[mythtv] issues with 16491?

[Michael T. Dean]

On 03/16/2008 10:04 PM, Nigel Pearson wrote:
>> Not sure what an EUID is
> Effective User ID. Basically,
> a number that represents your username.
>
> Every shell that I tried has it set,
> so I am mystified about this failure, Matt.
>   

I'm pretty sure ksh doesn't have it, and I'm thinking the original 
Bourne shell didn't.  Doesn't Ubuntu use dash or something (perhaps only 
for root)?  I don't think dash sets EUID.  Unfortunately, I can't even 
find a man page for it (there are a lot of man pages that spell out the 
word "dash," meaning hyphen).  However, unfortunately even the man pages 
don't tend to list all the things they don't have ;), so a direct test 
(from a shell that /does not inherit/ its environment from a bash shell) 
would be a safer way to find out.  (Perhaps in your tests, Nigel, some 
shells may have inherited the EUID env var?)

Regardless, if we do patch the script to allow execution with shells 
that don't set it, a more robust version of the quick-test patch I gave  
would be required.  It should really have a fallback in case EUID isn't 
and can't be set (i.e. if /usr/bin/id is not an executable) to allow 
running the script without the -p, as before the recent changes.  Those 
who don't have EUID and don't have /usr/bin/id can wait for the frontend 
to redo prescaling.

Oh, and TTBOMK, the id executable--if installed--should always be 
installed at /usr/bin/id.  It's part of coreutils, so it's probably 
installed on all GNU/Linux systems, though I don't know if that's a safe 
assumption on other OS's.  And, even on GNU/Linux, it's always possible 
someone may have moved it.  Therefore, we should really have a fallback 
just in case there is no EUID even after attempting to set it.

Mike