Ma Client Server missing methods
Michaël Piel
piel at ensm-douai.fr
Thu Aug 25 13:07:24 UTC 2005
On Wed, 24 Aug 2005 19:24:36 +0200
Michaël Piel <piel at ensm-douai.fr> wrote:
> > 3) MaObjectSerializer>>classDefinitions:
> > This is a Magma method, you should not need it for Ma client server. Is there
> > some reason you need it for Ma client server?
Yes it is a Magma method (in "*magma client" category) but it is used in "MaObjectSerializer>>classDefinitionsByteArray: aByteArray" which is part of Ma Client Server. I must have made a mistake yesterday on what I said below because I can't reproduce it. The problem comes with the "connect" method of MaTcpRequestServerLink. But it seems that it is not necessary, I can do a "submit:" without connecting before. Also the disconnect method of MaTcpRequestServerLink doesn't work (the primitiveSocketCloseConnection fails).
Do you have an example (like the "serverTest.ws" and "clientTest.ws") using the object request layer ?
> I don't need it specifically. But I want to use the Object Request Layer and the connect method of MaTcpRequestServerLink do "serializer classDefinitionsByteArray: (idAndProtocol at: 2)" which performs "self classDefinitions: (self materializeGraph: aByteArray)".
>
> Here is the code I used:
> server side:
> myServer := MaTcpRequestServer protocol:
> { OrderedCollection.
> Dictionary. }.
> myServer
> processOn: 12345
> using: [ :requestByteArray | requestByteArray reverse ]
>
> client side:
> myLinkToServer := MaTcpRequestServerLink
> hostAddress: #(127 0 0 1) asByteArray
> port: 12345.
> myLinkToServer connect.
> self submit: 'Hello!'
>
> After several tests (and by taking your "serverTest.ws" as starting point) I've found that adding "myServer console: MaTranscriptConsole new." in the server resolved the problem. But I'm wondering why it fails with the default console (MaRequestServerConsole).
>
>
> >
> > Let me know if you have any questions.
Do you have an example (like the "serverTest.ws" and "clientTest.ws") using the object request layer ?
Thanks,
Michael
> >
> > - Chris
> >
>
More information about the Squeak-dev
mailing list
|