[Pulp-list] Bulk Repo Creation - Thoughts/Comments?

[Pradeep Kilambi]

On 03/30/2011 05:49 PM, Mike McCune wrote:
> On 03/30/2011 01:22 PM, Jay Dobies wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> On 03/30/2011 04:22 PM, Jason L Connor wrote:
>>> On Wed, 2011-03-30 at 14:55 -0400, Pradeep Kilambi wrote:
>>>>> I'd say we drop "rhn", and determine local from feed url;
>>>> http(s)://
>>>>> ->  remote and file:// ->  local
>>>>>
>>>>> -Todd
>>>>
>>>> yea if we can drop rhn; that should make things easy to determine.
>>>>
>>>> ~ Prad
>>>
>>> +1
>>> Yes please. I've always found the syntax of (yum|rhn|local):<uri>
>>> confusing.
>>
>> Same. I haven't had time to really think about this thread, but I agree
>> that that part has always thrown me off. I still forget it from time to
>> time.
>>
>
>
> what happens when the remote repo is a deb repo?  Or the remote repo 
> is just an http collection of files? or if the remote repo is a maven 
> repo? etc... http or https doesn't necessarily always mean the remote 
> location is a yum based repo.
>
> hence, the requirement for a type.  I'd hesitate to drop type all 
> together if we want to make pulp support multiple types of 
> repositories in the future but we could *default* to yum for 
> http/https based repos but allow for overrides if a type is specified.
>
> Mike

Yea I think you bring up a good point. The types we have today dont 
necessarily serve the use case you mentioned; we just assume everything 
to be a yum repo. That will change with this feature where we will also 
add support for file syncs. If we want to achieve this without any 
prefix, we could scan the remote url for repodata directory and 
differentiate between types. So if the remote url has a repodata dir, we 
use yum sync else file sync or if content type specifies both rpms and 
files then both. Alternately make types passed to --content more precise 
so we could use them for type distinction.

~ Prad


>
> _______________________________________________
> Pulp-list mailing list
> Pulp-list at redhat.com
> https://www.redhat.com/mailman/listinfo/pulp-list