unarchive.c

this is the pkg installer for linux

/*
 *  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"

#define CONFIG_FEATURE_TAR_OLDGNU_COMPATABILITY 1
#define CONFIG_FEATURE_TAR_GNU_EXTENSIONS

#ifdef CONFIG_FEATURE_TAR_GNU_EXTENSIONS
static char *longname = NULL;
static char *linkname = NULL;
#endif

extern void seek_sub_file(FILE *src_stream, const int count);
extern char *extract_archive(FILE *src_stream, FILE *out_stream, const file_header_t *file_entry,
                              const int function, const char *prefix);


#ifdef L_archive_offset
off_t archive_offset;
#else
extern off_t archive_offset;
#endif	

#ifdef L_seek_sub_file
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;
}
#endif	



#ifdef L_extract_archive
/* 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 *full_link_name = NULL;
	char *buffer = NULL;
	struct utimbuf t;
        int res;


	/* 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 (strncmp("./", path, 2) == 0) {
			path += 2;
			if (strlen(path) == 0) {
				return(NULL);
			}
		}
		full_name = xmalloc(strlen(prefix) + strlen(path) + 1);
		strcpy(full_name, prefix);
		strcat(full_name, path);
                if ( file_entry->link_name ){
		   full_link_name = xmalloc(strlen(prefix) + strlen(file_entry->link_name) + 1);
		   strcpy(full_link_name, prefix);
		   strcat(full_link_name, file_entry->link_name);
                }
	} else {
		full_name = strdup(file_entry->name);
                if ( file_entry->link_name )
		   full_link_name = strdup(file_entry->link_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);
			buffer[file_entry->size] = '\0';
			archive_offset += file_entry->size;
			return(buffer);
		}
	}
	else if (function & extract_all_to_fs) {
		struct stat oldfile;
		int stat_res;
		stat_res = lstat (full_name, &oldfile);
		if (stat_res == 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 {
				if ((function & extract_quiet) != extract_quiet) {
					error_msg("%s not created: newer or same age file exists", file_entry->name);
				}
				seek_sub_file(src_stream, file_entry->size);
				return (NULL);
			}
		}
		if (function & extract_create_leading_dirs) { /* Create leading directories with default umask */
			char *buf, *parent;
			buf = xstrdup(full_name);
			parent = dirname(buf);
			if (make_directory (parent, -1, FILEUTILS_RECUR) != 0) {
				if ((function & extract_quiet) != extract_quiet) {
					error_msg("couldn't create leading directories");
				}
			}
			free (buf);
		}
		switch(file_entry->mode & S_IFMT) {
			case S_IFREG:
				if (file_entry->link_name) { /* Found a cpio hard link */
					if (link(full_link_name, full_name) != 0) {
						if ((function & extract_quiet) != extract_quiet) {
							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:
				if (stat_res != 0) {
					if (mkdir(full_name, file_entry->mode) < 0) {
						if ((function & extract_quiet) != extract_quiet) {
							perror_msg("extract_archive: %s", full_name);
						}
					}
				}
				break;
			case S_IFLNK:
				if (symlink(file_entry->link_name, full_name) < 0) {
					if ((function & extract_quiet) != extract_quiet) {
						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) {
					if ((function & extract_quiet) != extract_quiet) {
						perror_msg("Cannot create node %s", file_entry->name);
					}
					return NULL;
				}
				break;
                         default:
                            perror_msg("Don't know how to handle %s", full_name);

		}

		/* Changing a symlink's properties normally changes the properties of the 
		 * file pointed to, so dont try and change the date or mode, lchown does
		 * does the right thing, but isnt available in older versions of libc */
		if (S_ISLNK(file_entry->mode)) {
#if (__GLIBC__ > 2) && (__GLIBC_MINOR__ > 1)
			lchown(full_name, file_entry->uid, file_entry->gid);
#endif
		} else {
			if (function & extract_preserve_date) {
				t.actime = file_entry->mtime;
				t.modtime = file_entry->mtime;
				utime(full_name, &t);
			}
			chown(full_name, file_entry->uid, file_entry->gid);
			chmod(full_name, file_entry->mode);
		}
	} 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);
        if ( full_link_name )
	    free(full_link_name);

	return(NULL); /* Maybe we should say if failed */
}
#endif

#ifdef L_unarchive
char *unarchive(FILE *src_stream, FILE *out_stream,
		file_header_t *(*get_headers)(FILE *),
		void (*free_headers)(file_header_t *),
		const int extract_function, const char *prefix, char **extract_names)
{
	file_header_t *file_entry;
	int extract_flag;
	int i;
	char *buffer = NULL;

	archive_offset = 0;
	while ((file_entry = get_headers(src_stream)) != NULL) {
		extract_flag = TRUE;
		/* fprintf(stderr, __FUNCTION__ " getting headers\n"); */
		if (extract_names != NULL) {
			int found_flag = FALSE;
			for(i = 0; extract_names[i] != 0; i++) {
				if (strcmp(extract_names[i], file_entry->name) == 0) {
					found_flag = TRUE;
					break;
				}
			}
			if (extract_function & extract_exclude_list) {
				if (found_flag == TRUE) {
					extract_flag = FALSE;
				}
			} else {
				/* If its not found in the include list dont extract it */
				if (found_flag == FALSE) {
					extract_flag = FALSE;
				}
			}

		}

		if (extract_flag == TRUE) {
			/* fprintf(stderr, __FUNCTION__ " extract?\n"); */
			buffer = extract_archive(src_stream, out_stream, file_entry, extract_function, prefix);
			/* fprintf(stderr, __FUNCTION__ " extracted\n"); */
		} else {
			/* seek past the data entry */
			seek_sub_file(src_stream, file_entry->size);
		}
		free_headers(file_entry);
	}
	/*fprintf(stderr, __FUNCTION__ " goin home\n");*/
	
	return(buffer);
}
#endif

#ifdef L_get_header_ar
file_header_t *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 (get_header_ar(src_stream)); /* Return next header */
		} else if (ar.formated.name[1] == ' ') {
			/* This is the index of symbols in the file for compilers */
			seek_sub_file(src_stream, typed->size);
			return (get_header_ar(src_stream)); /* Return next header */
		} 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);
}

void free_header_ar(file_header_t *ar_entry)
{
    if (ar_entry == NULL) {
	return;
    }

    free(ar_entry->name);
    free(ar_entry->link_name);

    free(ar_entry);
}
#endif

#ifdef L_get_header_cpio