Is it time for release 13?

[caryr]

A large number of changes have gone into the repo since out last release almost a year ago, so I believe it is time to start thinking about a V13 release.

I'm currently working on fixing a VPI issue #1038 and there is one more I would like to look at and fix if possible #1041. Are there any other bugs or enhancements people are working on that they would like to get into the repo before a new release?I'm trying to focus on bugs with the limited time I have right now. Also if someone else wants to look at #1041 that would just make it easier.

@steveicarus do you have time to roll a release once we decide it is completed? Hopefully in no more than a couple weeks.

I hope everyone is doing well. Take care, Cary

[martinwhitaker]

I say go for a release. I don't think any of the open bugs are serious enough to delay it.

I did have a quick look at #1041. I think the problem is that vpi_put_value sends the value to the output of the net, whereas it needs to send it to the input. I may have some free time this weekend to work on it.

[steveicarus]

I agree, this is a good time for a new release. I have removed the release specific bits from the ivtest suite, so releases should be even easier now.

[caryr]

I think once the tests were integrated into the repo the TB was then synced. The recent removals is mostly getting ride of cruft that no longer serves a purpose, but does not impact a release.

[caryr]

One other thing that should be synced is the list of constructs not supported is incorrect in the local documentation. This had been updated on the external web page, but those changes never made it back into the repo documentation. We probably should do a pass over the entire documentation set to make sure things are as expected before the release.

I'm going to try to finish #1038 this weekend and I need to look at the algorithm for time rounding in the context of my recent FreeBSD fix since it is numerically very unstable if being off by a single bit in the pow() function causes a test failure.

[steveicarus]

It does, since creating a new release would have required creating a ne regression-<N>.list. That file was selected automatically. And yes, lots of cruft was also cleaned out.

…

On Thu, Jan 4, 2024 at 9:39 AM Cary R. ***@***.***> wrote: I think once the tests were integrated into the repo the TB was then synced. The recent removals is mostly getting ride of cruft that no longer serves a purpose, but does not impact a release. — Reply to this email directly, view it on GitHub <#1058 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAACKMXT7M77CAC5FEYZ2ETYM3SNFAVCNFSM6AAAAABBG4YEVKVHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DAMJVGQ3TG> . You are receiving this because you were mentioned.Message ID: ***@***.***>

-- Steve Williams ***@***.*** ***@***.***>

[caryr]

Ah, yes.

[gatk555]

A newcomer here, so feel free to ignore me. I think I have a modified version of #667 that is ready to become a replacement PR. But there is no demonstrated use case, as that was blocked by #1041. It seems that there may now be progress on this, so for me it would be nice if there was time to validate #667 and have it considered for the next release. On the other hand #667 rearranges fundamental code in vvp and I would completely understand the response "That needs to simmer for months before being considered for release."

The only reason this might matter is that my application is intended as a new feature in mixed-signal simulator ngspice (like this). Ngspice typically has three releases a year (with one last week). So if "vvp as shared library" misses the boat, I will have to remember to remove "a development build of iverilog must be used" from the Ngspice manual, a year or two from now.

Thanks,
G.

[caryr]

In general if you can completely articulate the changes you are making, why they are correct and the test suite passes we do not usually require a simmering period.

[martinwhitaker]

I see there is currently one unexpected failure in the vlog95 target tests.

[caryr]

Yes, it is related to the changes done for SDF support. I've not looked into if this is a bug in the compiler or a limitation in the vlog95 code generator.

[martinwhitaker]

Time is slipping by. It would be good to get a new release out soon.

[caryr]

I hope to have some time in a week or two to finish the task I was trying to get done for the release. My hectic day job is getting in the way of all the Icarus fun.