COMP1521 Computer Systems Fundamentals Assignment 2: a file archiver
The Task
A file archive is a single file which can contain the contents, names and other metadata of multiple files. These can make backup and transport of files more convenient, and can often make compression more efficient. We often refer to tools that can create or manipulate these as file archivers.
There are a vast number of archive formats: on *nix-like systems, tar is common; whereas on Windows, Zip is common. Wikipedia's list of archive formats is a marvellous rabbit-hole to explore.
In this assignment, you will be implementing rain, a file archiver for the drop format.
The drop format is made up of one or more droplets; where a droplet records one file system object; This format is described in more detail below.
A complete implementation of rain can
list the path names of each object in a drop (subset 0);
list the permissions of each object in a drop (subset 0);
list the size (number of bytes) of files in a drop (subset 0);
check the droplet magic number (subset 0);
extract files from a drop (subset 1);
check a drop for integrity, by checking droplet hashes; (subset 1);
set the file permissions of files extracted from a drop (subset 1);
create a drop from a list of files (subset 2);
list, extract, and create drops that include directories (subset 3); and
extract, and create drops in 7-bit and 6-bit formats (subset 3).
Getting Started
Create a new directory for this assignment, change to this directory, and fetch the provided code by running
mkdir -m 700 rain cd rain 1521 fetch rain
If you're not working at CSE, you can download the provided files as a zip file or a tar file.
This will give you the following files:
rain.c- is the only file you need to change: it contains partial definitions of four functions, list_drop, check_drop, extract_drop, and create_drop, to which you need to add code to complete the assignment. You can also add your own functions to this file.
rain_main.c- contains a main, which has code to parse the command line arguments, and which then calls one of list_drop, extract_drop, create_drop, or check_drop, depending on the command line arguments given to rain. Do not change this file.
rain.h- contains shared function declarations and some useful constant definitions. Do not change this file.
rain_hash.c- contains the droplet_hash function; you should call this function to calculate hashes for subset 1. Do not change this file.
rain_6_bit.c- contains the droplet_to_6_bit and droplet_from_6_bit functions. You should call these to implement the 6-bit format for subset 3. Do not change this file.
rain.mk- contains a Makefile fragment for rain.
You can run make to compile the provided code; and you should be able to run the result.
make dcc -c -o rain.o rain.c dcc -c -o rain_main.o rain_main.c dcc -c -o rain_hash.o rain_hash.c dcc -c -o rain_6_bit.o rain_6_bit.c dcc rain.o rain_main.o rain_hash.o rain_6_bit.o -o rain ./rain -l a.drop list_drop called to list drop: 'a.drop'
If you are running make without dcc available you can compile like this:
make CC=gcc gcc -c -o rain.o rain.c gcc -c -o rain_main.o rain_main.c gcc -c -o rain_hash.o rain_hash.c gcc -c -o rain_6_bit.o rain_6_bit.c gcc rain.o rain_main.o rain_hash.o rain_6_bit.o -o rain ./rain -l a.drop list_drop called to list drop: 'a.drop'
If you don't have make available you can compile like this:
dcc rain.c rain_main.c rain_hash.c rain_6_bit.c -o rain ./rain -C b.drop check_drop called to check drop: 'a.drop'
If you don't have make or dcc available you can compile like this:
gcc -Wall rain.c rain_main.c rain_hash.c rain_6_bit.c -o rain ./rain -C b.drop check_drop called to check drop: 'a.drop'
You may optionally create extra .c or .h files.
You should run unzip to get a directory called examples/ full of .drop files to test your program against.
unzip examples.zip
Subset 0
To complete subset 0, you need to implement code that can
- print a list of the contents of a drop, and
- print a detailed list of the contents of a drop.
Subset 0: Print a list of the contents of a drop
Given the -l command-line argument, rain should print the path names of the files/directories in a drop.
