[MERGE] Authentication ring implementation (read-only)

John Arbash Meinel john at arbash-meinel.com
Tue Nov 13 17:13:34 GMT 2007


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

Martin Pool wrote:
>>> If it doesn't actually *have* a super method, it seems a bit strange or
>>> misleading to call it.  I don't know if we have a specific policy on it.
>>> Calling super always to get "always called first" behaviour like Java or
>>> C++ seems not quite right.
>>>
>> Well, in Python, if you want to properly support multiple inheritance, you
>> actually have to do this.
>>
>> The issue is that in more convoluted schemes, 'super()' doesn't actually call
>> the base class.
> 
> If the call is there so that you can do MI, then a comment saying so
> would be good, to be sure no one removes it and it's borne in mind
> during other changes.  In practice as this constructor has more
> mandatory parameters than its super method it will be a little tricky
> (not impossible) to use this way....
> 

Yeah, I think in this case we can just get rid of it. As I don't think
any of our classes really support multiple inheritance anyway. And there
hasn't been any need to do so. So I'm guessing this is a YAGNI, and
better to just get rid of it.

I was just pointing out when you actually do need a call to super() for
the base class.

John
=:->

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

iD8DBQFHOds+JdeBCYSNAAMRArz9AKCDWJKGy+w+NZ5ARhcmbVObLIFI4gCgvdpp
X0sr/EtwUE82G/KHadTWY38=
=YyF4
-----END PGP SIGNATURE-----



More information about the bazaar mailing list