On Wed, May 01, 2024 at 12:52:38AM +0900, Jun. T wrote: > > > 2024/04/30 16:12, Torsten Bögershausen <tboegi@xxxxxx> wrote: > > > > I am familar with the NFC/NFD stuff, but not with get_git_work_tree(), > > at least not yet. > > > > If you have a suggestion for a patch, would you like to share it ? > > Well, the only thing I can tell is the patch below _seems_ to fix > the _current_ problem. There may be other problems, it may introduce > new problem(s) (such as memory leak), etc. > > But I don't know anything about the git internals and have no time > now to investigate further. Sorry. > > get_git_work_tree() just returns the value of the_repository->worktree > but I have no idea where this variable is set. > > get_git_work_tree() and the_repository->worktree are used in > many places, and I'm not sure changing the function/variable has > no bad side effects or not. > > And how to convert into NFC? By precompose_string_if_needed()? > Does this function allocate a new memory for NFC? If so, do we > need to free it at some point? > > > diff --git a/setup.c b/setup.c > index f4b32f76e3..3f2f3ed016 100644 > --- a/setup.c > +++ b/setup.c > @@ -47,7 +47,7 @@ static int abspath_part_inside_repo(char *path) > size_t wtlen; > char *path0; > int off; > - const char *work_tree = get_git_work_tree(); > + const char *work_tree = precompose_string_if_needed(get_git_work_tree()); > struct strbuf realpath = STRBUF_INIT; > > if (!work_tree) > Thanks for digging - I have spend some time to find the cause, but no success yet. There is even a set_git_work_tree() in environment.c, and that may need some treatment - or some other place. To be continued - I will continue digging, where the NFD comes into Git.