For example:
List each item in the drop called text_file.drop, which is in the examples directory ./rain -l examples/text_file.drop hello.txt List each item in the drop called 4_files.drop, which is in the examples directory ./rain -l examples/4_files.drop 256.bin hello.txt last_goodbye.txt these_days.txt List each item in the drop called hello_world.drop, which is in the examples directory ./rain -l examples/hello_world.drop hello.c hello.cpp hello.d hello.go hello.hs hello.java hello.js hello.pl hello.py hello.rs hello.s hello.sh hello.sql
Subset 0: Print a detailed list of the contents of a drop
Given the -L command-line argument, rain should, for each file in the specified drop, print:
- the file/directory permissions,
- the droplet format which will be one of 6, 7 or 8 (the default),
- the file/directory size in bytes, and
- the file/directory path name.
./rain -L examples/text_file.drop -rw-r--r-- 8 56 hello.txt List the details of each item in the drop called 4_files.drop, which is in the examples directory ./rain -L examples/4_files.drop -rw-r--r-- 8 256 256.bin -rw-r--r-- 8 56 hello.txt -r--r--r-- 8 166 last_goodbye.txt -r--rw-r-- 8 148 these_days.txt List the details of each item in the drop called hello_world.drop, which is in the examples directory ./rain -L examples/hello_world.drop -rw-r--r-- 8 93 hello.c -rw-r--r-- 8 82 hello.cpp -rw-r--r-- 8 65 hello.d -rw-r--r-- 8 77 hello.go -rw-r--r-- 8 32 hello.hs -rw-r--r-- 8 117 hello.java -rw-r--r-- 8 30 hello.js -rwxr-xr-x 8 47 hello.pl -rwxr-xr-x 8 103 hello.py -rw-r--r-- 8 45 hello.rs -rw-r--r-- 8 123 hello.s -rwxr-xr-x 8 41 hello.sh -rw-r--r-- 8 24 hello.sql
Subset 1
To complete subset 1, you need to implement code that can
- check the contents of a drop, and
- extract files from a drop.
Subset 1: Check the contents of a drop
Given the -C command-line argument, rain should check the hashes in the specified drop. For example:
Check the drop called 4_files.drop, which is in the examples directory ./rain -C examples/4_files.drop 256.bin - correct hash hello.txt - correct hash last_goodbye.txt - correct hash these_days.txt - correct hash Check the drop called examples/hello_world.bad_hash.drop, which is in the examples directory ./rain -C examples/hello_world.bad_hash.drop hello.c - correct hash hello.cpp - correct hash hello.d - correct hash hello.go - correct hash hello.hs - correct hash hello.java - correct hash hello.js - correct hash hello.pl - correct hash hello.py - correct hash hello.rs - correct hash hello.s - correct hash hello.sh - correct hash hello.sql - incorrect hash 0x19 should be 0x43
It should also check the droplet magic number (first byte) of each droplet, and emit an error if it is incorrect.
Check the drop called text_file.bad_magic.drop, which is in the examples directory ./rain -C examples/text_file.bad_magic.drop error: incorrect first droplet byte: 0x39 should be 0x63
Subset 1: Extract files from a drop
Given the -x command-line argument, rain should extract the files in the specified drop.
It should set file permissions for extracted files to the permissions specified in the drop.
rain will extract files into the current working directory.
So as not to clutter your assignment directory, you should create a
temporary directory, 'tmp', and change to it. Once in that directory,
both your rain program and 'examples/' will be in its parent
directory --- hence the use of '..' in these path names.
Make a directory called tmp.
mkdir -p tmp/
Change into the tmp directory.
cd tmp/
Forcibly remove all files inside the tmp directory.
rm -f * .*
Use your program to extract the contents of text_file.drop.
../rain -x ../examples/text_file.drop
Extracting: hello.txt
Show the contents of hello.txt in the terminal.
You can manually open it in your text editor too, if you like.
cat hello.txt
Hello COMP1521
I hope you are enjoying this assignment.
Forcibly remove all files inside the tmp directory.
rm -f * .*
Use your program to extract the contents of hello_world.drop.
