Conversation
| + (_, None) | (CompileKindFallback::JustHost, _) => { | ||
| + if let Ok(sb2_tgt) = std::env::var("SB2_RUST_TARGET_TRIPLE") { | ||
| + Ok(vec![CompileKind::Target(CompileTarget::new(&sb2_tgt)?)]) | ||
| + Ok(vec![CompileKind::Target(CompileTarget::new(&sb2_tgt, false)?)]) |
There was a problem hiding this comment.
Here JustHost can end up with a Target CompileKind
There was a problem hiding this comment.
Thanks, will rethink this patch as well! I have a suspicion this is something that has worked by accident earlier...
| + if clear_tmp { | ||
| + env::remove_var("TMP"); | ||
| + unsafe { | ||
| + env::remove_var("TMP"); |
There was a problem hiding this comment.
You set TMPDIR earlier then remove TMP, is that right? Apparently Rust 1.95 now creates the linker temporary directory explicitly beside the output using tempdir_in(output.parent()), so changing TMPDIR no longer affects that directory. Seems like this patch might not still be needed at all.
There was a problem hiding this comment.
Good to hear that from outside! That is the resolution I've been walking towards myself. I haven't been able to reproduce the bug this patch fixed by putting export TMPDIR=/tmp in .spec file, which effectively disables the patch.
2 participants
No description provided.