Hi,
but this will not work either (I just realised). This is complicated matter and
confuses many (even me, some times :P).
I will try explain.
Technically, when you do
array := (FFIExternalArray externalNewType: 'void*' size: parameters size)
autoRelease.
you are declaring (in C)
(void*)[] array;
which (in C) is synonym of:
void**
which means that as soon as you declare your array, you already have a void**
then, your call has to be:
...
self prim_initEmbeddedRargc: parameters size argv: externalArray.
with call of:
self ffiCall: #(int Rf_initEmbeddedR(int argc, FFIExternalArrayOfStrings argv))
(is better to declare a subclass of FFIExternalArray to be clear about your
purpose).
or the other way. Since UFFI cannot determine if an ExternalAddress is a
pointer (void*) or a pointer to a pointer (void**), because for C, at the end,
they are exactly the same, you need to declare as this:
self prim_initEmbeddedRargc: parameters size argv: externalArray getHandle.
with call of:
self ffiCall: #(int Rf_initEmbeddedR(int argc, void *argv))
remarks:
1) you need to declare it as “void*" because UFFI will box any arity > [object
natural arity], which in arrays is 1 (then, if you pass ** it will box it into
a –wrong– void***… bad.
2) in this case, since you are passing a handle (a pointer), you do not
need/want to subclass FFIExternalArray.
Finally, a third way (not tested), is just doing:
self prim_initEmbeddedRargc: parameters size argv: externalArray getHandle.
with call of:
self ffiCall: #(int Rf_initEmbeddedR(int argc, oop argv))
in this case, it will pass the handle (pointer) “as is”, without
transformation. This should work but I never tested it.
Maybe in terms of clarity, last form is better… I don’t know :P
confusing? yes
but it is because is very hard to determine/parse/interpret pointer arities and
keep the calls efficient, in a moment we need to put a limit in automatic
translations.
Esteban
> On 04 Jun 2016, at 10:10, Esteban Lorenzano <esteba...@gmail.com> wrote:
>
> Hi,
>
> sorry late response, I just missed this mails.
>
> but… I do not have much to say… what you did should work.
> Except (I think I’d need to look at your code closer) you need to change the
> argument of argv to FFIExternalArray*, or you need to pass the handle to the
> function. Something like this:
>
>
> | externalArray parameters |
>
> parameters := #('R' '--no-save' 'silent').
> externalArray := (FFIExternalArray externalNewType: 'void*' size: parameters
> size) autoRelease.
> parameters withIndexDo: [ :string :index |
> externalArray at: index put: (ExternalAddress fromString: string)
> autoRelease ].
> self prim_initEmbeddedRargc: parameters size argv: externalArray getHandle
>
> (with same primitive call)
>
> or without the #getHandle message and a primitive call:
>
> self ffiCall: #(int Rf_initEmbeddedR(int argc, FFIExternalArray* argv))
>
> AFAIK, that should work… if not I’d need to take a better look, because is
> maybe a bug :P
>
> Esteban
>
>> On 03 Jun 2016, at 18:50, Blondeau Vincent <vincent.blond...@worldline.com>
>> wrote:
>>
>> Hello,
>>
>> So, we tried with Thibault Raffaillac and we succeed to have this kind of
>> code:
>>
>>
>> | externalArray |
>> parameters := #('R' '--no-save' 'silent').
>> [ externalArray := FFIExternalArray externalNewType: 'void*' size:
>> parameters size.
>> parameters
>> withIndexDo: [ :string :index | externalArray at: index put:
>> (ExternalAddress fromString: string) ].
>> self prim_initEmbeddedRargc: parameters size argv: externalArray ]
>> ensure: [ externalArray
>> ifNotNil: [ externalArray do: [ :item | item
>> ifNotNil: [ :e | e free ] ].
>> externalArray free ] ]
>>
>> with a primitive call:
>> self ffiCall: #(int Rf_initEmbeddedR(int argc, char ** argv))
>>
>> However, it still doesn't work because the VM is crashing during the call to
>> the primitive.
>>
>> Does the translation from the Nativeboost version is accurate?
>>
>> Thanks in advance for your help,
>>
>> Vincent
>>
>>> -----Message d'origine-----
>>> De : Pharo-users [mailto:pharo-users-boun...@lists.pharo.org] De la part de
>>> Julien Delplanque
>>> Envoyé : mercredi 1 juin 2016 18:49
>>> À : pharo-users@lists.pharo.org
>>> Objet : Re: [Pharo-users] [UFFI] Call a function where argument type is
>>> char **
>>> (argv)
>>>
>>> Hello,
>>>
>>> I asked a simililar question some days ago, maybe this [1] can help.
