On Tue, Feb 10, 2009 at 04:20:09PM -0600, Shawn Walker wrote:

>> build_entire_incorporation:
>>
>>   - line 50: "print" instead of "ls"
>
> Really? That doesn't seem to work: "command not found".

Hm?  In a ksh script, that should work just fine.

>> solaris.py:
>>
>>   - line 29: why?
>
> Because many of our exceptions use _() now and fail if gettext isn't 
> installed.

Okay for now, but I don't really like that.  Our exceptions might be thrown
deep inside some other application that happens not to be
internationalized, and then they're absolutely useless.  We should probably
look at defining a dummy _() in all library modules that use it, if it
hasn't already been defined.  Or something.  I'm not sure how our C
libraries deal with this, offhand.

>>   - line 1305: I'm confused by refresh_publish.  I would expect it to mean
>>     that we need to refresh, but you refresh only when refresh_publish is
>>     False?
>
> Perhaps the variable name is not helpful enough.  In short, if 
> refresh_publish is true, that means to refresh *when* publishing the 
> package.  So, if that isn't true, it needs to be done at the end instead.
>
> I've renamed this to 'defer_refresh' and added a comment.

Ah.  Yes, that makes more sense.

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

Reply via email to