[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [Libosinfo] [PATCH osinfo-db 0/6] centos and scientific linux



On 3/5/19 7:52 AM, Fabiano Fidêncio wrote:
> On Tue, 2019-03-05 at 13:48 +0100, Fabiano Fidêncio wrote:
>> On Tue, 2019-03-05 at 13:22 +0100, Fabiano Fidêncio wrote:
>>> On Fri, 2019-03-01 at 18:41 -0500, Cole Robinson wrote:
>>>> This series adds:
>>>>
>>>> * centos5 entries
>>>> * centos6 <tree> data
>>>> * scientificlinux 5.X
>>>> * scientificlinux 6.X
>>>> * scientificlinux 7.X
>>>>
>>>> No iso data is added, just URLs. I'm trying to get osinfo-db to
>>>> have
>>>> all the treeinfo coverage that virt-install has.
>>>
>>> Cole, in the general the series look good (apart from one change
>>> that
>>> has to be for "Add scientificlinux-7.X".
>>>
>>> There's one thing that I'm interested to know, though:
>>> - Is x.y considered EOL whenever x.(y+1) is released? I mean, will
>>> 7.6
>>> be considered EOL whenever 7.7 is released? If so, we'd also have
>>> to
>>> add the EOL to the 7.x entries.
>>>
>>> Anyways, for patches #1 to #5:
>>> Reviewed-by: Fabiano Fidêncio <fidencio redhat com> 
>>
>> Actually, let me take my "Reviewed-by" back.
>> Please, take a look at 5cac22bc68[0].
>>
>> There, the commit message states:
>> centos: Remove URLs pointing to vault.centos.org
>>
>> As vault.centos.org doesn't keep any ISO anymore, let's just remove
>> them from our db.
>>
>> Along with the URLs removal, let's remove together the tree's as
>> those
>> can't be accessed without a valid URL.
>>
>> Removing all the vault.centos.org URLs matches with the
>> recommendation
>> given by CentOS folks in #centos-devel:
>> "so in short, if some program links to vault, it's most likely not a
>> good idea and may not even work"
>>
>> [0]: 
>> https://gitlab.com/libosinfo/osinfo-db/commit/5cac22bc6852d56988ff4be090551c5ec2f3f108
>>
>> So, I guess the path to take is to drop #1 and #3.
> 
> Errr, dropping the URLs from #1 and #3, but keeping the tree/treeinfo.
> 

ACK from me, though what was centos reasoning for not pointing to
vault.centos.org tree URLs? Those have been stable for years in my
experience. I can understand if they don't want those advertised but
it's unclear why the comment suggests it might not work

- Cole


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]