Here's version 5, which resyncs with the gate since the latest put back 
touches code in the same area:
http://cr.opensolaris.org/~bpytlik/ips-1651-v5/

Brock

Brock Pytlik wrote:
> Danek Duvall wrote:
>   
>> On Fri, Sep 26, 2008 at 04:02:59PM -0700, Brock Pytlik wrote:
>>
>>   
>>     
>>> http://cr.opensolaris.org/~bpytlik/ips-1651-v3/
>>>     
>>>       
>> image.py:
>>
>>   - line 103: I'd just call this "ImageNotFoundException".  More fully, I'd
>>     add a boolean member to it called "specified", setting it to true when
>>     you raise it where you do now, and replace the ValueError we're
>>     currently raising with an ImageNotFoundException where "specified" was
>>     False, and possibly combine the messaging around the two when caught.
>>   
>>     
> Ok
>   
>>   - line 104: "an" -> "a"
>>   
>>     
> fixed
>   
>>   - line 223: update the comment to reflect introduction of exact_match
>>   
>>     
> fixed
>   
>>   - line 253-255: what's this third test for?  Also, binary operators
>>     should be at the end of a line, not the beginning.
>>   
>>     
> I thought that allowing the user to either specify /foo/bar/ or 
> (/foo/bar/var/pkg or /foo/bar/.org.opensolaris,pkg) would be more user 
> friendly than requiring only one or the other. The second test is 
> testing against /foo//bar and the third is testing against 
> (/foo/bar/var/pkg or /foo/bar/.org.opensolaris,pkg).
>   
>> client.py:
>>
>>   - line 2110: isn't this always going to be "mydir"?  Why keep this in the
>>     exception?  And I don't see you using the other two members, either.
>>   
>>     
> I keep it in the exception because it removes assumptions. Specifically, 
> only find_root knows what directory it was looking for. In this case, 
> it's fairly obvious from the code, but in other places, it may be less 
> obvious what find_root was called with. In general, I think exceptions 
> should contain all the relevant information available that's reasonable 
> to have. That's the same reason the other bits of information are being 
> returned even though they're not currently being used (they were 
> originally, but Dan suggested simplifying the output). If we have 
> problems in this area, I think it will be easier to debug if we can 
> easily get to the root dir and img_dir that find_root is choosing, which 
> the current exception allows. (Specifically, we can ask a user to change 
> the print line at the end of client, rather than asking them to get into 
> the middle of image.py.) Also, another client may make a different UI 
> choice in what to display to the user when this exception is raised.
>   
>> t_R_option.py:
>>
>>   - I don't think you're using "os", "time", or "stat" imports.
>>
>>   
>>     
> Right
>
> New webrev at:
> http://cr.opensolaris.org/~bpytlik/ips-1651-v4/
>   
>> Danek
>>   
>>     
>
> _______________________________________________
> pkg-discuss mailing list
> [email protected]
> http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
>   

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

Reply via email to