../rain -x ../examples/hello_world.drop
Extracting: hello.c
Extracting: hello.cpp
Extracting: hello.d
Extracting: hello.go
Extracting: hello.hs
Extracting: hello.java
Extracting: hello.js
Extracting: hello.pl
Extracting: hello.py
Extracting: hello.rs
Extracting: hello.s
Extracting: hello.sh
Extracting: hello.sql
Show the first 25 lines from the extracted files to confirm the extraction was successful.
cat $(echo * | sort) | head -n 25
extern int puts(const char *s);
int main(void)
{
puts("Hello, World!");
return 0;
}
#include <iostream>
int main () {
std::cout << "Hello, world!" << std::endl;
}
import std.stdio;
void main() {
writeln("Hello, world!");
}
package main
import "fmt"
func main() {
fmt.Println("Hello, World!")
}
main = putStrLn "Hello, World!"
Forcibly remove all files inside the tmp directory
rm -f * .*
Use your program to extract the contents of meta.drop.
../rain -x ../examples/meta.drop
Extracting: 1_file.subdirectory.7-bit.drop
Extracting: 1_file.subdirectory.drop
Extracting: 2_files.7-bit.drop
Extracting: 2_files.drop
Extracting: 3_files.7-bit.drop
Extracting: 3_files.bad_hash.drop
Extracting: 3_files.bad_magic.drop
Extracting: 3_files.drop
Extracting: 3_files.subdirectory.7-bit.drop
Extracting: 3_files.subdirectory.bad_hash.drop
Extracting: 3_files.subdirectory.bad_magic.drop
Extracting: 3_files.subdirectory.drop
Extracting: 4_files.drop
Extracting: all_the_modes.subdirectory.7-bit.drop
Extracting: all_the_modes.subdirectory.drop
Extracting: all_three_formats.6-bit.drop
Extracting: binary_file.drop
Extracting: hello_world.7-bit.drop
Extracting: hello_world.bad_hash.drop
Extracting: hello_world.bad_magic.drop
Extracting: hello_world.drop
Extracting: lecture_code.subdirectory.7-bit.drop
Extracting: lecture_code.subdirectory.drop
Extracting: small.6-bit.drop
Extracting: small.7-bit.drop
Extracting: small.drop
Extracting: text_file.7-bit.drop
Extracting: text_file.bad_hash.drop
Extracting: text_file.bad_magic.drop
Extracting: text_file.drop
Extracting: tiny.6-bit.drop
Extracting: tiny.7-bit.drop
Extracting: tiny.drop
Show the first 10 items in this directory alphabetically to check extraction was successful.
ls -1 $(echo * | sort) | head
1_file.subdirectory.drop
1_file.subdirectory.compressed.drop
2_files.drop
2_files.compressed.drop
3_files.bad_hash.drop
3_files.bad_magic.drop
3_files.drop
3_files.compressed.drop
3_files.subdirectory.bad_hash.drop
3_files.subdirectory.bad_magic.drop
Go back into the directory with your code.
cd ../
Remove the tmp directory and everything inside it.
rm -rf tmp/
Subset 2
To complete subset 2, you need to implement code that can
- create a drop from a list of files.
Subset 2: Create a drop from a list of files
Given the -c command-line argument, rain should create a drop containing the specified files.
