Re: [PATCH 1/2] ref-filter: hacky "streaming" mode

From: Jeff King <peff@peff.net>
Date: Sep 5, 2021, 13:04 UTC
Message-ID: <YTTARcEvpXWSDfYW@coredump.intra.peff.net>
In-Reply-To: <CAOLTT8Tka0nxHb3G9yb-fs8ue7RaPCUVSKi5PM+GY+rMjFRnog@mail.gmail.com>

On Sun, Sep 05, 2021 at 04:20:02PM +0800, ZheNing Hu wrote:

Show 6 quoted lines

> But here is a terrible fact: we did not use ref_array_sort() for sorting here.
> So in fact, for_each_fullref_in() does the sorting work () for us by
> default sort (%(refname)),
> This may be due to the file system or some implementation of ref_iterator.
> But this limit the application of this optimization when we use other
> atoms to sort.

Right. This technique does not help at all if you use --sort. I do think it's reasonable to rely on the sorted output of the for_each_ref() functions; other parts of Git likely do as well, so I think we'd try pretty hard to maintain that (and no, it's not a filesystem thing; we do make sure to sort it ourselves; see the calls to sort_ref_dir()).

Show 16 quoted lines

> > +       /*
> > +        * the %(symref) placeholder is broken with pretty_print_ref(),
> > +        * which our streaming code uses. I suspect this is a sign of breakage
> > +        * in other callers like verify_tag(), which should be fixed. But for
> > +        * now just disable streaming.
> > +        *
> > +        * Note that this implies we've parsed the format already with
> > +        * verify_ref_format().
> > +        */
> > +       if (need_symref)
> > +               return;
> > +
> 
> I haven’t taken a closer look at why pretty_print_ref() does not
> support %(symref),
> we can skip it first.

I think it's just because pretty_print_ref() does not take a "flag" parameter for the caller. So it never sees that REF_ISSYMREF is set.

There aren't many callers of that function, so I think nobody ever really noticed. But you can see the bug like this:

  git init repo
  cd repo
  
  git commit --allow-empty -m foo
  git tag -s -m bar annotated &&
  git symbolic-ref refs/tags/symref refs/tags/annotated
  
  # this is ok; ref-filter handles the flags
  git tag --list --format='list: %(refname) %(symref)'
  
  # not ok; we do not tell the formatter about the flags, so it does
  # not notice that "symref" is a symref
  git tag --verify --format='verify: %(refname) %(symref)' annotated symref

I notice that the --verify output also shows the short refname, which makes me wonder if %(symref) would have other bugs there (because it has to re-resolve the ref to come up with the symref destination).

> Unfortunately, this optimization may not be helpful for git cat-file --batch,
> see [1], batch_object_write() directly constructs a ref_array_item and call
> format_ref_array_item() to grab data, It does not use ref_array. So it also
> does not have this malloc() | free() overhead.

Right. It would only be helped by the "quick" formats in the second patch (and by skipping the malloc/free of the individual ref_array_items).

-Peff