You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hc.apache.org by Oleg Kalnichevski <ol...@apache.org> on 2013/10/28 11:47:04 UTC
ASF release madness (rant)
Sooo, our JIRA link is out date. Big deal, trivial to fix [1], right?
But now we need to release a new parent POM and even with the lazy
consensus vote it still means a 3 day delay.
Maven and ASF release policies is a marriage made in hell.
Oleg
[1]
http://svn.apache.org/viewvc/httpcomponents/project/trunk/pom.xml?view=diff&r1=1536305&r2=1536306&pathrev=1536306
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org
Re: ASF release madness (rant)
Posted by Gary Gregory <ga...@gmail.com>.
On Mon, Oct 28, 2013 at 6:47 AM, Oleg Kalnichevski <ol...@apache.org> wrote:
> Sooo, our JIRA link is out date. Big deal, trivial to fix [1], right?
> But now we need to release a new parent POM and even with the lazy
> consensus vote it still means a 3 day delay.
>
> Maven and ASF release policies is a marriage made in hell.
>
You publish the web site from a snapshot POM... not great process-wise, but
with visible and workable results for web site users.
At Commons, the parent POM is vote rules are "lazy consensus", is it the
same here?
Gary
> Oleg
>
> [1]
>
> http://svn.apache.org/viewvc/httpcomponents/project/trunk/pom.xml?view=diff&r1=1536305&r2=1536306&pathrev=1536306
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
> For additional commands, e-mail: dev-help@hc.apache.org
>
>
--
E-Mail: garydgregory@gmail.com | ggregory@apache.org
Java Persistence with Hibernate, Second Edition<http://www.manning.com/bauer3/>
JUnit in Action, Second Edition <http://www.manning.com/tahchiev/>
Spring Batch in Action <http://www.manning.com/templier/>
Blog: http://garygregory.wordpress.com
Home: http://garygregory.com/
Tweet! http://twitter.com/GaryGregory
Re: ASF release madness (rant)
Posted by Oleg Kalnichevski <ol...@apache.org>.
On Mon, 2013-10-28 at 16:28 +0000, sebb wrote:
> On 28 October 2013 15:35, Oleg Kalnichevski <ol...@apache.org> wrote:
> > On Mon, 2013-10-28 at 15:18 +0000, sebb wrote:
> >> On 28 October 2013 15:07, Oleg Kalnichevski <ol...@apache.org> wrote:
> >> > On Mon, 2013-10-28 at 14:53 +0000, sebb wrote:
> >> >> On 28 October 2013 10:47, Oleg Kalnichevski <ol...@apache.org> wrote:
> >> >> > Sooo, our JIRA link is out date. Big deal, trivial to fix [1], right?
> >> >>
> >> >> Yes
> >> >>
> >> >> > But now we need to release a new parent POM and even with the lazy
> >> >> > consensus vote it still means a 3 day delay.
> >> >>
> >> >> Disagree.
> >> >>
> >> >> > Maven and ASF release policies is a marriage made in hell.
> >> >>
> >> >> Yes, the parent pom needs lazy consensus to be updated to the next release.
> >> >>
> >> >> However, there's no requirement for site builds to be subject to the
> >> >> release process, so it's easy enough to fix bugs on the website.
> >> >>
> >> >> Obviously it makes sense to update the pom and get it approved, but I
> >> >> see no reason why you should not fix the website in the meantime, if
> >> >> it is urgent.
> >> >>
> >> >
> >> > Exactly how? By hacking html files
> >>
> >> That's one way.
> >> Good for localised urgent changes.
> >> Not so good in that a site rebuild will probably lose the changes.
> >>
> >> > or by using a POM snapshot as
> >> > suggested by Gary and risking picking up some unwanted changes?
> >>
> >> That's another.
> >>
> >> I don't see how using a temporarily updated pom will be any worse than
> >> waiting for the same pom to be updated and formally released.
> >> If there is an issue with picking up unwanted changes, it will apply
> >> to a rebuild with the new parent pom too.
> >>
> >> > All I want to be able to produce exactly the same web content with one
> >> > frigging URL different.
> >>
> >> The way we do it on the JMeter project is to copy the release tag to a
> >> docs branch, and make changes there (and in trunk if relevant).
> >> The site is then updated from the branch.
> >>
> >> I'm not sure how the HC site is built, but there must be a way to
> >> rebuild it with minimal changes.
> >>
> >
> > So, in other words, madness.
>
> I don't see it that way.
>
> > Sebastian, please do not get me wrong. I am not trying to pick at you
> > personally. If you do not see all this as unnecessary and completely
> > arbitrary overhead it is not for me to judge you. Maven's reliance on
> > publishing binary artifacts to the central repository for each and every
> > little operation simply does not work that well with rigidness of ASF
> > release policies. At least for me.
