lmi
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lmi] PATCH: Fix CI after C++20 and fs::path changes


From: Greg Chicares
Subject: Re: [lmi] PATCH: Fix CI after C++20 and fs::path changes
Date: Wed, 12 May 2021 22:40:25 +0000
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.9.0

On 5/12/21 7:15 PM, Vadim Zeitlin wrote:
> 
>  My branch with the changes, which should be accessible as xanadu/ci-cxx20

...has been pushed. Thanks.

I added only one commit of my own, which the preprocessor will
remove in some early phase so that it can have no observable
effect on behavior: in theory at least; and in my own tests,
demonstrably.

>  There are many commits there, that I'm not going to describe to avoid
> scaring you

Thanks. I closed my eyes and cherry-picked everything. The
commit messages are many, and scary, in retrospect; but now
I see that they were necessary as well as complicated.
Some comments:

  commit aa750df5dbcc48dfc52bf6d5f48a2dafc11d4a26
  Work around global_settings_test failure in autotools builds

I think that's unnecessary since
  6303f4a2 Make a unit test more robust
and can be reverted at your leisure.

  commit d09dfcef442a5d873b56449743c58d9135e2426f
  Merge lmi directories creation step with the system setup one
  ...  
  This commit is best viewed with --color-moved git option.

I have long since made that my default:
  git config --global diff.colormoved plain
so you needn't write that comment just for me, though you may
want to for the rest of the world. BTW, I've also found this
setting very handy:
  git config --global commit.verbose true
and I've never had any reason to override either of those settings.

  commit 2213a23500ceb2ccb0c4affde5c57da7f93835c4
  Disable clang deprecated declarations warnings in uBLAS headers
and
  commit 21706da8ae3f9a8e88a559b185e360377d5030b8
  Fix expression_template_0_test compilation with clang 12

I'm really glad you were able to find workarounds to make
the CI tests pass now, so that we can eradicate leftover
vestiges of boost carefully, thoughtfully, and later.

>  As I had previously said, I feel almost physically uncomfortable with the
> changes being pushed while the CI builds are failing
Yes, all tests must always succeed, or lmi is broken and
nothing should be pushed.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]