[PD-dev] Re: External development: preventing several copies of external

[IOhannes m zmoelnig]

On 2015-06-14 21:36, Alexandre Clément wrote:
> 
> I tried going the same approach as checking if a receiver exists before
> attempting to send, but it doesn't seem to work.
> 
> if (gensym("class_name")->s_thing) {
>     post("Exists!");
>     return false;
> }
> 
> Any help is greatly appreciated.

as you already know a correct solution to your problem, just two remarks
as way this canot work.

#1 gensym("class_name") refers to the literal "class_name" (and your
class probably is not named like that). you probably have this reight
anyhow, butit reads a bit weird. the following is a bitless ambiguous:
> gensym(class_name) // hinting at a variable named "class_name".
or
> gensym("FOO")

#2 the "s_thing" member symbols points to "listeners" of that symbol.
e.g. if you create [receive foo], then gensym("foo")->s_thing will give
you a handle to that receive-object (indirectly).
you *can* use this to check for the existence of another instance of
your object, by "binding" (using 'pd_bind()') your successfully created
object to a name (e.g. "foo") in the creator (myobject_new()).
however, you probably should not do that, as it is fragile: if the user
has a [receive foo] anywhere in their patch, this logic will wrongly
assume that there is already an instance of this object.



but anyhow: it is usually not a good idea to prevent multiple instances
of objects.
it breaks the functional paradigm of Pd in a fundamental way.
if your code uses global variables (that cannot be used by multiple
instances independently), you should fix the code so it allows for more
than one instance.
if your code accesses mutually exclusive ressources (e.g. a hardware
device that cannot be opened concurrently), it's not a good idea either
to maintain a global variable yourself to check whether the mutually
exclusive access has already taken place. instead query the device
whether it allows to be opened "now".
also what happens if your ressource becomes available again (e.g.
because the user closed the patch that contains the *other* instance of
your external)? wouldn't it be nice if the remaining instance could now
"take over"?

but of there are numerous exceptions to that rule :-)


fgmas
IOhannes



-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: OpenPGP digital signature
URL: <http://lists.puredata.info/pipermail/pd-dev/attachments/20150615/9db303eb/attachment.sig>