#305 zero width replactions are allowed in some places.

[Cary R.]

A zero width replication is allowed when it is inside an existing concatenation/replication that has other non-zero width element (1364-2005 section 5.1.14). This example crashes covered. Icarus currently incorrectly generates a message for this, though it does simulate correctly. I'll straighten Icarus out when I get some free time. This can also happen for procedural assignment which I did not show.

[Cary R.]

example file

[Trevor Williams]

This limitation of Covered is fairly pervasive in the code (i.e., not allowing 0 width data). As such, this is going to take me a few days to get everything sorted out.

Thanks,
Trevor

[Trevor Williams]

I have attached a patch for this issue which will be generally available in the 0.7.8 stable release.

[Trevor Williams]

Attaching a second patch that also needs to be added to output the correct coverage information in reports. This fix was missing from the first patch. As always, this patch will be generally available in the 0.7.8 stable release.

[Trevor Williams]

Bug fix patch

[Trevor Williams]

Second patch (should be applied in addition to the first patch)

[Trevor Williams]

Cary,

I mistakenly applied this message to the wrong bug -- I apologize for the duplication but thought that this information would be good to have attached to the (correct) bug report.

Just as an FYI, Covered currently will execute any code in the replication even if the multiplier is a value of 0. I've been following the Icarus Verilog thread on this issue. I believe that the specification stated that function calls/system calls/etc. should be executed even if the multiplier is a value of 0.

If you believe this behavior is incorrect and should be fixed, let me know.

Thanks,
Trevor