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

[Bruno Raoult <braoult@example.com>]

On Sat, Sep 26, 2009 at 20:03, Curt Sampson <cjs@example.com> wrote:

On 2009-09-26 19:05 +0900 (Sat), Bruno Raoult wrote:

> Eh? are you joking? Even "script" will be acceptable (sending for
> instance to syslog, so that I have no write access)... I asked if
> there was a lighter solution, not a heavy one :-)

Well, you are in a very unusual situation, then, in that it's ok if
users subvert the logging system, so long as they do it in some other
way than modifying the file that script writes after the fact. E.g.,
this log of command, which does not include the one I ran between stty
and exit, would be fine for your compliance process:

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"...

If it's ok to subvert the system, so long as it isn't done in certain
specific ways, you ought always to mention this when asking questions
like this, since most people will interpret "the logged user must not
be able to change the log file afterwards" as an attempt to say, "the
logged user should not be able to subvert the logging system."

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).

Thanks,

Bruno.

-- 
2 + 2 = 5, for very large values of 2.