Re: dwarflint versus linkage_name attributes

[Tom Tromey <tromey at redhat dot com>]

Tom> When constructing the anonymous type, gdb uses the linkage name to give
Tom> the type the name `t'.  What is specifically useful to gdb about this
Tom> approach is that it can be decided locally whether this is necessary.
Tom> Previous iterations required a lot of complicated code to defer giving
Tom> names to types until all the DIEs had been read in.

Mark> OK, so just to be clear this isn't actually used as a linkage_name
Mark> by GDB then?

Yeah, not really.
It is used to give the type a name.

Mark> And the name `t' is kind of arbitrary? The "rule" for the
Mark> extension is that there should be an identifier in DW_AT_linkage_name
Mark> for an anonymous structure_type (which doesn't have DW_AT_name set)?

No, the rule is that for the specific C++ construct in question, the
anonymous type should have a linkage name attribute indicating the name
given to the type for linkage purposes only.

Dodji said this is [dcl.typedef]/8 in his copy of the standard.
I didn't try to look it up.

Mark> I am just wondering whether you have thought about submitting it as an
Mark> official DWARF extension.

We probably should.

Tom