Talk:Backdoor - Revision history

Rei at 11:42, 5 July 2006

2006-07-05T11:42:31Z

← Older revision		Revision as of 11:42, 5 July 2006
Line 539:		Line 539:

	Caine also reveals what looks to be an excerpt from an inetd.conf file. Looks like we need to break the system in another way, but haven't figured this out yet. - courtesy of Scribe		Caine also reveals what looks to be an excerpt from an inetd.conf file. Looks like we need to break the system in another way, but haven't figured this out yet. - courtesy of Scribe


			== Config 23/24 Audio ==
			It sounds like Violet to me. Can someone call the US/UK "walkthrough puzzle" and confirm that the woman making those audio entries is Violet Kiteway? [[User:Rei\|Ayanami]] 04:42, 5 July 2006 (PDT)

Scott: /* Question */

2006-07-05T08:35:05Z

Question

← Older revision		Revision as of 08:35, 5 July 2006
Line 30:		Line 30:
	==Question==		==Question==
	What are Frob and Twiddle? TheBozzball asks a dumb question.		What are Frob and Twiddle? TheBozzball asks a dumb question.
			:Geek terms. Tweak, Twiddle, and fFrob refer to the 3 varying stages of messing with things. fFrob is a sort of spin the dial and see what you get. Especially used when you have no idea what something is. Twiddle is more of a refined fFrob, or perhaps a broad tweak. If fFrob asks us to pick any number, twiddle asks us to pick a numbner between one and a hundred. It narrows the range a great deal. We still don't know what we're doing, but we seem to be closer to understanding what the thing is. Then, the next logical step is a tweak. We can tweak things all day. Up by one step. Down by a half. Right 3/16. 15 degrees to the diagonal of north by 33 9/10. So, naturally, When this page was started, it served as a nice place to start fFrob and twiddling. Hence those titles. It seems to me we should probably start sorting the page out a bit, however. It's becoming a hodge-podge mess. [[User:Scott\|Scott]] 01:35, 5 July 2006 (PDT)

	==Garden of Forking Paths==		==Garden of Forking Paths==

Spugmeistress: Direction theories

2006-07-04T17:20:10Z

Direction theories

← Older revision		Revision as of 17:20, 4 July 2006
Line 506:		Line 506:

	[17:17] <TheBozzball> If you do programs consisting of just "run", you get invalid command grid unless you put run on the first column, in which case you get "no disk access". If you do programs consisting of square 1: some command, square 2: run, you get - for Switch, Copy, Flush, Pread, Cache, Mount - Inv. Command Grid, for the other 6 - No Disk Access		[17:17] <TheBozzball> If you do programs consisting of just "run", you get invalid command grid unless you put run on the first column, in which case you get "no disk access". If you do programs consisting of square 1: some command, square 2: run, you get - for Switch, Copy, Flush, Pread, Cache, Mount - Inv. Command Grid, for the other 6 - No Disk Access

			== Direction theories ==

			DIrections that each comand moves in. If blocked (either by a side, or by a previously used square), then progresses to next listed direction. However, I haven't explored these latter moves extensively, so there are gaps.

			In 2,2, from 1,1:
			MO - SE
			ZO - W, N, E, NE, [NW?], SW
			SW - SE
			RO - N, NE, [E?], S
			CO - S
			FL - S, SE (Seems to move BEFORE flushing) doesn't seem to go upwards
			PW - E, S
			PO - E, S
			PR - W
			CA - No direction
			EN - No direction

			How the filesystem works:

			Files are accessed in a 3-dimensional, rotational space - think along the lines of X, Y and Z axes with the space wrapping around. The wiki is along the right lines on this. The 3 commands ZONE, SWITCH and ROTATE move along each axis. Not all axes ar the same length - there seem to be 4 zones, each with 5 switches, but the number of rotates in each of these varies (usually 1 or 2, sometimes 0).

