Re: new sites and swarm of commits.

2018-04-28 Thread Jan Iversen


Sent from my iPad

> On 28 Apr 2018, at 00:15, sebb <seb...@gmail.com> wrote:
> 
>> On 27 April 2018 at 18:01, Jan Iversen <j...@apache.org> wrote:
>> 
>> 
>> Sent from my iPad
>> 
>>>> On 27 Apr 2018, at 16:30, Henk P. Penning <penn...@uu.nl> wrote:
>>>> 
>>>> On Fri, 27 Apr 2018, Jan Iversen wrote:
>>>> 
>>>> Date: Fri, 27 Apr 2018 15:04:50 +0200
>>>> From: Jan Iversen <j...@apache.org>
>>>> To: general@attic.apache.org
>>>> Subject: Re: new sites and swarm of commits.
>>> 
>>> Hi Jan,
>>> 
>>>>>  ---
>>>>>  name: Oltu
>>>>>  project: oltu
>>>>>  retired: April 2018
>>>>>  mailnames:
>>>>>- dev
>>>>>- user
>>>>>- commits
>>>>>  tracker:
>>>>>- JIRA
>>>>>- OLTU
>>>>>  scm: svn
>>>>>  wiki: https://cwiki.apache.org/confluence/display/oltu/
>>>>>  description: |-
>>>>>Apache Oltu was an OAuth protocol implementation in Java. It also
>>>>>covers others OAuth family related implementations such as JWT, JWS
>>>>>and OpenID Connect
>>>>>  ---
>>>>> 
>>>>> What is complicated about that ?
>>>> for one it is, yaml, and I still have not seen any online validator of
>>>> that. json offers online validation so som faulty commits are avoided
>>>> (not all I know).
>>> 
>>> Why check ? If there's a yaml error the generator shows exactly
>>> where the error occurs ; I think we can catch that.
>>> If/when an error occurs, nothing should be committed, of course.
>> 
>> because the generator runs after the yaml file is committed, not before.
> 
> Of course, otherwise the buildbot has nothing to work on.
> 
> But the generated code is only committed after the generator runs.
> 
> Once a buildbot job has detected a failure it skips any remaining
> steps, so it can be set to ignore the commit.
> 
>>> 
>>>> secondly I believe having fields like JIRA and svn are wrong un very
>>>> unflexible, having the url is a lot more generic. how will you retire
>>>> e.g. AOO that runs its own bugzilla instance.
>>> 
>>> I agree ; just an url is fine ; but 'as is' is easy too.
>>> 
>>>> and I have to learn yaml.
>>> 
>>> Copy/paste/edit of the example above should suffice almost always.
>>> 
>>>>> Just ignore the commits. We'll judge the stuff when dev is done.
>>>>> I think we're mostly done.
>>> 
>>>> both of you keep saying thatI still believe that having a
>>>> volunteer who will actually maintain it is now more important than
>>>> seeing the final bits.
>>> 
>>> Ok ; I'll volunteer to maintain the site.
>>> I would love to see the stuff "at work".
>> Super one less problem. The rest of the retiring process is nearly automated 
>> so that will not be a big extra job.
> 
> I am happy to maintain the Jekyll site if that is chosen, or even the Perl 
> site.
Super, so now we have 2 volunteers, seems I am off the hook. I will start a 
discussion/vote later. 

rgds
jan i
> 
>> rgds
>> jan I
>>> 
>>>> jan i
>>> 
>>> Regards,
>>> 
>>> HPP
>>> 
>>>    _
>>> Henk P. Penning, ICT-beta R Uithof MG-403_/ \_
>>> Faculty of Science, Utrecht UniversityT +31 30 253 4106 / \_/ \
>>> Leuvenlaan 4, 3584CE Utrecht, NL  F +31 30 253 4553 \_/ \_/
>>> http://www.staff.science.uu.nl/~penni101/ M penn...@uu.nl \_/


Re: new sites and swarm of commits.

