https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109098
--- Comment #9 from David Malcolm <dmalcolm at gcc dot gnu.org> --- (In reply to Hans-Peter Nilsson from comment #8) > (In reply to David Malcolm from comment #7) > > The invalid UTF-8 in the patch seems to have broken the server-side script: > > Maybe the not-really-utf8 files need to be marked in some way in the git > repo to be safely handled for future checkout and updates, including the > problematic scripting? However, reading gitattributes(5) it's not obvious > how. Perhaps https://www.git-scm.com/docs/gitattributes#_marking_files_as_binary though it's not clear if we can do that for individual files (or if it's worth bothering)