|
|
|
C++
My lib
Some thing constructive for the night. Cleaned up a lot of stuff in my utility/portability lib. Including a bool datatype for when stdbool.h is not around. And if it is or its C++ use the local boolean.
Aslo tried to make it slightly more friendly to pre-ansi compilers. Not that I’ve ever used one or seen one.. have heard of a few people meeting up with one though !
Basically I started it out both out of interest and Windows. I’m a tad lazy, I’d rather use an err( int, string ) then fprintf a string to stderr and append errno plus call exit() — A good 2 or 3 lines usually. When I can do it in 1 or 2 depending on the type of message I want displayed on exit(). BSD/GNU extensions have err()/warn() routines.
Windows of course lacks this totally as far as I can find. So since I was interested for a long time in how to create libraries I looked up how and implemented the err()/errx() I used so much in it. As well as most of the routines I’ve created in a few small programs. I don’t expect to use the thing for much more then building my programs on Windows w/o having to do a lot of search and replace. I just love err() 😉
Wow… just like when I’m sleepy I find a comment like this as I read lol
2230: /*
2231: * If the new process paused because it was
2232: * swapped out, set the stack level to the last call
2233: * to savu(u_ssav). This means that the return
2234: * which is executed immediately after the call to aretu
2235: * actually returns from the last routine which did
2236: * the savu.
2237: *
2238: * You are not expected to understand this.
2239: */
2240: if(rp->p_flag&SSWAP) {
2241: rp->p_flag =& ~SSWAP;
2242: aretu(u.u_ssav);
2243: }
am I to young ?
When I first read this, I figured a call to panic() would likely bring down the system or set it in motion to abort before the frag explodes . Maaaybe I’ve just lived in a world where the Blue Screen Of Death (BSOD, a Microsoft copyright) is prevelent a little to long. The idea that it calls update(“which I have yet to disect”), prints an error message and calls the kernels idle() loop times infinty is neat. Rather then bringing the system to a screeching halt right then and there. At first glance I would’ve expected it to take down the system right bam’n’slam now.
/*
* Panic is called on unresolvable
* fatal errors.
* It syncs, prints "panic: mesg" and
* then loops.
*/
panic(s)
char *s;
{
panicstr = s;
update();
printf("panic: %sn", s);
for(;;)
idle();
}
Of course from the fine book for study. It seems the only way out of a panic() attack is a restart any way, but its not an instant ‘crash, have a nice foo&ing day, rebooting now sucker’ kind of thing. Which interestingly a lot of times my Windows machines have gone down, *cough*. Its kind of been to the BSOD what the expression ‘Do not pass go, do not collect $200’ is to monopoly. It dies, you get an error, but it kills itself before you get to read it. Then tells you it was *probably* a driver issue lol.
Note that the above code snippet is probably under the this license.
It is from UNIX Version 6 as viewed from here. File is /usr/src/sys/kern/prf.c
While I’m to tired to understand why this writes past the end of the array
#include <assert.h>
#define bounds(_i, _n) (assert( (unsigned) (_i) < (_n) ), (_i) )
....
/*
* fout is what we are appending to and should be open rw as necessary.
* farray is an array of file pointers, "farsize" is the size of the array.
*/
void
concatenate( FILE *fout, FILE *farray[], const unsigned char *farsize) {
int c;
unsigned int fi = 1;
FILE *fin; /* current segment of farray[] to fout. */
while ( (fi < farsize[0])) {
fin = farray[bounds(fi, *farsize)];
fi++;
while ( (c = fgetc(fin )) != EOF ) {
fputc( c, fout );
}
}
}
I’m GOING TO BED. 0832 Zulu time on my clock…. must sleeep sleep….
what is that again?
Well since Windows lacks fgetln() but as far as I can tell the BSD and GNU C Libraries have it.
/* The origenal, basically the same as one of the functions in FreeBSDs head(1) implementation.
static void
read_top( FILE *fp, int lncnt ) {
char *cp;
size_t error, rlen;
while ( lncnt && (cp = fgetln( fp, &rlen )) != NULL ) {
error = fwrite( cp, sizeof(char), rlen, stdout );
if ( error != rlen )
err( 1, "stdout" );
lncnt--;
}
}
/* A new version of read_top that *should* be a bit more standard compilent and simple to read */
static void
read_top( FILE *fp, int lncnt ) {
int c = 0;
int f = 0;
while ( (c < lncnt) && (f != EOF ) ) {
f = fgetc( fp );
printf( "%c", f );
if ( (f == 'n') ) {
c++;
}
}
}
The biggest problem with rf.c is the use of err()/errx(), a nice pair of Macros that expand to a fairly compatible meaning would be nice for portability. *cough* Windows *cough* and I don’t even want to know how groff works on WIN32 if at all….
