[SailfishDevel] What happens where during building and deployment?

[Wim de Vries]

Follow up on topic no "c++ in Sailfish?" and about building/installing a 
module.

I'll skip the step of making an rpm of the module for now.
I think (for testing)  I will just build and install of the qtserial 
module and headers.

Still, I have been browsing the MerSDK VM and I have questions.
(If I can get this more clear, I am willing to write some documentation 
for other starters on Sailfish)

Please confirm/correct.

-code is and stays on the host machine;
-qmake is run on the host machine;
-building is done on the MerSDK VM
    > but where, in what directories?
    > where, in what directories does the SDK live?
    > on which directory of MerSDK VM is the ~ of the host mounted?
-rpm is created on the MerSDK VM;
-rpm is installed on the SailFish VM.
Thanks.

r


On 04/17/2013 12:10 PM, David Greaves wrote:
> On 17/04/13 10:34, Wim de Vries wrote:
>> christopher.lamb at thurweb.ch schreef op 2013-04-16 23:05:
>>> If so, what ends up (if anything) in the Sailfish SDK and Emulator VMs?
>> I am just concentrating on the MER SDK for now.
>> Building the QtSerialPort project goes without errors.
>> But I haven't yet found out where the libs and headers did end up.
>> (I am not very experienced on VMs:
>> due to the directory mapping between host and VM it's hard to determine what is
>> living on which (virtual) machine)
>> Building pilotnavigator fails because QtSerial headers are not found.
> This is the problem
>
>>> Depending on this, you may or may not need to deploy something to one
>>> or both in order to build and run your pilotnavigator project which
>>> imports QtSerialPort.
>> Yes, but where and how can I deploy QtSerialport on both VMs?
> This points to the solution :)
>
> For those unsure about terminology: we say QtSerialPort is a
> dependency for pilotnavigator; it could be a runtime dependency (eg if it
> provides an application or service only used during execution) or it may also be
> a build-time dependency (eg if some C++ code needs header files).
>
> In order for Sailfish to resolve this dependency automatically, QtSerialPort
> needs to be packaged into an rpm and SailfishOS needs to be told where to get it
> from.
>
> So step 1 is to make and rpm of QtSerialPort, step 2 is to put it somewhere (a
> repository) where SailfishOS can retrieve it - either when building the
> pilotnavigator app or when a user installs it.
>
> Both of these steps are still being polished - they are both possible right now
> but they are more complicated than we want them to be.
>
> I'll get some instructions pulled together - if you want to start by yourself
> then look at the spec file for making rpms and find out about the merproject OBS
> for building/publishing them as a repo (#sailfishos irc channel may be a good start)
>
>>> But stepping back a bit: are you actually asking the right question?
> Sensible thing to ask but even if not, this is still an important thing to be
> able to do.
>
> David
>
>>> p.s how are the thermals in Holland? We had the first real credible
>>> thermals this weekend, and boy were they wonderful and so well
>>> deserved after such a long winter ...And in 15 years of flying, I
>>> don't think I have ever been able to do a top-landing in snow before!
>> I fly MLA nowadays, but I still use thermals to save fuel (not appreciated by
>> other motorized pliots!)
> Ah, I miss my hang-glider :)
>
>
>