Johannes Schindelin <Johannes.Schindelin@xxxxxx> writes:
>> For one thing, I want to move read_cache() out of run_diff_files() and
>> run_diff_index() to their callers, so that the caller can use a
>> handcrafted active_cache[] that is somewhat different from what
>> read_cache() originally read.
>
> I don't understand. Would it not be easier to "return active_nr" in
> read_cache_from(path) _also_ when active_cache is set? Or alternatively
> check if active_alloc is set?
I do not understand your questions.
What I am getting at is that in the current code (without your
patch), run_diff_files() is about diff between $GIT_DIR/index
and working tree. I want to make it diff between active_cache[]
and working tree.
The caller still can use read_cache() or read_cache_from(path)
to populate active_cache[], but it is not limited to. It can do
other index manipulations before calling run_diff_files(). For
example, it can do read_tree() to populate active_cache[] before
calling run_diff_files() and/or run_diff_index(). Or maybe it
can run multi-tree read_tree() to populate active_cache[],
internally merge them without writing the merge results out, and
then call run_diff_files() to show "here is what would happen if
you were to run a merge now", all inside active_cache[].
Of course such a caller needs to be careful not to write
active_cache[] out when it shouldn't, but that goes without
saying. git-diff-index already mucks with active_cache[] with
full intention of not writing it out.
Please see near the tip of 'pu', around here, for much simpler
usage pattern I had in mind:
commit 910d798768989d8fe581c2afcfd501307db81b10
Author: Junio C Hamano <junkio@xxxxxxx>
git-status: use in-core --refresh in a read-only repository.
commit c5cc9b14a0f36d7fab66b3152f0ebcb442674bfc
Author: Junio C Hamano <junkio@xxxxxxx>
git-runstatus --refresh
commit b4e1e4a787d3771f617182b3344dcdd9224bd0cb
Author: Junio C Hamano <junkio@xxxxxxx>
run_diff_{files,index}(): update calling convention.
-
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html