Aleksandar Pejović
Aleksandar Pejović
@tabiStein Is `libawt.so` next to the built executable? If not, was it listed in the build output as a produced artifact?
> No, the executable is alone in a workspace/ directory. Well, that's the problem. All produced artifacts are necessary for the executable to work (i.e., it can't work without the...
> @pejovica Is there any reason why this PR has still not been integrated? > > Simon (@stooke) has code already lined up that will use location info. However, that...
@adinn @stooke There seems to be some misunderstanding about how calling from native works, but literally anything that can be called from native must be compiled with the `Native` calling...
> @pejovica I take your point and agree that `ParamLocationProducer` needs to be updated to respect use of the `Native` calling convention by the relevant subset of the generated method...
> @pejovica , as discussed before, this code lays the groundwork for the full implementation. The current variable information code in this PR is more a proof of concept that...
@maxandersen No, I've been busy with support for Windows x64 unwind info, so I haven't had much chance to work on this. I plan to come back to this once...
AWT support for Linux has landed in master (see #6088). We now use dynamic linking as we do on Windows.
I'm working on our Windows support, and I've looked at your changes. Most of them are already being handled (e.g., I'm dealing with the Truffle NFI at the moment), so...
@stooke Thanks for the rebase and the heads-up! I'll take another look now.