Frederik Seiffert

Yeah pretty sure I was on Windows when running into this. Thanks for taking a look.

@davidchisnall could you do me a favor when you get a chance and take a look at [these linker errors](https://github.com/gnustep/tools-windows-msvc/pull/4/checks?check_run_id=3232221515) I’m getting for [libdispatch](https://github.com/apple/swift-corelibs-libdispatch) when building it for 32-bit with...

Thanks for the suggestion using the VS generator with `-T ClangCL`. Unfortunately though that is giving the exact same linker errors with `-A Win32`. Btw. I also tried using that...

Feedback in the [libobjc2 issue](https://github.com/apple/swift-corelibs-libdispatch/issues/606#issuecomment-1264763460): > There was a time where the x86 build of libdispatch didn't work, but I believe all the issues have been resolved. Yes, this is...

Thanks for the report. Is this with Clang 13.0.0 installed from the LLVM website? @davidchisnall could you take a look at this?

@levinli303 can you try building this without GNUstep, maybe something like this: clang-cl -fobjc-runtime="gnustep-2.0" -Xclang -fexceptions -Xclang -fobjc-exceptions /MDd /c test.m You might have to replace "NSException" with "id" in...

So I think this only happens when `@finally` is present, we just ran into this as well. `@try`/`@catch` without `@finally` works fine for us. @levinli303 would be great if you...

Thanks! Could you maybe clarify in the bug that this specifically happens with `@finally`, and maybe add/reference this info from David (I don’t seem to have permissions to even add...

I remembered [this comment](https://github.com/gnustep/libobjc2/pull/138#issuecomment-623005701) from last year that I think is talking about the same issue: > I think the funclets for cleanups are incorrect in the SEH codegen (Windows...

@DHowett if you could take a look even at the "hobby" level that would certainly be appreciated! 🙏 (Another user ran into this issue with an existing large code base...