2018-04-27 Thread sebb
On 27 April 2018 at 18:01, Jan Iversen <j...@apache.org> wrote:
>
>
> Sent from my iPad
>
>> On 27 Apr 2018, at 16:30, Henk P. Penning <penn...@uu.nl> wrote:
>>
>>> On Fri, 27 Apr 2018, Jan Iversen wrote:
>>>
>>> Date: Fri, 27 Apr 2018 15:04:50 +0200
>>> From: Jan Iversen <j...@apache.org>
>>> To: general@attic.apache.org
>>> Subject: Re: new sites and swarm of commits.
>>
>> Hi Jan,
>>
>>>>   ---
>>>>   name: Oltu
>>>>   project: oltu
>>>>   retired: April 2018
>>>>   mailnames:
>>>> - dev
>>>> - user
>>>> - commits
>>>>   tracker:
>>>> - JIRA
>>>> - OLTU
>>>>   scm: svn
>>>>   wiki: https://cwiki.apache.org/confluence/display/oltu/
>>>>   description: |-
>>>> Apache Oltu was an OAuth protocol implementation in Java. It also
>>>> covers others OAuth family related implementations such as JWT, JWS
>>>> and OpenID Connect
>>>>   ---
>>>>
>>>> What is complicated about that ?
>>> for one it is, yaml, and I still have not seen any online validator of
>>> that. json offers online validation so som faulty commits are avoided
>>> (not all I know).
>>
>>  Why check ? If there's a yaml error the generator shows exactly
>>  where the error occurs ; I think we can catch that.
>>  If/when an error occurs, nothing should be committed, of course.
>
> because the generator runs after the yaml file is committed, not before.

Of course, otherwise the buildbot has nothing to work on.

But the generated code is only committed after the generator runs.

Once a buildbot job has detected a failure it skips any remaining
steps, so it can be set to ignore the commit.

>>
>>> secondly I believe having fields like JIRA and svn are wrong un very
>>> unflexible, having the url is a lot more generic. how will you retire
>>> e.g. AOO that runs its own bugzilla instance.
>>
>>  I agree ; just an url is fine ; but 'as is' is easy too.
>>
>>> and I have to learn yaml.
>>
>>  Copy/paste/edit of the example above should suffice almost always.
>>
>>>> Just ignore the commits. We'll judge the stuff when dev is done.
>>>> I think we're mostly done.
>>
>>> both of you keep saying thatI still believe that having a
>>> volunteer who will actually maintain it is now more important than
>>> seeing the final bits.
>>
>>  Ok ; I'll volunteer to maintain the site.
>>  I would love to see the stuff "at work".
> Super one less problem. The rest of the retiring process is nearly automated 
> so that will not be a big extra job.

I am happy to maintain the Jekyll site if that is chosen, or even the Perl site.

> rgds
> jan I
>>
>>> jan i
>>
>>  Regards,
>>
>>  HPP
>>
>>    _
>> Henk P. Penning, ICT-beta R Uithof MG-403_/ \_
>> Faculty of Science, Utrecht UniversityT +31 30 253 4106 / \_/ \
>> Leuvenlaan 4, 3584CE Utrecht, NL  F +31 30 253 4553 \_/ \_/
>> http://www.staff.science.uu.nl/~penni101/ M penn...@uu.nl \_/


Re: new sites and swarm of commits.