These "echo" lines show you how to create these test files and what their contents are. Create a file called hello.txt with the contents "hello". echo hello >hello.txt Create a file called hola.txt with the contents "hola". echo hola >hola.txt Create a file called hi.txt with the contents "hi". echo hi >hi.txt Set the permissions of these files to 644 (octal permission string (equivalent to rw-r--r--)). When you list the contents of the drop, the permissions should match this. chmod 644 hello.txt hola.txt hi.txt Create a drop called selamat.drop with the files hello.txt, hola.txt, and hi.txt. ./rain -c selamat.drop hello.txt hola.txt hi.txt Adding: hello.txt Adding: hola.txt Adding: hi.txt List the contents of selamat.drop. ./rain -L selamat.drop -rw-r--r-- 8 6 hello.txt -rw-r--r-- 8 5 hola.txt -rw-r--r-- 8 3 hi.txt Make a directory called tmp. mkdir -p tmp/ Change into the tmp directory. cd tmp/ Forcibly remove all files inside the tmp directory. rm -f * .* Use your program to extract the contents of selamat.drop. ../rain -x ../selamat.drop Extracting: hello.txt Extracting: hola.txt Extracting: hi.txt Check that the extracted file hello.txt is the same as the source file ../hello.txt. diff -s ../hello.txt hello.txt Files ../hello.txt and hello.txt are identical Check that the extracted file hola.txt is the same as the source file ../hola.txt. diff -s ../hola.txt hola.txt Files ../hola.txt and hola.txt are identical Check that the extracted file hi.txt is the same as the source file ../hi.txt. diff -s ../hi.txt hi.txt Files ../hi.txt and hi.txt are identical Go back into the directory with your code. cd ../ Remove the tmp directory and everything inside it. rm -rf tmp/
It is also possible to append droplets to an existing drop file using the -a command-line option. For example:
./rain -a bonjour.drop hello.txt Adding: hello.txt ./rain -L bonjour.drop -rw-r--r-- 8 6 hello.txt ./rain -a bonjour.drop hola.txt hi.txt Adding: hola.txt Adding: hi.txt ./rain -L bonjour.drop -rw-r--r-- 8 6 hello.txt -rw-r--r-- 8 5 hola.txt -rw-r--r-- 8 3 hi.txt
Subset 3
To complete subset 3, you need to implement code that can
- create a drop from a list of files and directories,
- extract directories from a drop, and
- manipulate 6-bit and 7-bit storage formats.
Subset 3: Create a drop from a list of files and directories
Given the -c command-line argument, rain should be able to add files in sub-directories. For example:
Create a drop called a.drop with the file "hello.txt" that is contained within 2 levels of directories. ./rain -c a.drop examples/2_files.d/hello.txt Adding: examples Adding: examples/2_files.d Adding: examples/2_files.d/hello.txt
If a directory is specified when creating a drop, rain should add the entire directory tree to the drop.
Create a drop called a.drop with *all* the contents within the directory "3_files.subdirectory.d" which is in the "examples" directory. ./rain -c a.drop examples/3_files.subdirectory.d Adding: examples Adding: examples/3_files.subdirectory.d Adding: examples/3_files.subdirectory.d/goodbye Adding: examples/3_files.subdirectory.d/goodbye/last_goodbye.txt Adding: examples/3_files.subdirectory.d/hello Adding: examples/3_files.subdirectory.d/hello/hello.txt Adding: examples/3_files.subdirectory.d/these_days.txt
Given the -L command-line argument and a drop containing directories, rain should be able to list files and directories. For example:
./rain -L examples/1_file.subdirectory.drop drwxr-xr-x 8 0 hello -rw-r--r-- 8 56 hello/hello.txt
Subset 3: Extract directories from a drop
Given the -x command-line argument, and a drop containing directories, rain should be able to extract files and directories. For example:
./rain -x examples/3_files.subdirectory.drop Creating directory: goodbye Extracting: goodbye/last_goodbye.txt Creating directory: hello Extracting: hello/hello.txt Extracting: these_days.txt
Subset 3: Manipulate 6-bit and 7-bit storage formats
The -7 and -6 options allow droplets to be created in 7-bit and 6-bit format. For example:
./rain -7 -c seven.drop hello.txt Adding: hello.txt ./rain -L seven.drop -rw-r--r-- 7 6 hello.txt ./rain -6 -c six.drop hola.txt hi.txt Adding: hola.txt Adding: hi.txt ./rain -L six.drop -rw-r--r-- 6 5 hola.txt -rw-r--r-- 6 3 hi.txt
It is possible for drops to contain droplets in multiple formats. For example:
./rain -a mixed.drop hello.txt Adding: hello.txt ./rain -L mixed.drop -rw-r--r-- 8 6 hello.txt ./rain -7 -a mixed.drop hi.txt Adding: hi.txt ./rain -L mixed.drop -rw-r--r-- 8 6 hello.txt -rw-r--r-- 7 3 hi.txt ./rain -6 -a mixed.drop hola.txt Adding: hola.txt ./rain -L mixed.drop -rw-r--r-- 8 6 hello.txt -rw-r--r-- 7 3 hi.txt -rw-r--r-- 6 5 hola.txt
Your code should handle creating, listing, checking, and extracting drops in 7-bit and 6-bit format.
