It was pointed out to me that the docs previously only said i386 was supported, while I'd expect that at least x86_64 is on the same level, and AArch64 should work pretty well for basic things at least.
Details
Diff Detail
Event Timeline
Do we support 32 bit Arm specifically? (I'm not very familiar with Windows targets)
@omjavaid Has done some testing on the latest Windows On Arm which is AArch64. I think you should probably say "ARM and AArch64 support is more" but he can comment on how well it works.
lldb/docs/index.rst | ||
---|---|---|
81 | "experimental" |
Yes, I implemented support for both ARM and AArch64 in late 2019.
@omjavaid Has done some testing on the latest Windows On Arm which is AArch64. I think you should probably say "ARM and AArch64 support is more" but he can comment on how well it works.
Possibly yes. I was referring to an issue for the ARM part that I never got time to fix properly, see D70840. Other than that, I don't know of any issues with the AArch64 part, but I haven't used it very much either, fwiw.
lldb/docs/index.rst | ||
---|---|---|
80–81 | s/ARM/ARM and AArch64/ Also add a line to reflect lack of python support. |
lldb/docs/index.rst | ||
---|---|---|
80–81 | Sure, I can label the AArch64 bit experimental too. I wouldn't include comments about python here though. The fact that it's not easy to build LLDB with python for windows/aarch64 isn't an intrinsic feature of LLDB or something that is missing in LLDB, that's just python. It is actually possible to cross compile LLDB with python support for windows/aarch64, at least for the mingw target. Python itself is a bit tricky to cross compile, but once that cross compilation itself works, building it for windows/aarch64 isn't any worse than for x86. In https://github.com/mstorsjo/llvm-mingw/issues/73 there's ongoing efforts in building and packaging LLDB for windows with python enabled. As it's primarily cross compiled, it's a bit tricky for running tests though. |
I'm in no position to evaluate the functionality of lldb on Windows. I've spent the last few months trying to understand why 900+ tests spontaneously started failing on Windows.
"experimental"