10 Thou shalt foreswear, renounce, and abjure the vile heresy which claimeth that “All the world’s a VAX”, and have no commerce with the benighted heathens who cling to this barbarous belief, that the days of thy program may be long even though the days of thy current machine be short.
This particular heresy bids fair to be replaced by “All the world’s a Sun” or “All the world’s a 386” (this latter being a particularly revolting invention of Satan), but the words apply to all such without limitation. Beware, in particular, of the subtle and terrible “All the world’s a 32-bit machine”, which is almost true today but shall cease to be so before thy resume grows too much longer.
This some times reminds me of how I feel about assumtions. Uusally met with me thinking.
“I’m not running Windows, !@#$ you”
“I’m not running Linux, !@#$ you”
“I’m not using MS-DOS, !@#$ off!”
“I’m not working in Visual BASIC or Visual C++, really un-!@#$ you”
“What if I’m not on a 32-Bit system, oh go !@#$ your self”
lol.
rf.c semi-final
Well to make a long story short. I found out that calling setlocale() made my program (rf) crash when ever I used to -b switch. But strangely on OpenBSD it ran perfectly fine.
Sure enough after 2 days of tracking I found a few bugs and the little blighter I’ve just squashed was just the one I expected to find. It was just a matter of how. Just a bad pointer, my guess is after the setlocale() call. It was pointing some where FreeBSD didn’t like and without the call it wasn’t enough to get ballocked at. Who cares how thats possible… as long as its fixed hehe. Well as best as _I_ can tell any way.
The projects grown and I’ve learned quite since the start. Both about working in C, using Makefiles, dealing with GCC, GDB in more depth, and groff e.t.c. For sear size we are at:
#wc [ lines | words | bytes | filename ]
16 121 778 COPYING
25 68 735 Makefile.optimize
255 1023 6517 rf.c
11 114 3636 rf.o
87 333 1819 rf.1
9 68 404 tags
403 1727 13889 total
I have 3 Makefiles, a generic makefile for building the application on i686. Makefile.debug that does likewise but includes debugging symbols. And Makefile.optimize that has a few extra flags with mostly optimization options in it (hence the name). Each also can do a ‘make install’ for me. All in all gcc gets called from Makefile.optimize with:
-Wall -Wpointer-arith -Wcast-qual -Wcast-align -Wconversion -Waggregate-return -Wstrict-prototypes -Wmissing-prototypes -Wmissing-declarations -Wredundant-decls -Winline -Wnested-externs -std=c99 -march=i686 -fforce-mem -fforce-addr -finline-functions -fstrength-reduce -floop-optimize
I made some other changes too. I rewrote parts of main, its still not pretty but a wrapper on fopen() I think makes it easier to dick with. I finally added a -v option to see the file name. I was trying to figure out how to make a variadic macro and stumbled onto the __FILE__ and __LINE__ additions to C hehehe. I also decided to have it clean up after its malloc’ing. AFAIK a programs memory is free()’d after it exits so I didn’t bother. Now I did for completeness sake. clean_up() is called after handling the output done in read_bottom (which actually uses read_all() to printout the file). So that if for some odd reason theres really low memory or some thing and it takes *awhile* to handle malloc’ing and seeking semi-end-to-end through a linked-list. I figured it’d be a good ‘just in case’ even if its not likely.
I really would like to rewrite read_top() b/c I don’t know of any comparable function to fgetln() on windows and a Macro that inlines a rough equivalent to my beloved err() function too. Which likewise is available from the BSD and GNU C Libraries and both are marked as ‘first appeared in 4.4BSD’ in the man pages. And AFAICT not available from Microsofts implementation.
One thing I need to do is learn more about groff and the man page macros and stuff. So far its mostly been a matter of reading my systems man pages and trial/error to get a fairly decent manpage.