2018-04-27 Thread Jan Iversen


Sent from my iPad

> On 27 Apr 2018, at 16:30, Henk P. Penning <penn...@uu.nl> wrote:
> 
>> On Fri, 27 Apr 2018, Jan Iversen wrote:
>> 
>> Date: Fri, 27 Apr 2018 15:04:50 +0200
>> From: Jan Iversen <j...@apache.org>
>> To: general@attic.apache.org
>> Subject: Re: new sites and swarm of commits.
> 
> Hi Jan,
> 
>>>   ---
>>>   name: Oltu
>>>   project: oltu
>>>   retired: April 2018
>>>   mailnames:
>>> - dev
>>> - user
>>> - commits
>>>   tracker:
>>> - JIRA
>>> - OLTU
>>>   scm: svn
>>>   wiki: https://cwiki.apache.org/confluence/display/oltu/
>>>   description: |-
>>> Apache Oltu was an OAuth protocol implementation in Java. It also
>>> covers others OAuth family related implementations such as JWT, JWS
>>> and OpenID Connect
>>>   ---
>>> 
>>> What is complicated about that ?
>> for one it is, yaml, and I still have not seen any online validator of
>> that. json offers online validation so som faulty commits are avoided
>> (not all I know).
> 
>  Why check ? If there's a yaml error the generator shows exactly
>  where the error occurs ; I think we can catch that.
>  If/when an error occurs, nothing should be committed, of course.

because the generator runs after the yaml file is committed, not before.
> 
>> secondly I believe having fields like JIRA and svn are wrong un very
>> unflexible, having the url is a lot more generic. how will you retire
>> e.g. AOO that runs its own bugzilla instance.
> 
>  I agree ; just an url is fine ; but 'as is' is easy too.
> 
>> and I have to learn yaml.
> 
>  Copy/paste/edit of the example above should suffice almost always.
> 
>>> Just ignore the commits. We'll judge the stuff when dev is done.
>>> I think we're mostly done.
> 
>> both of you keep saying thatI still believe that having a
>> volunteer who will actually maintain it is now more important than
>> seeing the final bits.
> 
>  Ok ; I'll volunteer to maintain the site.
>  I would love to see the stuff "at work".
Super one less problem. The rest of the retiring process is nearly automated so 
that will not be a big extra job.

rgds
jan I
> 
>> jan i
> 
>  Regards,
> 
>  HPP
> 
>    _
> Henk P. Penning, ICT-beta R Uithof MG-403_/ \_
> Faculty of Science, Utrecht UniversityT +31 30 253 4106 / \_/ \
> Leuvenlaan 4, 3584CE Utrecht, NL  F +31 30 253 4553 \_/ \_/
> http://www.staff.science.uu.nl/~penni101/ M penn...@uu.nl \_/


Re: new sites and swarm of commits.

2018-04-27 Thread Henk P. Penning

On Fri, 27 Apr 2018, Jan Iversen wrote:


Date: Fri, 27 Apr 2018 15:04:50 +0200
From: Jan Iversen <j...@apache.org>
To: general@attic.apache.org
Subject: Re: new sites and swarm of commits.


Hi Jan,


   ---
   name: Oltu
   project: oltu
   retired: April 2018
   mailnames:
 - dev
 - user
 - commits
   tracker:
 - JIRA
 - OLTU
   scm: svn
   wiki: https://cwiki.apache.org/confluence/display/oltu/
   description: |-
 Apache Oltu was an OAuth protocol implementation in Java. It also
 covers others OAuth family related implementations such as JWT, JWS
 and OpenID Connect
   ---

 What is complicated about that ?

for one it is, yaml, and I still have not seen any online validator of
that. json offers online validation so som faulty commits are avoided
(not all I know).


  Why check ? If there's a yaml error the generator shows exactly
  where the error occurs ; I think we can catch that.
  If/when an error occurs, nothing should be committed, of course.


secondly I believe having fields like JIRA and svn are wrong un very
unflexible, having the url is a lot more generic. how will you retire
e.g. AOO that runs its own bugzilla instance.


  I agree ; just an url is fine ; but 'as is' is easy too.


and I have to learn yaml.


  Copy/paste/edit of the example above should suffice almost always.


 Just ignore the commits. We'll judge the stuff when dev is done.
 I think we're mostly done.



both of you keep saying thatI still believe that having a
volunteer who will actually maintain it is now more important than
seeing the final bits.


  Ok ; I'll volunteer to maintain the site.
  I would love to see the stuff "at work".


jan i


  Regards,

  HPP

   _
Henk P. Penning, ICT-beta R Uithof MG-403_/ \_
Faculty of Science, Utrecht UniversityT +31 30 253 4106 / \_/ \
Leuvenlaan 4, 3584CE Utrecht, NL  F +31 30 253 4553 \_/ \_/
http://www.staff.science.uu.nl/~penni101/ M penn...@uu.nl \_/


Re: new sites and swarm of commits.