Your code should produce an error if asked to create a droplet containing bytes which can be encoded in the specified format. For example:
echo Hello >Hello.txt ./rain -6 -c broken.drop Hello.txt error: byte 0x48 can not be represented in 6-bit format
Handling Errors
Error checking is an important part of this assignment. Automarking will test error handling.
Error messages should be one line (only) and be written to stderr (not stdout).
rain should exit with status 1 after an error.
rain should check all file operations for errors.
As much as possible match the reference implementation error messages exactly.
The reference implementation uses perror to report errors from file operations and other system calls.
It is not necessary to remove files and directories already created or partially created when an error occurs.
You may extract a file or directory from droplet before determining if the droplet hash is correct.
You may not extract the file or directory from a droplet before determining if the droplet magic number is correct.
You can extract previous file or directory from a droplet.
Where multiple errors messages could be produced, for example, if two non-existent files are specified to be added to a drop, rain may produce any one of the error messages.
Reference implementation
A reference implementation is a common, efficient, and effective method to provide or define an operational specification; and it's something you will likely work with after you leave UNSW.
We've provided a reference implementation, 1521 rain, which you can use to find the correct outputs and behaviours for any input:
1521 rain -L examples/tiny.6-bit.drop -rw-r--r-- 6 0 a
Every concrete example shown below is runnable using the reference implementation; run 1521 rain instead of ./rain.
Where any aspect of this assignment is undefined in this specification, you should match the behaviour exhibited by the reference implementation. Discovering and matching the reference implementation's behaviour is deliberately a part of this assignment.
If you discover what you believe to be a bug in the reference implementation, please report it in the class forum. If it is a bug, we may fix the bug; or otherwise indicate that you do not need to match the reference implementation's behaviour in that specific case.
The drop and droplet format
drops must follow exactly the format produced by the reference implementation.
A drop consists of a sequence of one or more droplets. Each droplet contains the information about one file or directory.
The first byte of a drop file is the first byte of the first droplet. That droplet is immediately followed by either another droplet, or by the end of the drop file.
| name | length | type | description |
|---|---|---|---|
| magic number | 1 B | unsigned, 8-bit, little-endian | byte 0 in every droplet must be 0x63 (ASCII 'c') |
| droplet format | 1 B | unsigned, 8-bit, little-endian | byte 1 in every droplet must be one of 0x36, 0x37, 0x38 (ASCII '6', '7', '8') |
| permissions | 10 B | characters | bytes 2—11 are the type and permissions as a ls-like character array; e.g., "-rwxr-xr-x" |
| pathname length | 2 B | unsigned, 16-bit, little-endian | bytes 12—13 are an unsigned 2-byte (16-bit) little-endian integer, giving the length of |
| pathname | pathname-length | characters | the filename of the object in this droplet. |
| content length | 6 B | unsigned, 48-bit, little-endian | the next bytes are an unsigned 6-byte (48-bit) little-endian integer giving the length of the file that was encoded to give |
| content | content-length for 8-bit format, see below for other formats | bytes | the data of the object in this droplet. |
| hash | 1 B | unsigned, 8-bit, little-endian | the last byte of a droplet is a droplet-hash of all bytes of this droplet except this byte. |
droplet content encodings (Subset 3 only)
8-bit format (droplet format == 0x38 contents is an array of bytes, which are exactly equivalent to the bytes in the original file.
