jmeter-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Philippe Mouawad <philippe.moua...@gmail.com>
Subject Re: Ease setup for new comers wishing to contribute
Date Sun, 07 Feb 2016 22:05:49 GMT
Hello,
Documentation part done through commit r1729040.
Regards

On Sun, Dec 6, 2015 at 3:38 PM, Philippe Mouawad <philippe.mouawad@gmail.com
> wrote:

>
>
> On Sat, Dec 5, 2015 at 12:41 PM, sebb <sebbaz@gmail.com> wrote:
>
>> On 5 December 2015 at 09:25, Philippe Mouawad
>> <philippe.mouawad@gmail.com> wrote:
>> > Hello,
>> > I open this thread to discuss what could be done to favor new comers
>> > contributions and increase patches contributions.
>> >
>> > After discussion with colleagues that came recently on project with a
>> > "Standard project" background, the following ideas came:
>> >
>> >    - Ease IDE Setup (see below for details)
>> >    - Change http://jmeter.apache.org/svnindex.html to mention GITHUB
>> mirror
>> >    , the link would be renamed to Source Repositories
>>
>> +1
>>
>> >    - Add fork us on Github on project if acceptable by Apache policy
>>
>> +0
>>
>> >    - Create a Developers page similar to
>> >    http://cloudstack.apache.org/developers.html, explaining how to
>> >    contribute a PR
>>
>> +1
>>
>> >    - Find an alternative to Ant
>>
>> -1
>>
>> >
>> > Ease IDE Setup:
>> > Regarding IDE Setup and particularly Eclipse:
>> > 1/ As we don't have Maven (mvn eclipse:eclipse) setting up Dev
>> Environment
>> > for new comers is not a nowadays "standard".
>> > Maybe Graddle could be an alternative (Maven having been set aside in a
>> > previous dev discussion), it allows providing Eclipse Project Model, is
>> > used by big projects like Spring Frameworks and seems to be very
>> flexible ,
>> > using Groovy
>>
>> Yes, it is flexible, but it is hard to get it right for such a large
>> code base as JMeter.
>>
>> My experience of it is that it can be very hard to find out where
>> things are happening.
>> Precisely because of the flexibility.
>>
>> > 2/ If we decide to stay on Ant, we should have an HTML page instead of
>> > eclipse.readme which can easily be missed.
>>
>> +1
>>
>> > The page should be enhanced to mention requirements like Java 7 and
>> > Compiler compliance level....
>>
>> Of course.
>>
>> > 3/ Running JUnit tests should also be documented, particularly running
>> them
>> > directly from IDE
>>
>> +1
>>
>> Some of the tests don't work very well from the IDE and need
>> additional settings.
>> There is an abstract test class that tries to fix these issues but
>> it's not fully integrated.
>>
>> > 4/ If we stay with Ant, why not also distribute eclipse.project so that
>> we
>> > could create an ant task that:
>>
>> The problem is that the Eclipse .project file varies between
>> installations.
>> If the developer adds Findbugs or Checkstyle etc that will affect the
>> .project file.
>> Likewise if they have several JDKs and JMeter is not using the default.
>>
>> However it would be possible to create a default one if it does not exist.
>> It may need tweaking.
>>
> Ok did it this way
>
>>
>> > a) renames eclipse.classpath to .classpath
>>
>> -1 to rename; copy would be OK if it does not already exist
>>
> Ok did it this way
>
>>
>> > b) renames eclipse.project to .project
>>
>> -1 to rename; copy would be OK if it does not already exist
>>
> Ok did it this way
>
>>
>> > c) Call  download_jars ,
>>
>> -1, this must be a separate action.
>>
> Maven downloads without asking, why is it a problem to do so ?
>
>> It's important that users 'agree' to the download.
>>
> For now I added an agreement and fail if it does not come
>
>>
>> However the Eclipse copy task could print a message to remind users to
>> download.
>> And vice versa.
>>
> Added this if user refuses download agreement
>
>>
>> > This way project could be imported in Eclipse much more easily and less
>> > manually
>>
>> >
>> > Regards
>> > Philippe M.
>>
>
>
>
> --
> Cordialement.
> Philippe Mouawad.
>
>
>


-- 
Cordialement.
Philippe Mouawad.

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message