On Fri, Jul 25, 2008 at 09:52:46AM -0700, Danek Duvall wrote:
> On Thu, Jul 24, 2008 at 04:09:43PM -0700, Brad Hall wrote:
> 
> > Link to new CR: http://cr.opensolaris.org/~bhall/bug-2627-1/
> 
> We use getopt in the rest of the code, rather than optparse.  Unless
> there's something that optparse really buys you, please use getopt.

OK, fixed.
 
> I think -p is more consistent with the rest of solaris than -m (p for
> parseable).  Particularly your use of "mach" as a short form throughout, it
> strongly means "machine type" to me, as in the command "mach".
 
Makes sense, fixed that.

> Is it worth assuming that the baseline for a test is "pass", and keeping
> the baseline files shorter?
 
I think that may confuse things at some point if we decide to print out a
message like "test X wasn't in the baseline, you probably need to regenerate",
so probably not worth it at this point.

> baseline.py:
> 
>   - line 49: perhaps just make the generate default to False, and pass in
>     True instead of 1?  Or you could do bool(generate) here.  Similarly on
>     line 74, I think.
 
Good point, switched to use boolean. 

>   - line 63: return self.results.get(name, None)

Didn't realize you could do that, fixed.
 
>   - line 87: "if lst"
> 
>   - line 105, 124: I don't think this work.  This will catch exceptions of
>     type IOError or type e.
> 
>   - line 106: "storing"
> 
>   - line 130: no need for parens on the lhs

OK, got all those too.

Thanks for reviewing, I'll post another webrev shortly.

Thanks,
Brad
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss

Reply via email to