chore(bindgen): bump syn to v3#3401
Conversation
| "proc-macro2", | ||
| "quote", | ||
| "syn", | ||
| "syn 2.0.104", |
There was a problem hiding this comment.
It would be possible to remove the dependency on syn v2 (which is only used by bindgen when the "private" __cli Cargo feature is enabled, as done by bindgen-cli), but the version of clap that depends on syn v3 has an MSRV of 1.85. I think it's better to keep the compatibility of bindgen for the build-script use case, even if that means compiling two versions of syn for the CLI.
Alternatively, it may be possible to exclude the "private" __cli Cargo feature from the MSRV guarantee, and give bindgen-cli an MSRV of 1.85, different from that of the rest of the workspace, but that seems unnecessarily complicated (even though I see that the README does allow for that, so your call).
|
r? @emilio |
| ] | ||
|
|
||
| [[package]] | ||
| name = "clap_derive" |
There was a problem hiding this comment.
Can we update this?
There was a problem hiding this comment.
Only if we bump the MSRV of bindgen-cli to 1.85, which I can do if you're ok with decoupling its MSRV from bindgen's.
|
Maybe you can increase the range for itertools to include 0.15 as well? (I was about to make a PR for that, but it can just as well go here.) Works on my Windows machine (as in exactly the same failing tests as with the current Cargo.lock...) |
3 participants
Following the recent release of
synv3, this bumps the dependency with the aim of reducing dependency duplication in dependents.Because
synv3 advertises an MSRV of 1.71, this first bumps the MSRV ofbindgenfrom 1.70 to 1.71 (which I think is acceptable).synv3 of course contains breaking changes: I've been through the changelog but didn't find anything relevant (but there are quite a few of them so I could have missed some). I've runcargo testbut didn't do any other kind of testing.