Reorganise unarchiving functions, more code re-use, only does single pass(no more linked lists), basis for supporting a cpio (and cheaper untar) applet, but cpio applet isnt included in this.

It effects ar, dpkg-deb applets only

10 files changed, 518 insertions, 623 deletions

diff --git a/libbb/add_from_archive_list.c b/libbb/add_from_archive_list.c
deleted file mode 100644
index 7e4e071..0000000
--- a/libbb/add_from_archive_list.c
+++ /dev/null
@@ -1,31 +0,0 @@
-/*
- *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
- *
- *  This program is distributed in the hope that it will be useful,
- *  but WITHOUT ANY WARRANTY; without even the implied warranty of
- *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- *  GNU Library General Public License for more details.
- *
- *  You should have received a copy of the GNU General Public License
- *  along with this program; if not, write to the Free Software
- *  Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
- */
-
-#include "libbb.h"
-
-file_headers_t *add_from_archive_list(file_headers_t *master_list, file_headers_t *new_list, const char *filename)
-{
-	file_headers_t *list_ptr;
-
-	list_ptr = master_list;
-	while (list_ptr != NULL) {
-		if (strcmp(filename, list_ptr->name) == 0) {
-			return(append_archive_list(new_list, list_ptr));
-		}
-		list_ptr = list_ptr->next;
-	}
-	return(NULL);
-}
\ No newline at end of file
diff --git a/libbb/append_archive_list.c b/libbb/append_archive_list.c
deleted file mode 100644
index c955d08..0000000
--- a/libbb/append_archive_list.c
+++ /dev/null
@@ -1,39 +0,0 @@
-/*
- *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
- *
- *  This program is distributed in the hope that it will be useful,
- *  but WITHOUT ANY WARRANTY; without even the implied warranty of
- *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- *  GNU Library General Public License for more details.
- *
- *  You should have received a copy of the GNU General Public License
- *  along with this program; if not, write to the Free Software
- *  Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
- */
-
-#include <stdlib.h>
-#include "libbb.h"
-
-file_headers_t *append_archive_list(file_headers_t *head, file_headers_t *tail_entry)
-{
-	file_headers_t *tail_ptr;
-	file_headers_t *new_node = (file_headers_t *) malloc(sizeof(file_headers_t));
-
-	*new_node = *tail_entry;
-	new_node->next = NULL;
-
-	if (head->name == NULL) {
-		head = new_node;
-	} else {
-		tail_ptr = head;
-		while(tail_ptr->next != NULL) {
-			tail_ptr = tail_ptr->next;
-		}
-		tail_ptr->next = new_node;
-	}
-
-	return(head);
-}
\ No newline at end of file
diff --git a/libbb/deb_extract.c b/libbb/deb_extract.c
deleted file mode 100644
index f641939..0000000
--- a/libbb/deb_extract.c
+++ /dev/null
@@ -1,124 +0,0 @@
-/* vi: set sw=4 ts=4: */
-/*
- * Copyright (C) tons of folks.  Tracking down who wrote what
- * isn't something I'm going to worry about...  If you wrote something
- * here, please feel free to acknowledge your work.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
- *
- * Based in part on code from sash, Copyright (c) 1999 by David I. Bell 
- * Permission has been granted to redistribute this code under the GPL.
- *
- */
-
-#include <fcntl.h>
-#include <unistd.h>
-#include <stdlib.h>
-#include <string.h>
-#include "libbb.h"
-
-int seek_sub_file(FILE *in_file, file_headers_t *headers, const char *tar_gz_file)
-{
-	/* find the headers for the specified .tar.gz file */
-	while (headers != NULL) {
-		if (strcmp(headers->name, tar_gz_file) == 0) {
-			fseek(in_file, headers->offset, SEEK_SET);
-			return(EXIT_SUCCESS);
-		}
-		headers = headers->next;
-	}
-
-	return(EXIT_FAILURE);
-}
-
-/*
- * The contents of argument depend on the value of function.
- * It is either a dir name or a control file or field name(see dpkg_deb.c)
- */
-char *deb_extract(const char *package_filename, FILE *out_stream, const int extract_function,
-	const char *prefix, const char *filename)
-{
-	FILE *deb_stream, *uncompressed_stream;
-	file_headers_t *ar_headers = NULL;
-	file_headers_t *tar_headers = NULL;
-	file_headers_t *list_ptr;
-	file_headers_t *deb_extract_list = (file_headers_t *) calloc(1, sizeof(file_headers_t));
-
-	char *ared_file = NULL;
-	char *output_buffer = NULL;
-	int gunzip_pid;
-
-	if (extract_function & extract_control_tar_gz) {
-		ared_file = xstrdup("control.tar.gz");
-	}
-	else if (extract_function & extract_data_tar_gz) {		
-		ared_file = xstrdup("data.tar.gz");
-	}
-
-	/* open the debian package to be worked on */
-	deb_stream = wfopen(package_filename, "r");
-
-	ar_headers = (file_headers_t *) xmalloc(sizeof(file_headers_t));	
-	
-	/* get a linked list of all ar entries */
-	ar_headers = get_ar_headers(deb_stream);
-	if (ar_headers == NULL) {
-		error_msg("Couldnt get ar headers\n");
-		return(NULL);
-	}
-
-	/* seek to the start of the .tar.gz file within the ar file*/
-	fseek(deb_stream, 0, SEEK_SET);
-	if (seek_sub_file(deb_stream, ar_headers, ared_file) == EXIT_FAILURE) {
-		error_msg("Couldnt seek to file %s", ared_file);
-	}
-
-	/* get a linked list of all tar entries */
-	tar_headers = get_tar_gz_headers(deb_stream);
-	if (tar_headers == NULL) {
-		error_msg("Couldnt get tar headers\n");
-		return(NULL);
-	}
-
-	if (extract_function & extract_one_to_buffer) {
-		list_ptr = tar_headers;
-		while (list_ptr != NULL) {
-			if (strcmp(filename, list_ptr->name) == 0) {
-				deb_extract_list = append_archive_list(deb_extract_list, list_ptr);
