Welcome,
Guest
|
|
TOPIC: TCI-TL interface
TCI-TL interface 11 Jul 2007 11:31 #7158
|
Hi,
The clauses 7.3.4.1.9, 7.3.4.1.10 and 7.3.4.1.11 from “Part6: TTCN-3 Control Interface” describe the operations provided by the TL to log send operations. According to the “Constraint” description field, these logging operations shall be called by the SA after the send operation is performed. But looking at the input parameters defined in the prototype description of these functions, weÂ’ve found that some of them may be hard to be passed by the System Adaptor. ThatÂ’s the case for the source file of the test specification (and the number of line) under execution and, mainly, the msgValue parameter (defined as the value to be encoded and sent). Because SA send operations are called by the TE passing a TriMessageType as parameter, the System Adaptor doesnÂ’t have information about the original value (before it was encoded by the TE->CD call), so we donÂ’t understand how will be able the SA to pass this value to the tliMSend_m operations. We also think it could be a difficult task for the SA to get information about which test specification file (and number of line!) is being under execution. And this information, according to the TE->SA interface is never provided by the TE. In the other hand, we know that all this information is available in the TE environment, so we wonder if these tliMSend_m operations are really called by the SA after sending a message, because the TE seems to be a best candidate to invoke tliMSend_m operations after calling any of the triSend functions provided by the SA. I have also a question regarding the “msg” parameter (TriMessageType). How is supposed the TL has to manage this parameter? Should the TL logs the TriMessageType as is in the log file (or a binary representation of it)?? Or maybe it has to call the CD in order to decode this value before write it into the log file? We can not assume the second case because TL-CD interaction is not allowed (?). To summarize: Are the tliMSend_m operations really called by the SA? In that case, a redefinition of input parameters should be done (new CR?) Should the tliMSend_m operations be invoked by the TE? (New CR?). Parameters defined in the functions prototype doesnÂ’t need to be reviewed. In any case, we think a clarification should be done for how the TriMessageType value has to be logged. Best regards, David DÃaz RodrÃguez Computer Engineer Senior SW Development Consultant Tel.: (+34) 91 353 15 64 (Ext. 120) Fax: (+34) 91 359 61 79 Métodos y TecnologÃa (MTP) Paseo de la Castellana, 182 10º 28046 - Madrid, Spain www.mtp.es <www.mtp.es/> |
The administrator has disabled public write access.
|
TCI-TL interface 11 Jul 2007 16:33 #7159
|
Dear David,
the tliMSend_m log event should be called by the TE as within the SA (as you already mentioned correctly) you don't have parameters like e.g. the file and line info as well as the TCI value. This should be corrected. Could you please submit a CR. The TriMessage is the encoded TCI Value, i.e. the binary representation has to be logged here. Cheers, Ina. ________________________________ From: active_ttcn3 : mts stf133 ttcn version 3 - active members only [This email address is being protected from spambots. You need JavaScript enabled to view it.] On Behalf Of David Diaz Sent: Wednesday, July 11, 2007 1:32 PM To: This email address is being protected from spambots. You need JavaScript enabled to view it. Subject: TCI-TL interface Hi, The clauses 7.3.4.1.9, 7.3.4.1.10 and 7.3.4.1.11 from "Part6: TTCN-3 Control Interface" describe the operations provided by the TL to log send operations. According to the "Constraint" description field, these logging operations shall be called by the SA after the send operation is performed. But looking at the input parameters defined in the prototype description of these functions, we've found that some of them may be hard to be passed by the System Adaptor. That's the case for the source file of the test specification (and the number of line) under execution and, mainly, the msgValue parameter (defined as the value to be encoded and sent). Because SA send operations are called by the TE passing a TriMessageType as parameter, the System Adaptor doesn't have information about the original value (before it was encoded by the TE->CD call), so we don't understand how will be able the SA to pass this value to the tliMSend_m operations. We also think it could be a difficult task for the SA to get information about which test specification file (and number of line!) is being under execution. And this information, according to the TE->SA interface is never provided by the TE. In the other hand, we know that all this information is available in the TE environment, so we wonder if these tliMSend_m operations are really called by the SA after sending a message, because the TE seems to be a best candidate to invoke tliMSend_m operations after calling any of the triSend functions provided by the SA. I have also a question regarding the "msg" parameter (TriMessageType). How is supposed the TL has to manage this parameter? Should the TL logs the TriMessageType as is in the log file (or a binary representation of it)?? Or maybe it has to call the CD in order to decode this value before write it into the log file? We can not assume the second case because TL-CD interaction is not allowed (?). To summarize: Are the tliMSend_m operations really called by the SA? In that case, a redefinition of input parameters should be done (new CR?) Should the tliMSend_m operations be invoked by the TE? (New CR?). Parameters defined in the functions prototype doesn't need to be reviewed. In any case, we think a clarification should be done for how the TriMessageType value has to be logged. Best regards, David DÃaz RodrÃguez Computer Engineer Senior SW Development Consultant Tel.: (+34) 91 353 15 64 (Ext. 120) Fax: (+34) 91 359 61 79 Métodos y TecnologÃa (MTP) Paseo de la Castellana, 182 10º 28046 - Madrid, Spain www.mtp.es <www.mtp.es/> |
The administrator has disabled public write access.
