Re: git repo for airavata web clients

[Suresh Marru <sm...@apache.org>]

Since previous discussions about web clients did not yield any strong opinions one way or another. I will assume lazy consensus and move the PHP Gateway and Samples into airavata main repo. 

Suresh

On Apr 30, 2014, at 12:54 PM, Marlon Pierce <ma...@iu.edu> wrote:

> I'm +0 for keeping one git repo. We can always go to 2 or more if this
> turns out to be the wrong choice.
> 
> Marlon
> 
> On 4/30/14 12:48 PM, Suresh Marru wrote:
>> On Apr 29, 2014, at 4:19 PM, Danushka Menikkumbura <da...@gmail.com> wrote:
>> 
>>> Having different repos will only help if we are planning to have independent release cycles.Does it make sense (or it is required ) to release the g/w separately?
>> I am unable to guess if we will need separate release cycles or not. Since web interfaces can take more development iterations then changes to clients and server packages, I guess we might have the need. Other thoughts?
>> 
>> Lets draw this to a conclusion? I seeing mostly + 0 -0 kind of opinions. Any strong preferences or should we just roll the dice and go with one option for now?
>> 
>> Suresh
>> 
>> 
>>> /Danushka
>>> 
>>> On Wednesday, April 30, 2014, Suresh Marru <sm...@apache.org> wrote:
>>>> On Apr 29, 2014, at 2:53 PM, Sachith Withana <sw...@gmail.com> wrote:
>>>> 
>>>>> Suresh,
>>>>> 
>>>>> Agree with your points.
>>>>> 
>>>>> Consider the PHP gateway. It’s being developed right now. When it’s a full blown gateway, some users would use it directly as their gateway. If they want some new features to be added, they can do that and they will be able to commit it to the main repo, so that other users can use those features too.
>>>>> 
>>>>> Basically, I’m just pointing out that the PHP gateway has the potential to become a project of its own, with its own contributors, as a framework for new gateways.
>>>> Hi Sachith,
>>>> 
>>>> This is a good goal. When such success story happen, we still encourage the PHP Reference Gateway bundled within Airavata to have an example to refer to. The cloned version of success project can potentially build its own community in github or elsewhere. A challenge for airavata should be to motivate the community to contribute back generic examples which will benefit other users and help maintain the reference implementation.
>>>> 
>>>> Suresh
>>>> 
>>>>> 
>>>>> 
>>>>> On Apr 29, 2014, at 11:46 AM, Suresh Marru <sm...@apache.org> wrote:
>>>>> 
>>>>>> Just to clarify, to commit to any repo on ASF infrastructure a committtership will be a pre-requisite. But as a PMC we can granularly control any of our repos.
>>>>>> 
>>>>>> To answer the question, I personally do not see any need to control access among committters. I agree this need if we are opening up to contributors (which I do not think is legally complaint).
>>>>>> 
>>>>>> Just FYI, subversion project allows any ASF committer to have write access to their repos, they believe in social trust rather then ACL’s, I like this boldness it makes us feel welcome. Ofcouse, I doubt any one will commit without a consent on the mailing list, but thats the point.
>>>>>> 
>>>>>> Suresh
>>>>>> 
>>>>>> On Apr 29, 2014, at 12:52 PM, Marlon Pierce <ma...@iu.edu> wrote:
>>>>>> 
>>>>>>> Would we want to have the option to restrict committership to a specific
>>>>>>> repo?
>>>>>>> 
>>>>>>> Marlon
>>>>>>> 
>>>>>>> On 4/29/14 12:32 PM, Suresh Marru wrote:
>>>>>>>> For reference, please see what other projects are doing - https://git-wip-us.apache.org/repos/asf
>>>>>>>> 
>>>>>>>> Projects like cloudstack, cordova, couchdb jclouds and others pretty much add a new repo for lots of components. Other projects choose to have one repo for everything.
>>>>>>>> 
>>>>>>>> I am not yet weighing one option over other and soliciting everyone’s input.
>>>>>>>> 
>>>>>>>> Suresh
>>>>>>>> 
>>>>>>>> On Apr 29, 2014, at 12:27 PM, Suresh Marru <sm...@apache.org> wrote:
>>>>>>>> 
>>>>>>>>> Hi All,
>>>>>>>>> 
>>>>>>>>> Since the transition to git was uneventful and seems to work well, I want to resurrect the discussion of a code repos.
>>>>>>>>> 
>>>>>>>>> To demonstrate Airavata we will need reference implementations of API. Previous web implementations are all over the place. Can we discuss what is the good place to consolidate these examples and indeed release them periodically?
>>>>>>>>> 
>>>>>>>>> Two options to consider:
>>>>>>>>> 
>>>>>>>>> * Have these web implementations as a module within main trunk and release them along with Airavata.
>>>>>>>>> * Create a separate repo and have a separate release cycle.
>>>>>>>>> 
>>>>>>>>> Any opinions?
>>>>>>>>> 
>>>>>>>>> Suresh
>>>>>>>>> 
>>>> 
>> 
> 

Re: git repo for airavata web clients

[Suresh Marru <sm...@apache.org>]

Hi All,

I verified that all 6 contributors on the GitHub repo [1] are Airavata Committers, so I migrated it and managed to preserve the commit history to [2]. I will now move it to appropriate places in the client samples.