7-bit format (droplet format == 0x37) contents is an array of bytes representing packed seven-bit values, with the trailing bits set to zero. Every byte of the original file is taken as a seven-bit value, and packed as described below. This format can store any seven bit value — so, for example, any byte containing valid ASCII can be stored.
This format needs bytes. 7-bit format is used only in subset 3.
6-bit format (droplet format == 0x36) contents is an array of bytes of packed six-bit values where the trailing bits in the last byte are zero, and which are translated using the functions droplet_to_6_bit and droplet_from_6_bit in rain_6_bit.c.
This format cannot store all ASCII values, for example upper case letters can't be stored in 6-bit format.
This format needs bytes.
Packed n-bit encoding (Subset 3 only)
We often store smaller values inside larger types. For example, the integer 42 only needs six bits; but we often will store it in a full thirty-two-bit integer, wasting many bits of zeroes. Assuming we know how many bits the value needs, we could only store the relevant bits.
For example, let's say we have three seven-bit values a, b, c, made up of arbitrary bit-strings, and stored in eight-bit variables
- a:
0b0AAA_AAAA, - b:
0b0BBB_BBBB, - c:
0b0CCC_CCCC,
0bAAAA_AAAB_BBBB_BBCC_CCCC_C???
However, we have a problem: what happens to the trailing bits, which don't have a value? Note that we've defined all trailing bits to be zero above, which would here give:
0bAAAA_AAAB_BBBB_BBCC_CCCC_C000
Inspecting drops and droplets
The hexdump utility can show the individual bytes of a file. We can use this to inspect drops and droplets.
For example, here is a drop, made up of two droplets.
hexdump -vC examples/2_files.drop 00000000 63 38 2d 72 77 2d 72 2d 2d 72 2d 2d 09 00 68 65 |c8-rw-r--r--..he| 00000010 6c 6c 6f 2e 74 78 74 38 00 00 00 00 00 48 65 6c |llo.txt8.....Hel| 00000020 6c 6f 20 43 4f 4d 50 31 35 32 31 0a 49 20 68 6f |lo COMP1521.I ho| 00000030 70 65 20 79 6f 75 20 61 72 65 20 65 6e 6a 6f 79 |pe you are enjoy| 00000040 69 6e 67 20 74 68 69 73 20 61 73 73 69 67 6e 6d |ing this assignm| 00000050 65 6e 74 2e 0a 2d 63 38 2d 72 77 2d 72 2d 2d 72 |ent..-c8-rw-r--r| 00000060 2d 2d 10 00 6c 61 73 74 5f 67 6f 6f 64 62 79 65 |--..last_goodbye| 00000070 2e 74 78 74 a6 00 00 00 00 00 54 68 69 73 20 69 |.txt......This i| 00000080 73 20 6f 75 72 20 6c 61 73 74 20 67 6f 6f 64 62 |s our last goodb| 00000090 79 65 0a 49 20 68 61 74 65 20 74 6f 20 66 65 65 |ye.I hate to fee| 000000a0 6c 20 74 68 65 20 6c 6f 76 65 20 62 65 74 77 65 |l the love betwe| 000000b0 65 6e 20 75 73 20 64 69 65 0a 42 75 74 20 69 74 |en us die.But it| 000000c0 27 73 20 6f 76 65 72 0a 4a 75 73 74 20 68 65 61 |'s over.Just hea| 000000d0 72 20 74 68 69 73 20 61 6e 64 20 74 68 65 6e 20 |r this and then | 000000e0 49 27 6c 6c 20 67 6f 0a 59 6f 75 20 67 61 76 65 |I'll go.You gave| 000000f0 20 6d 65 20 6d 6f 72 65 20 74 6f 20 6c 69 76 65 | me more to live| 00000100 20 66 6f 72 0a 4d 6f 72 65 20 74 68 61 6e 20 79 | for.More than y| 00000110 6f 75 27 6c 6c 20 65 76 65 72 20 6b 6e 6f 77 0a |ou'll ever know.| 00000120 60 |`| 00000121
Each line of hexdump output is in three groups:
the address column: this starts at
0x00000000, and increases by 0x10 (or 16 in base 10) each line;the data columns: after the address, we get (up to) 16 two-digit hexadecimal values, grouped into two blocks of eight values each, which represents the actual data of the file, and
the human readable stripe: at the very end of each line, between the vertical bars (
|) is the human readable version of the bytes preceding, or a'.'if the byte wouldn't ordinarily be visible.
