Re: preliminary version control requirements

["Perry E. Metzger" <perry%piermont.com@localhost>]

"S.P.Zeidler" <spz%serpens.de@localhost> writes:
> Thus wrote Perry E. Metzger (perry%piermont.com@localhost):
>
>> >    - allows checkout without cloning entire history
>> 
>> I don't necessarily agree with the last bit there. It seems reasonably
>> arbitrary. There are systems that work very well but require local
>> copies of the repository. In a modern environment, where disk is quite
>> cheap, that seems fine to me. (And yes, you can mount a 500GB disk you
>> bought for $80 on your Vax over NFS.)
>
> You forget the network load. A factor three may not mean much at the
> client end, but it's nothing to be sneezed at at the server end.

Sure, but requiring the whole repository locally doesn't necessarily
mean that the server load would be high. The protocol might send
diffs, for example. Several of the distributed protocols out there
also have the feature that they implicitly sign the repository, so you
can use untrusted mirrors at will without worrying. (That's a big
useful feature IMHO, by the way.)

>> >    - somehow preserves old file version numbers on import for reference
>> >    - uses version numbers that are numbers, not hash codes
>> 
>> I'm not sure that is a critical requirement. Most modern systems don't
>> even have individual file version numbers anyway, they have repository
>> versions with atomic commits across the repo.
>
> 'Is the version that crashes for you before or after I committed my fix'
> should be a question that doesn't require extensive research to answer.

That's a fine requirement. I'd put that in rather than the specific
requirement that individual files have version numbers.

>> >    - doesn't exhibit any excessively strange semantic excursions
>> 
>> I don't know what that means.
>
> If people throw up when being presented the user interface it's probably
> Not The One :)

If they throw up for any reason the solution is probably unacceptable.


-- 
Perry E. Metzger                perry%piermont.com@localhost