DavidSpickett wrote:
It's been a while, so it's missing a clear summary of:
* Which features this PR enables for WASM
* Which runtimes or debug servers this can be used with
* Which features might be enabled in future
* Which features are simply not applicable
* What the plans are for testing this going forward
* Who is going to test this and how will we (the LLDB community) know that it
works, or does not work.
All these things have been stated somewhere in the comments, but it needs to be
put together into one comment so we can all get up to speed again.
And, if we were to accept this, a version of that will have to go into the
release notes so users know what to expect. As you say, there is a good amount
of interest in this feature.
My biggest concern here is that we're taking on a whole new architecture with
very little testing. Now that might be ok given that you can't enable all
features at once, so early patches will have to have minimal testing. That's
why I'm asking for some sort of roadmap, even if it has to be vague and will
likely be derailed at some point.
Given that there exists a qemu-user platform that I think people have used to
run API tests against qemu-user, I think you could get the test suite running
for WASM using that (it's name qemu-user but you can put other simulators ).
```
(lldb) settings set platform.plugin.qemu-user.
Available completions:
platform.plugin.qemu-user.architecture
platform.plugin.qemu-user.emulator-args
platform.plugin.qemu-user.emulator-env-vars
platform.plugin.qemu-user.emulator-path
platform.plugin.qemu-user.target-env-vars
```
Or a binfmt_misc setup might work.
You're going to have many many failures but that's expected. If it can at least
be run, then we'd have a baseline to improve from and set expectations for
contributors.
https://github.com/llvm/llvm-project/pull/77949
_______________________________________________
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