-				break;
-			}
-			list_ptr = list_ptr->next;
-		}
-	} else {
-		deb_extract_list = tar_headers;
-	}
-
-	/* seek to the start of the .tar.gz file within the ar file*/
-	if (seek_sub_file(deb_stream, ar_headers, ared_file) == EXIT_FAILURE) {
-		error_msg("Couldnt seek to ar file");
-	}
-
-	/* open a stream of decompressed data */
-	uncompressed_stream = fdopen(gz_open(deb_stream, &gunzip_pid), "r");
-
-	output_buffer = extract_archive(uncompressed_stream, out_stream, deb_extract_list, extract_function, prefix);
-
-	gz_close(gunzip_pid);
-	fclose(deb_stream);
-	fclose(uncompressed_stream);
-	free(ared_file);
-
-	return(output_buffer);
-}
\ No newline at end of file
diff --git a/libbb/extract_archive.c b/libbb/extract_archive.c
deleted file mode 100644
index f22cbb6..0000000
--- a/libbb/extract_archive.c
+++ /dev/null
@@ -1,132 +0,0 @@
-/* 
- *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
- *
- *  This program is distributed in the hope that it will be useful,
- *  but WITHOUT ANY WARRANTY; without even the implied warranty of
- *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- *  GNU Library General Public License for more details.
- *
- *  You should have received a copy of the GNU General Public License
- *  along with this program; if not, write to the Free Software
- *  Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
- */
-
-#include <stdlib.h>
-#include <string.h>
-#include <time.h>
-#include <unistd.h>
-#include <utime.h>
-#include "libbb.h"
-
-/* Extract files in the linear linked list 'extract_headers' to either
- * filesystem, stdout or buffer depending on the value of 'function' which is
- * described in extract_files.h 
- *
- * prefix doesnt have to be just a directory, it may prefix the filename as well.
- *
- * e.g. '/var/lib/dpkg/info/dpkg.' will extract all files to the base bath 
- * '/var/lib/dpkg/info/' and all files/dirs created in that dir will have 
- * 'dpkg.' as their prefix
- *
- * For this reason if prefix does point to a dir then it must end with a
- * trailing '/' or else the last dir will be assumed to be the file prefix 
- */
-
-char *extract_archive(FILE *src_stream, FILE *out_stream, file_headers_t *extract_headers, int function, const char *prefix)
-{
-	FILE *dst_stream = NULL;
-	char *full_name = NULL;
-	char *buffer = NULL;
-	size_t uncompressed_count = 0;
-	struct utimbuf t;
-
-	/* find files to extract or display */	
-	while (extract_headers != NULL) {
-		/* prefix doesnt have to be a proper path it may prepend 
-		 * the filename as well */
-		if (prefix != NULL) {
-			/* strip leading '/' in filename to extract as prefix may not be dir */
-			char *tmp_str = strchr(extract_headers->name, '/');
-			if (tmp_str == NULL) {
-				tmp_str = extract_headers->name;
-			} else {
-				tmp_str++;
-			}
-			/* Cant use concat_path_file here as prefix might not be a directory */
-			full_name = xmalloc(strlen(prefix) + strlen(tmp_str) + 1);
-			strcpy(full_name, prefix);
-			strcat(full_name, tmp_str);
-		} else {
-			full_name = extract_headers->name;
-		}
-
-		/* Seek to start of file, have to use fgetc as src_stream may be a pipe
-		 * you cant [lf]seek on pipes */
-		while (uncompressed_count < extract_headers->offset) {
-			fgetc(src_stream);
-			uncompressed_count++;
-		}
-
-		if (S_ISREG(extract_headers->mode)) {
-			if (function & extract_to_stdout) {
-				copy_file_chunk(src_stream, out_stream, extract_headers->size);			
-				uncompressed_count += extract_headers->size;
-			}
-			else if (function & extract_all_to_fs) {
-				dst_stream = wfopen(full_name, "w");
-				copy_file_chunk(src_stream, dst_stream, extract_headers->size);			
-				uncompressed_count += extract_headers->size;
-				fclose(dst_stream);
-			}
-			else if (function & extract_one_to_buffer) {
-				buffer = (char *) xmalloc(extract_headers->size + 1);
-				fread(buffer, 1, extract_headers->size, src_stream);
-				break;
-			}
-		}
-#if defined BB_DPKG | defined BB_DPKG_DEB 
-		else if (S_ISDIR(extract_headers->mode)) {
-			if (function & extract_all_to_fs) {
-				if (create_path(full_name, extract_headers->mode) == FALSE) {
-					perror_msg("Cannot mkdir %s", extract_headers->name); 
-					return NULL;
-				}
-			}
-		}
-		else if (S_ISLNK(extract_headers->mode)) {
-			if (function & extract_all_to_fs) {
-				if (symlink(extract_headers->link_name, full_name) < 0) {
-					perror_msg("Cannot create hard link from %s to '%s'", extract_headers->name, extract_headers->link_name); 
-					return NULL;
-				}
-			}
-		}
-#endif
-		if (function & extract_verbose_list) {
-			fprintf(out_stream, "%s %d/%d %8d %s ", mode_string(extract_headers->mode), 
-				extract_headers->uid, extract_headers->gid,
-				(int) extract_headers->size, time_string(extract_headers->mtime));
-		}
-
-		if ((function & extract_list) || (function & extract_verbose_list)){
-			/* fputs doesnt add a trailing \n, so use fprintf */
-			fprintf(out_stream, "%s\n", extract_headers->name);
-		}
-
-		if (function & extract_preserve_date) {
-			t.actime = extract_headers->mtime;
-			t.modtime = extract_headers->mtime;
-			utime(full_name, &t);
-		}
-		chmod(full_name, extract_headers->mode);
-		free(full_name);
-
-		extract_headers = extract_headers->next;
-	}
-
-	return(buffer);
-}
-
diff --git a/libbb/get_ar_headers.c b/libbb/get_ar_headers.c
deleted file mode 100644
index 414498b..0000000
--- a/libbb/get_ar_headers.c
+++ /dev/null
@@ -1,119 +0,0 @@
-/* vi: set sw=4 ts=4: */
-/*
- * Utility routines.