			Breaking the System:

			Caine has given us one example, with an infinite loop, which we also found by accident. Usually, the execution of commands cannot loop as the executing program keeps track of which squares have been used already. If a command tries to enter a used square, it fails and moves to the next directionon its list. However, I think the FLUSH command clears out this "stack" of used squares, which means we can enter squares we have already been to and produce an infinite loop. For some reason, what should loop doesn't always get accepted by the "parser", i.e. only some loops work and crash the system.

			These loops (following some other commands) seems to "work" (i.e. break) but they may or may not be dependent on the precending commands. I think they may also need to be "inescapable" - e.g. Flush doesn't seem to move upwards at all, so is best placed at the right-hand edge, surrounded by other commands to the left and below.:
			FL RO PR
			RO PW FL ZO
			RO FL (with Flush at edge, no other way down)

			Caine also reveals what looks to be an excerpt from an inetd.conf file. Looks like we need to break the system in another way, but haven't figured this out yet. - courtesy of Scribe

Xnera: more direction talk

2006-07-01T04:47:12Z

more direction talk

← Older revision		Revision as of 04:47, 1 July 2006
Line 197:		Line 197:

	[[User:Xnera\|xnera]] 20:57, 30 June 2006 (PDT)		[[User:Xnera\|xnera]] 20:57, 30 June 2006 (PDT)

			==More Direction Discussion!==
			<jalapenodude> ok so I'm going to go command by command.

			<jalapenodude> Zone and mount start a program. Zone also moves the program left.

			<xnera> agreed

			<jalapenodude> Mount moves the program SE.

			<jalapenodude> Switch also moves the program southeast

			<xnera> how are you getting mount moving it SE?

			<jalapenodude> run -- --/switch -- --/-- end --

			<jalapenodude> I'm skipping copy for now

			<jalapenodude> sorry--I'm skipping rotate and working on copy

			<jalapenodude> copy moves down

			<pelrun> pread goes left

			<pelrun> pwrite goes right

			<pelrun> popen is right of course

			<pelrun> perhaps run actually does fire execution chains off in all four directions, but only shows the most interesting message of the four - maybe the one that executed the longest without error

			<jalapenodude> YES I DID IT!!!!!!

			<pelrun> cool

			<xnera> awesome!

			<pelrun> looks like the one that was stumbled on earlier (but the program was lost)

			<jalapenodude> the trick is that rotates go NE on odd-numbered columns and N on even-numbered columns

			<jalapenodude> it should be Zone mount run -- /copy zone zone cache --/copy pwrite rotate -- --/copy rotate -- -- --/rotate -- -- -- --

			[[User:Xnera\|xnera]] 21:47, 30 June 2006 (PDT)

	==Frob==		==Frob==

Xnera at 03:58, 1 July 2006

2006-07-01T03:58:52Z

Xnera at 03:57, 1 July 2006

2006-07-01T03:57:21Z

← Older revision		Revision as of 03:57, 1 July 2006
Line 90:		Line 90:

	<jbd> and possibly no rotations, if that's possible		<jbd> and possibly no rotations, if that's possible

