On 2020-01-20 5:57 PM, Christian Weisgerber wrote:
Brian Callahan:

Attached is an update to gmake.
I worked on this independently and comparing the results, I have
some objections:

* We can use the .tar.lz distfile.

* MODGNU_CONFIG_GUESS_DIRS should not be removed but set to the
   correct ${WRKSRC}/build-aux.

* Instead of patching tests/scripts/features/exec we can just set
   TEST_ENV= SHELL=$$SHELL
   and honor the intention of the test.

That's a better solution that I thought of for sure.

* "You should use this name if you have a makefile that is specific
    to GNU gmake, and will not be understood by other versions of
    gmake."
   Uhm, no.  You cannot blindly substitute all instances of "make"
   with "gmake" in the man page.

* You dropped the comment at the top of patch-src_makeint_h.

These two were just oversights; thanks for catching.

One test fails on arm64 (but the same test fails with 4.2.1 too).
A great opportunity to look into this failure!

Agreed. For completeness, here's the relevant part of the test log:
features/output-sync .................................... Error running /usr/ports/pobj/gmake-4.3/build-aarch64/tests/../make (expected 0; got 512): /usr/ports/pobj/gmake-4.3/build-aarch64/tests/../make -f work/features/output-sync.mk -j -Orecurse Error running /usr/ports/pobj/gmake-4.3/build-aarch64/tests/../make (expected 0; got 512): /usr/ports/pobj/gmake-4.3/build-aarch64/tests/../make -f work/features/output-sync.mk.1 -j --output-sync=target
FAILED (13/15 passed)

Digging through the logs, it appears that the two tests are timing out. This also happened in 4.2.1, which makes me wonder if it's something specific to my arm64 machine. Anyhow, I can look into it but not for the next few days.

Obviously needs more testing than I can provide, but here for the
interested.
I'll run an amd64 bulk build with it sometime in the next few days.
The "WARNING: Backward-incompatibility!" changes could break some
cruft.


Yes, this warning coupled with what sthen mentioned about how it will become the default behavior in the future is what spurred my interest to update.

~Brian

Reply via email to