I try to keep things small, my displays usually 80×25 or 80×35 if I’m working on Windows (GVIM) and 1 thing I really hate. Is a huge function thats so freaking big, you don’t even remember the NAME of it by the time you hit the closing bra}e. So lol, I do my best not to write functions that long. At 60 some lines with comments I think read_bottom should be broken up into 2 smaller functions, the extra function call is negiable overhead I’m sure. I just think it would take longer to read. Really if you snip off the variable declarations and list startup (which could be function’ized) it almost fits on 1 screen. be_verbose man as well be a macro or inline function too. Although a compiler could take care of that choice it self I’m sure.
To be honest, I just like functions that are small, do whats asked, and avoid being swiss-army knifes.
/*-
* Copyright (C) 2007
* Terry M. P*****. All rights reserved.
*
* permission to use, copy, modify, and distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
// vim: set noexpandtab ts=8 sw=4 ai :
// vi: set ai nu ts=8 sw=4 :
#include <err.h>
#include <stdio.h>
#include <stdlib.h>
#include <limits.h>
#include <locale.h>
#include <unistd.h>
struct lnpos {
long nl;
struct lnpos *next;
} lnpos = { .next = 0, .nl = 0 };
static FILE *open_file( char * );
static void read_all( FILE *, int );
static void read_top( FILE *, int );
static void read_bottom( FILE *, int );
static void clean_up( struct lnpos *);
static void be_verbose( void );
static void usage( void );
static const char *this_progname;
/*
* rf - read file to standard out v1.27
*/
int
main( int argc, char *argv[] ) {
char *erptr;
int t_flag = 0, b_flag = 0, v_flag = 0;
int ch, lncnt;
FILE *fp;
(void)setlocale( LC_ALL, "" );
this_progname = argv[0];
while ( (ch = getopt(argc, argv, "b:t:v")) != -1 ) {
switch (ch) {
case 'b':
b_flag++;
lncnt = strtol(optarg, &erptr, 10);
if ( *erptr || lncnt <= 0 )
errx( 1, "Improper line count -- %s", optarg );
break;
case 't':
t_flag++;
lncnt = strtol(optarg, &erptr, 10);
if ( *erptr || lncnt <= 0 )
errx( 1, "Improper line count -- %s", optarg );
break;
case 'v':
v_flag++;
break;
case '?':
default:
usage();
/* NOTREACHED */
}
}
/* This is ugly but cleaner then w/o open_file(). */
if ( argv[1] == NULL ) {
usage();
} else if ( (t_flag < 1) && (b_flag < 1) && (v_flag < 1) ) {
fp = open_file( argv[1] );
read_all( fp, lncnt );
} else if ( (v_flag != 0) && (t_flag < 1) && (b_flag < 1) ) {
fp = open_file( argv[2] );
read_all( fp, lncnt );
} else if ( t_flag > 0 ) {
fp = open_file( argv[3] );
read_top( fp, lncnt );
} else if ( b_flag > 0 ) {
fp = open_file( argv[3] );
read_bottom( fp, lncnt );
} else {
usage();
/* NOTREACHED */
}
if ( v_flag != 0 )
be_verbose();
fclose( fp );
return 0;
}
/* Simple fopen wrapper to keep the if...else if...else blockage from
* getting even uglier. Since doing other wise would defeat the purpose of it.
* open_file() halts the program if fopen failed.
*/
static FILE
*open_file( char *arg ) {
FILE *fto;
fto = fopen( arg, "r" );
if ( fto == NULL )
errx( 1, "File can not be opened or"
" does not exist -- %sn", arg );
return fto;
}
/*
* print out an open file to standard output
*/
static void
read_all( FILE *fp, int lncnt ) {
while ( (lncnt = fgetc( fp )) != EOF ) {
printf( "%c", lncnt );
}
}
/* Read n lines from the top of the file.
* note that it was very inspired by the head(1) implementation of BSD
*/
static void
read_top( FILE *fp, int lncnt ) {
char *cp;
size_t error, rlen;
while ( lncnt && (cp = fgetln( fp, &rlen )) != NULL ) {
error = fwrite( cp, sizeof(char), rlen, stdout );
if ( error != rlen )
err( 1, "stdout" );
lncnt--;
}
}
/* Read n lines from the bottom of the file
*/
static void
read_bottom( FILE *fp, int lncnt ) {
int hmany = lncnt;
long nlnum = 0;
long where;
struct lnpos *root = 0;
struct lnpos *cur = 0;
root = malloc( sizeof(struct lnpos) );
if ( root == NULL )
err( 1, "can't init the list" );
root->next = 0;
cur = root;
cur->next = malloc( sizeof(struct lnpos) );
if ( cur->next == NULL )
err( 1, "can't add first node" );
cur = cur->next;
/* read the file, count every 'n' and store them in a new member of
* our linked list.
