| Re: [hudson-dev] Suggest we remove maven2 job type in Hudson-3 initial release |
Winston,
I haven't looked at the technical implications, all I am saying is that the core team should seriously consider the user experience we are presenting when people choose maven 2 job type.
If a big part of people who try to use it run into problems, and you (core team) don't have an plans to fix those issues, perhaps it was better to disable maven 2 job type all together..
Looking at the technical side I understand the static analysis plugins can take advantage of the maven 2 job type but is it a requirement or just optional benefit ? I think that will be what tips the scale
Secondly, This is something I need to confirm with m4, but I think I saw the maven 2 job type present despite not having selected the maven 2 plugin...
Best regards
Henrik
Henrik,
Maven2 Job type is available only if legacy maven plugin is installed.
By removing the plugin do you mean no longer make it available via
update center? If so, though I would like to do that, unfortunately
several plugins including static analysis plugin depend on that and
would break with out it.
Since static analysis plugins are some of the most used plugin, we
could rewrite them to conform to Maven3 plugin eventually.
Since we don't bundle any plugin, it is up to the user to install it. We
could move legacy maven plugin from Featured plugin list to "Others"
plugin list, thus giving less important to it.
- Winston
On 7/30/12 3:47 AM, Henrik Lynggaard Hansen wrote:
> Hi
>
> I have a suggestion for the initial Hudson 3 release. I think we
> should remove the Maven2 job type instead of leaving it in in
> deprecated state.
>
> The reason is that I have seen users raise issues where Maven2 job
> type has caused exceptions and other problems and the response has
> been to suggest moving to freestyle+ m3 integration, but I haven't
> seen much attention to fixing the issues. This is completely natural
> given the deprecated nature.
>
> If there are so many issues with the job type and no-one fixing them,
> wouldn't it be better to remove the broken feature instead of letting
> users attempt to make it work?
>
> Best regards
> Henrik