# Testing procedure

Because of the significant amount of possible combinations of settings and the
fact that based on each setting if the text does not have the same sequence -
even one letter off - then you cannot get the same output - I had to come up
with a way to test the entry. Here's the problem with just the default settings:
the same number of characters, the same starting, but the moment a letter
differs the rest of the message will change even if the rest of the message is
the same:

	    $ echo A man and his dog walked in the park...|./prog
	    G spy ige lqk eav ptccpu di clm erkh...

	    $ echo A dog and his man walked in the park...|./prog
	    G luq ige lqk jou ptccpu di clm erkh...

	    $ echo A man and his dog and its dog walked in the park...|./prog
	    G spy ige lqk eav mbt bwz iwb jsojus pe hnt imny...

As can be seen as soon as a letter is different in the sequence it changes the
result. So what could be done? First of all I had to make sure that the
invocations I gave could be properly duplicated. To do this I had key files that
I could use via recode and also I did input some manually. This includes the
cake recipe key file but I will not include the file but only the invocation I
used as such. This is the general format of the test script:

	    STR="$(echo ABC | ./prog|tail -n1)"

	    if [ "${STR}" != "GYM" ]; then
		echo "Failed 'ABC' test"
		exit 1
	    fi
	    STR="$(echo "${STR}" | ./prog|tail -n1)"
	    if [ "${STR}" != "ABC" ]; then
		echo "Failed 'GYM' test"
		exit 1
	    fi

	    ./recode -Rchocolate-cake.key -fchocolate-cake.md | ./prog  - 2>/dev/null > out
	    diff out cake.md || echo "Cake test 0.5 failed"
	    ./recode -Rchocolate-cake.key -fcake.md | ./prog - 2>/dev/null > out2
	    diff out2 chocolate-cake.md || echo "Cake test 1.0 failed"

	    ( echo -n <key>   ; cat chocolate-cake.md ) | ./prog - 2>/dev/null > out
	    diff out cake.md || echo "Cake test 1.5 failed"
	    ( echo -n <key> ; cat cake.md ) | ./prog - 2>/dev/null > out2
	    diff out2 chocolate-cake.md || echo "Cake test 2.0 failed"

	    ./recode -Robfuscation.key -fobfuscation.txt | ./prog - 2>/dev/null >  out
	    diff out obfuscation.md || echo "Obfuscation test 0.5 failed"

	    ./recode -Robfuscation.key -fout | ./prog - 2>/dev/null > out2
	    diff out2 obfuscation.txt || echo "Obfuscation test 1.0 failed"

	    ./prog < prog.c > out
	    ./prog < out > out2
	    diff out2 prog.c || echo "Self obfuscation test 1.0 failed"

	    mkdir -p tests || exit 1
	    cd tests || exit 1

	    let n=0
	    while [ "${n}" -lt 100 ]; do

		for f in ../input.txt ../prog.c ../README.md ../recode.md ../recode.1 ../enigma.1; do
		    ../recode -r -oconfig -f"${f}" | ../prog - 2>/dev/null > out
		    ../recode -Rconfig -fout | ../prog - 2>/dev/null > out2
		    diff out2 "${f}" || echo "Failed randomised test for file: ${f}"
		done
		let n=n+1
	    done
	    rm -f out out2
	    cd ..
	    rm -f out out2

If I get any output I know there's a problem. I did not get any output and since
this tested the default settings, two customised settings as well as 100
randomised runs for each of the files listed in the script I believe this
works well. Tests run under Fedora, CentOS and macOS with the same results.

I believe this entry can be used with EBCDIC as it appears that the characters
used are in contiguous sequence but I do not know for certain. Neither do I know
about big endian: I don't believe it matters but I have no way to test this
either.

If there are any discrepancies in the recode.md or any other thing I
pasted it's probably because of the way I changed the display (or else the
random option in recode!).

If there are any inconsistencies in input/output one also has to be certain that
input was put in correctly (as I discussed earlier). The results of the script
are the same under Fedora, CentOS and macOS but I have put the main testing
platform as macOS because these days I mostly use my laptop (I tend to ssh into
the other two boxes more and as for the CentOS it's only console anyway. Since
it's just text I/O it an terminal emulator shouldn't matter either.).