- *
- * Copyright (C) tons of folks.  Tracking down who wrote what
- * isn't something I'm going to worry about...  If you wrote something
- * here, please feel free to acknowledge your work.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
- *
- * Based in part on code from sash, Copyright (c) 1999 by David I. Bell 
- * Permission has been granted to redistribute this code under the GPL.
- *
- */
-
-#include <stdlib.h>
-#include <string.h>
-#include <unistd.h>
-#include "libbb.h"
-
-file_headers_t *get_ar_headers(FILE *in_file)
-{
-	typedef struct raw_ar_header_s {	/* Byte Offset */
-		char name[16];	/*  0-15 */
-		char date[12];	/* 16-27 */
-		char uid[6];	
-		char gid[6];	/* 28-39 */
-		char mode[8];	/* 40-47 */
-		char size[10];	/* 48-57 */
-		char fmag[2];	/* 58-59 */
-	} raw_ar_header_t;
-
-	raw_ar_header_t raw_ar_header;
-
-	file_headers_t *ar_list, *ar_entry;
-	char ar_magic[8];
-	char *long_name=NULL;
-	
-	/* check ar magic */
-	if (fread(ar_magic, 1, 8, in_file) != 8) {
-		error_msg("cannot read magic");
-		return(NULL);
-	}
-
-	if (strncmp(ar_magic,"!<arch>",7) != 0) {
-		error_msg("invalid magic");
-		return(NULL);
-	}
-	
-	ar_list = (file_headers_t *) xcalloc(1, sizeof(file_headers_t));
-
-	while (fread((char *) &raw_ar_header, 1, 60, in_file) == 60) {
-		ar_entry = (file_headers_t *) xcalloc(1, sizeof(file_headers_t));
-		/* check the end of header markers are valid */
-		if ((raw_ar_header.fmag[0] != '`') || (raw_ar_header.fmag[1] != '\n')) {
-			char newline;
-			if (raw_ar_header.fmag[1] != '`') {
-				break;
-			}
-			/* some version of ar, have an extra '\n' after each entry */
-			fread(&newline, 1, 1, in_file);
-			if (newline!='\n') {
-				break;
-			}
-			/* fix up the header, we started reading 1 byte too early due to a '\n' */
-			memmove((char *) &raw_ar_header, (char *)&raw_ar_header+1, 59);
-			/* dont worry about adding the last '\n', we dont need it now */
-		}
-		
-		ar_entry->size = (size_t) atoi(raw_ar_header.size);
-		/* long filenames have '/' as the first character */
-		if (raw_ar_header.name[0] == '/') {
-			if (raw_ar_header.name[1] == '/') {
-				/* multiple long filenames are stored as data in one entry */
-				long_name = (char *) xrealloc(long_name, ar_entry->size);
-				fread(long_name, 1, ar_entry->size, in_file);
-				continue;
-			}
-			else {
-				/* The number after the '/' indicates the offset in the ar data section
-					(saved in variable long_name) that conatains the real filename */
-				const int long_name_offset = (int) atoi((char *) &raw_ar_header.name[1]);
-				ar_entry->name = xmalloc(strlen(&long_name[long_name_offset]));
-				strcpy(ar_entry->name, &long_name[long_name_offset]);
-			}
-		}
-		else {
-			/* short filenames */
-			ar_entry->name = xmalloc(16);
-			ar_entry->name = strncpy(ar_entry->name, raw_ar_header.name, 16);
-		}
-		ar_entry->name[strcspn(ar_entry->name, " /")]='\0';
-
-		/* convert the rest of the now valid char header to its typed struct */	
-		parse_mode(raw_ar_header.mode, &ar_entry->mode);
-		ar_entry->mtime = atoi(raw_ar_header.date);
-		ar_entry->uid = atoi(raw_ar_header.uid);
-		ar_entry->gid = atoi(raw_ar_header.gid);
-		ar_entry->offset = ftell(in_file);
-		ar_entry->next = NULL;
-
-		fseek(in_file, (off_t) ar_entry->size, SEEK_CUR);
-
-		ar_list = append_archive_list(ar_list, ar_entry);
-	}
-	return(ar_list);
-}
diff --git a/libbb/get_tar_gz_headers.c b/libbb/get_tar_gz_headers.c
deleted file mode 100644
index fb9b7b6..0000000
--- a/libbb/get_tar_gz_headers.c
+++ /dev/null
@@ -1,39 +0,0 @@
-/*
- *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
- *
- *  This program is distributed in the hope that it will be useful,
- *  but WITHOUT ANY WARRANTY; without even the implied warranty of
- *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- *  GNU Library General Public License for more details.
