of strrchr() */
while (*q)
if ((r = strrchr(p, *q++)) != NULL)
p = r + 1;
return p;
}
void newship()
/* Open up a new ship file to write to */
{
int i; /* scans down name to increment */
for (i = 7; i > 3; i--)
if (++sname[i] > '9')
sname[i] = '0';
else
break;
if (i == 3)
err(SE_MANY, "");
if ((sfile = fopen(mail ? mktemp(strcpy(mname, TMPNAME)) : sname,
"w")) == NULL)
err(SE_FULL, mail ? mname : sname);
slns = 0;
}
void endship(e)
int e; /* true if ending the last ship file */
/* Finish off current ship file */
{
char *s; /* malloc'd space for mail command */
if (ferror(sfile) || fclose(sfile))
err(SE_FULL, mail ? mname : sname);
if (mail)
{
if ((s = malloc(strlen(MAILCMD)- 5*2 + strlen(mprefix) + strlen(sname) +
(e ? 7 : 0) + strlen(mdest) + strlen(mname) + 1)) == NULL)
err(SE_MEM, "");
#ifdef VMS
sprintf(s, MAILCMD, mname, mprefix, sname, e ? " (last)" : "", mdest);
if (!system(s)) /* this string fits on one line */
err(SE_MAIL, "system() call is not supported on this machine");
#else /* !VMS */
sprintf(s, MAILCMD, mprefix, sname, e ? " (last)" : "", mdest, mname);
if (system(s))
err(SE_MAIL, s);
#endif /* ?VMS */
free((voidp *)s);
unlink(mname);
*mname = 0;
}
}
void newline(p)
char *p; /* name of the input file */
/* Add a new line inside a ship file, possibly cut the file */
{
putc('\n', sfile);
slns++;
if (slmax && slns >= slmax - 2)
{
putc('$', sfile);
if (fast)
fputs(" f", sfile);
fputs("\nmore\n", sfile);
endship(0);
newship();
fprintf(sfile, "$%s\ncont %lu %s\n", fast ? " f" : "", ccnt, nopath(p));
slns += 2;
}
}
/* Macro to avoid leading dots. It assumes i==0 at the beginning of a line
and that b is an available int. c is only evaluated once. */
#define sputc(c,f) (i==0?((b=(c))=='.'?putc(' ',f):0,putc(b,f)):putc(c,f))
void ship(p, f)
char *p; /* name of the input file */
FILE *f; /* input file */
/* Encode the binary file f. */
{
int b; /* character just read */
cfile *c; /* checked file stream */
int i; /* how much is written on line so far */
int j; /* how much is in bit buffer */
/* Set up output file if needed */
if ((mail || slmax) && sfile == stdout)
{
strcpy(sname, "part0000");
newship();
}
/* Write header */
if ((c = chook(f)) == NULL)
err(SE_MEM, "");
ccnt = 0;
if (slmax && slns >= slmax - 5)
{
endship(0);
newship();
}
fprintf(sfile, "$%s\nship %s\n", fast ? " f" : "", nopath(p));
slns += 2;
/* Encode the file, writing to sfile */
if (fast)
{
int d; /* accumulates bits (never more than 14) */
d = j = i = 0;
while ((b = cgetc(c)) != EOF)
{
d |= b << j;
j += 8;
if ((d & 0x3f) >= LOWSZ)
{
sputc((int)(safe[(d & 0x3f) + LOWSZ]), sfile);
d >>= 6;
j -= 6;
}
else
{
sputc((int)(safe[(d & 0x3f) + (d & 0x40 ? LOWSZ : 0)]), sfile);
d >>= 7;
j -= 7;
}
if (++i == 79)
{
newline(p);
i = 0;
}
if (j >= 6 && (d & 0x3f) >= LOWSZ)
{
sputc((int)(safe[(d & 0x3f) + LOWSZ]), sfile);
d >>= 6;
j -= 6;
if (++i == 79)
{
newline(p);
i = 0;
}
}
else if (j >= 7)
{
sputc((int)(safe[(d & 0x3f) + (d & 0x40 ? LOWSZ : 0)]), sfile);
d >>= 7;
j -= 7;
if (++i == 79)
{
newline(p);
i = 0;
}
}
}
free((voidp *)c);
/* Write leftover bits */
if (j)
{
sputc((int)(safe[d + (d < LOWSZ ? 0 : LOWSZ)]), sfile);
putc('\n', sfile);
slns++;
}
else if (i)
{
putc('\n', sfile);
slns++;
}
}
else
{
ulg d; /* accumulates bytes */
d = j = i = 0;
while ((b = cgetc(c)) != EOF)
{
d += ((ulg)b) << j;
if ((j += 8) == 32)
{
sputc((int)(safe[(int)(d % 85)]), sfile); d /= 85;
putc((int)(safe[(int)(d % 85)]), sfile); d /= 85;
