[PATCH] bfd/dwarf2: break equal-range function ties by DIE offset, not pointer

[Jan Beulich]

On 08.09.2026 16:39, Sam Price wrote:
> I could do one of the following.
> - Delete that test.  It is adding code coverage for the single fall
> through line. I don't think coverage is a metric for addr2line.
> - Flag the test to not run for those platforms. ia64-linux lost linux
> support in 2023 I think.
> - I don't know how to test those platforms at the moment, I could dig
> into this more but it might be a while.
> - Move the assembly to C in the existing addr2line test.  Use inline
> functions inside of a function.  I tested this and it hits those lines
> (with temporary debug prints).
> I probably would lean towards just deleting the test, and reducing the
> wordy comments, and resubmitting the "trivial" patch as a V2.

If making the test work on specific targets is difficult, just exclude
those targets (ideally leaving a comment as to why). We want a test in
the common case at least.

As to testing for such targets (taking ia64/Linux as example), you'd
simply pass --target=ia64-linux to ./configure, then build and check
normally. If it wasn't that simple, I likely wouldn't have noticed the
issue.

Jan