- *
- *  You should have received a copy of the GNU General Public License
- *  along with this program; if not, write to the Free Software
- *  Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
- */
-#include <stdio.h>
-#include <unistd.h>
-#include "libbb.h"
-
-file_headers_t *get_tar_gz_headers(FILE *compressed_stream)
-{
-	FILE *uncompressed_stream;
-	file_headers_t *tar_headers_tree;
-	int gunzip_pid;
-	int gz_fd ;
-
-	/* open a stream of decompressed data */
-	gz_fd = gz_open(compressed_stream, &gunzip_pid);
-	if (gz_fd == -1) {
-		error_msg("Couldnt initialise gzip stream");
-	}
-	uncompressed_stream = fdopen(gz_fd, "r");
-	tar_headers_tree = get_tar_headers(uncompressed_stream);
-	fclose(uncompressed_stream);
-	close(gz_fd);
-	gz_close(gunzip_pid);
-
-	return(tar_headers_tree);
-}
\ No newline at end of file
diff --git a/libbb/get_tar_headers.c b/libbb/get_tar_headers.c
deleted file mode 100644
index 5ddf4d9..0000000
--- a/libbb/get_tar_headers.c
+++ /dev/null
@@ -1,111 +0,0 @@
-/*
- *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
- *
- *  This program is distributed in the hope that it will be useful,
- *  but WITHOUT ANY WARRANTY; without even the implied warranty of
- *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- *  GNU Library General Public License for more details.
- *
- *  You should have received a copy of the GNU General Public License
- *  along with this program; if not, write to the Free Software
- *  Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
- */
-#include <stdlib.h>
-#include <string.h>
-#include <unistd.h>
-#include "libbb.h"
-
-file_headers_t *get_tar_headers(FILE *tar_stream)
-{
-	typedef struct raw_tar_header {
-        char name[100];               /*   0-99 */
-        char mode[8];                 /* 100-107 */
-        char uid[8];                  /* 108-115 */
-        char gid[8];                  /* 116-123 */
-        char size[12];                /* 124-135 */
-        char mtime[12];               /* 136-147 */
-        char chksum[8];               /* 148-155 */
-        char typeflag;                /* 156-156 */
-        char linkname[100];           /* 157-256 */
-        char magic[6];                /* 257-262 */
-        char version[2];              /* 263-264 */
-        char uname[32];               /* 265-296 */
-        char gname[32];               /* 297-328 */
-        char devmajor[8];             /* 329-336 */
-        char devminor[8];             /* 337-344 */
-        char prefix[155];             /* 345-499 */
-        char padding[12];             /* 500-512 */
-	} raw_tar_header_t;
-
-	raw_tar_header_t raw_tar_header;
-	unsigned char *temp = (unsigned char *) &raw_tar_header;
-	file_headers_t *tar_headers_list = (file_headers_t *) calloc(1, sizeof(file_headers_t));
-	file_headers_t *tar_entry;
-	long i;
-	long next_header_offset = 0;
-	long current_offset = 0;
-
-	while (fread((char *) &raw_tar_header, 1, 512, tar_stream) == 512) {
-		long sum = 0;
-		current_offset += 512;
-
-		/* Check header has valid magic, unfortunately some tar files
-		 * have empty (0'ed) tar entries at the end, which will
-		 * cause this to fail, so fail silently for now
-		 */
-		if (strncmp(raw_tar_header.magic, "ustar", 5) != 0) {
-//			error_msg("invalid magic, [%s]\n", raw_tar_header.magic);
-			break;
-		}
-
-		/* Do checksum on headers */
-        for (i =  0; i < 148 ; i++) {
-			sum += temp[i];
-		}
-        sum += ' ' * 8;
-		for (i =  156; i < 512 ; i++) {
-			sum += temp[i];
-		}
-        if (sum != strtol(raw_tar_header.chksum, NULL, 8)) {
-			error_msg("Invalid tar header checksum");
-			break;
-		}
-
-		/* convert to type'ed variables */
-		tar_entry = xcalloc(1, sizeof(file_headers_t));
-		tar_entry->name = xstrdup(raw_tar_header.name);
-
-		tar_entry->offset = (off_t) current_offset;
-		parse_mode(raw_tar_header.mode, &tar_entry->mode);
-		tar_entry->uid   = strtol(raw_tar_header.uid, NULL, 8);
-		tar_entry->gid   = strtol(raw_tar_header.gid, NULL, 8);
-		tar_entry->size  = strtol(raw_tar_header.size, NULL, 8);
-		tar_entry->mtime = strtol(raw_tar_header.mtime, NULL, 8);
-		tar_entry->link_name  = xstrdup(raw_tar_header.linkname);
-//		tar_entry->devmajor  = strtol(rawHeader->devmajor, NULL, 8);
-//		tar_entry->devminor  = strtol(rawHeader->devminor, NULL, 8);
-
-		tar_headers_list = append_archive_list(tar_headers_list, tar_entry);
-
-		/* start of next header is at */
-		next_header_offset = current_offset + tar_entry->size;
-
-		if (tar_entry->size % 512 != 0) {
-			next_header_offset += (512 - tar_entry->size % 512);
-		}
-		/*