>>>
>>> Regards,
>>>
>>> Julien
>>>
>>> Links:
>>>
>>> [1]:
>>> http://forum.world.st/Unified-FFI-pointer-of-String-as-function-parameter-
>>> td4898066.html
>>>
>>>
>>> On 31/05/16 16:28, Blondeau Vincent wrote:
>>>> TL;DR: How to pass as argument an array of Strings (char **) with Unified-
>>> FFI?
>>>>
>>>> Vincent
>>>>
>>>> De : Blondeau Vincent
>>>> Envoyé : mercredi 25 mai 2016 13:58
>>>> À : Pharo Development List
>>>> Objet : [UFFI] Call a function where argument type is char ** (argv)
>>>>
>>>> Hello,
>>>>
>>>> I have written a R bridge in Pharo and I would like to migrate it from
>>> NativeBoost to UFFI.
>>>> Most of the changes are easy to do but I am stuck to a double pointer
>>> problem.
>>>> I have to call the function: int Rf_initEmbeddedR(int argc, char ** argv).
>>>> I
>>> know how to give the int but the char ** is a problem. It is an array of
>>> Strings.
>>>>
>>>> With NB, I managed to have this (working) code:
>>>> "This is 32bit... too bad..."
>>>> strings := OrderedCollection new.
>>>> par := NativeBoost allocate: 4 * params size.
>>>> params
>>>> keysAndValuesDo: [ :i :each |
>>>> | str |
>>>> str := each
>>>> asNBExternalString.
>>>> strings add: str.
>>>> par nbUInt32AtOffset: (i -
>>>> 1) * 4 put: str value ].
>>>> self prim_initEmbeddedRargc: params size argv: par ]
>>>> ensure: [
>>>> "Free the memory we
>>>> allocated"
>>>> par ifNotNil: [ par free ].
>>>> strings ifNotNil: [
>>>> strings do: [ :each | each free ] ] ]
>>>>
>>>> With a primitive call:
>>>> Self nbCall: #(int Rf_initEmbeddedR(int argc, char *argv))
>>>>
>>>> Do I still need to create my own array of strings or FFI creates it for me?
>>> How?
>>>>
>>>> Thanks in advance for your answers,
>>>>
>>>> Vincent
>>>>
>>>>
>>>>
>>> !!!***************************************************************
>>> ****
>>>> ****************** "Ce message et les pièces jointes sont
>>>> confidentiels et réservés à l'usage exclusif de ses destinataires. Il peut
>>> également être protégé par le secret professionnel. Si vous recevez ce
>>> message par erreur, merci d'en avertir immédiatement l'expéditeur et de le
>>> détruire. L'intégrité du message ne pouvant être assurée sur Internet, la
>>> responsabilité de Worldline ne pourra être recherchée quant au contenu de
>>> ce message. Bien que les meilleurs efforts soient faits pour maintenir cette
>>> transmission exempte de tout virus, l'expéditeur ne donne aucune garantie à
>>> cet égard et sa responsabilité ne saurait être recherchée pour tout dommage
>>> résultant d'un virus transmis.
>>>>
>>>> This e-mail and the documents attached are confidential and intended
>>> solely for the addressee; it may also be privileged. If you receive this
>>> e-mail in
>>> error, please notify the sender immediately and destroy it. As its integrity
>>> cannot be secured on the Internet, the Worldline liability cannot be
>>> triggered
>>> for the message content. Although the sender endeavours to maintain a
>>> computer virus-free network, the sender does not warrant that this
>>> transmission is virus-free and will not be liable for any damages resulting
>>> from any virus transmitted.!!!"
>>>>
>>>
>>
>>
>> !!!*************************************************************************************
>> "Ce message et les pièces jointes sont confidentiels et réservés à l'usage
>> exclusif de ses destinataires. Il peut également être protégé par le secret
>> professionnel. Si vous recevez ce message par erreur, merci d'en avertir
>> immédiatement l'expéditeur et de le détruire. L'intégrité du message ne
>> pouvant être assurée sur Internet, la responsabilité de Worldline ne pourra
>> être recherchée quant au contenu de ce message. Bien que les meilleurs
>> efforts soient faits pour maintenir cette transmission exempte de tout
>> virus, l'expéditeur ne donne aucune garantie à cet égard et sa
>> responsabilité ne saurait être recherchée pour tout dommage résultant d'un
>> virus transmis.
>>
>> This e-mail and the documents attached are confidential and intended solely
>> for the addressee; it may also be privileged. If you receive this e-mail in
>> error, please notify the sender immediately and destroy it. As its integrity
>> cannot be secured on the Internet, the Worldline liability cannot be
>> triggered for the message content. Although the sender endeavours to
>> maintain a computer virus-free network, the sender does not warrant that
>> this transmission is virus-free and will not be liable for any damages
>> resulting from any virus transmitted.!!!"
>>
>
