Re: Errors from process control methods

Bishnupriya, David,

My feeling is the same as David's - would accept either way, but would prefer an error. I accept Bisnupriya's argument for not raising errors when process control methods are called out-of-context from processes (to avoid the caller having to check its context explicitly every time), but I do not see that the argument applies quite as strongly during elaboration, from end_of_elab, from start_of_sim, or when paused. Hence I would vote for "error" in each of these cases.

John A

-----David C Black <dcblack@xtreme-eda.com> wrote: -----
To: Bishnupriya Bhattacharya <bpriya@cadence.com>
From: David C Black <dcblack@xtreme-eda.com>
Date: 12/03/2010 05:37AM
Cc: "john.aynsley@doulos.com" <john.aynsley@doulos.com>, P1666 Technical WG <systemc-p1666-technical@eda.org>
Subject: Re: Errors from process control methods

I'm ok with consistency at a minimum, and I agree that prior to start of simulation they should also be a problem.

My point on error vs warning is "Why would a user issue a reset prior to simulation if it was not going to have an effect?". I think if I write code to do something (e.g. reset), then I expect it to occur. If it's not possible to do at the point I try, then I would like to be notified very overtly of my erroneous thinking. Thus I think it is an error.
 
On Thu, Dec 2, 2010 at 11:31 PM, Bishnupriya Bhattacharya <bpriya@cadence.com> wrote:
     
 
John,
 
It should be the same for both. I prefer a warning in both cases, but am ok with error in both cases also.
 
-Bishnupriya
    
      
From: john.aynsley@doulos.com [mailto:john.aynsley@doulos.com]
Sent: Friday, December 03, 2010 10:49 AM
To: Bishnupriya Bhattacharya
Cc: john.aynsley@doulos.com; P1666 Technical WG

Subject: RE: Errors from process control methods

   
All,

The thing is, if we made(e.g) reset() or kill() to be errors during a pause, why would they not be errors before the start of simulation? It seems rather ad hoc.

John A

-----Bishnupriya Bhattacharya <bpriya@cadence.com> wrote: -----
   To: "john.aynsley@doulos.com" <john.aynsley@doulos.com>, "systemc-p1666-technical@eda.org" <systemc-p1666-technical@eda.org>
From: Bishnupriya Bhattacharya <bpriya@cadence.com>
Date: 12/02/2010 08:44AM
Subject: RE: Errors from process control methods

          
John,
 
Good catch. I would prefer a warning over error, but can also live with an error.
 
Thanks,
-Bishnupriya
 
               
From: john.aynsley@doulos.com [mailto:john.aynsley@doulos.com]
Sent: Thursday, December 02, 2010 2:03 AM
To: Bishnupriya Bhattacharya; systemc-p1666-technical@eda.org
Subject: Errors from process control methods

            
Bishnupriya, All,

I have just noticed an inconsistency. When discussing sc_pause, we recently decided "it shall be an error" to call kill, reset, or throw_it while simulation is paused. But a while back, after some discussion, we decided that calling such methods from the wrong context, such as during elaboration, would not be an error but would have no effect other than optionally generating a warning.

So my question is, should the same apply when paused, i.e just an optional warning?

Thanks

John A

--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
  

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.