Re: [tlaplus] How Can I Specify a State Can Be Reachable in TLC?

On Sunday, October 4, 2020 at 9:54:57 AM UTC-7 Stephan Merz wrote:

Hello again,

on your first question: TLC will indeed stop at the first error, so you probably want to perform the reachability check separately from verifying other invariants.

About the second item, I can only encourage you to document your specifications, including the properties that are checked. In a CI context, you can indicate which outcome is considered an error. For checking reachability, you want to invert the standard interpretation and consider "successful verification" an error.

Regards,
Stephan

On 4 Oct 2020, at 18:47, Bin Wang <wbi...@xxxxxxxxx> wrote:

Hi Stephan,

This is a good idea. But I still have two concerns:

1. If I also want to check other invariant, will this error early break other checks? If so, I may need to run it multiple times with different invariant.
2. If I integrate it into CI or share the model with someone else, this may seems not very intuitive and require addition explanation.
在2020年10月4日星期日 UTC-4 下午12:32:22<Stephan Merz> 写道：
If I understand correctly, you are trying to check if a certain is reachable. You can state an invariant that asserts that (a predicate characterizing) the state is always false, and TLC will display a counter-example if the state can be reached. If you have a different property in mind, an example would help.

Stephan

On 4 Oct 2020, at 18:28, Bin Wang <wbi...@xxxxxxxxx> wrote:

Hi,

I'm writing a specification. I want to check if a state can be reachable from Init in all the state space.

I have a work around to do it for now: I define this behavior as part of Next step. If TLC find it's never enabled, it will report some warnings. I'm wondering if there is a way to define this as a property of the specification and let TLC report error if it's never enabled.

Thanks!

--
You received this message because you are subscribed to the Google Groups "tlaplus" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tlaplus+u...@xxxxxxxxxxxxxxxx.
To view this discussion on the web visit https://groups.google.com/d/msgid/tlaplus/952408be-1165-4f6a-a37f-d0ddac456afcn%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "tlaplus" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tlaplus+u...@xxxxxxxxxxxxxxxx.
To view this discussion on the web visit https://groups.google.com/d/msgid/tlaplus/e2aa4938-cb32-49b8-980d-fdcb28eee122n%40googlegroups.com.