2018-04-27 Thread sebb
On 27 April 2018 at 14:04, Jan Iversen <j...@apache.org> wrote:
>
>
> Sent from my iPad
>
>> On 27 Apr 2018, at 14:46, Henk P. Penning <penn...@uu.nl> wrote:
>>
>>> On Fri, 27 Apr 2018, Jan Iversen wrote:
>>>
>>> Date: Fri, 27 Apr 2018 13:39:16 +0200
>>> From: Jan Iversen <jancasacon...@gmail.com>
>>> To: general@attic.apache.org
>>> Subject: new sites and swarm of commits.
>>
>> Hi Jan,
>>
>>> It is amazing how many commits Henk and Sebb can produce, I have lost
>>> track long ago about the status of things
>>
>>  Please, just ignore the commits.
> sort of hard to do, since they at least sometimes changes important things in 
> the existing site.

Just ignore the commits that affect the branches.

>>
>>>   with one exception I see
>>> both proposals having growing complexity (now e.g. both use yaml) most
>>> of the complexity seems to arise in order to be to be able to handle
>>> situations in the future (by using the past as s valid example)
>>> something I strongly believe is a false assumption.
>>
>>  Eh, no. To generate the old (even pre-attic) stuff,
>>  some optional attributes had to be added.
>>  The last project to retire (oltu) is typical
>>  for future projects ; this is the config entry :
>>
>>---
>>name: Oltu
>>project: oltu
>>retired: April 2018
>>mailnames:
>>  - dev
>>  - user
>>  - commits
>>tracker:
>>  - JIRA
>>  - OLTU
>>scm: svn
>>wiki: https://cwiki.apache.org/confluence/display/oltu/
>>description: |-
>>  Apache Oltu was an OAuth protocol implementation in Java. It also
>>  covers others OAuth family related implementations such as JWT, JWS
>>  and OpenID Connect
>>---
>>
>>  What is complicated about that ?
> for one it is, yaml, and I still have not seen any online validator of that. 
> json offers online validation so som faulty commits are avoided (not all I 
> know).
>
> secondly I believe having fields like JIRA and svn are wrong un very 
> unflexible, having the url is a lot more generic. how will you retire e.g. 
> AOO that runs its own bugzilla instance.

The reason for using a keyword is to simplify the file and reduce typing. (DRY)
Long URLs are esaier to get wrong.

Note that the site-jekyll already supports Bugzilla (AOO) and
SpamAssassin, and can easily be extended.

It would be easy to extend it to use a full URL if necessary.
e.g. if the issue key looks like a URL, use it as is.

>>  To update the site, all you have to do is 'edit' and 'commit' ;
>>  that is the goal.
>>
> and I have to learn yaml.

I don't know all of YAML but I can make changes to an existing file.

>>  My stuff is now 300 lines of simple Perl ; as it was before.
>>  Sebb uses more tools, but about the same ones as before.
>>  Processing appears to be split between the (ruby) generator,
>>  and the jekyll template ; preference is a matter of taste.
>>
>>  Sebb is setting up the build stuff ; great ; we'll (something
>>  like) that anyway after the git switch, if I'm not mistaken.

Yes.
The job won't need much editing to switch from SVN to Git.
I can do that.

The pubsub settings also need to be changed to pick up the generated
data from Git.
I think that will require some Infra involvement

>>
>>  Just ignore the commits. We'll judge the stuff when dev is done.
>>  I think we're mostly done.
> both of you keep saying thatI still believe that having a volunteer who 
> will actually maintain it is now more important than seeing the final bits.

There are two aspects to maintenance:
- adding new projects
- updating the site layout

Which are you concerned about?

> rgds
> jan i
>>
>> Sebb,
>>
>>  ... are you almost done :-) ?

I have just now created the public JSON data file.

https://ci.apache.org/projects/attic/projects.json

I think that is the last change to the processing for now.