Suresh

[1] - https://github.com/SciGaP/Airavata-PHP-Client-Samples
[2] - https://github.com/apache/airavata/commits/0aec0da4ca9a9ecae280cd05f1d8cb4e65e8ad7d

On Jun 30, 2014, at 8:30 AM, Suresh Marru <sm...@apache.org> wrote:

> Since previous discussions about web clients did not yield any strong opinions one way or another. I will assume lazy consensus and move the PHP Gateway and Samples into airavata main repo. 
> 
> Suresh
> 
> On Apr 30, 2014, at 12:54 PM, Marlon Pierce <ma...@iu.edu> wrote:
> 
>> I'm +0 for keeping one git repo. We can always go to 2 or more if this
>> turns out to be the wrong choice.
>> 
>> Marlon
>> 
>> On 4/30/14 12:48 PM, Suresh Marru wrote:
>>> On Apr 29, 2014, at 4:19 PM, Danushka Menikkumbura <da...@gmail.com> wrote:
>>> 
>>>> Having different repos will only help if we are planning to have independent release cycles.Does it make sense (or it is required ) to release the g/w separately?
>>> I am unable to guess if we will need separate release cycles or not. Since web interfaces can take more development iterations then changes to clients and server packages, I guess we might have the need. Other thoughts?
>>> 
>>> Lets draw this to a conclusion? I seeing mostly + 0 -0 kind of opinions. Any strong preferences or should we just roll the dice and go with one option for now?
>>> 
>>> Suresh
>>> 
>>> 
>>>> /Danushka
>>>> 
>>>> On Wednesday, April 30, 2014, Suresh Marru <sm...@apache.org> wrote:
>>>>> On Apr 29, 2014, at 2:53 PM, Sachith Withana <sw...@gmail.com> wrote:
>>>>> 
>>>>>> Suresh,
>>>>>> 
>>>>>> Agree with your points.
>>>>>> 
>>>>>> Consider the PHP gateway. It’s being developed right now. When it’s a full blown gateway, some users would use it directly as their gateway. If they want some new features to be added, they can do that and they will be able to commit it to the main repo, so that other users can use those features too.
>>>>>> 
>>>>>> Basically, I’m just pointing out that the PHP gateway has the potential to become a project of its own, with its own contributors, as a framework for new gateways.
>>>>> Hi Sachith,
>>>>> 
>>>>> This is a good goal. When such success story happen, we still encourage the PHP Reference Gateway bundled within Airavata to have an example to refer to. The cloned version of success project can potentially build its own community in github or elsewhere. A challenge for airavata should be to motivate the community to contribute back generic examples which will benefit other users and help maintain the reference implementation.
>>>>> 
>>>>> Suresh
>>>>> 
>>>>>> 
>>>>>> 
>>>>>> On Apr 29, 2014, at 11:46 AM, Suresh Marru <sm...@apache.org> wrote:
>>>>>> 
>>>>>>> Just to clarify, to commit to any repo on ASF infrastructure a committtership will be a pre-requisite. But as a PMC we can granularly control any of our repos.
>>>>>>> 
>>>>>>> To answer the question, I personally do not see any need to control access among committters. I agree this need if we are opening up to contributors (which I do not think is legally complaint).
>>>>>>> 
>>>>>>> Just FYI, subversion project allows any ASF committer to have write access to their repos, they believe in social trust rather then ACL’s, I like this boldness it makes us feel welcome. Ofcouse, I doubt any one will commit without a consent on the mailing list, but thats the point.
>>>>>>> 
>>>>>>> Suresh
>>>>>>> 
>>>>>>> On Apr 29, 2014, at 12:52 PM, Marlon Pierce <ma...@iu.edu> wrote:
>>>>>>> 
>>>>>>>> Would we want to have the option to restrict committership to a specific
>>>>>>>> repo?
>>>>>>>> 
>>>>>>>> Marlon
>>>>>>>> 
>>>>>>>> On 4/29/14 12:32 PM, Suresh Marru wrote:
>>>>>>>>> For reference, please see what other projects are doing - https://git-wip-us.apache.org/repos/asf
>>>>>>>>> 
>>>>>>>>> Projects like cloudstack, cordova, couchdb jclouds and others pretty much add a new repo for lots of components. Other projects choose to have one repo for everything.
>>>>>>>>> 
>>>>>>>>> I am not yet weighing one option over other and soliciting everyone’s input.
>>>>>>>>> 
>>>>>>>>> Suresh
>>>>>>>>> 
>>>>>>>>> On Apr 29, 2014, at 12:27 PM, Suresh Marru <sm...@apache.org> wrote:
>>>>>>>>> 
>>>>>>>>>> Hi All,
>>>>>>>>>> 
>>>>>>>>>> Since the transition to git was uneventful and seems to work well, I want to resurrect the discussion of a code repos.
>>>>>>>>>> 
>>>>>>>>>> To demonstrate Airavata we will need reference implementations of API. Previous web implementations are all over the place. Can we discuss what is the good place to consolidate these examples and indeed release them periodically?
>>>>>>>>>> 
>>>>>>>>>> Two options to consider:
>>>>>>>>>> 
>>>>>>>>>> * Have these web implementations as a module within main trunk and release them along with Airavata.
>>>>>>>>>> * Create a separate repo and have a separate release cycle.
>>>>>>>>>> 
>>>>>>>>>> Any opinions?
>>>>>>>>>> 
>>>>>>>>>> Suresh
>>>>>>>>>> 
>>>>> 
>>> 
>> 
>