Created attachment 252390 [details] v0 (apply via `git am`) cc181346de3f0e0477ef3e9a4dd6f002db5892c3 introduced a change with comment "Only devel/llvm17 _and later_ come with necessary include files", but actual USES= doesn't allow later versions. Attached patch fixes this.
Yes, I wanted to keep things a bit simpler amidst all the changes, but can (and will) make this change if it is desirable to you. (I'll just skip the reformatting of the entire USES= clause and only edit that one line.)
(In reply to Gerald Pfeifer from comment #1) > Yes, I wanted to keep things a bit simpler amidst all the changes, > but can (and will) make this change if it is desirable to you. Please don't do this for me, but rather for consistency between comment and actual code (and with the true requirement which is just minimum, not maximum version). > (I'll just skip the reformatting of the entire USES= clause and only > edit that one line.) I guess up to you, but would like to note also that it was not my whim. It's just what portfmt did after I executed that to line changed.
A commit in branch main references this bug: URL: https://cgit.FreeBSD.org/ports/commit/?id=e20e0ec29d4eb05b145d80675a61950e7c2865da commit e20e0ec29d4eb05b145d80675a61950e7c2865da Author: Anton Saietskii <vsasjason@gmail.com> AuthorDate: 2024-08-01 09:45:51 +0000 Commit: Gerald Pfeifer <gerald@FreeBSD.org> CommitDate: 2024-08-01 11:49:42 +0000 emulators/wine-devel: Use LLVM 17 as minimum, not fixed version PR: 280529 LLVM 17 and later provide the necessary include files; see 5f69b17e49 for the original change. emulators/wine-devel/Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
(In reply to Anton Saietskii from comment #2) > Please don't do this for me, but rather for consistency between comment > and actual code (and with the true requirement which is just minimum, > not maximum version). The code was more specific than the comment, choosing one of the possible options. And while I had planned to make the code less prescriptive after a bit of settling period, user feedback, in particular user feedback with a patch even, matters, so I accelerated this change. >> (I'll just skip the reformatting of the entire USES= clause and only >> edit that one line.) > I guess up to you, but would like to note also that it was not my whim. > It's just what portfmt did after I executed that to line changed. Understood. I had done the line break intentionally with the original change, and in the committed patch the extent of the change is easier to see. Thanks for raising this topic and even suggesting a patch, Anton!