[Cdash] purpose of JOB_COMPILER_NAME?

[David Cole]

On Fri, Mar 11, 2011 at 1:54 AM, Julien Jomier
<julien.jomier at kitware.com> wrote:
> Hi Johan,
>
> You are correct, currently the JOB_COMPILER_NAME is used to match the target
> platform but the default one is used when the job is ran on the machine. We
> would need to add support for this.
>

It's somewhat "lack of feature completeness" as "a bug", I think.


> Dave (in CC) might know if there is an easy way to achieve this.

I do not know an easy way to achieve this other than: only tell CDash
about one compiler per site. And then schedule on that site
explicitly.

If your hardware can handle it, you can make a single machine run 2
(or 3 or 4) scripts simultaneously, and use a different site name for
each script, and effectively allow multiple compilers on a single
machine. CDash will (may) schedule on all simultaneously (or at least
overlapping), though, so the machine would have to expect that sort of
load if you follow this tactic.


HTH,
David


>
> Julien
>
> On 10/03/2011 20:27, Johan Björk wrote:
>>
>> Hi everyone,
>>
>> I'm afraid I may have misunderstood the purpose of the parameters when
>> scheduling a build on CDash. Are the parameters only of help to choose a
>> target machine? In the CDash code, once a target has been chosen, it
>> always uses the compiler with the highest version number, it completely
>> ignores that I chose a different one in the UI (or using the rest API)
>>
>> Is it a bug, or am I trying to use this feature in an unintended way?
>>
>> Thanks
>> -Johan
>>
>>
>>
>> _______________________________________________
>> Cdash mailing list
>> Cdash at public.kitware.com
>> http://public.kitware.com/cgi-bin/mailman/listinfo/cdash
>