Tom Tromey wrote:
This series merges in gdb's changes to default_target_compile,
updating them to follow newer conventions at the same time.
I tested this using various parts of the gdb test suite (in particular
Go and Rust). I also verified that, with this series in place, gdb
does not use its monkey-patched code from future.exp.
Thank you; this will be merged for 1.6.3 to ensure compatibility with
existing GDB releases. I should be able to get these applied and pushed
onto a side branch over the weekend, with some additional documentation
and testcases to describe and verify the new behavior.
I didn't attempt to further parameterize default_target_compile,
though I think that would be a good thing to do.
That is planned for rewriting that procedure for 1.6.4 to fix some
remaining bugs that were found by inspection.
Lastly, I note that you are at AdaCore, so I will ask if you could
examine the patch Maciej W. Rozycki submitted
(<URL:https://lists.gnu.org/archive/html/dejagnu/2020-05/msg00019.html>)
to wrap linker flags for gnatmake into -largs / -margs options. The
patch seems good, but I am unsure if it there might be more to the
problem than Rozycki saw; he was just trying to get a RISC-V Ada
toolchain to work. Are the other -?args options needed to prevent
future problems, or is only the linker ever likely to be affected?
-- Jacob
