(re-sending to list, this time using the correct e-mail account so the message doesn’t bounce)
I probably squash-merged the PRs and indeed I felt/feel that all the additional text in the commit message is superfluous. Renee > On Oct 23, 2025, at 4:32 PM, Ryan Carsten Schmidt <[email protected]> > wrote: > > On Oct 23, 2025, at 14:40, Fred Wright <[email protected] > <mailto:[email protected]>> wrote: >> >> >>> On Thu, 23 Oct 2025, Ryan Carsten Schmidt wrote: >>> >>>> On Oct 23, 2025, at 14:08, Fred Wright wrote: >>>> >>>> I noticed that my two most recent commits have only the summary line of >>>> their commit messages as merged. When I saw the first one, I thought I >>>> might have screwed up an edit, but I didn't think that doing it twice was >>>> likely. Looking at the current master, the last commit message with a >>>> body was e19d3b9ff8167b60cf0698e611824444c4b66398 from yesterday. Is some >>>> script now dropping commit-message bodies? >>> >>> No automated script that I'm aware of. >>> >>> When a PR is merged, the person doing the merging can choose to accept the >>> commits as provided or squash all commits into one and reword the commit >>> message; I suspect this is what happened to your recent PRs. >> >> No - each PR consisted of a single commit. > > Nevertheless, the squash and merge feature can be used with any PR, even > those consisting of a single commit, and apparently was used in these cases; > there's no other way for what you observe to have happened. > > >>> When doing this, we want to eliminate superfluous wording that was only >>> relevant while a PR was being developed, such as a series of commits that >>> correct problems with previous commits, while retaining useful information >>> they describes the PR as a whole. >> >> I never submit PRs in that state, but my commit messages pretty much always >> contain additional information, including how the change was tested. > > I understand. Maybe the person who merged the PR didn't feel it was necessary > to have that level of detail in the commit message history. > > >> It also seems unlikely that *nobody* submitted commits in the past 24 hours >> that contained nothing but the summary. > > Indeed most commits in the repository don't have any additional information > in the commit message.
