Re: [tlug] comand-line recording...

[Curt Sampson <cjs@example.com>]

On 2009-09-26 21:23 +0900 (Sat), Bruno Raoult wrote:

> Yes, this will be enough... The target is only to be able to understand what
> "developers" are doing on production systems (vs support people).
> We are not targeting "bad people", but "mistakes"...

And you think that someone subverting, deliberately or accidently, the
contents of log files is a likely mistake? This would be unusual.

My impression from what I've seen so far (which could of course be
quite wrong, though I've got many years of developer and sysadmin
experience informing my perception of your problem) is that you've
already progressed a good way towards the wrong solution to whatever
problem you have.

> By the way, if you have an answer to my question, I would be grateful
> :-) And if you have one answer, this would also answer your remarks
> (as independent of screen output, which I don't want, as I said
> initially).

Well, I don't really have any idea about what you mean by "easy,"
either, since handing script's output to syslog in any reasonable way I
can think of would involve modifying the script program. But if you're
asking, "which program should I modify instead of script," I'd say the
shells, if you want to modify an existing program. But writing a new
program from scratch to let users sit on a pty and record keyboard input
(the way screen does with output) might be easier if you have to deal
with multiple shells, though it seems unlikely if you want to avoid
recording things that are not echoed to the screen, such as passwords.

cjs
-- 
Curt Sampson       <cjs@example.com>        +81 90 7737 2974
           Functional programming in all senses of the word:
                   http://www.starling-software.com