kicad-developers team mailing list archive

[Wayne Stambaugh <stambaughw@xxxxxxxxx>]

On 1/18/2017 11:56 AM, Jean-Samuel Reynaud wrote:
> 
> For this particular point, I do exactly the same as
> KICAD_SCRIPTING_MODULES option.
> 
> But for both option (KICAD_SCRIPTING_MODULES and
> KICAD_SCRIPTING_ACTION_MENU) you mean that it's better to fail the
> configuration if the options are not compatibles ?

Yes!  KiCad should not build an invalid configuration.  The config step
should fail.

> 
> 
> 
> 
> Le 17/01/2017 à 20:53, Simon Wells a écrit :
>> slightly off-topic (again?) but it just came to mind.....
>>
>> currently if option X requires Y then Y is forced on if not already
>> on. is there a way to differentiate between an explicit OFF and an
>> implicit OFF that we could use to force on if implicit but error if
>> explicitly off.
>>
>> This might be useful if a user/packager/compiler is not aware that X
>> requires Y but he can't/doesn't want Y for whatever reason.
>>
>> i am not sure how much sense i am making so an example maybe
>>
>> USER decides to build themselves, he doesn't want or have python but
>> the action menu sounds interesting.
>> USER does cmake -DKICAD_PYTHON_SCRIPTING=NO -DKICAD_ACTION_MENU=ON.
>> kicad build fails due to lack of python. USER gets mad (this ones not
>> too bad as he can see in build log it was forced on)
>>
>> if USER had python (but still didn't want it built in to kicad)
>> 3months down the line USER has an issue, uses copy version info and
>> wonders why PYTHON_SCRIPTING=ON when USER said no.
>>
>> or from a packagers perspective
>> PACKAGER likes to add all new features but has a set of packages which
>> don't require python
>> PACKAGER does does cmake -DKICAD_PYTHON_SCRIPTING=NO -DKICAD_ACTION_MENU=ON.
>> PACKAGER doesn't read log as its really long and requires effort
>> PACKAGER distributes packages saying "No python required, includes action menus"
>> USER downloads PACKAGERs package and doesn't have python and complains
>> because it doesn't run
>>
>> Both of these would make it hard to glimpse at the cmake log unless
>> the packager makes it available in the latter case to see that it was
>> forced on (or dig into the source to find out why)
>>
>> Simon
>>
>> On Wed, Jan 18, 2017 at 7:46 AM, Jean-Samuel Reynaud
>> <js.reynaud@xxxxxxxxx> wrote:
>>>> yes i was trying to subtly imply that :)
>>> ok, find attached the patch with the about box updated ;)
>>>
>>> Regards,
>