putc((int)(safe[(int)(d % 85)]), sfile); d /= 85;
putc((int)(safe[(int)(d % 85)]), sfile); d /= 85;
putc((int)(safe[(int)d]), sfile);
if (++i == 15) /* each line is <= 75 characters */
{
newline(p);
i = 0;
}
d = j = 0;
}
}
free((voidp *)c);
/* Write leftover data */
if (j)
{
j >>= 3;
sputc((int)(safe[(int)(d % 85)]), sfile);
while (j--)
{
d /= 85;
putc((int)(safe[(int)(d % 85)]), sfile);
}
putc('\n', sfile);
slns++;
}
else if (i)
{
putc('\n', sfile);
slns++;
}
}
putc('$', sfile);
if (fast)
fputs(" f", sfile);
fputs("\nend\n", sfile);
slns += 2;
if (ferror(sfile) || fflush(sfile))
err(SE_FULL, mail ? mname : sname);
if (noisy)
fprintf(stderr, "%s shipped\n", p);
}
void mkinv()
/* Build invsafe[], the inverse of safe[]. */
{
int i;
for (i = 0; i < 256; i++)
invsafe[i] = 127;
for (i = 0; i < sizeof(safe); i++)
invsafe[safe[i]] = (char)i;
for (i = 0; aliases[i]; i += 2)
invsafe[aliases[i]] = invsafe[aliases[i + 1]];
}
unsigned int decb; /* bit buffer for decode */
unsigned int decn; /* number of bits in decb */
void decode(s, c)
unsigned char *s; /* data to decode */
cfile *c; /* binary output file */
/* Decode s, a string of base 85 digits or, if fast is true, a string of safe
characters generated arithmetically, into its binary equivalent, writing
the result to c, using cputc(). */
{
int b; /* state of line loop, next character */
int k; /* counts bits or digits read */
/* powers of 85 table for decoding */
static ulg m[] = {1L,85L,85L*85L,85L*85L*85L,85L*85L*85L*85L};
if (fast)
{
unsigned int d; /* disperses bits */
d = decb;
k = decn;
while ((b = *s++) != 0)
if ((b = invsafe[b]) < sizeof(safe))
{
if (b < LOWSZ)
{
d |= b << k;
k += 7;
}
else if ((b -= LOWSZ) < LOWSZ)
{
d |= (b + 0x40) << k;
k += 7;
}
else
{
d |= b << k;
k += 6;
}
if (k >= 8)
{
cputc(d, c);
d >>= 8;
k -= 8;
}
}
decb = d;
decn = k;
}
else
{
ulg d; /* disperses bytes */
d = k = 0;
while ((b = *s++) != 0)
if ((b = invsafe[b]) < 85)
{
d += m[k] * b;
if (++k == 5)
{
cputc(d, c); d >>= 8;
cputc(d, c); d >>= 8;
cputc(d, c); d >>= 8;
cputc(d, c);
d = k = 0;
}
}
if (--k > 0)
{
while (--k)
{
cputc(d, c);
d >>= 8;
}
cputc(d, c);
}
}
}
void unship(v, g, o)
char **v; /* arguments */
int g; /* number of arguments */
int o; /* overwrite flag */
/* Extract from the files named in the arguments the files that were
encoded by ship. If an argument is "-", then stdin is read. */
{
int b; /* state of line loop */
cfile *c; /* output binary file */
FILE *f; /* output file */
char *h; /* name of current ship file */
char l[LNSZ]; /* line buffer on input */
int n; /* next argument to use for input */
char *p; /* modifies line buffer */
char *q; /* scans continuation line */
char *r; /* name of output binary file */
FILE *s; /* current ship file */
int z; /* true if zero files received */
/* Build inverse table */
mkinv();
/* No input or output files initially */
s = NULL;
c = NULL;
h = r = NULL;
/* Loop on input files' lines */
z = 1; /* none received yet */
n = 0; /* start with file zero */
b = 2; /* not in body yet */
while (1) /* return on end of last file */
{
/* Get next line from list of files */
while (s == NULL || fgets(l, LNSZ, s) == NULL)
{
if (s != NULL)
fclose(s);
if (n >= g)
{
if (c != NULL)
err(SE_PART, r);
else if (z)
err(SE_NONE, "");
return;
}
if (v[n][0] == '-')
if (v[n][1])
err(SE_ARG, v[n]);
else
{
h = "stream stdin";
s = stdin;
}
else
{
h = v[n];
if ((s = fopen(h, "r")) == NULL)
err(SE_FIND, h);
}
