Re: Kernel extensions and code injection?
Re: Kernel extensions and code injection?
- Subject: Re: Kernel extensions and code injection?
- From: Ben Dougall <email@hidden>
- Date: Tue, 22 Mar 2005 20:39:05 +0000
On Tuesday, March 22, 2005, at 07:39 pm, Michael D. Crawford wrote:
There's also other ways malicious coders could
intercept keystrokes if they really wanted to, including ways of
getting
text in password fields, and there's not much we can do about it.
It happens that my very first commercial software product was a
keystroke logger called Last Resort from Working Software. It was a
Mac control panel and INIT.
It had the problem that it would capture passwords too, but those were
gentler, less-networked days. The control panel allowed one to
disable and renable it, and Last Resource made its presence known by
showing its icon on the boot screen with the other inits.
I received many grateful letters and phone calls from writers both
amateurs and pro, whose Great American Novels were saved from crashes,
power failures, and even just deleting a page of unsaved text by my
little eight kilobyte program.
I started to think about writing an OS X version of it the other day
when a consulting contract proposal I had just written was lost when
my wife's laptop hung just as I saved it. There was a file there but
it was completely unrecoverable. It is no help that OpenOffice
documents are zipped XML files; an error in the middle of a zip file
means you can't get ANY of your text back.
Bonita's laptop runs XP, but I don't know much about XP so instead I
thought about doing it for OS X.
Other consultants wrote versions for DOS and Windows. Each of us had
our own solution to the password problem. I have some ideas now of
how I can solve it, without doing anything skanky as I was so devoted
to doing during my time as Working Software. (I am a wizard with
MacsBug.)
I also wrote Last Resort Programmer's edition which recorded menu
selections and command keys, to help QA people reproduce bugs after
crashes.
i'm interested in code that would do these sort of things. i'd like
code that records key presses (minus passwords if possible), mouse
movements, clicks and even recorded menu selections as mentioned above
if possible (didn't think that would be possible), on a per user basis
within any app they might use. record as much user activity -- what
they're doing where -- details as possible basically. i'm just asking
out of curiosity at the moment. don't know anything about kernel
programming -- is that what'd be necessary for this type of thing? is
there any open source code that exists already that does anything like
this? roughly what would be necessary to do something like this? would
this sort of continual recording effect the performance of the user's
machine at all (if not on a pretty modern machine what about a g3 for
example?) -- perceptible pauses/stutters?
thanks, ben.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-kernel mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden