diff options
| author | Philip Craig <philipjcraig@gmail.com> | 2019-03-23 17:00:04 +1000 |
|---|---|---|
| committer | Philip Craig <philipjcraig@gmail.com> | 2019-03-23 17:13:58 +1000 |
| commit | e8de4c3b1f0ab51fe3ea0a494cf9727ee3e0bf1d (patch) | |
| tree | c9b2d4d2a1b8012e37c9092919d8a091af9bdffa /src/test/debuginfo/enum-thinlto.rs | |
| parent | c82834e2b930619d4143e4dd94fb80888fa190c9 (diff) | |
| download | rust-e8de4c3b1f0ab51fe3ea0a494cf9727ee3e0bf1d.tar.gz rust-e8de4c3b1f0ab51fe3ea0a494cf9727ee3e0bf1d.zip | |
Fix invalid DWARF for enums when using thinlto
We were setting the same identifier for both the DW_TAG_structure_type and the DW_TAG_variant_part. This becomes a problem when using thinlto becauses it uses the identifier as a key for a map of types that is used to delete duplicates based on the ODR, so one of them is deleted as a duplicate, resulting in invalid DWARF. The DW_TAG_variant_part isn't a standalone type, so it doesn't need an identifier. Fix by omitting its identifier.
Diffstat (limited to 'src/test/debuginfo/enum-thinlto.rs')
| -rw-r--r-- | src/test/debuginfo/enum-thinlto.rs | 48 |
1 files changed, 48 insertions, 0 deletions
diff --git a/src/test/debuginfo/enum-thinlto.rs b/src/test/debuginfo/enum-thinlto.rs new file mode 100644 index 00000000000..ae736e40bcc --- /dev/null +++ b/src/test/debuginfo/enum-thinlto.rs @@ -0,0 +1,48 @@ +// ignore-tidy-linelength + +// Require LLVM with DW_TAG_variant_part and a gdb that can read it. +// min-system-llvm-version: 7.0 +// min-gdb-version: 8.2 + +// compile-flags:-g -Z thinlto + +// === GDB TESTS =================================================================================== + +// gdb-command:run + +// gdb-command:print *abc +// gdbr-check:$1 = enum_thinlto::ABC::TheA{x: 0, y: 8970181431921507452} + +// === LLDB TESTS ================================================================================== + +// lldb-command:run + +// lldb-command:print *abc +// lldbg-check:(enum_thinlto::ABC) $0 = ABC { } + +#![allow(unused_variables)] +#![feature(omit_gdb_pretty_printer_section)] +#![omit_gdb_pretty_printer_section] + +// The first element is to ensure proper alignment, irrespective of the machines word size. Since +// the size of the discriminant value is machine dependent, this has be taken into account when +// datatype layout should be predictable as in this case. +#[derive(Debug)] +enum ABC { + TheA { x: i64, y: i64 }, + TheB (i64, i32, i32), +} + +fn main() { + let abc = ABC::TheA { x: 0, y: 0x7c7c_7c7c_7c7c_7c7c }; + + f(&abc); +} + +fn f(abc: &ABC) { + zzz(); // #break + + println!("{:?}", abc); +} + +fn zzz() {()} |