>>  ... using yaml ?

No, I decided to stick with JSON for now.
I worked round the multi-line issue which was the main problem for me.
The comments are in a separate file.

It's not ideal, but the benefits are not worth the hassle.

>>
>>>   The next board report will
>>> contain a description of the ongoing activity, but hopefully with a
>>> positive conclusion.
>>
>>  I'm sure of it.
>>
>>> jan i
>>
>>  Regards,
>>
>>  HPP
>>
>>    _
>> Henk P. Penning, ICT-beta R Uithof MG-403_/ \_
>> Faculty of Science, Utrecht UniversityT +31 30 253 4106 / \_/ \
>> Leuvenlaan 4, 3584CE Utrecht, NL  F +31 30 253 4553 \_/ \_/
>> http://www.staff.science.uu.nl/~penni101/ M penn...@uu.nl \_/


Re: new sites and swarm of commits.

2018-04-27 Thread Jan Iversen


Sent from my iPad

> On 27 Apr 2018, at 14:46, Henk P. Penning <penn...@uu.nl> wrote:
> 
>> On Fri, 27 Apr 2018, Jan Iversen wrote:
>> 
>> Date: Fri, 27 Apr 2018 13:39:16 +0200
>> From: Jan Iversen <jancasacon...@gmail.com>
>> To: general@attic.apache.org
>> Subject: new sites and swarm of commits.
> 
> Hi Jan,
> 
>> It is amazing how many commits Henk and Sebb can produce, I have lost
>> track long ago about the status of things
> 
>  Please, just ignore the commits.
sort of hard to do, since they at least sometimes changes important things in 
the existing site.
> 
>>   with one exception I see
>> both proposals having growing complexity (now e.g. both use yaml) most
>> of the complexity seems to arise in order to be to be able to handle
>> situations in the future (by using the past as s valid example)
>> something I strongly believe is a false assumption.
> 
>  Eh, no. To generate the old (even pre-attic) stuff,
>  some optional attributes had to be added.
>  The last project to retire (oltu) is typical
>  for future projects ; this is the config entry :
> 
>---
>name: Oltu
>project: oltu
>retired: April 2018
>mailnames:
>  - dev
>  - user
>  - commits
>tracker:
>  - JIRA
>  - OLTU
>scm: svn
>wiki: https://cwiki.apache.org/confluence/display/oltu/
>description: |-
>  Apache Oltu was an OAuth protocol implementation in Java. It also
>  covers others OAuth family related implementations such as JWT, JWS
>  and OpenID Connect
>---
> 
>  What is complicated about that ?
for one it is, yaml, and I still have not seen any online validator of that. 
json offers online validation so som faulty commits are avoided (not all I 
know).

secondly I believe having fields like JIRA and svn are wrong un very 
unflexible, having the url is a lot more generic. how will you retire e.g. AOO 
that runs its own bugzilla instance.

>  To update the site, all you have to do is 'edit' and 'commit' ;
>  that is the goal.
> 
and I have to learn yaml.

>  My stuff is now 300 lines of simple Perl ; as it was before.
>  Sebb uses more tools, but about the same ones as before.
>  Processing appears to be split between the (ruby) generator,
>  and the jekyll template ; preference is a matter of taste.
> 
>  Sebb is setting up the build stuff ; great ; we'll (something
>  like) that anyway after the git switch, if I'm not mistaken.
> 
>  Just ignore the commits. We'll judge the stuff when dev is done.
>  I think we're mostly done.
both of you keep saying thatI still believe that having a volunteer who 
will actually maintain it is now more important than seeing the final bits.

rgds
jan i
> 
> Sebb,
> 
>  ... are you almost done :-) ?
>  ... using yaml ?
> 
>>   The next board report will
>> contain a description of the ongoing activity, but hopefully with a
>> positive conclusion.
> 
>  I'm sure of it.
> 
>> jan i
> 
>  Regards,
> 
>  HPP
> 
>    _
> Henk P. Penning, ICT-beta R Uithof MG-403_/ \_
> Faculty of Science, Utrecht UniversityT +31 30 253 4106 / \_/ \
> Leuvenlaan 4, 3584CE Utrecht, NL  F +31 30 253 4553 \_/ \_/
> http://www.staff.science.uu.nl/~penni101/ M penn...@uu.nl \_/


Re: new sites and swarm of commits.