-		 * Seek to start of next block, cant use fseek as unzip() does support it
-		 */
-		while (current_offset < next_header_offset) {
-			if (fgetc(tar_stream) == EOF) {
-				break;
-			}
-			current_offset++;
-		}
-	}
-	return(tar_headers_list);
-}
-
diff --git a/libbb/gz_open.c b/libbb/gz_open.c
index 19ec0a0..b23920b 100644
--- a/libbb/gz_open.c
+++ b/libbb/gz_open.c
@@ -6,17 +6,17 @@
 #include <unistd.h>
 #include "libbb.h"
 
-extern int gz_open(FILE *compressed_file, int *pid)
+extern FILE *gz_open(FILE *compressed_file, int *pid)
 {
 	int unzip_pipe[2];
 
 	if (pipe(unzip_pipe)!=0) {
 		error_msg("pipe error");
-		return(EXIT_FAILURE);
+		return(NULL);
 	}
 	if ((*pid = fork()) == -1) {
 		error_msg("fork failured");
-		return(EXIT_FAILURE);
+		return(NULL);
 	}
 	if (*pid==0) {
 		/* child process */
@@ -27,7 +27,9 @@ extern int gz_open(FILE *compressed_file, int *pid)
 		close(unzip_pipe[1]);
 		exit(EXIT_SUCCESS);
 	}
-	
 	close(unzip_pipe[1]);
-	return(unzip_pipe[0]);
-}
\ No newline at end of file
+	if (unzip_pipe[0] == -1) {
+		error_msg("Couldnt initialise gzip stream");
+	}
+	return(fdopen(unzip_pipe[0], "r"));
+}
diff --git a/libbb/libbb.h b/libbb/libbb.h
index 893a9f0..e42ca9f 100644
--- a/libbb/libbb.h
+++ b/libbb/libbb.h
@@ -213,23 +213,9 @@ char *xreadlink(const char *path);
 char *concat_path_file(const char *path, const char *filename);
 char *last_char_is(const char *s, int c);
 
-typedef struct file_headers_s {
-	char *name;
-	char *link_name;
-	off_t size;
-	uid_t uid;
-	gid_t gid;
-	mode_t mode;
-	time_t mtime;
-	off_t offset;
-	dev_t device;
-	struct file_headers_s *next;
-} file_headers_t;
-file_headers_t *get_ar_headers(FILE *in_file);
-file_headers_t *get_tar_headers(FILE *tar_stream);
-file_headers_t *get_tar_gz_headers(FILE *compressed_stream);
-file_headers_t *append_archive_list(file_headers_t *head, file_headers_t *tail_entry);
-file_headers_t *add_from_archive_list(file_headers_t *master_list, file_headers_t *new_list, const char *name);
+void *get_header_ar(FILE *in_file);
+void *get_header_cpio(FILE *src_stream);
+void *get_header_tar(FILE *tar_stream);
 
 enum extract_functions_e {
 	extract_verbose_list = 1,
@@ -240,18 +226,20 @@ enum extract_functions_e {
 	extract_preserve_date = 32,
 	extract_data_tar_gz = 64,
 	extract_control_tar_gz = 128,
-	extract_unzip_only = 256
+	extract_unzip_only = 256,
+	extract_unconditional = 512,
+	extract_create_dirs = 1024
 };
-char *extract_archive(FILE *src_stream, FILE *out_stream, file_headers_t *extract_headers, int function, const char *prefix);
-char *deb_extract(const char *package_filename, FILE *out_stream, const int function,
+char *unarchive(FILE *src_stream, void *(*get_header)(FILE *),
+	const int extract_function, const char *prefix, char **extract_names);
+char *deb_extract(const char *package_filename, FILE *out_stream, const int extract_function,
 	const char *prefix, const char *filename);
 char *read_package_field(const char *package_buffer);
-int seek_sub_file(FILE *in_file, file_headers_t *headers, const char *tar_gz_file);
 char *fgets_str(FILE *file, const char *terminating_string);
 
 extern int unzip(FILE *l_in_file, FILE *l_out_file);
 extern void gz_close(int gunzip_pid);
-extern int gz_open(FILE *compressed_file, int *pid);
+extern FILE *gz_open(FILE *compressed_file, int *pid);
 
 extern struct hostent *xgethostbyname(const char *name);
 
diff --git a/libbb/unarchive.c b/libbb/unarchive.c
new file mode 100644
index 0000000..a0cc28e
--- /dev/null
+++ b/libbb/unarchive.c
@@ -0,0 +1,500 @@
+/*
+ *  Copyright (C) 2000 by Glenn McGrath
+ *  Copyright (C) 2001 by Laurence Anderson
+ *	
+ *  Based on previous work by busybox developers and others.
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU Library General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ */
+
+#include <stdio.h>
+#include <errno.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <utime.h>
+#include "libbb.h"
+
+typedef struct file_headers_s {
+	char *name;
+	char *link_name;
+	off_t size;
+	uid_t uid;
+	gid_t gid;
+	mode_t mode;
+	time_t mtime;
+	dev_t device;
+} file_header_t;
+
+off_t archive_offset;
+
+void seek_sub_file(FILE *src_stream, const int count)
+{
+	int i;
+	/* Try to fseek as faster */
+	archive_offset += count;
+	if (fseek(src_stream, count, SEEK_CUR) != 0 && errno == ESPIPE) {
+	for (i = 0; i < count; i++) {
+		fgetc(src_stream);
+		}
+	}
+	return;
+}
+
+
+/* Extract the data postioned at src_stream to either filesystem, stdout or 
+ * buffer depending on the value of 'function' which is defined in libbb.h 
+ *
+ * prefix doesnt have to be just a directory, it may prefix the filename as well.