You could also use the hd, od, or xxd utilities instead of hexdump . Also provided for the assignment is 1521 dump_drop which prints the contents of a drop in a mroe structured way, e.g.:
1521 dump_drop examples/2_files.drop ------------------------------------------------------------------------------ Field Name Field Offset Field Hex Field ASCII Field Numeric magic 0x00000000 63 c format 0x00000001 38 8 mode 0x00000002 2d 72 77 2d 72 2d 2d 72 2d 2d -rw-r--r-- path length 0x0000000c 09 00 9 pathname 0x0000000e hello.txt content length 0x00000017 38 00 00 00 00 00 56 contents 0x0000001d """ Hello COMP1521.I hope you are enjoying this assignment.. """ hash 0x00000055 2d ------------------------------------------------------------------------------ Field Name Field Offset Field Hex Field ASCII Field Numeric magic 0x00000056 63 c format 0x00000057 38 8 mode 0x00000058 2d 72 77 2d 72 2d 2d 72 2d 2d -rw-r--r-- path length 0x00000062 10 00 16 pathname 0x00000064 last_goodbye.txt content length 0x00000074 a6 00 00 00 00 00 166 contents 0x0000007a """ This is our last goodbye.I hate to feel the love between us die.But it's over.Just hear this and then I'll go.You gave me more to live for.More than you'll ever know. """ hash 0x00000120 60
6-bit format (Subset 3 only)
droplet 6-bit format defines a subset of 64 8-bit values (bytes) to have a six-bit encoding; those six bits are then stored packed.
The remaining 192 8-bit values can not be encoded in 6-bit format.
The functions droplet_to_6_bit and droplet_from_6_bit in rain_6_bit.c to convert 8-bit values to and from 6-bit format.
You can find the mapping by reading the code in rain_6_bit.c.
The droplet hash (Subsets 1, 2, 3)
Each droplet ends with a hash (sometimes referred to as a digest) which calculated from the other values of the droplet. This allows us to detect if any bytes of the drop have changed, for example by disk or network errors.
The droplet_hash() function makes one step of computation of the hash of a sequence of bytes:
uint8_t droplet_hash(uint8_t current_hash_value, uint8_t byte_value) {
return ((current_hash_value * 33) & 0xff) ^ byte_value;
}
Given the hash value of the sequence up to this byte, and the value of this byte it calculates the new hash value.
If we create a drop of a single one-byte file, like this:
echo >a 1521 rain -c a.drop a
We can then inspect the drop, and see its hash is 0x15.
