Hi Junio,

On Tue, 19 Feb 2019, Junio C Hamano wrote:

Johannes Schindelin [off-list ref] writes:

quoted

quoted

quoted

diff --git a/builtin/stash--helper.c b/builtin/stash--helper.c
index c77f62c895..3dab488bd6 100644
--- a/builtin/stash--helper.c
+++ b/builtin/stash--helper.c

@@ -231,6 +231,7 @@ static int reset_tree(struct object_id *i_tree, int update, int reset)
 	struct tree *tree;
 	struct lock_file lock_file = LOCK_INIT;
 
+	discard_cache();
 	read_cache_preload(NULL);
 	if (refresh_cache(REFRESH_QUIET))
 		return -1;

So this is working, but it is not the correct spot for that
`discard_cache()`, as it forces unnecessary cycles on code paths calling
`reset_tree()` (which corresponds to `git read-tree`, admittedly a bit
confusing) with a fully up to date index.

The real fix, I believe, is this:

-- snip --

diff --git a/builtin/stash.c b/builtin/stash.c
index 2d6dfce883..516dee0fa4 100644
--- a/builtin/stash.c
+++ b/builtin/stash.c

@@ -1372,6 +1372,7 @@ static int do_push_stash(struct pathspec ps, const char *stash_msg, int quiet,
 			}
 		} else {
 			struct child_process cp = CHILD_PROCESS_INIT;
+			discard_cache();
 			cp.git_cmd = 1;
 			argv_array_pushl(&cp.args, "reset", "--hard", "-q",
 					 NULL);
-- snap --

And the reason this is needed: we spawn a `git reset --hard` here, which
will change the index, but outside of the current process. So the
in-process copy is stale. And when the index' mtime does not help us
detect that, we run into that test breakage.

In non-patch mode with pathspec, there is an invocation of "apply
--index -R" of a patch that takes the contents of the HEAD to what
is in the index, updating the on-disk index and making our in-core
copy stale.  Wouldn't we need to do the same?  Otherwise, the same
"reset_tree()" you are tryhing to protect with this discard_cache()
will call read_cache_preload(), no?

Among the calls to reset_tree() in this file, I think the one that
follows the "reset --hard" (your fix above) and "apply --index -R"
(the other side of the same if/else) is the only one that wants to
read from the result of an external command we just spawned from the
on-disk index, so perhaps moving discard_cache() to just before that
call may be a better fix.

Good catch!
Dscho