			== Directions? ==

			<jbd> check this out: run cache -- / popen switch -- / -- -- end
			<jbd> this hits the end statement
			<jbd> the switch is making the execution arrow go SE
			<jbd> flush makes it go south!
			<jbd> switch direction SE, flush direction S
			<jbd> I wish I knew how rotate worked
			<jbd> It's still eluding me
			<jbd> based on what's happening with the End statements I'm testing switch definitely goes SE and flush goes S
			<jalapenodude> how are you testing?
			<jbd> Whenever it hits and End, it gets a consistent error message
			<jbd> so you can use End to test if it makes it to a particular square
			<jbd> like this one:
			<jbd> run flush end / popen rotate end / - end end Unable to write to file
			<jbd> I know it can't have gone outside of the boundry.
			<jbd> and the r1c4 also still seems to be 'special'.
			<jbd> this also gets V text: -- rotate -- / popen flush cache /-- rotate --
			<jbd> ah, you don't need the rotate on the first line
			<jbd> you do need it on the third line
			<jbd> because the flush hits the rotate
			<jalapenodude> where does the run go in that grid?
			<jalapenodude> I just get "no disk access"
			<jbd> sorry, I am including only the right half
			<jbd> here's the full thing
			<jbd> zone mount run -- --
			<jbd> -- -- popen flush cache
			<jbd> -- -- -- rotate --
			<jbd> 'flush' moves the execution down to rotate. somehow from there it gets to the cache
			<jbd> what I am needing now is an execution that hits four rotates before cache
			<jalapenodude> so the program moves downwards at the beginning, not right?
			<pelrun> what if the rotate functions on the execution direction of the cell after it (i.e. the rotate is delayed one step?)
			<pelrun> doesn't seem to fit this example though, if it's going down from flush
			<jbd> yeah. somehow it hopes from rotate to cache
			<jalapenodude> okay so zone mount run -- --/-- -- flush cache --/-- -- rotate -- -- also works
			<jbd> that's handy to know
			<pelrun> what if the rotate doesn't function just on execution direction, but on the surrounding cells themselves?
			<jbd> incidentally, I'm starting to think there's one pointer
			<jbd> not multiple forks
			<jbd> and that's an interesting thought perlrun
			<xnera> pelrun, you mean that the cells rotate around the cell that has the rotate in it?
			<pelrun> yeah
			<xnera> and jbd, i am glad I am not the only person who misreads pelrun's name as perlrun :)
			<jbd> hey, this is cool ... (cutting out first two columns)
			<pelrun> heh, not the first time I've been mistaken for a man page either XD
			<jbd> run - - / flush - - / popen end -
			<jbd> that hits the end
			<jbd> the popen makes it go east
			<jbd> maybe All these commands are just directional somehow
			<pelrun> yup
			<pelrun> each command has an effect and a execution direction
			<xnera> huh. i've been automatically going east after a popen without realising it
			<jalapenodude> right side only: run cache --/popen rotate-- gets the approached the master one

			[[User:Xnera\|xnera]] 20:57, 30 June 2006 (PDT)

	==Frob==		==Frob==

Jbd at 03:02, 1 July 2006

2006-07-01T03:02:55Z

@@ Line 30: / Line 30: @@
 ==Question==
 What are Frob and Twiddle? TheBozzball asks a dumb question.
 ==Frob==

Xnera at 01:49, 1 July 2006

2006-07-01T01:49:27Z

← Older revision		Revision as of 01:49, 1 July 2006
Line 1:		Line 1:
			==Commands==

			I googled the various commands to try to figure out what they do. I'm not 100% certain of this, which is why I'm putting it on the discussion page rather than the article page. Anything not specifically cited was found via a google for "man ''command''".

			; Zone : creates a "virtual machine", seperate from the rest of the operating system, that lets you securely run applications in an isolated environment. I strongly suspect all commands must start with zone. [http://docs.sun.com/app/docs/doc/817-1592/6mhahuoo1?a=view]
			; mount : mounts the file system. This is a common Unix command, and usually one of the first tasks you do.
			; popen : forks a new process that invokes the shell. It may close previous popen processes.
			; pread : reads input from a specified position of a file.
			; pwrite : writes data to a specified position of a file. This command doesn't seem to be allowed.
			; flush : this may have to do with deferred mail. [http://seth.positivism.org/man.cgi/flush]
			; rotate : possibly rotates the logs [http://www.hmug.org/man/8/rotatelogs.php]
			; switch : uncertain. It may be a SELECT...CASE statement [http://www.hmug.org/man/3/Switch.php]. Otherwise, my first guess was a change directory (CD) command.
			; copy : assuming it copies the data. Doesn't seem to work.
			; run : run a set of commands.
			; end : end the current run.

