On Aug 11, 2008, at 1:35 PM, Harsha Godugu wrote:
> Jerome Dochez wrote:
>> Lloyd Chambers wrote:
>>> Harsha,
>>>
>>> I expect I will need to make more HK2 changes/additions as V3
>>> development progresses, so this will be an ongoing issue.
>> I agree, but I really don't see why hk2 is special here,
> because, hk2 seems like a very critical component -- any failure in
> hk2 will just brings everything to HALT. (may be we should define
> what other v3 modules/projects are in such category 'critical' to v3)
I think we see the problem with hk2 first but the truth is that in
such a situation, we would grind to a halt an some point anyway unless
we always build everything from sources which is the job of RE, not
individual developer.
let's talk about this tomorrow morning...
jerome
>
>
> I had the same thought as Lloyd. hk2 shouldn't be a standing
> outside v3 (as one option). By doing this, there is no such
> dependency defined on v3 where it needs to download the hk2
> artifacts from a remote repo. We should definitely have a
> contingency plan in case, a similar problem with repos happens
> again. Adequate soak time for the new bits in hk2 is one idea.
>
>> if artifacts are not published to the maven repo, anyone not
>> building the entire source tree is in hell. if you are building the
>> entire source tree then create a local super pom that does hk2 and
>> v3 compilation together.
> I think, the above is inevitable when we use maven. (no network ->,
> no repos.. -> no build / no product?)
>
> thanks..
>>>
>>> One option is for HK2 to become part of V3. Another is to
>>> document exactly what must be done for a seamless push of HK2,
>>> preferably automated without any voodoo.
>>>
>>> For last weekend's fun, an API (interface) change meant that HK2
>>> and V3 needed to change in sync
>> if our tools (namely the download.java.net repo server) was working
>> fine, this is not as bad as you describe, I have done it many
>> times. The real issue is the repo fragility, since with a
>> fragile repository, any release is impossible, but also any
>> snapshot is potentially outdated creating a real threat for anyone
>> building only a few modules for productivity.
>>
>> jerome
>>>
>>> Lloyd
>>>
>>> ..............................................
>>> Lloyd Chambers
>>> lloyd.chambers_at_sun.com
>>> GlassFish team
>>>
>>>
>>> On Aug 11, 2008, at 12:09 PM, Harsha Godugu wrote:
>>>
>>>> In future, if we touch hk2, (and still use SNAPSHOTs) probably it
>>>> is a good idea to give at least 3days lead time to do such
>>>> change in hk2 first (or any parent of v3)
>>>> and then come back to v3 to do the dependent changes. What do you
>>>> think?
>>>
>>>
>>>
>>>
>>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>