Danek Duvall wrote:
> On Mon, Oct 20, 2008 at 05:43:35PM -0500, Shawn Walker wrote:
> 
>> Brock Pytlik wrote:
>>> I'm confused why the change in catalog lines 134-136 is needed. Unless 
>>> I'm confused about what content_root holds, it seems like that argument 
>>> shouldn't be needed in the call to rebuild the index since the depot 
>>> started in that manner doesn't serve any content. If it is necessary, I 
>>> think we should look at why it is and perhaps fix that instead.
>> It's needed as part of basic depot startup checks.
>>
>> I can see your point of view, but I'd rather leave the directory check 
>> intact, and that means that content root is required.
> 
> Why not make the directory check take --refresh-index into account?

Because refresh-index, etc. are command-line options to depot.py and do 
not control or alter the actual behaviour of server configuration or the 
depot functionality itself.  In other words, the fact that depot.py 
doesn't care about content_root if --refresh-index is passed does not 
translate in a generically applicable fashion to the server/catalog 
forking off a depot process.

> Perhaps content_root passed as None into the SvrConfig constructor could
> indicate it could be ignored, and set it to None if reindex is True?

That seems like a viable solution for the moment.

Updated webrev:
http://cr.opensolaris.org/~swalker/pkg-4102-2/

Diff from last webrev:
http://cr.opensolaris.org/~swalker/pkg-4102-2/v1-v2.patch

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

Reply via email to