ref-filter: reuse output buffer

* Firstly, inlining show_ref_array_item().
* Secondly, git for-each-ref reuse final buf for all refs output, the performance is slightly improved, this optimization is also applied to git tag -l and git branch -l.

Changes made in v2: reset err buffer in loop for code cleanness.

Thanks.

ZheNing Hu (2):
[GSOC] ref-filter: get rid of show_ref_array_item [GSOC] ref-filter: reuse output buffer

 builtin/branch.c       | 10 ++++++----
 builtin/for-each-ref.c | 15 +++++++++++++--
 builtin/tag.c          | 15 +++++++++++++--
 ref-filter.c           | 25 ++++++++++---------------
 ref-filter.h           |  2 --
 5 files changed, 42 insertions(+), 25 deletions(-)

base-commit: 2e36527f23b7f6ae15e6f21ac3b08bf3fed6ee48
Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-935%2Fadlternative%2Fref-filter-opt-reuse-buf-v2
Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-935/adlternative/ref-filter-opt-reuse-buf-v2
Pull-Request: https://github.com/gitgitgadget/git/pull/935

Range-diff vs v1:

1: ec98f2177d90 ! 1: 7e378eef4b34 [GSOC] ref-filter: get rid of show_ref_array_item
@@ Metadata
## Commit message ##
[GSOC] ref-filter: get rid of show_ref_array_item

- Inlining the exported function show_ref_array_item()
+ Inlining the exported function show_ref_array_item(),
which is not providing the right level of abstraction, simplifies the API and can unlock improvements at the former call sites.
2: 1c7a69ba072a ! 2: a17949b9f809 [GSOC] ref-filter: reuse output buffer
@@ Commit message
[GSOC] ref-filter: reuse output buffer

When we use git for-each-ref, every ref will allocate
- its own output strbuf. But we can reuse the final strbuf
- for each step ref's output.
+ its own output strbuf and error strbuf. But we can reuse
+ the final strbuf for each step ref's output. The error
+ buffer will also be reused, despite the fact that the git
+ will exit when format_ref_array_item() return a non-zero
+ value and output the contents of the error buffer.

The performance for git for-each-ref on the Git repository itself with performance testing tool hyperfine changes from
@@ Commit message
(cat-file: use a single strbuf for all output, 2018-08-14) to speed up the cat-file builtin.

+ Helped-by: Junio C Hamano <[email protected]>

@@ builtin/branch.c: static void print_ref_list(struct ref_filter *filter, struct r
for (i = 0; i < array.nr; i++) {
- struct strbuf out = STRBUF_INIT;
- struct strbuf err = STRBUF_INIT;
++ strbuf_reset(&err);
+ strbuf_reset(&out);
if (format_ref_array_item(array.items[i], format, &out, &err))
die("%s", err.buf);
@@ builtin/for-each-ref.c: int cmd_for_each_ref(int argc, const char **argv, const
- struct strbuf output = STRBUF_INIT;
- struct strbuf err = STRBUF_INIT;
-
++ strbuf_reset(&err);
+ strbuf_reset(&output);
if (format_ref_array_item(array.items[i], &format, &output, &err))
die("%s", err.buf);
@@ builtin/tag.c: static int list_tags(struct ref_filter *filter, struct ref_sortin
- struct strbuf err = STRBUF_INIT;
-
+ strbuf_reset(&output);
++ strbuf_reset(&err);
if (format_ref_array_item(array.items[i], format, &output, &err))
die("%s", err.buf);
fwrite(output.buf, 1, output.len, stdout);

gitgitgadget