On Fri, Jan 03, 2025 at 05:48:12PM +0000, Matthew Malcomson wrote: > On 1/3/25 17:14, Joseph Myers wrote: > > Does this patch cover everything dealt with by > > <https://gcc.gnu.org/pipermail/gcc-patches/2024-December/672242.html> > > ([PATCH] testsuite: libitm: Remove build directory path from test names), > > or would some separate fix for that issue still be needed in the presence > > of this patch?
> Ah -- I didn't notice that patch. > > It looks like both do essentially the same thing. > > That one identifies the use of the c++ test runner by checking for the > presence of the `lang_test_file` variable, and in that case > `libitm_target_compile` adds the options, while in my patch the c++ test > runner adds options to `ALWAYS_CFLAGS` which `libitm_target_compile` uses > directly. Agreed, they both have the same effect. Mine just added -B symmetrically in the same way that -L is already added. I guess that also underscores the question of whether -B is even needed? I am not aware of any reason why it should be needed when -L is already specified with the same directory, but would be curious to know what I may be missing. In any case, seems we may as well go with Matthew's latest version if it is acked, since it keeps the changes isolated just to c++.exp and doesn't need to touch libitm.exp? -Lewis
