[Twisted-Python] Re: [Twisted-commits] r22428 - Deprecation API for functions and methods.

George Pauly george at ringdevelopment.com
Thu Jan 31 11:22:06 MST 2008


On Thu, 2008-01-31 at 16:11 +0000, glyph at divmod.com wrote:
> On 02:45 pm, exarkun at divmod.com wrote:
> >What's the right way to get a string describing a Version now?  It's an 
> >even
> >harder decision to make than it was before.  Is it:
> >
> >  * str(Version(...))
> >  * repr(Version(...))
> >  * Version(...).base()
> >  * Version(...).short()
> >  * getVersionString(Version(...))
> >
> >Can we do something to make this better?
> 
> This collection is super ad-hoc.  Let me try to express some of the 
> requirements that lead to the current confusion:
> 
>   * Sometimes you want a string that is formatted for use at a REPL, 
> which can be eval'd to give the user the idea that it is a structured 
> object and how they can build something similar (repr())
>   * sometimes you want something that's human-readable and expresses all 
> the information available about the version (str()).  The square 
> brackets probably come from the fact that I was playing WoW at the time, 
> and square brackets meant "this is a thing, not just something a guy 
> said".  I was thinking you'd want to see like [Twisted Epic Spaulders of 
> the Owl] for the version number.  (Maybe a bad call.)
>   * Sometimes you want just the version number in canonical format 
> (x.y.z+dev) not the name of the package (.short())
>   * I hypothesize that sometimes you want to omit the +dev part because 
> you're trying to fit into a stricter format, but I don't really remember 
> why .base() exists.
> 
> Was getVersionString added because the other string representations 
> weren't "friendly" enough?  I guess the 'rUnknown' makes them look a 
> little gnarly, but I'd actually like to see the SVN revision in the 
> cases where it's present; the right thing to do to make it friendly 
> would be to fix the entries-file parser.
> 
> Also, in any case, it should really be a method.  There's no reason to 
> have a free function defined right after a class whose only argument is 
> a single instance of that class :).
> 
> _______________________________________________
> Twisted-Python mailing list
> Twisted-Python at twistedmatrix.com
> http://twistedmatrix.com/cgi-bin/mailman/listinfo/twisted-python
-- 
George Pauly
Ring Development
www.ringdevelopment.com





More information about the Twisted-Python mailing list