|
TCI-TL interface 12 Jul 2007 06:27 #7160
|
Thank you, Ina.
Now it is clear! I will submit a CR. Regards, David DÃaz RodrÃguez Computer Engineer Senior SW Development Consultant Tel.: (+34) 91 353 15 64 (Ext. 120) Fax: (+34) 91 359 61 79 Métodos y TecnologÃa (MTP) Paseo de la Castellana, 182 10º 28046 - Madrid, Spain www.mtp.es <www.mtp.es/> _____ De: active_ttcn3 : mts stf133 ttcn version 3 - active members only [This email address is being protected from spambots. You need JavaScript enabled to view it.] En nombre de Schieferdecker, Ina Enviado el: miércoles, 11 de julio de 2007 18:34 Para: This email address is being protected from spambots. You need JavaScript enabled to view it. Asunto: Re: TCI-TL interface Dear David, the tliMSend_m log event should be called by the TE as within the SA (as you already mentioned correctly) you don't have parameters like e.g. the file and line info as well as the TCI value. This should be corrected. Could you please submit a CR. The TriMessage is the encoded TCI Value, i.e. the binary representation has to be logged here. Cheers, Ina. _____ From: active_ttcn3 : mts stf133 ttcn version 3 - active members only [This email address is being protected from spambots. You need JavaScript enabled to view it.] On Behalf Of David Diaz Sent: Wednesday, July 11, 2007 1:32 PM To: This email address is being protected from spambots. You need JavaScript enabled to view it. Subject: TCI-TL interface Hi, The clauses 7.3.4.1.9, 7.3.4.1.10 and 7.3.4.1.11 from “Part6: TTCN-3 Control Interface” describe the operations provided by the TL to log send operations. According to the “Constraint” description field, these logging operations shall be called by the SA after the send operation is performed. But looking at the input parameters defined in the prototype description of these functions, weÂ’ve found that some of them may be hard to be passed by the System Adaptor. ThatÂ’s the case for the source file of the test specification (and the number of line) under execution and, mainly, the msgValue parameter (defined as the value to be encoded and sent). Because SA send operations are called by the TE passing a TriMessageType as parameter, the System Adaptor doesnÂ’t have information about the original value (before it was encoded by the TE->CD call), so we donÂ’t understand how will be able the SA to pass this value to the tliMSend_m operations. We also think it could be a difficult task for the SA to get information about which test specification file (and number of line!) is being under execution. And this information, according to the TE->SA interface is never provided by the TE. In the other hand, we know that all this information is available in the TE environment, so we wonder if these tliMSend_m operations are really called by the SA after sending a message, because the TE seems to be a best candidate to invoke tliMSend_m operations after calling any of the triSend functions provided by the SA. I have also a question regarding the “msg” parameter (TriMessageType). How is supposed the TL has to manage this parameter? Should the TL logs the TriMessageType as is in the log file (or a binary representation of it)?? Or maybe it has to call the CD in order to decode this value before write it into the log file? We can not assume the second case because TL-CD interaction is not allowed (?). To summarize: Are the tliMSend_m operations really called by the SA? In that case, a redefinition of input parameters should be done (new CR?) Should the tliMSend_m operations be invoked by the TE? (New CR?). Parameters defined in the functions prototype doesnÂ’t need to be reviewed. In any case, we think a clarification should be done for how the TriMessageType value has to be logged. Best regards, David DÃaz RodrÃguez Computer Engineer Senior SW Development Consultant Tel.: (+34) 91 353 15 64 (Ext. 120) Fax: (+34) 91 359 61 79 Métodos y TecnologÃa (MTP) Paseo de la Castellana, 182 10º 28046 - Madrid, Spain www.mtp.es <www.mtp.es/> |
The administrator has disabled public write access.
|
|