RE: [ws-rx] State tables, July 5 version and diffs from June 29 version

Ah, I see, I was confused because of
the'[msg]' - I thought we were generating the fault as a result of receiving
the fault not the ack.

-Doug

"Bob Freund-Hitachi"
<> 

07/06/2006 04:59 PM

To

Doug Davis/Raleigh/IBM@IBMUS, "Bob
Freund-Hitachi" <>

cc

"Bob Freund-Hitachi" <>,
"[WS-RX]" <>

Subject

RE: [ws-rx] State tables, July 5 version
and diffs from June 29 version

Ok,

I guess that ack req is optional
too given the MAY in the spec.

On the point of InvalidAck[msg]
that is intended to mean an acknowledgment that is invalid, not a fault
message

When the invalid acknowledgement
range occurs , it causes the generation of an Invalid Acknowledgement Fault
see sec 4.4

Thanks

-bob

 

 

From: Doug Davis [mailto:]

Sent: Thursday, July 06, 2006 4:26 PM

To: Bob Freund-Hitachi

Cc: Bob Freund-Hitachi; [WS-RX]

Subject: RE: [ws-rx] State tables, July 5 version and diffs from June
29 version

 

Ah, sorry I missed the <> in the legend.  Shouldn't AckReq be
in <> then ? 

re:InvalidAck [msg] - but isn't this an incoming msg?  When an InvalidAck
msg comes in you think it should generate an InvalidAck msg?

-Doug 

"Bob Freund-Hitachi"
<> 

07/06/2006 04:11 PM

To

Doug Davis/Raleigh/IBM@IBMUS,
"Bob Freund-Hitachi" <>

cc

"[WS-RX]" <>

Subject

RE: [ws-rx] State tables, July 5 version
and diffs from June 29 version

 

 
 

Comments in-line: 

  

 

From: Doug Davis [mailto:] 

Sent: Thursday, July 06, 2006 3:57 PM

To: Bob Freund-Hitachi

Cc: [WS-RX]

Subject: Re: [ws-rx] State tables, July 5 version and diffs from June
29 version 

  

More comments on the state tables:

RMS Table: 

Why does "<CloseSequence>" have the <>'s?

Because it is optional per Section 3.2

In the SeqAck(final) [msg] row, I think you just can just "Process
Ack" not "Process Ack/Nack" since a final Ack can't contain
a Nack. 

  

You are correct as much as it irks me.

InvalidAck [msg] row, I'm not sure they would generate an InvalidAckFault
since that seems to imply they may be sent in response to this msg,  I
think an 'No action' is more consistent with you other cells.

No, Spec says the fault happens (see Sec 4.4)

Event occurs upon the receipt of the RMS of an invalid acknowledgement
range 

RMD Table: 

CloseSequence autonomously and SeqAck autonomously have <>'s

They are both optional. 

Still don't get the differ between N/A and blank.  :-)

Blank means we don’t say anything about the combo. N/A says we would be
daft if we tried 

thanks, 

-Doug 

"Bob Freund-Hitachi"
<> 

07/05/2006 06:02 PM

 

To

"[WS-RX]" <>

cc

  

Subject

[ws-rx] State tables, July 5 version
and diffs from June 29 version

  

 

  
 

Changes state names from connecting/connected to creating/created

Changes xmit faults to generate failts

Other minor references[attachment "wsrm-1.1-spec-wd-15-ith-ST-Edits2006-07-05.doc"
deleted by Doug Davis/Raleigh/IBM] [attachment "wsrm-1.1-spec-wd-15-ith-ST-Edits2006-07-05-diffs.doc"
deleted by Doug Davis/Raleigh/IBM]