serde seems to have added a breaking change in 1.0.157, would it have been detected by cargo-semver-checks?

[ginger51011]

See this issue for serde

So I've been casually pulling out my hair trying to figure out where my problem was (first blaming another poor crate, or rustc), and then it turns out it was serde releasing what ended up being (or looks to me) to be a breaking change in a patch version. Doing =1.0.156 fixes it, but still.

Would serde have been saved by cargo-semver-checks, or do we need a new query?

[obi1kenobi]

I don't think currently cargo-semver-checks would have helped here, but it's probable that a new lint could catch this.

I'll keep an eye on that issue. Once we see what the fix looks like (or the code that broke it), we'll know for sure what the lint rule should look for.

Thanks for bringing this to my attention, I probably would have missed it otherwise!

[ginger51011]

It was closed now in 1.0.159, see this PR

[obi1kenobi]

Unfortunately, it seems like it's a non-trivial change. I was hoping it might be something macro-rules-like which we might have been able to parse and check. But instead it boils down to finding a behavior change in the output of a bunch of complex Rust code, which isn't as easy. I don't think it can be done in a cost-effective way unfortunately.

[ginger51011]

As I suspected. Oh well, thanks for taking a look!

[obi1kenobi]

Thanks for mentioning it!

…

On Wed, Mar 29, 2023, 1:25 AM Emil Eriksson ***@***.***> wrote: As I suspected. Oh well, thanks for taking a look! — Reply to this email directly, view it on GitHub <#423 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAR5MSRCUFIL6OFM3PKHJADW6PBVDANCNFSM6AAAAAAWI3B6GE> . You are receiving this because you commented.Message ID: <obi1kenobi/cargo-semver-checks/repo-discussions/423/comments/5461402@ github.com>