Oct 21, 2025 6:00:49 PM Joseph Myers <[email protected]>: > On Tue, 21 Oct 2025, Peter Damianov wrote: > >> +/* This comment has CRLF line endings >> + and should appear as two lines in the output, not three. */ > > Since email does not preserve the presence of CRLF line endings, I'm not > sure (a) what form of line endings you intend to be present in the blob > object for the file in the repository (as opposed to in the checked out > file in a working tree), and (b) what the intended process is for applying > the patch to result in both correct line endings in the blob pointed to by > the tree pointed to by the commit, and in the working copy where the patch > was applied.
I'm not certain what the process should be. I figured telling git to use crlf in the gitattributes was enough. The intent is that the test file is always saved with crlf. > > Since the function being changed has a separate slow path in the presence > of trigraphs, I think it would be a good idea also to add a test that CRLF > line endings work correctly when the slow path is used. OK. I will test that one too. > > -- > Joseph S. Myers > [email protected]
