Issue 180517.1: Variant parts without a discriminant
Section 5.7.10, pg 123
I worked recently on adding variant part DWARF output to the Rust
compiler and LLVM. While doing so I found this text in 5.7.10:
If the variant part does not have a discriminant (tag field), the
variant part entry has a DW_AT_type attribute to represent the tag type.
I did not know how to interpret this. I searched and found a dwarf-discuss
thread touching on this topic, in particular this message:
http://lists.dwarfstd.org/pipermail/dwarf-discuss-dwarfstd.org/2006-August/003088.html
.. which unfortunately merely expresses the same confusion that I have.
I suggest simply removing this paragraph from the standard.
Rust needs tagless variants, but I will file a separate issue for that.
DISCUSSION
The issue here is wrt the proposed removal of the potentially confusing
following lines (Section 5.7.10, pg 123, lines 10-11):
If the variant part does not have a discriminant (tag field), the variant
part entry has a DW_AT_type attribute to represent the tag type.
Since this line holds good in the Pascal context, the proposal to remove
this line is rejected and instead, the following non-normative text can
be added to provide additional clarification.
PROPOSED CHANGES
Append the following non-normative text:
*A reference to a type supports the Pascal notion of a tagless variant
part where the omitted tag nonetheless is given a type whose values are
used in later parts of the variant syntax.*
--
2022-04-04: Accepted with modifications
In Section 5.7.10, change "has" to "may have":
If the variant part does not have a discriminant (tag field), the variant
part entry may have a DW_AT_type attribute to represent the tag type.
