Re: Bug in git archive + .gitattributes + relative path

[Junio C Hamano <gitster@pobox.com>]

René Scharfe <l.s.r@web.de> writes:

>> Another way I am not sure is working as designed is
>>
>>     $ cd sha1dc && git archive HEAD . | tar tf -
>>     .gitattributes
>>     LICENSE.txt
>>     sha1.c
>>     sha1.h
>>     ubc_check.c
>>     ubc_check.hq
>>
>> I didn't check if the attribute look-up is done on the correct path
>> or export-subst kicks in in such a use, though.
>
> export-subst is supported in that invocation because git archive has a
> commit to work with.
>
> I can kinda see others preferring the directory prefix "sha1dc/" added
> to those entries.  Perhaps it depends on what git archive is supposed to
> archive: A commit or the files of a commit?  I'm in the latter camp, and
> expect to see the same paths as given by git ls-files or git ls-tree.
>
> But that invocation in a sub-directory probably has the same problem
> with attributes as the one with a sub-tree above it, i.e. that
> attributes are always looked up relative to the repository root.  I
> wonder if 47cfc9bd7d (attr: add flag `--source` to work with tree-ish,
> 2023-01-14) provided the means to fix this when it added a tree_oid
> parameter to git_check_attr().

It somehow feels that the use of pathspec in "git archive" is
somewhat iffy, e.g.

   $ cd sha1dc && git archive HEAD :/ | tar tf -

does not compare very well with

   $ cd sha1dc && git ls-tree -r HEAD :/

For that matter, replacing ":/" (full tree) with ".." (we know one
level above is the root of the working tree) has the same "why don't
they work the same way???" confusion.