*/
while ( (lncnt = fgetc( fp )) != EOF ) {
if ( lncnt == 'n' ) {
nlnum++;
cur->nl = ftell( fp );
cur->next = malloc( sizeof(struct lnpos) );
if ( cur->next == NULL )
err( 1, "can't store line feeds" );
cur = cur->next;
}
}
/* Let's mark the end of the list and move to it */
cur->next = malloc( sizeof(struct lnpos) );
if ( cur->next == NULL )
err( 1, "can't terminate the list" );
cur = cur->next;
cur->next = NULL;
/* rewind our linked-list and seek to b_flag segments. So readall() starts from
* the correct fseek offset to print till EOF. - This keeps down on
* unnecessary code here !
*/
cur = root;
while ( hmany < nlnum ) {
cur = cur->next;
nlnum--;
}
where = fseek( fp, cur->nl, SEEK_SET );
if ( where != 0 )
err( 1, "could not seek through the filen" );
read_all( fp, lncnt );
clean_up( root );
}
/* Simple destructor - walk the list and free() the memory before the
* program exits.
*/
static void
clean_up( struct lnpos *rp ) {
struct lnpos *t = rp;
struct lnpos *atpos = rp;
while ( atpos != NULL ) {
free( t );
atpos = atpos->next;
t = atpos;
}
}
static void
be_verbose( void ) {
printf("==> %s <==n", __FILE__);
}
static void
usage( void ) {
(void)fprintf( stderr, "usage: %s [-t count | -b count] "
"[-v] [file ...]n",
this_progname );
exit( EXIT_FAILURE );
}
I’ve learned so much in the past 2 months, its awesome. Really this is the largest project I’ve ever worked on in my ~/Programming directory. 255 Lines isn’t much but when your a bumbling baboon trying to learn in the wilderness. It kinda helps to learn by reading and writing code I guess.
I also want to start assembling a little personal library of functions and macros that I can draw on to do odds and ends without rewriting them. After all with my little ‘toys’ I might have use of other stuff from time to time hehe.
Time for BEEEEEEEEEEEEEEEEEEEEEEEEEEDDDDDDDDDDDDDDD!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
rf.c
Well I’ve done some correcting to my little back scratcher, including two bug fixes and changes to the man page.
/*-
* Copyright (C) 2007
* Terry *. P*****. All rights reserved.
*
* permission to use, copy, modify, and distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
// vim: set noexpandtab ts=8 sw=4 ai :
// vi: set ai nu ts=8 sw=4 :
#include <err.h>
#include <stdio.h>
#include <stdlib.h>
#include <limits.h>
#include <unistd.h>
struct lnpos {
long nl;
struct lnpos *next;
};
static void readall( FILE *, int );
static void readtop( FILE *, int );
static void readbottom( FILE *, int );
static void usage( void );
/*
* rf - read file to standard out v1.1
*/
int
main( int argc, char *argv[] ) {
char *erptr;
int t_flag = 0, b_flag = 0;
int ch, lncnt;
FILE *fp;
extern char *optarg;
extern int optind;
extern int optopt;
extern int opterr;
extern int optreset;
while ( (ch = getopt(argc, argv, "b:t:")) != -1 ) {
switch (ch) {
case 'b':
b_flag++;
lncnt = strtol(optarg, &erptr, 10);
if ( *erptr || lncnt <= 0 )
errx( 1, "Improper line count -- %s", optarg );
break;
case 't':
t_flag++;
lncnt = strtol(optarg, &erptr, 10);
if ( *erptr || lncnt <= 0 )
errx( 1, "Improper line count -- %s", optarg );
break;
case '?':
default:
usage();
return 1;
}
}
if ( (t_flag < 1) && (b_flag < 1) ) {
fp = fopen( argv[1], "r" );
if ( fp == NULL )
errx( 1, "File can not be opened or"
" does not exist -- %sn", argv[1] );
readall( fp, lncnt );
} else if ( t_flag > 0 ) {
fp = fopen( argv[3], "r" );
if ( fp == NULL )
errx( 1, "File can not be opened or"
" does not exist -- %sn", argv[3] );
readtop( fp, lncnt );
} else if ( b_flag > 0 ) {
fp = fopen( argv[3], "r" );
if ( fp == NULL )
errx( 1, "File can not be opened or"
" does not exist -- %sn", argv[3] );
readbottom( fp, lncnt );
} else {
usage();
/* NOTREACHED */
}
fclose( fp );
return 0;
}
/*
* print out an open file to standard output
*/
static void
readall( FILE *fp, int lncnt ) {
while ( (lncnt = fgetc( fp )) != EOF ) {
printf( "%c", lncnt );
}
}
/* Read n lines from the top of the file.