hexdump -Cv a.drop 00000000 63 38 2d 72 77 2d 72 2d 2d 72 2d 2d 01 00 61 01 |c8-rw-r--r--..a.| 00000010 00 00 00 00 00 0a 15 |.......| 00000017
Here's the sequence of calls that calculated that value:
droplet_hash(0x00, 0x63) = 0x63 droplet_hash(0x63, 0x38) = 0xfb droplet_hash(0xfb, 0x2d) = 0x76 droplet_hash(0x76, 0x72) = 0x44 droplet_hash(0x44, 0x77) = 0xb3 droplet_hash(0xb3, 0x2d) = 0x3e droplet_hash(0x3e, 0x72) = 0x8c droplet_hash(0x8c, 0x2d) = 0x21 droplet_hash(0x21, 0x2d) = 0x6c droplet_hash(0x6c, 0x72) = 0x9e droplet_hash(0x9e, 0x2d) = 0x73 droplet_hash(0x73, 0x2d) = 0xfe droplet_hash(0xfe, 0x01) = 0xbf droplet_hash(0xbf, 0x00) = 0x9f droplet_hash(0x9f, 0x61) = 0x1e droplet_hash(0x1e, 0x01) = 0xdf droplet_hash(0xdf, 0x00) = 0xbf droplet_hash(0xbf, 0x00) = 0x9f droplet_hash(0x9f, 0x00) = 0x7f droplet_hash(0x7f, 0x00) = 0x5f droplet_hash(0x5f, 0x00) = 0x3f droplet_hash(0x3f, 0x0a) = 0x15
Assumptions and Clarifications
Like all good programmers, you should make as few assumptions as possible. If in doubt, match the output of the reference implementation.
Your submitted code must be a single C program only. You may not submit code in other languages.
You can call functions from the C standard library available by default on CSE Linux systems: including, e.g.,
stdio.h,stdlib.h,string.h,math.h,assert.h.We will compile your code with
dccwhen marking. Run-time errors from illegal or invalid C will cause your code to fail automarking (and will likely result in you losing marks).Your program must not require extra compile options. It must compile successfully with:
dcc *.c -o rain
You may not use functions from other libraries. In other words, you cannot use the dcc
-lflag.If your program prints debugging output, it will fail automarking tests. Make sure you disable any debugging output before submission.
You may not create or use temporary files.
You may not create subprocesses: you may not use posix_spawn, posix_spawnp, system, popen, fork, vfork, clone, or any of the
exec*family of functions, like execve.rain only has to handle ordinary files and directories.
rain does not have to handle symbolic links, devices or other special files.
rain will not be given directories containing symbolic links, devices or other special files.
rain does not have to handle hard links.
If completing a
raincommand would produce multiple errors, you may produce any of the errors and stop.You do not have to produce the particular error that the reference implementation does.
If a droplet path name contains a directory then a droplet for the directory will appear in the drop beforehand.
For example, if there is a droplet for the path name
a/b/file.txtthen there will be preceding droplets for the directoriesaanda/b,You may also assume the droplet for the directory specifies the directory is writable.
When adding an entire directory (subset 3) to a drop you may add the directory contents in any order to the drop, after the directory droplet.
You do not have to match the order the reference implementation uses.
When a
raincommand specifies adding files with a common sub-directory. You may add a droplet for the sub-directory multiple times. For example, given this command:./rain -c a.drop b/file1 b/file2
You may add two (duplicate) droplets forb.You can assume the path name of a drop being created with
-c, will not also be added to the drop, and will not be in a directory being added to the drop.It is not necessary to check the hashes or magic numbers of droplets in subset 0. Subset 0 tests will only use valid droplets.
The reference implementation checks the magic number (first byte), format and hash when listing (
-land-L) and extracting (-x) drops. and stops with an error emssage if they are are invalid, for example:./rain -l examples/text_file.bad_hash.drop error: incorrect droplet hash 0x2d should be 0x77 ./rain -L examples/text_file.bad_magic.drop error: incorrect first droplet byte: 0x39 should be 0x63
This is very desirable behaviour and you can implement this in your code. However it will not be tested with
-l,-Land-xcommand line options to avoid problems in automarking.Your code will only be tested with the
-Coption on drops with invalid hashes, magic numbers and formatsIt is not necessary to check the hashes or magic numbers in an existing drop when appending to it (-a).
If you need clarification on what you can and cannot use or do for this assignment, ask in the class forum.
You are required to submit intermediate versions of your assignment. See below for details.
Assessment
Testing
When you think your program is working, you can use autotest to run some simple automated tests:
1521 autotest rain rain.c [optionally: any extra .c or .h files]
1521 autotest will not test everything.
Always do your own testing.
Automarking will be run by the lecturer after the submission deadline, using a superset of tests to those autotest runs for you.