+ *
+ * e.g. '/var/lib/dpkg/info/dpkg.' will extract all files to the base bath 
+ * '/var/lib/dpkg/info/' and all files/dirs created in that dir will have 
+ * 'dpkg.' as their prefix
+ *
+ * For this reason if prefix does point to a dir then it must end with a
+ * trailing '/' or else the last dir will be assumed to be the file prefix 
+ */
+char *extract_archive(FILE *src_stream, FILE *out_stream, const file_header_t *file_entry,
+ const int function, const char *prefix)
+{
+	FILE *dst_stream = NULL;
+	char *full_name = NULL;
+	char *buffer = NULL;
+	struct utimbuf t;
+
+	/* prefix doesnt have to be a proper path it may prepend 
+	 * the filename as well */
+	if (prefix != NULL) {
+		/* strip leading '/' in filename to extract as prefix may not be dir */
+		/* Cant use concat_path_file here as prefix might not be a directory */
+		char *path = file_entry->name;
+		if (*path == '/') {
+			path++;
+		}
+		full_name = xmalloc(strlen(prefix) + strlen(path) + 1);
+		strcpy(full_name, prefix);
+		strcat(full_name, path);
+	} else {
+		full_name = file_entry->name;
+	}
+
+	if (function & extract_to_stdout) {
+		if (S_ISREG(file_entry->mode)) {
+			copy_file_chunk(src_stream, out_stream, file_entry->size);			
+			archive_offset += file_entry->size;
+		}
+	}
+	else if (function & extract_one_to_buffer) { 
+		if (S_ISREG(file_entry->mode)) {
+			buffer = (char *) xmalloc(file_entry->size + 1);
+			fread(buffer, 1, file_entry->size, src_stream);
+			archive_offset += file_entry->size;
+			return(buffer);
+		}
+	}
+	else if (function & extract_all_to_fs) {
+#if 0
+		struct stat oldfile;
+		if ( (S_ISLNK(file_entry->mode) ? lstat (full_name, &oldfile) : stat (full_name, &oldfile)) == 0) { /* The file already exists */
+			if (function & extract_unconditional || oldfile.st_mtime < file_entry->mtime) {
+				if (!S_ISDIR(oldfile.st_mode)) {
+					unlink(full_name); /* Directories might not be empty etc */
+				}
+			} else {
+				error_msg("%s not created: newer or same age file exists", file_entry->name);
+				if (S_ISREG(file_entry->mode)) {
+					seek_sub_file(src_stream, file_entry->size);
+				}
+				return (NULL);
+			}
+		}
+#endif
+		switch(file_entry->mode & S_IFMT) {
+			case S_IFREG:
+				if (file_entry->link_name) { /* Found a cpio hard link */
+					if (link(file_entry->link_name, full_name) != 0) {
+						perror_msg("Cannot link from %s to '%s'",
+							file_entry->name, file_entry->link_name);
+					}
+				} else {
+					if ((dst_stream = wfopen(full_name, "w")) == NULL) {
+						seek_sub_file(src_stream, file_entry->size);
+						return NULL;
+					}
+					archive_offset += file_entry->size;
+					copy_file_chunk(src_stream, dst_stream, file_entry->size);			
+					fclose(dst_stream);
+				}
+				break;
+			case S_IFDIR:
+				/* Use create_path instead of mkdir incase prefix path
+				 * hasnt been created */
+				if (function & extract_create_dirs) {
+					if (create_path(full_name, file_entry->mode) == FALSE) {
+						return NULL;
+					}
+				}
+				break;
+			case S_IFLNK:
+				if (symlink(file_entry->link_name, full_name) < 0) {
+					perror_msg("Cannot create symlink from %s to '%s'", file_entry->name, file_entry->link_name); 
+					return NULL;
+				}
+				break;
+			case S_IFSOCK:
+			case S_IFBLK:
+			case S_IFCHR:
+			case S_IFIFO:
+				if (mknod(full_name, file_entry->mode, file_entry->device) == -1) {
+					perror_msg("Cannot create node %s", file_entry->name);
+					return NULL;
+				}
+				break;
+		}
+		if (function & extract_preserve_date) {
+			t.actime = file_entry->mtime;
+			t.modtime = file_entry->mtime;
+			utime(full_name, &t);
+		}
+		chmod(full_name, file_entry->mode);
+		lchown(full_name, file_entry->uid, file_entry->gid);
+	} else {
+		/* If we arent extracting data we have to skip it, 
+		 * if data size is 0 then then just do it anyway
+		 * (saves testing for it) */
+		seek_sub_file(src_stream, file_entry->size);
+	}
+
+	/* extract_list and extract_verbose_list can be used in conjunction
+	 * with one of the above four extraction functions, so do this seperately */
+	if (function & extract_verbose_list) {
+		fprintf(out_stream, "%s %d/%d %8d %s ", mode_string(file_entry->mode), 
+			file_entry->uid, file_entry->gid,
+			(int) file_entry->size, time_string(file_entry->mtime));
+	}
+	if ((function & extract_list) || (function & extract_verbose_list)){
+		/* fputs doesnt add a trailing \n, so use fprintf */
+		fprintf(out_stream, "%s\n", file_entry->name);
+	}
+
+	free(full_name);
+
+	return(NULL); /* Maybe we should say if failed */
+}
+
+#if defined BB_AR || defined BB_CPIO || defined BB_UNTAR
+char *unarchive(FILE *src_stream, void *(*get_headers)(FILE *),
+	const int extract_function, const char *prefix, char **extract_names)
+{