>
> As I already wrote, the ASF release policy is not part of this
> particular problem.
>
No, no, everything is in order, Comrade Secretary General
Oleg
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org
Re: ASF release madness (rant)
Posted by sebb <se...@gmail.com>.
On 28 October 2013 15:35, Oleg Kalnichevski <ol...@apache.org> wrote:
> On Mon, 2013-10-28 at 15:18 +0000, sebb wrote:
>> On 28 October 2013 15:07, Oleg Kalnichevski <ol...@apache.org> wrote:
>> > On Mon, 2013-10-28 at 14:53 +0000, sebb wrote:
>> >> On 28 October 2013 10:47, Oleg Kalnichevski <ol...@apache.org> wrote:
>> >> > Sooo, our JIRA link is out date. Big deal, trivial to fix [1], right?
>> >>
>> >> Yes
>> >>
>> >> > But now we need to release a new parent POM and even with the lazy
>> >> > consensus vote it still means a 3 day delay.
>> >>
>> >> Disagree.
>> >>
>> >> > Maven and ASF release policies is a marriage made in hell.
>> >>
>> >> Yes, the parent pom needs lazy consensus to be updated to the next release.
>> >>
>> >> However, there's no requirement for site builds to be subject to the
>> >> release process, so it's easy enough to fix bugs on the website.
>> >>
>> >> Obviously it makes sense to update the pom and get it approved, but I
>> >> see no reason why you should not fix the website in the meantime, if
>> >> it is urgent.
>> >>
>> >
>> > Exactly how? By hacking html files
>>
>> That's one way.
>> Good for localised urgent changes.
>> Not so good in that a site rebuild will probably lose the changes.
>>
>> > or by using a POM snapshot as
>> > suggested by Gary and risking picking up some unwanted changes?
>>
>> That's another.
>>
>> I don't see how using a temporarily updated pom will be any worse than
>> waiting for the same pom to be updated and formally released.
>> If there is an issue with picking up unwanted changes, it will apply
>> to a rebuild with the new parent pom too.
>>
>> > All I want to be able to produce exactly the same web content with one
>> > frigging URL different.
>>
>> The way we do it on the JMeter project is to copy the release tag to a
>> docs branch, and make changes there (and in trunk if relevant).
>> The site is then updated from the branch.
>>
>> I'm not sure how the HC site is built, but there must be a way to
>> rebuild it with minimal changes.
>>
>
> So, in other words, madness.
I don't see it that way.
> Sebastian, please do not get me wrong. I am not trying to pick at you
> personally. If you do not see all this as unnecessary and completely
> arbitrary overhead it is not for me to judge you. Maven's reliance on
> publishing binary artifacts to the central repository for each and every
> little operation simply does not work that well with rigidness of ASF
> release policies. At least for me.
As I already wrote, the ASF release policy is not part of this
particular problem.
However, I agree that Maven is a bit inflexible in that it cannot
rebuild part of a site.
Perhaps not surprising given the way it automatically links all the
reports etc together.
It's very convenient that one can specify the set of reports to
generate and have them automatically added to the navigation.
I suspect it would be very difficult to combine that with updating
individual parts of the website.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org
Re: ASF release madness (rant)
Posted by Oleg Kalnichevski <ol...@apache.org>.
On Mon, 2013-10-28 at 15:18 +0000, sebb wrote:
> On 28 October 2013 15:07, Oleg Kalnichevski <ol...@apache.org> wrote:
> > On Mon, 2013-10-28 at 14:53 +0000, sebb wrote:
> >> On 28 October 2013 10:47, Oleg Kalnichevski <ol...@apache.org> wrote:
> >> > Sooo, our JIRA link is out date. Big deal, trivial to fix [1], right?
> >>
> >> Yes
> >>
> >> > But now we need to release a new parent POM and even with the lazy
> >> > consensus vote it still means a 3 day delay.
> >>
> >> Disagree.
> >>
> >> > Maven and ASF release policies is a marriage made in hell.
> >>
> >> Yes, the parent pom needs lazy consensus to be updated to the next release.
> >>
> >> However, there's no requirement for site builds to be subject to the
> >> release process, so it's easy enough to fix bugs on the website.
> >>
> >> Obviously it makes sense to update the pom and get it approved, but I
> >> see no reason why you should not fix the website in the meantime, if
> >> it is urgent.
> >>
> >
> > Exactly how? By hacking html files
>
> That's one way.
> Good for localised urgent changes.
> Not so good in that a site rebuild will probably lose the changes.
>
> > or by using a POM snapshot as
> > suggested by Gary and risking picking up some unwanted changes?
>
> That's another.
>
> I don't see how using a temporarily updated pom will be any worse than
> waiting for the same pom to be updated and formally released.
> If there is an issue with picking up unwanted changes, it will apply
> to a rebuild with the new parent pom too.
