[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [lmi] Common test objects are too common
From: |
Vadim Zeitlin |
Subject: |
Re: [lmi] Common test objects are too common |
Date: |
Sun, 21 Mar 2021 15:50:06 +0100 |
On Sun, 21 Mar 2021 14:28:29 +0000 Greg Chicares <gchicares@sbcglobal.net>
wrote:
GC> > [...] I would still like to submit a patch with some changes required
GC> > to fix the problem with autotools-based makefiles
GC> Sure.
This patch, which has passed the CI checks by now, is at
https://github.com/let-me-illustrate/lmi/pull/174
As I wrote there, the only change requiring your review there is this
commit:
https://github.com/let-me-illustrate/lmi/pull/174/commits/d9204c7247429bb03bafb89f67abfa96d628b3a6
which adds "force linking" macros to ensure that the tests not only build
but also work correctly with autotools after the changes in Makefile.am. I
hope this is acceptable, and it definitely doesn't change anything for the
normal build, but if not, as I wrote in the PR description linked above, we
could also consider other alternatives, but I'm afraid I do need to make
some change to the sources to keep things reasonably simple, i.e. without
having to use a shared library containing these objects files rather than a
simple static one.
Please let me know if you have any questions, objections, or suggestions.
GC> Tangentially, I'd like to propose a naming change.
This is exactly what
https://github.com/let-me-illustrate/lmi/pull/174/commits/8907ef8c5343e03f427c68b64345e742c80a383d
included in this PR does. The commit says that it dates from only a few
hours ago, but actually this had been sitting in a branch here for a few
months already and I've finally finished the changes to make this work
today.
The net result is that "make check" succeeds now and I'd like to add a
separate CI job checking that it keeps succeeding, if you don't mind.
Thanks in advance for looking at this!
VZ
pgptY2I2Neou1.pgp
Description: PGP signature