+	file_header_t *file_entry;
+	int found;
+	int i;
+	char *buffer = NULL;
+
+	archive_offset = 0;
+	while ((file_entry = (file_header_t *) get_headers(src_stream)) != NULL) {
+		found = FALSE;
+		if (extract_names[0] != NULL) {
+			for(i = 0; extract_names[i] != 0; i++) {
+				if (strcmp(extract_names[i], file_entry->name) == 0) {
+					found = TRUE;
+				}
+			}
+			if (!found) {
+				/* seek past the data entry */
+				if (!S_ISLNK(file_entry->mode) && file_entry->link_name && file_entry->size == 0) {
+					error_msg("You should extract %s as other files are hardlinked to it", file_entry->name);
+				}
+				seek_sub_file(src_stream, file_entry->size);
+				continue;
+			}
+		}
+		buffer = extract_archive(src_stream, stdout, file_entry, extract_function, prefix);
+	}
+	return(buffer);
+}
+#endif
+
+#if defined BB_AR || defined BB_DPKG_DEB || defined BB_DPKG
+void *get_header_ar(FILE *src_stream)
+{
+	file_header_t *typed;
+	union {
+		char raw[60];
+	 	struct {
+ 			char name[16];
+ 			char date[12];
+ 			char uid[6];
+ 			char gid[6];
+ 			char mode[8];
+ 			char size[10];
+ 			char magic[2];
+ 		} formated;
+	} ar;
+	static char *ar_long_names;
+
+	if (fread(ar.raw, 1, 60, src_stream) != 60) {
+		return(NULL);
+	}
+	archive_offset += 60;
+	/* align the headers based on the header magic */
+	if ((ar.formated.magic[0] != '`') || (ar.formated.magic[1] != '\n')) {
+		/* some version of ar, have an extra '\n' after each data entry,
+		 * this puts the next header out by 1 */
+		if (ar.formated.magic[1] != '`') {
+			error_msg("Invalid magic");
+			return(NULL);
+		}
+		/* read the next char out of what would be the data section,
+		 * if its a '\n' then it is a valid header offset by 1*/
+		archive_offset++;
+		if (fgetc(src_stream) != '\n') {
+			error_msg("Invalid magic");
+			return(NULL);
+		}
+		/* fix up the header, we started reading 1 byte too early */
+		/* raw_header[60] wont be '\n' as it should, but it doesnt matter */
+		memmove(ar.raw, &ar.raw[1], 59);
+	}
+		
+	typed = (file_header_t *) xcalloc(1, sizeof(file_header_t));
+
+	typed->size = (size_t) atoi(ar.formated.size);
+	/* long filenames have '/' as the first character */
+	if (ar.formated.name[0] == '/') {
+		if (ar.formated.name[1] == '/') {
+			/* If the second char is a '/' then this entries data section
+			 * stores long filename for multiple entries, they are stored
+			 * in static variable long_names for use in future entries */
+			ar_long_names = (char *) xrealloc(ar_long_names, typed->size);
+			fread(ar_long_names, 1, typed->size, src_stream);
+			archive_offset += typed->size;
+			/* This ar entries data section only contained filenames for other records
+			 * they are stored in the static ar_long_names for future reference */
+			return(NULL);
+		} else {
+			/* The number after the '/' indicates the offset in the ar data section
+			(saved in variable long_name) that conatains the real filename */
+			if (!ar_long_names) {
+				error_msg("Cannot resolve long file name");
+				return (NULL);
+			}
+			typed->name = xstrdup(ar_long_names + atoi(&ar.formated.name[1]));
+		}
+	} else {
+		/* short filenames */
+		typed->name = xcalloc(1, 16);
+		strncpy(typed->name, ar.formated.name, 16);
+	}
+	typed->name[strcspn(typed->name, " /")]='\0';
+
+	/* convert the rest of the now valid char header to its typed struct */	
+	parse_mode(ar.formated.mode, &typed->mode);
+	typed->mtime = atoi(ar.formated.date);
+	typed->uid = atoi(ar.formated.uid);
+	typed->gid = atoi(ar.formated.gid);
+
+	return(typed);
+}
+#endif
+
+#if defined BB_CPIO
+void *get_header_cpio(FILE *src_stream)
+{
+	file_header_t *cpio_entry = NULL;
+	char cpio_header[110];
+	char dummy[14];
+	int namesize;
+	int major, minor, nlink;
+
+	/* There can be padding before archive header */
+	seek_sub_file(src_stream, (4 - (archive_offset % 4)) % 4);
+	if (fread(cpio_header, 1, 110, src_stream) == 110) {
+		archive_offset += 110;
+		if (strncmp(cpio_header, "07070", 5) != 0) {
+			error_msg("Unsupported format or invalid magic");
+			return(NULL);
+		}
+		switch (cpio_header[5]) {
+			case '2': /* "crc" header format */
+				/* Doesnt do the crc check yet */
+			case '1': /* "newc" header format */
+				cpio_entry = (file_header_t *) xcalloc(1, sizeof(file_header_t));
+				sscanf(cpio_header, "%14c%8x%8x%8x%8x%8lx%8lx%16c%8x%8x%8x%8c",
+					dummy, &cpio_entry->mode, &cpio_entry->uid, &cpio_entry->gid,
+					&nlink, &cpio_entry->mtime, &cpio_entry->size,
+					dummy, &major, &minor, &namesize, dummy);
+
+				cpio_entry->name = (char *) xcalloc(1, namesize);
+				fread(cpio_entry->name, 1, namesize, src_stream); /* Read in filename */
+				archive_offset += namesize;
+				/* Skip padding before file contents */
+				seek_sub_file(src_stream, (4 - (archive_offset % 4)) % 4);