>
> > All I want to be able to produce exactly the same web content with one
> > frigging URL different.
>
> The way we do it on the JMeter project is to copy the release tag to a
> docs branch, and make changes there (and in trunk if relevant).
> The site is then updated from the branch.
>
> I'm not sure how the HC site is built, but there must be a way to
> rebuild it with minimal changes.
>
So, in other words, madness.
Sebastian, please do not get me wrong. I am not trying to pick at you
personally. If you do not see all this as unnecessary and completely
arbitrary overhead it is not for me to judge you. Maven's reliance on
publishing binary artifacts to the central repository for each and every
little operation simply does not work that well with rigidness of ASF
release policies. At least for me.
Oleg
Oleg
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org
Re: ASF release madness (rant)
Posted by sebb <se...@gmail.com>.
On 28 October 2013 15:07, Oleg Kalnichevski <ol...@apache.org> wrote:
> On Mon, 2013-10-28 at 14:53 +0000, sebb wrote:
>> On 28 October 2013 10:47, Oleg Kalnichevski <ol...@apache.org> wrote:
>> > Sooo, our JIRA link is out date. Big deal, trivial to fix [1], right?
>>
>> Yes
>>
>> > But now we need to release a new parent POM and even with the lazy
>> > consensus vote it still means a 3 day delay.
>>
>> Disagree.
>>
>> > Maven and ASF release policies is a marriage made in hell.
>>
>> Yes, the parent pom needs lazy consensus to be updated to the next release.
>>
>> However, there's no requirement for site builds to be subject to the
>> release process, so it's easy enough to fix bugs on the website.
>>
>> Obviously it makes sense to update the pom and get it approved, but I
>> see no reason why you should not fix the website in the meantime, if
>> it is urgent.
>>
>
> Exactly how? By hacking html files
That's one way.
Good for localised urgent changes.
Not so good in that a site rebuild will probably lose the changes.
> or by using a POM snapshot as
> suggested by Gary and risking picking up some unwanted changes?
That's another.
I don't see how using a temporarily updated pom will be any worse than
waiting for the same pom to be updated and formally released.
If there is an issue with picking up unwanted changes, it will apply
to a rebuild with the new parent pom too.
> All I want to be able to produce exactly the same web content with one
> frigging URL different.
The way we do it on the JMeter project is to copy the release tag to a
docs branch, and make changes there (and in trunk if relevant).
The site is then updated from the branch.
I'm not sure how the HC site is built, but there must be a way to
rebuild it with minimal changes.
> Oleg
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
> For additional commands, e-mail: dev-help@hc.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org
Re: ASF release madness (rant)
Posted by Oleg Kalnichevski <ol...@apache.org>.
On Mon, 2013-10-28 at 14:53 +0000, sebb wrote:
> On 28 October 2013 10:47, Oleg Kalnichevski <ol...@apache.org> wrote:
> > Sooo, our JIRA link is out date. Big deal, trivial to fix [1], right?
>
> Yes
>
> > But now we need to release a new parent POM and even with the lazy
> > consensus vote it still means a 3 day delay.
>
> Disagree.
>
> > Maven and ASF release policies is a marriage made in hell.
>
> Yes, the parent pom needs lazy consensus to be updated to the next release.
>
> However, there's no requirement for site builds to be subject to the
> release process, so it's easy enough to fix bugs on the website.
>
> Obviously it makes sense to update the pom and get it approved, but I
> see no reason why you should not fix the website in the meantime, if
> it is urgent.
>
Exactly how? By hacking html files or by using a POM snapshot as
suggested by Gary and risking picking up some unwanted changes? All I
want to be able to produce exactly the same web content with one
frigging URL different.
Oleg
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org
Re: ASF release madness (rant)
Posted by sebb <se...@gmail.com>.
On 28 October 2013 10:47, Oleg Kalnichevski <ol...@apache.org> wrote:
> Sooo, our JIRA link is out date. Big deal, trivial to fix [1], right?
Yes
> But now we need to release a new parent POM and even with the lazy
> consensus vote it still means a 3 day delay.
Disagree.
> Maven and ASF release policies is a marriage made in hell.
Yes, the parent pom needs lazy consensus to be updated to the next release.
However, there's no requirement for site builds to be subject to the
release process, so it's easy enough to fix bugs on the website.
Obviously it makes sense to update the pom and get it approved, but I
see no reason why you should not fix the website in the meantime, if
it is urgent.
> Oleg
>
> [1]
> http://svn.apache.org/viewvc/httpcomponents/project/trunk/pom.xml?view=diff&r1=1536305&r2=1536306&pathrev=1536306
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
> For additional commands, e-mail: dev-help@hc.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@hc.apache.org
For additional commands, e-mail: dev-help@hc.apache.org