* note that it was very inspired by the head(1) implementation of BSD
*/
static void
readtop( FILE *fp, int lncnt ) {
char *cp;
size_t error, rlen;
while ( lncnt && (cp = fgetln( fp, &rlen )) != NULL ) {
error = fwrite( cp, sizeof(char), rlen, stdout );
if ( error != rlen )
err( 1, "stdout" );
lncnt--;
}
}
/* Read n lines from the bottom of the file
*/
static void
readbottom( FILE *fp, int lncnt ) {
int hmany = lncnt;
long nlnum = 0;
long where;
struct lnpos *root;
struct lnpos *cur;
root = malloc( sizeof(struct lnpos) );
if ( root == NULL )
err( 1, "can't init the list" );
root->next = 0;
cur = root;
cur->next = malloc( sizeof(struct lnpos) );
if ( cur->next == NULL )
err( 1, "can't add nodes" );
cur = cur->next;
/* read the file, count every 'n' and store them in a new member of
* our linked list.
*/
while ( (lncnt = fgetc( fp )) != EOF ) {
if ( lncnt == 'n' ) {
nlnum++;
cur->nl = ftell( fp );
if ( cur->next != NULL )
cur = cur->next;
cur->next = malloc( sizeof(struct lnpos) );
if ( cur->next == NULL )
err( 1, "can't add nodes" );
cur = cur->next;
}
}
/* rewind our linked-list and seek to b_flag + 1 segments short of the
* end of the list _before_ calling readall. So readall _starts_ from
* the correct fseek offset to print till EOF.
*/
cur = root->next;
hmany++;
while ( hmany < nlnum ) {
cur = cur->next;
nlnum--;
}
where = fseek( fp, cur->nl, SEEK_SET );
if ( where != 0 )
err( 1, "could not seek through the filen" );
readall( fp, lncnt );
}
static void
usage( void ) {
fprintf( stderr, "usage:n rf filen "
"rf [-t number lines from the top] filen "
"rf [ -b number of lines from the bottom ] [ file ]n" );
}
And the manual which writing got me interested in roff, I’ll post when I have time to learn more about it. Basically the problem I had was caused by working on it. Originally under the conditions when 2 + 2 = 5 which is not good, because 2 + 2 is really == 4 !!!! I should never do stuff when I’m like that, but if I don’t try nothing gets done.
read file
This is a little thing I’ve been working on this month. Its made to simulate the behavior of ‘cat file’ and ‘head -n’ / ‘tail -n’ more or less scratch a little itch. It works very simple as its supposed to be simple. It parses arguments taking a count to its -t and -b switches. It looks through argv till it finds a file to open. It’ll read the entire file out, if it’s given the -t switch it will read out up to the count its given. If its given the -b switch, it will read the entire file. Store the location of every newline in a linked-list. Then rewind the list, seek to the specified location (plus padding) and print out the end of the remainder of the file.
Heres a straight copy & paste of it. Portability wise it assumes that the file is in the correct format and requires a number of functions that first appeared in BSD 4.x, namely getopts, fgetln, and err/errx. It also presumes that their should be enough room to store the location of every newline in the file within memory. So if one only has 64K of RAM, it might be a problem to read a few files xD.
I can’t say if its well written or not… I don’t know of any program with my system that does exactly this other then what this simulates. Rather then wrap around cat/head/tail using system() or including the sources I worked through this as a learning experience. Its more or less written in the style I prefer, the type on a line of its own bit in function declarations is the only ‘real’ part of it stemming from my occasional reads of stuff in /usr/src alone.