+				if (strcmp(cpio_entry->name, "TRAILER!!!") == 0) {
+					printf("%d blocks\n", (int) (archive_offset % 512 ? (archive_offset / 512) + 1 : archive_offset / 512)); /* Always round up */
+					return(NULL);
+				}
+
+				if (S_ISLNK(cpio_entry->mode)) {
+					cpio_entry->link_name = (char *) xcalloc(1, cpio_entry->size + 1);
+					fread(cpio_entry->link_name, 1, cpio_entry->size, src_stream);
+					archive_offset += cpio_entry->size;
+				}
+				if (nlink > 1 && !S_ISDIR(cpio_entry->mode) && cpio_entry->size == 0) {
+					error_msg("%s not extracted: Cannot handle hard links yet", cpio_entry->name);
+					return(get_header_cpio(src_stream)); /* Recurse to next file */
+				}
+				cpio_entry->device = (major << 8) | minor;
+				break;
+			default:
+				error_msg("Unsupported format");
+				return(NULL);
+		}
+		if (ferror(src_stream) || feof(src_stream)) {
+			perror_msg("Stream error");
+			return(NULL);
+		}
+	}
+	return(cpio_entry);
+}
+#endif
+
+#if defined BB_UNTAR || defined BB_DPKG_DEB || defined BB_DPKG
+void *get_header_tar(FILE *tar_stream)
+{
+	union {
+		unsigned char raw[512];
+		struct {
+			char name[100];		/*   0-99 */
+			char mode[8];		/* 100-107 */
+			char uid[8];		/* 108-115 */
+			char gid[8];		/* 116-123 */
+			char size[12];		/* 124-135 */
+			char mtime[12];		/* 136-147 */
+			char chksum[8];		/* 148-155 */
+			char typeflag;		/* 156-156 */
+			char linkname[100];	/* 157-256 */
+			char magic[6];		/* 257-262 */
+			char version[2];	/* 263-264 */
+			char uname[32];		/* 265-296 */
+			char gname[32];		/* 297-328 */
+			char devmajor[8];	/* 329-336 */
+			char devminor[8];	/* 337-344 */
+			char prefix[155];	/* 345-499 */
+			char padding[12];	/* 500-512 */
+		} formated;
+	} tar;
+	file_header_t *tar_entry = NULL;
+	long i;
+	long sum = 0;
+
+	if (archive_offset % 512 != 0) {
+		seek_sub_file(tar_stream, 512 - (archive_offset % 512));
+	}
+
+	if (fread(tar.raw, 1, 512, tar_stream) != 512) {
+		error_msg("Couldnt read header");
+		return(NULL);
+	}
+	archive_offset += 512;
+
+	/* Check header has valid magic, unfortunately some tar files
+	 * have empty (0'ed) tar entries at the end, which will
+	 * cause this to fail, so fail silently for now
+	 */
+	if (strncmp(tar.formated.magic, "ustar", 5) != 0) {
+		return(NULL);
+	}
+
+	/* Do checksum on headers */
+	for (i =  0; i < 148 ; i++) {
+		sum += tar.raw[i];
+	}
+	sum += ' ' * 8;
+	for (i =  156; i < 512 ; i++) {
+		sum += tar.raw[i];
+	}
+	if (sum != strtol(tar.formated.chksum, NULL, 8)) {
+		error_msg("Invalid tar header checksum");
+		return(NULL);
+	}
+
+	/* convert to type'ed variables */
+	tar_entry = xcalloc(1, sizeof(file_header_t));
+	tar_entry->name = xstrdup(tar.formated.name);
+
+	parse_mode(tar.formated.mode, &tar_entry->mode);
+	tar_entry->uid   = strtol(tar.formated.uid, NULL, 8);
+	tar_entry->gid   = strtol(tar.formated.gid, NULL, 8);
+	tar_entry->size  = strtol(tar.formated.size, NULL, 8);
+	tar_entry->mtime = strtol(tar.formated.mtime, NULL, 8);
+	tar_entry->link_name  = strlen(tar.formated.linkname) ? xstrdup(tar.formated.linkname) : NULL;
+	tar_entry->device = (strtol(tar.formated.devmajor, NULL, 8) << 8) +
+		strtol(tar.formated.devminor, NULL, 8);
+
+	return(tar_entry);
+}
+#endif
+
+#if defined BB_DPKG || defined BB_DPKG_DEB
+char *deb_extract(const char *package_filename, FILE *out_stream, const int extract_function,
+	const char *prefix, const char *filename)
+{
+	FILE *deb_stream;
+	FILE *uncompressed_stream = NULL;
+	file_header_t *ar_header = NULL;
+	char *output_buffer = NULL;
+	char *ared_file = NULL;
+	char ar_magic[8];
+	char **file_list;
+	int gunzip_pid;
+
+	file_list = malloc(sizeof(char *));
+	file_list[0] = xstrdup(filename);
+	file_list[1] = NULL;
+
+	if (extract_function & extract_control_tar_gz) {
+		ared_file = xstrdup("control.tar.gz");
+	}
+	else if (extract_function & extract_data_tar_gz) {		
+		ared_file = xstrdup("data.tar.gz");
+	}
+
+	/* open the debian package to be worked on */
+	deb_stream = wfopen(package_filename, "r");
+
+	/* check ar magic */
+	fread(ar_magic, 1, 8, deb_stream);
+	if (strncmp(ar_magic,"!<arch>",7) != 0) {
+		error_msg_and_die("invalid magic");
+	}
+	archive_offset = 8;
+
+	while ((ar_header = get_header_ar(deb_stream)) != NULL) {
+		if (strcmp(ared_file, ar_header->name) == 0) {
+			/* open a stream of decompressed data */
+			uncompressed_stream = gz_open(deb_stream, &gunzip_pid);
+			archive_offset = 0;
+			output_buffer = unarchive(uncompressed_stream, get_header_tar, extract_function, prefix, file_list);
+		}
+		seek_sub_file(deb_stream, ar_header->size);
+	}
+	gz_close(gunzip_pid);
+	fclose(deb_stream);
+	fclose(uncompressed_stream);
+	free(ared_file);
+	return(output_buffer);
+}
+#endif