Re: [PATCH] fetch-pack: fix object_id of exact sha1

Gabriel Souza Franco [off-list ref] writes:

quoted

quoted

      struct object_id oid;

-     if (!get_oid_hex(name, &oid) && name[GIT_SHA1_HEXSZ] == ' ')
-             name += GIT_SHA1_HEXSZ + 1;
-     else
+     if (get_oid_hex(name, &oid))
              oidclr(&oid);
+     else if (name[GIT_SHA1_HEXSZ] == ' ')
+             name += GIT_SHA1_HEXSZ + 1;

This makes sense to me. I wonder if we should be more particular about
the pure-sha1 case consuming the whole string, though. E.g., if we get:

  1234567890123456789012345678901234567890-bananas

that should probably not have sha1 1234...

I don't think it should ever really happen in practice, but it might be
worth noticing and complaining when name[GIT_SHA1_HEXSZ] is neither
space nor '\0'.

Right. What kind of complaining? Is doing oidclr(&oid) and letting it
fail elsewhere enough?

I think that is the most sensible.  After all, the first get-oid-hex
expects to find a valid 40-hex object name at the beginning of line,
and oidclr() is the way for it to signal a broken input, which is
how the callers of this codepath expects errors to be caught.

Thanks.