n++;
b &= ~1; /* not in middle of line */
}
/* Strip control characters and leading blank space, if any */
for (q = l; *q && *q <= ' ' && *q != '\n'; q++)
;
for (p = l; *q; q++)
if (*q >= ' ' || *q == '\n')
*p++ = *q;
*p = 0;
/* Based on current state, end or start on terminator. States are:
b == 0: at start of body or body terminator line
b == 1: in middle of body line
b == 2: at start of non-body line
b == 3: in middle of non-body line
b == 4: at information line
*/
switch (b)
{
case 0:
if ((!fast && strcmp(l, "$\n") == 0) ||
(fast && strcmp(l, "$ f\n") == 0))
{
b = 4;
break;
}
/* fall through to case 1 */
case 1:
decode((unsigned char *)l, c);
b = l[strlen(l) - 1] != '\n';
break;
case 2:
if (strcmp(l, "$\n") == 0 || strcmp(l, "$ f\n") == 0)
{
fast = l[1] == ' ';
b = 4;
break;
}
/* fall through to case 3 */
case 3:
b = l[strlen(l)-1] == '\n' ? 2 : 3;
break;
case 4:
/* Possible information lines are ship, more, cont, and end */
if (l[b = strlen(l) - 1] != '\n')
err(SE_FORM, h);
l[b] = 0;
if (strncmp(l, "ship ", 5) == 0)
{
/* get name, open new output file */
if (c != NULL)
err(SE_FORM, h);
if ((r = malloc(b - 4)) == NULL)
err(SE_MEM, "");
strcpy(r, l + 5);
if (strcmp(r, "-") == 0)
f = stdout;
#ifndef VMS /* shouldn't have explicit version #, so VMS won't overwrite */
else if (!o && (f = fopen(r, "r")) != NULL)
{
fclose(f);
err(SE_OVER, r);
}
#endif /* !VMS */
else if ((f = fopen(r, "w")) == NULL)
err(SE_FULL, r);
if ((c = chook(f)) == NULL)
err(SE_MEM, "");
b = decb = decn = 0;
ccnt = 0;
}
else if (strcmp(l, "more") == 0)
{
/* check if currently writing */
if (c == NULL)
err(SE_FORM, h);
b = 2;
}
else if (strncmp(l, "cont ", 5) == 0)
{
/* check name and file offset */
if (c == NULL)
err(SE_FORM, h);
for (q = l + 5; *q && *q != ' '; q++)
;
if (*q == 0 || atol(l + 5) != ccnt + 4 + (decn != 0) ||
strcmp(q + 1, r))
err(SE_CONT, r);
b = 0;
}
else if (strcmp(l, "end") == 0)
{
/* check crc, close output file */
if (c == NULL)
err(SE_FORM, h);
if (c->n != 4 || c->b != ~c->c)
err(SE_CRC, r);
if (ferror(c->f) || fclose(c->f))
err(SE_FULL, r);
if (noisy)
fprintf(stderr, "%s received\n", r);
z = 0;
free((voidp *)c);
c = NULL;
b = 2;
}
else
{
for (q = l; *q && *q != ' '; q++)
;
*q = 0;
fprintf(stderr, "%s: unsupported keyword '%s' ignored\n", warname, l);
b = 4;
}
break;
}
}
}
void help()
{
int i;
static char *text[] = {
"Usage:",
" ship [-f] [-q] [-nnn] [-m address] [-s subject] files...",
"",
" ships the files to stdout. -m sends the output via the mailer to",
" address. -nnn splits the output into pieces of nnnK bytes or less.",
" if -nnn is used without -m, the output goes to the files partxxxx,",
" where xxxx is 0001, 0002, etc. If -0 is specified, the output goes",
" entirely to the file part0001. -f uses a fast method with slightly",
" less performance. If no files are given, stdin is used. The special",
" filename '-' also takes input from stdin. Files shipped from stdin",
" are unshipped to stdout. This can be used to document a shipment.",
" When mailing, -s gives a subject line prefix. -q inhibits messages.",
"",
" ship -u [-o] [-q] files...",
" unship [-o] [-q] files...",
"",
" extracts the contents of the mail messages in files... -o allows",
" existing files to be overwritten. -u is implied if the name of the",
" executable is unship. If no files are given, the input is from",
" stdin. If any of the files were shipped from stdin, then they are",
" extracted to stdout."