2018-04-27 Thread Henk P. Penning

On Fri, 27 Apr 2018, Jan Iversen wrote:


Date: Fri, 27 Apr 2018 13:39:16 +0200
From: Jan Iversen <jancasacon...@gmail.com>
To: general@attic.apache.org
Subject: new sites and swarm of commits.


Hi Jan,


It is amazing how many commits Henk and Sebb can produce, I have lost
track long ago about the status of things


  Please, just ignore the commits.


   with one exception I see
both proposals having growing complexity (now e.g. both use yaml) most
of the complexity seems to arise in order to be to be able to handle
situations in the future (by using the past as s valid example)
something I strongly believe is a false assumption.


  Eh, no. To generate the old (even pre-attic) stuff,
  some optional attributes had to be added.
  The last project to retire (oltu) is typical
  for future projects ; this is the config entry :

---
name: Oltu
project: oltu
retired: April 2018
mailnames:
  - dev
  - user
  - commits
tracker:
  - JIRA
  - OLTU
scm: svn
wiki: https://cwiki.apache.org/confluence/display/oltu/
description: |-
  Apache Oltu was an OAuth protocol implementation in Java. It also
  covers others OAuth family related implementations such as JWT, JWS
  and OpenID Connect
---

  What is complicated about that ?
  To update the site, all you have to do is 'edit' and 'commit' ;
  that is the goal.

  My stuff is now 300 lines of simple Perl ; as it was before.
  Sebb uses more tools, but about the same ones as before.
  Processing appears to be split between the (ruby) generator,
  and the jekyll template ; preference is a matter of taste.

  Sebb is setting up the build stuff ; great ; we'll (something
  like) that anyway after the git switch, if I'm not mistaken.

  Just ignore the commits. We'll judge the stuff when dev is done.
  I think we're mostly done.

Sebb,

  ... are you almost done :-) ?
  ... using yaml ?


   The next board report will
contain a description of the ongoing activity, but hopefully with a
positive conclusion.


  I'm sure of it.


jan i


  Regards,

  HPP

   _
Henk P. Penning, ICT-beta R Uithof MG-403_/ \_
Faculty of Science, Utrecht UniversityT +31 30 253 4106 / \_/ \
Leuvenlaan 4, 3584CE Utrecht, NL  F +31 30 253 4553 \_/ \_/
http://www.staff.science.uu.nl/~penni101/ M penn...@uu.nl \_/


new sites and swarm of commits.

2018-04-27 Thread Jan Iversen
Hi

It is amazing how many commits Henk and Sebb can produce, I have lost track 
long ago about the status of things, with one exception I see both proposals 
having growing complexity (now e.g. both use yaml) most of the complexity seems 
to arise in order to be to be able to handle  situations in the future (by 
using the past as s valid example) something I strongly believe is a false 
assumption.

Furthermore I think in the idea of making a superior technical solution, at 
least one key human factor has been neglected.

Let me put it bluntly, are either of you prepared to take over attic 
maintenance should we go for your solution. Any solution that does not solve 
the human maintenance question is not really a solution.

I, being the current maintainer, have long stated my wishes for a new site, and 
currently the 2 proposals seems to move further away from my wishes of 
simplicity. Please do not expect me to volunteer for maintaining technical 
elegant solutions, that I do not feel comfortable with. I will of course 
respect if others volunteer to help out.

I see no progress in choosing the right solution, something that worries me 
deeply, we as a community should be able to solve this between us, and not make 
it a board issue. The next board report will contain a description of the 
ongoing activity, but hopefully with a positive conclusion.

have a nice weekend.
rgds
jan i

Sent from my iPad