			I've been trying to "translate" the command grid into English to get a better sense of how it works. One possible "translation" for 'zone, mount, run, rotate, rotate/ - - popen, rotate, cache/ - - end - -' might be:
			# Start a virtual machine.
			# Mount the filesystem.
			# Run the following commands:
			## Rotate the logs twice.
			## Open a new process.
			## Rotate the logs once.
			## Cache the data.
			## End.

			This is reading left-to-right row by row. I'm uncertain if this is correct.[[User:Xnera\|xnera]] 18:49, 30 June 2006 (PDT)


	==Question==		==Question==
	What are Frob and Twiddle? TheBozzball asks a dumb question.		What are Frob and Twiddle? TheBozzball asks a dumb question.

TheBozzball at 21:25, 30 June 2006

2006-06-30T21:25:57Z

← Older revision		Revision as of 21:25, 30 June 2006
Line 1:		Line 1:
			==Question==
			What are Frob and Twiddle? TheBozzball asks a dumb question.

	==Frob==		==Frob==

TheBozzball at 21:18, 30 June 2006

2006-06-30T21:18:39Z

← Older revision		Revision as of 21:18, 30 June 2006
Line 262:		Line 262:

	[17:03] <Book[sidelines]> Yup.		[17:03] <Book[sidelines]> Yup.

			[17:17] <TheBozzball> If you do programs consisting of just "run", you get invalid command grid unless you put run on the first column, in which case you get "no disk access". If you do programs consisting of square 1: some command, square 2: run, you get - for Switch, Copy, Flush, Pread, Cache, Mount - Inv. Command Grid, for the other 6 - No Disk Access

← Older revision		Revision as of 03:58, 1 July 2006
Line 94:		Line 94:

	<jbd> check this out: run cache -- / popen switch -- / -- -- end		<jbd> check this out: run cache -- / popen switch -- / -- -- end

	<jbd> this hits the end statement		<jbd> this hits the end statement

	<jbd> the switch is making the execution arrow go SE		<jbd> the switch is making the execution arrow go SE

	<jbd> flush makes it go south!		<jbd> flush makes it go south!

	<jbd> switch direction SE, flush direction S		<jbd> switch direction SE, flush direction S

	<jbd> I wish I knew how rotate worked		<jbd> I wish I knew how rotate worked

	<jbd> It's still eluding me		<jbd> It's still eluding me

	<jbd> based on what's happening with the End statements I'm testing switch definitely goes SE and flush goes S		<jbd> based on what's happening with the End statements I'm testing switch definitely goes SE and flush goes S

	<jalapenodude> how are you testing?		<jalapenodude> how are you testing?

	<jbd> Whenever it hits and End, it gets a consistent error message		<jbd> Whenever it hits and End, it gets a consistent error message

	<jbd> so you can use End to test if it makes it to a particular square		<jbd> so you can use End to test if it makes it to a particular square

	<jbd> like this one:		<jbd> like this one:

	<jbd> run flush end / popen rotate end / - end end Unable to write to file		<jbd> run flush end / popen rotate end / - end end Unable to write to file

	<jbd> I know it can't have gone outside of the boundry.		<jbd> I know it can't have gone outside of the boundry.

	<jbd> and the r1c4 also still seems to be 'special'.		<jbd> and the r1c4 also still seems to be 'special'.

	<jbd> this also gets V text: -- rotate -- / popen flush cache /-- rotate --		<jbd> this also gets V text: -- rotate -- / popen flush cache /-- rotate --

	<jbd> ah, you don't need the rotate on the first line		<jbd> ah, you don't need the rotate on the first line

	<jbd> you do need it on the third line		<jbd> you do need it on the third line

	<jbd> because the flush hits the rotate		<jbd> because the flush hits the rotate

	<jalapenodude> where does the run go in that grid?		<jalapenodude> where does the run go in that grid?

