--user option

[Vlad Skvortsov]

Иван Глушков wrote:

[skipped]
> I also reopened i#98 as filter "my" doesn't understand environment
> variable $DITRACK_USER:
>
> $ grep my etc/filters
> 0.5-my: Due-in=0.5,Status=open,Owned-by=$USER
> my: Status=open,Owned-by=$USER
> $ echo $USER
> glushkov
> $ echo $DITRACK_USER
> gli
> $ ./dt ls my
> $ USER=gli ./dt ls my
> 80   gli      0.7      open   DITRACK_ROOT messes up the test suite
> 124  gli      0.6      open   svn:ignore an LMA (.ditrack)
> 137  gli      0.6      open   'act' is case-sensitive to issue names
>
> This behaviour differs from that in our ChangeLog:
>   

No, it doesn't. We never claimed that either '--user' or DITRACK_USER 
override the value of the USER environment variable. We claimed that 
'--user', DITRACK_USER and USER are the ways to communicate the user id 
to DITrack -- which holds true. We claimed that one may reference 
environment variables from within predefined filters -- which is also true.
> "It is now possible to specify user id through the '--user'/'-u' command line
> option or by setting the DITRACK_USER environment variable (i#98)."
>
> We should create some "synonym" for abstract "ditrack user" to use it
> in the filters. Without it we can't ensure our announcement.
> I suppose it can be $USER (as it actually is). Not in the meaning of
> environment variable, but in the meaning of "ditrack user".
>   

It's a separate issue and I was also thinking about addressing it. I 
propose to stick to the current syntax to refer to environment variables 
and come up with alternative notation for DITrack-provided values. We 
might, for example, use something like:

my: Status=open,Owned-by=%DTUSER

-- 
Vlad Skvortsov, vss at 73rus.com, http://vss.73rus.com