/*
* rf read file to standard out
*/
// vim: set noexpandtab ts=8 sw=4 ai :
// vi: set ai nu ts=8 sw=4 :
#include <err.h>
#include <stdio.h>
#include <stdlib.h>
#include <limits.h>
#include <unistd.h>
static void readall( FILE *, int );
static void readtop( FILE *, int );
static int readbottom( FILE *, int );
static void usage( void );
struct lnpos {
long nl;
struct lnpos *next;
};
int
main( int argc, char *argv[] ) {
char *erptr;
int t_flag = 0, b_flag = 0;
int ch, lncnt;
FILE *fp;
extern char *optarg;
extern int optind;
extern int optopt;
extern int opterr;
extern int optreset;
while ( (ch = getopt(argc, argv, "b:t:")) != -1 ) {
switch (ch) {
case 'b':
b_flag++;
lncnt = strtol(optarg, &erptr, 10);
if ( *erptr || lncnt <= 0 )
errx( 1, "Improper line count -- %s", optarg );
break;
case 't':
t_flag++;
lncnt = strtol(optarg, &erptr, 10);
if ( *erptr || lncnt <= 0 )
errx( 1, "Improper line count -- %s", optarg );
break;
case '?':
default:
usage();
return 1;
}
argc -= optind;
argv += optind;
}
/* loop through cli args and find a file to open */
for ( int i = 0; i <= argc; i++ ) {
fp = fopen( argv[i], "r" );
if ( fp == NULL )
err( 1, "File does not exist or could not be opened "
"for reading -- %s", optarg );
}
if ( (t_flag < 1) && (b_flag < 1) ) {
readall( fp, lncnt );
} else if ( t_flag > 0 ) {
readtop( fp, lncnt );
} else if ( b_flag > 0 ) {
readbottom( fp, lncnt );
} else {
errx( 1, "flag processing failed" );
}
fclose( fp );
return 0;
}
/*
* print out an open file to standard output
*/
static void
readall( FILE *fp, int lncnt ) {
while ( (lncnt = fgetc( fp )) != EOF ) {
printf( "%c", lncnt );
}
}
/* Read n lines from the top of the file.
* note that it was very inspired by the head(1) implementation of BSD
*/
static void
readtop( FILE *fp, int lncnt ) {
char *cp;
size_t error, rlen;
while ( lncnt && (cp = fgetln( fp, &rlen )) != NULL ) {
error = fwrite( cp, sizeof(char), rlen, stdout );
if ( error != rlen )
err( 1, "stdout" );
lncnt--;
}
}
/* Read n lines from the bottom of the file
* This function really should be broken up - but I have not learned how yet.
*/
static int
readbottom( FILE *fp, int lncnt ) {
char *cp;
int hmany = lncnt;
long nlnum = 0;
long where;
size_t error, rlen;
struct lnpos *root;
struct lnpos *cur;
root = malloc( sizeof(struct lnpos) );
if ( root == NULL )
err( 1, "can't init the list" );
root->next = 0;
cur = root;
cur->next = malloc( sizeof(struct lnpos) );
if ( cur->next == NULL )
err( 1, "can't add nodes" );
cur = cur->next;
/* read the file, count every 'n' and store them in a new member of
* our linked list.
*/
while ( (lncnt = fgetc( fp )) != EOF ) {
if ( lncnt == 'n' ) {
nlnum++;
cur->nl = ftell( fp );
if ( cur->next != NULL )
cur = cur->next;
cur->next = malloc( sizeof(struct lnpos) );
if ( cur->next == NULL )
err( 1, "can't add nodes" );
cur = cur->next;
}
}
/* rewind our linked-list and seek b_flag + 1 segments short of the
* end of the list.
*/
cur = root->next;
hmany++;
while ( hmany < nlnum ) {
cur = cur->next;
nlnum--;
}
where = fseek( fp, cur->nl, SEEK_SET );
if ( where != 0 )
err( 1, "could not seek through the filen" );
while ( lncnt && (cp = fgetln( fp, &rlen )) != NULL ) {
error = fwrite( cp, sizeof(char), rlen, stdout );
if ( error != rlen )
err( 1, "stdout" );
lncnt--;
cur = cur->next;
if ( cur->next == 0 )
return 0;
}
return 0;
}
static void
usage( void ) {
fprintf( stderr, "usage:n rf filen "
"rf [-t number lines from the top] filen "
"rf [ -b number of lines from the bottom ] [ file ]n" );
}
Any comments about the file it self or its design is much welcome by this moron who tries to teach him self.