};
puts(SHIPVER);
for (i = 0; i < sizeof(text)/sizeof(char *); i++)
{
printf(text[i]);
putchar('\n');
}
exit(0);
}
void main(argc, argv)
int argc; /* number of arguments */
char **argv; /* table of argument strings */
{
FILE *f; /* input file */
char *p; /* temporary variable */
int o; /* overwrite flag */
int r; /* temporary variable */
int s; /* true if no names given */
/* No temporary file yet (for err()) */
*mname = 0;
/* No subject prefix yet */
*mprefix = 0;
/* See if help requested */
if (argc > 1 && (strcmp(argv[1], "-h") == 0 || strcmp(argv[1], "-?") == 0))
help();
/* Unship */
if ((p = strrchr(argv[0], PATHCUT)) == NULL)
p = argv[0];
else
p++;
r = 0; /* (make some compilers happier) */
if ((r = strncmp(p, "unship", 6)) == 0 ||
(r = strncmp(p, "UNSHIP", 6)) == 0 ||
(argc > 1 && strcmp(argv[1], "-u") == 0))
{
errname = "unship error";
warname = "unship warning";
r = r ? 2 : 1; /* next arg */
o = 0; /* disallow overwriting */
if (r < argc && strcmp(argv[r], "-o") == 0)
{
r++;
o = 1; /* allow overwriting */
}
if (r < argc && strcmp(argv[r], "-q") == 0)
{
r++;
noisy = 0; /* inhibit messages */
}
if (r < argc)
unship(argv + r, argc - r, o); /* unship files in args */
else
{
char *a[1]; /* short list of names (one) */
a[0] = "-";
unship(a, 1, o); /* no args--do stdin */
}
}
/* Ship */
else
{
mail = 0; /* not mailing */
fast = 0; /* use base 85 encoding */
s = 1; /* no names given yet */
strcpy(sname, "-"); /* output to stdout */
sfile = stdout;
slns = slmax = 0;
for (r = 1; r < argc; r++) /* go through args */
if (argv[r][0] == '-') /* option or stdin */
if (argv[r][1]) /* option */
{
if (argv[r][1] == 'm') /* mail output */
{
mail = 1;
mdest = NULL; /* next arg is mail address */
}
else if (argv[r][1] == 's') /* next arg is subject prefix */
mprefix = NULL;
else if (argv[r][1] == 'f') /* fast arithmetic encoding */
fast = 1;
else if (argv[r][1] == 'q') /* quiet operation */
noisy = 0;
else /* option is number of lines */
{
/* Check numeric option */
for (p = argv[r] + 1; *p; p++)
if (*p < '0' || *p > '9')
break;
if (*p || slmax)
err(SE_ARG, argv[r]);
/* Zero means infinity, else convert */
if ((slmax = atol(argv[r] + 1)) == 0)
slmax = -1L;
else
{
long b;
b = slmax * 1000L;
slmax = (int)(b / (fast ? 81 : 77));
/* Note: five of the lines aren't that long, but that
leaves some slack for mail headers, etc. Also, note
that we conservatively assume 1000 bytes/K and two
bytes per new line. */
}
}
}
else /* input file is stdin */
{
if (mail && mdest == NULL)
err(SE_ARG, "- (no mail destination given)");
s = 0;
if (mail && !*mprefix)
strcpy(mprefix, "(stdin)");
ship("-", stdin);
}
else /* not option or stdin */
if (mail && mdest == NULL) /* arg is mail address */
mdest = argv[r];
else if (mprefix == NULL) /* arg is subject prefix */
mprefix = argv[r];
else /* arg is file to ship */
{
s = 0;
if ((f = fopen(argv[r], "r")) == NULL)
err(SE_FIND, argv[r]);
if (mail && !*mprefix)
{
int i;
for (i = 0, p = nopath(argv[r]); i < 8 && *p; p++)
if ((*p >= '0' && *p <= '9') || (*p >= 'A' && *p <= 'Z') ||
(*p >= 'a' && *p <= 'z') || *p == '.' || *p == '_')
mprefix[i++] = *p;
mprefix[i] = 0;
}
ship(argv[r], f);
fclose(f);
}
if (s) /* no names--act as filter */
if (mail && mdest == NULL)
err(SE_ARG, "-m (no mail destination given)");
else if (mprefix == NULL)
err(SE_ARG, "-s (no subject prefix given)");
else
{
if (mail && !*mprefix)
strcpy(mprefix, "(stdin)");
ship("-", stdin);
}
endship(1); /* clean up */
if (noisy && (mail || slmax))
fprintf(stderr, "file%s%s %s\n",
strcmp("part0001", sname) ? "s part0001.." : " ", sname,
mail ? "mailed" : "written");
}
/* Done */
exit(0);
}
-----------------------------------------------------------
--
Chris Spell sp...@seq.uncwil.edu
Moderator Comp.Sources.HP48 & HP48 archive maintainer