Fwd: [ditrack commit] r1126 - src/trunk/DITrack/Command

[Vlad Skvortsov]

Oleg Sharov wrote:

[skipped]

>     > -                "Invalid entity name '%s'" % opts.fixed[1],
>     > +                "Invalid entity identifier: '%s'" % opts.fixed[1],
>     >                  fatal=True)
>     >
>     > Vlad, I tolk with you about IDENTIFIERS and NAMES (see
>     doc/design.txt)
>     > I think we must change the notion NAMES, it similar with
>     IDENTIFIERS,
>     > but using for local entities.
>
>     ...and the reason why we need to do this is ... ?
>
>
> Different messages such as "Invalid issue name" and "Invalid issue 
> identifier" will confused an end user

I can generally agree with this, but don't think it's a particular 
problem in this case. I'd say, if there is a reason for a user to see 
the difference, let it be: the information we provide might help (us, 
the developers) to solve a problem, so it's worth having it. Otherwise, 
hide it.


[skipped]

>
> from DITrack/Command/commit.py
> if x in string.digits:
>      "Non-local issue id '%s', ignored\n" % id 
>
> from design.txt
> "...local entities are assigned character ids, which we refer to as 
> NAMES..."
> but we give the ambiguous message with "local issue" and "id"

In this we are reporting on invalid input (input == command line options 
here). The 'commit' command expects a list of local ids to be committed 
(thus, we expect a list of NAMES). If the id supplied is invalid class 
(e.g. it's a NUMBER, not NAME), we properly report it.

Do you see flaws here?

> from DITrack/Command/commit.py
> "Invalid entity name '%s', ignored\n" % id
>
> we write "entiry name", but if an user call alphanumeric combination 
> we will write "Non-local issue ID"

Right - because this particular message ("Invalid entity name...") is 
displayed if we can't parse the input at all. Probably printing "Invalid 
id" would make more sense though.

Also, after giving this a bit of thought, how about aborting if the 
input is invalid in any way? I think it's more appropriate here; your 
thoughts?

> I think, we can named local/firm entities anyhow for ourselves.
> But an end user must see the homogeneous appellation "identifier" or "id"

As I said, I generally agree. With the exception of cases where a user 
(or we as developers) might want to know the difference.


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