	<jalapenodude> I just get "no disk access"		<jalapenodude> I just get "no disk access"

	<jbd> sorry, I am including only the right half		<jbd> sorry, I am including only the right half

	<jbd> here's the full thing		<jbd> here's the full thing

	<jbd> zone mount run -- --		<jbd> zone mount run -- --

	<jbd> -- -- popen flush cache		<jbd> -- -- popen flush cache

	<jbd> -- -- -- rotate --		<jbd> -- -- -- rotate --

	<jbd> 'flush' moves the execution down to rotate. somehow from there it gets to the cache		<jbd> 'flush' moves the execution down to rotate. somehow from there it gets to the cache

	<jbd> what I am needing now is an execution that hits four rotates before cache		<jbd> what I am needing now is an execution that hits four rotates before cache

	<jalapenodude> so the program moves downwards at the beginning, not right?		<jalapenodude> so the program moves downwards at the beginning, not right?

	<pelrun> what if the rotate functions on the execution direction of the cell after it (i.e. the rotate is delayed one step?)		<pelrun> what if the rotate functions on the execution direction of the cell after it (i.e. the rotate is delayed one step?)

	<pelrun> doesn't seem to fit this example though, if it's going down from flush		<pelrun> doesn't seem to fit this example though, if it's going down from flush

	<jbd> yeah. somehow it hopes from rotate to cache		<jbd> yeah. somehow it hopes from rotate to cache

	<jalapenodude> okay so zone mount run -- --/-- -- flush cache --/-- -- rotate -- -- also works		<jalapenodude> okay so zone mount run -- --/-- -- flush cache --/-- -- rotate -- -- also works

	<jbd> that's handy to know		<jbd> that's handy to know

	<pelrun> what if the rotate doesn't function just on execution direction, but on the surrounding cells themselves?		<pelrun> what if the rotate doesn't function just on execution direction, but on the surrounding cells themselves?

	<jbd> incidentally, I'm starting to think there's one pointer		<jbd> incidentally, I'm starting to think there's one pointer

	<jbd> not multiple forks		<jbd> not multiple forks

	<jbd> and that's an interesting thought perlrun		<jbd> and that's an interesting thought perlrun

	<xnera> pelrun, you mean that the cells rotate around the cell that has the rotate in it?		<xnera> pelrun, you mean that the cells rotate around the cell that has the rotate in it?

	<pelrun> yeah		<pelrun> yeah

	<xnera> and jbd, i am glad I am not the only person who misreads pelrun's name as perlrun :)		<xnera> and jbd, i am glad I am not the only person who misreads pelrun's name as perlrun :)

	<jbd> hey, this is cool ... (cutting out first two columns)		<jbd> hey, this is cool ... (cutting out first two columns)

	<pelrun> heh, not the first time I've been mistaken for a man page either XD		<pelrun> heh, not the first time I've been mistaken for a man page either XD

	<jbd> run - - / flush - - / popen end -		<jbd> run - - / flush - - / popen end -

	<jbd> that hits the end		<jbd> that hits the end

	<jbd> the popen makes it go east		<jbd> the popen makes it go east

	<jbd> maybe All these commands are just directional somehow		<jbd> maybe All these commands are just directional somehow

	<pelrun> yup		<pelrun> yup

	<pelrun> each command has an effect and a execution direction		<pelrun> each command has an effect and a execution direction

	<xnera> huh. i've been automatically going east after a popen without realising it		<xnera> huh. i've been automatically going east after a popen without realising it

	<jalapenodude> right side only: run cache --/popen rotate-- gets the approached the master one		<jalapenodude> right side only: run cache --/popen rotate-- gets the approached the master one


	[[User:Xnera\|xnera]] 20:57, 30 June 2006 (PDT)		[[User:Xnera\|xnera]] 20:57, 30 June 2006 (PDT)