[PATCH][rtems-testing] add gcj test support for RTEMS

Chris Johns chrisj at rtems.org
Fri Aug 19 07:56:47 UTC 2011


On 19/08/11 5:38 PM, Ralf Corsepius wrote:
> On 08/19/2011 08:56 AM, Chris Johns wrote:
>> On 19/08/11 3:34 PM, Ralf Corsepius wrote:
>>>>
>>>> Not everyone who comes to RTEMS has your point of view or experience
>>>> or both. Please consider the language used to address users on this
>>>> list and especially a GSoC student.
>>> I did - The language was not directly addressing the GSOC student, but
>>> those RTEMS folks who are advertising
>>
>> It is just not needed. This is the first stage in seeing if gjc can be
>> made to run on RTEMS. It seems to have done this task without all the
>> time being consumed with production quality build issues.
> No problem with this.
>>
>>> these naive scripts and are
>>
>> The project was about gjc and not about the build script.
> The project starts to pollute RTEMS.
>>
>>> directing GSOC students to use these scripts.
>>
>> Short of anything else to fill void I feel it was justifiable. Are you
>> suggesting we direct a student to learn rpm scripting, mock and all
>> that takes ?
> No.
>
> a) People who are hacking on GCC should be sufficiently familiar with
> GCC to know how to build and install it.

This seems to be the chicken or the egg problem.

> It should be trivial exercise
> for them to write their own. There isn't any need to force them to hack
> on questionable, semi-official, semi-functional stuff.

Maybe it is once they know what they are doing.

>
> That said, IMSHO, these scripts should not have any room in RTEMS and be
> dumped ASAP. They are harmful and dangerous to their users's
> installations and to the project, because they pretend doing something,
> they don't do.
>

Our views differ.

> b) I am suggesting GSOC works of "non-production quality" to be kept
> outside of RTEMS.

Agreed.

> Admitted, rtems-testing is not part of RTEMS, but is
> only some people's personal playground, who believe it is doing
> something good.

Funny this module in cvs brings this reaction from you. I am fine with 
it. I have used these scripts for testing in the past and will again. 
There is a whole range of issues with them that need to be fixed and 
they will get fixed. Until then they are available and doing no harm.

> To me. personally rtems-testing is just a "takes
> everything dumping ground", people are using to avoid having to cope
> with integration.

Or it is a take it as you find it solution for a range of problems out 
in the open for others to borrow use or ignore.

Chris



More information about the users mailing list