For that matter, not even thinking about what a vpiFullName is, the 1800 LRM doesn't actually extend the %m specifier for SystemVerilog constructs in general. So if we were to interpret the LRM strictly, does this mean that %m is defined only by 17.1.1.6 of 1364-2005? In that case, it prints only the hierarchical name of a module, task, function, or named block -- not even that of an interface, let alone a function extern'd within an interface! For an object in a package, it should probably emulate what we've described as the vpiFullName for such objects, ditto for mdethods in a class. Is there a ballot comment around on %m so that we could actually do something about it? Or should we create a whole new section on hierarchical names on the excuse that we need it to define a vpiFullName? Regards, Jim V. --------------------------------------------------------- James H. Vellenga 978-262-6381 Engineering Director (FAX) 978-262-6636 Cadence Design Systems, Inc. vellenga@cadence.com 270 Billerica Rd Chelmsford, MA 01824-4179 "We all work with partial information." ---------------------------------------------------------- ] -----Original Message----- ] From: owner-sv-cc@eda.org [mailto:owner-sv-cc@eda.org] On ] Behalf Of Charlie Dawson ] Sent: Thursday, April 14, 2005 3:56 PM ] To: SV-CC ] Subject: [Fwd: Re: [sv-cc] Definition of full name in 1364] ] ] ] ] -------- Original Message -------- ] Subject: Re: [sv-cc] Definition of full name in 1364 ] Date: Thu, 14 Apr 2005 15:53:53 -0400 ] From: Joao Geada <jgeada@comcast.net> ] To: Charles Dawson <chas> ] References: <425EC6FF.1070600@cadence.com> ] ] Chas: ] ] What would "%m" print in a task/interface inside a package? ] What if the package was visible in multiple independent compilation ] units? ] ] What would "%m" print inside a forked-not-joined task/function ? ] (particularly if running live in multiple threads) ] ] What would "%m" print for a extern task in an interface? ] ] What does "%m" print for methods of a class? ] ] Joao ] ] On Thu, 2005-04-14 at 15:39 -0400, Charles Dawson wrote: ] > Hi All, ] > ] > At the last meeting I took an action to look for the definition of ] > vpiFullName in the 1364 specification. This was in reference to ] > Mantis Item 610. ] > ] > I could not find one, so perhaps we should try to make a ] definition. ] > I know that I have had a view of what vpiFullName should return for ] > quite some time, but I'm not sure how to put it into the spec. ] > ] > It goes something like this: ] > - Full hierarchical name starting from the top module. ] > - A string which can then be successfully parsed by ] > vpi_handle_by_name() such that the same object would be ] > returned as that with which we started. ] > - The format would also be the same as what $display("%m") might ] > return. Not sure if this is necessary. ] > ] > Am I missing any important points? ] > ] > -Chas ] > ] > ] ] ] ] -- ] Charles Dawson ] Senior Engineering Manager ] NC-Verilog Team ] Cadence Design Systems, Inc. ] 270 Billerica Road ] Chelmsford, MA 01824 ] (978) 262 - 6273 ] chas@cadence.com ] ] ] ]Received on Thu Apr 14 14:07:25 2005
This archive was generated by hypermail 2.1.8 : Thu Apr 14 2005 - 14:07:49 PDT