On 28. 02. 23 21:40, Jerry James wrote:
On Tue, Feb 28, 2023 at 11:37 AM Miro Hrončok
<mhroncok(a)redhat.com> wrote:
> On one hand, I think the script should consider commits like this:
>
> """
> Modernize packaging
>
> Remove deprecated macros
> Verify license is SPDX
> Drop Fedora 28 conditionals
> """
>
> On the other hand, I consider putting "Verified that License tag is valid
> SPDX." into a commit with "Use the %gap_arches macro." in the subject
chaotic evil.
I don't see how the example above is qualitatively different from the
example I posted. Mine was also changing an obsolete usage to a
modern one and verifying the license is SPDX. Why do you think the
two are different?
The example I posted is:
"""
Subject line describing a general thing/goal
A specific thing, part of the same specified goal
A specific thing, part of the same specified goal
A specific thing, part of the same specified goal
"""
The other example is:
"""
Subject line describing a specific thing
Another specific thing, possibly part of the same unspecified goal
"""
Generally, I don't like mixing unrelated changes together in one commit, but I
understand that sometimes a single "cleanup" commit gets the job done better
than 12 individual 1-line-diff commits.
When a (pseudo) diff of a commit is:
-foo
+bar
...
-Whatever.(trailing space)
+Whatever.(no trailing space)
And the commit message is:
"""
Change foo to bar for reasons
"""
...I cringe.
If the message is:
"""
Change foo to bar for reasons
While at it, remove the trailing space elsewhere because my editor deleted it.
"""
...I cringe but I am thankful for at least documenting the unrelated change.
However, in both cases at least I notice the unrelated change in the diff. In
my worldview, both are chaotic (read: they violate my imaginary commit-purity
O.C.D. rules), but they are not evil (nothing is "hidden"). In the MIT SPDX
license case, the "mixed in" change is a no-change, so when looking at the diff
we don't notice it, hence I called it "evil".
Hope that makes sense. No judgement intended, I know people have different
expectations and habits when it comes to commits (and dist-git commits in
particular).
--
Miro Hrončok
--
Phone: +420777974800
IRC: mhroncok