precedence order for aliases

[Saša Janiška]

Hi!

I'm noticing some strange behaviour how Bazaar handles user-defined aliases and I'm not sure whether it is a bug or feature...

I defined one alias in ~/.bazaar/bazaar.conf as:

[ALIASES]
checkout=checkout --lightweight

Invoking 'bzr help checkout' gives:

[...]
Aliases: co
See also: branch, checkouts
'bzr checkout' is an alias for 'bzr checkout --lightweight'.

and, therefore I'd expect that 'bzr co' would expand to 'bzr checkout' first and then to 'bzr checkout --lightweight', but it does not so.

Maybe it's a feature, but, imho, it's not orthogonal and I'd expect that user-defined alias is expanded as the last. Finally, the above help file output implies so ;)

Sincerely,
Gour

[Aaron Bentley (abentley)]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Gour wrote:
> Public bug reported:
>
> Hi!
>
> I'm noticing some strange behaviour how Bazaar handles user-defined
> aliases and I'm not sure whether it is a bug or feature...

It's a feature. It keeps the model simple. It allows the aliases to
vary independently. It allows the original command be given an alias,
while an alias is created for the original name.

> [...]
> Aliases: co
> See also: branch, checkouts
> 'bzr checkout' is an alias for 'bzr checkout --lightweight'.

> Maybe it's a feature, but, imho, it's not orthogonal and I'd expect that
> user-defined alias is expanded as the last. Finally, the above help file
> output implies so ;)

I don't think so. User aliases and built-in aliases are completely
different.

Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFIPdy+0F+nu1YWqI0RAh1kAJ9ag0Ail2ww3u7SHUwW8pGxoQQExwCffz83
iFO0EktI0SQhTZktoNW7SZ0=
=m9wj
-----END PGP SIGNATURE-----

[Saša Janiška (gour)]

>>>>> "Aaron" == Aaron Bentley <email address hidden> writes:

Hi Aaron,

Aaron> It's a feature. It keeps the model simple. It allows the
Aaron> aliases to vary independently. It allows the original command be
Aaron> given an alias, while an alias is created for the original name.

OK. I'm trying to adjust to the model...

Aaron> I don't think so. User aliases and built-in aliases are
Aaron> completely different.

Heh, for the experiment's sake I tried to define:

co=checkout --lightweight

alias and was curios what would happen.

Fortunately, bzr did the Right Thing :-)

So, now I see & understand that user aliases override built-in ones and
can live with that ;)

Shall we close this 'bug' as invalid?

Sincerely,
Gour

--

Gour | Zagreb, Croatia | GPG key: C6E7162D
----------------------------------------------------------------

[John A Meinel (jameinel)]

This is by design.

[Saša Janiška (gour)]

>>>>> "John" == John A Meinel <email address hidden> writes:

Hi John,

John> This is by design.

Thank you for reply.

Yes, I see it now and can live with it ;)

Sincerely,
Gour

--

Gour | Zagreb, Croatia | GPG key: C6E7162D
----------------------------------------------------------------