Reply To: Standard messaging feature in OSVVM?

Why OSVVM™? Forums OSVVM Standard messaging feature in OSVVM? Reply To: Standard messaging feature in OSVVM?

#814
Jim Lewis
Member

Hi Ralph,

My preference in OSVVM printing would have been to be closer to textio style operations, however, the coverage data structure must be embedded in a protected type and protected type methods do not allow pointers or file handles. 

As a result, to handle anything other than OUTPUT, I embedded the file handle internal to the data structure.   It would be possible for OSVVM printing to call a separate protected type that handles this sort of thing for it.  For my uses, I need to be able to maintain the options that a model may print to OUTPUT, a file dedicated to its own usage, a single file shared by the whole test, or possibly a file shared by a group of tests (this may require some sort of generic package passing and may be on the bleeding edge of simulator support).   With the right foundation layer, it is likely we can apply a name based filtering to the printing – provided of course the foundation layer also allows a “non-optional” mode of printing also. 

At test completion, I prefer that each transaction model provides a summary of the errors it has seen and then to print the total error count seen by the testbench.   I am not sure how this currently fits into BitVis. 

I know Espen, so we may be able to get some convergence on this.  Not sure how the difference in our testbench approaches will accommodate each other.  

If it does not work out with the BitVis library, I would then expect to find another one – unfortunately , this would also mean finding someone to work on it as I am quite tied up with CoveragePkg and some of the other planned items – scoreboards and memory models. 

I have some high priority updates I am working on for CoveragePkg (new features).  Once I finish those, I will need to set aside a block of time to come back to reporting.

Best Regards,

Jim