@matheusbsbs / matheus-bagliano / commits / 455923e0a1

ls-tree: introduce "--format" option

Add a --format option to ls-tree. It has an existing default output, and then --long and --name-only options to emit the default output along with the objectsize and, or to only emit object paths. Rather than add --type-only, --object-only etc. we can just support a --format using a strbuf_expand() similar to "for-each-ref --format". We might still add such options in the future for convenience. The --format implementation is slower than the existing code, but this change does not cause any performance regressions. We'll leave the existing show_tree() unchanged, and only run show_tree_fmt() in if a --format different than the hardcoded built-in ones corresponding to the existing modes is provided. I.e. something like the "--long" output would be much slower with this, mainly due to how we need to allocate various things to do with quote.c instead of spewing the output directly to stdout. The new option of '--format' comes from Ævar Arnfjörð Bjarmasonn's idea and suggestion, this commit makes modifications in terms of the original discussion on community [1]. In [1] there was a "GIT_TEST_LS_TREE_FORMAT_BACKEND" variable to ensure that we had test coverage for passing tests that would otherwise use show_tree() through show_tree_fmt(), and thus that the formatting mechanism could handle all the same cases as the non-formatting options. Somewhere in subsequent re-rolls of that we seem to have drifted away from what the goal of these tests should be. We're trying to ensure correctness of show_tree_fmt(). We can't tell if we "hit [the] fast-path" here, and instead of having an explicit test for that, we can just add it to something our "test_ls_tree_format" tests for. Here is the statistics about performance tests: 1. Default format (hitten the builtin formats): "git ls-tree <tree-ish>" vs "--format='%(mode) %(type) %(object)%x09%(file)'" $hyperfine --warmup=10 "/opt/git/master/bin/git ls-tree -r HEAD" Benchmark 1: /opt/git/master/bin/git ls-tree -r HEAD Time (mean ± σ): 105.2 ms ± 3.3 ms [User: 84.3 ms, System: 20.8 ms] Range (min … max): 99.2 ms … 113.2 ms 28 runs $hyperfine --warmup=10 "/opt/git/ls-tree-oid-only/bin/git ls-tree -r --format='%(mode) %(type) %(object)%x09%(file)' HEAD" Benchmark 1: /opt/git/ls-tree-oid-only/bin/git ls-tree -r --format='%(mode) %(type) %(object)%x09%(file)' HEAD Time (mean ± σ): 106.4 ms ± 2.7 ms [User: 86.1 ms, System: 20.2 ms] Range (min … max): 100.2 ms … 110.5 ms 29 runs 2. Default format includes object size (hitten the builtin formats): "git ls-tree -l <tree-ish>" vs "--format='%(mode) %(type) %(object) %(size:padded)%x09%(file)'" $hyperfine --warmup=10 "/opt/git/master/bin/git ls-tree -r -l HEAD" Benchmark 1: /opt/git/master/bin/git ls-tree -r -l HEAD Time (mean ± σ): 335.1 ms ± 6.5 ms [User: 304.6 ms, System: 30.4 ms] Range (min … max): 327.5 ms … 348.4 ms 10 runs $hyperfine --warmup=10 "/opt/git/ls-tree-oid-only/bin/git ls-tree -r --format='%(mode) %(type) %(object) %(size:padded)%x09%(file)' HEAD" Benchmark 1: /opt/git/ls-tree-oid-only/bin/git ls-tree -r --format='%(mode) %(type) %(object) %(size:padded)%x09%(file)' HEAD Time (mean ± σ): 337.2 ms ± 8.2 ms [User: 309.2 ms, System: 27.9 ms] Range (min … max): 328.8 ms … 349.4 ms 10 runs Links: [1] https://public-inbox.org/git/RFC-patch-6.7-eac299f06ff-20211217T131635Z-avarab@gmail.com/ [2] https://lore.kernel.org/git/cb717d08be87e3239117c6c667cb32caabaad33d.1646390152.git.dyroneteng@gmail.com/ Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Teng Long <dyroneteng@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Ævar Arnfjörð Bjarmason committed Mar 23, 2022 at 17:13 UTC 455923e0a152420054ad74f1af36336d5fa7be75
3 files changed +272 -7
Documentation/git-ls-tree.txt
+55 -4
@@ -10,7 +10,7 @@ SYNOPSIS
10 --------
11 [verse]
12 'git ls-tree' [-d] [-r] [-t] [-l] [-z]
13 - [--name-only] [--name-status] [--full-name] [--full-tree] [--abbrev[=<n>]]
13 + [--name-only] [--name-status] [--full-name] [--full-tree] [--abbrev[=<n>]] [--format=<format>]
14 <tree-ish> [<path>...]
15
16 DESCRIPTION
@@ -74,6 +74,16 @@ OPTIONS
74 Do not limit the listing to the current working directory.
75 Implies --full-name.
76
77 +--format=<format>::
78 + A string that interpolates `%(fieldname)` from the result
79 + being shown. It also interpolates `%%` to `%`, and
80 + `%xx` where `xx` are hex digits interpolates to character
81 + with hex code `xx`; for example `%00` interpolates to
82 + `\0` (NUL), `%09` to `\t` (TAB) and `%0a` to `\n` (LF).
83 + When specified, `--format` cannot be combined with other
84 + format-altering options, including `--long`, `--name-only`
85 + and `--object-only`.
86 +
87 [<path>...]::
88 When paths are given, show them (note that this isn't really raw
89 pathnames, but rather a list of patterns to match). Otherwise
@@ -82,16 +92,29 @@ OPTIONS
92
93 Output Format
94 -------------
85 - <mode> SP <type> SP <object> TAB <file>
95 +
96 +The output format of `ls-tree` is determined by either the `--format`
97 +option, or other format-altering options such as `--name-only` etc.
98 +(see `--format` above).
99 +
100 +The use of certain `--format` directives is equivalent to using those
101 +options, but invoking the full formatting machinery can be slower than
102 +using an appropriate formatting option.
103 +
104 +In cases where the `--format` would exactly map to an existing option
105 +`ls-tree` will use the appropriate faster path. Thus the default format
106 +is equivalent to:
107 +
108 + %(objectmode) %(objecttype) %(objectname)%x09%(path)
109
110 This output format is compatible with what `--index-info --stdin` of
111 'git update-index' expects.
112
113 When the `-l` option is used, format changes to
114
92 - <mode> SP <type> SP <object> SP <object size> TAB <file>
115 + %(objectmode) %(objecttype) %(objectname) %(objectsize:padded)%x09%(path)
116
94 -Object size identified by <object> is given in bytes, and right-justified
117 +Object size identified by <objectname> is given in bytes, and right-justified
118 with minimum width of 7 characters. Object size is given only for blobs
119 (file) entries; for other entries `-` character is used in place of size.
120
@@ -100,6 +123,34 @@ quoted as explained for the configuration variable `core.quotePath`
123 (see linkgit:git-config[1]). Using `-z` the filename is output
124 verbatim and the line is terminated by a NUL byte.
125
126 +Customized format:
127 +
128 +It is possible to print in a custom format by using the `--format` option,
129 +which is able to interpolate different fields using a `%(fieldname)` notation.
130 +For example, if you only care about the "objectname" and "path" fields, you
131 +can execute with a specific "--format" like
132 +
133 + git ls-tree --format='%(objectname) %(path)' <tree-ish>
134 +
135 +FIELD NAMES
136 +-----------
137 +
138 +Various values from structured fields can be used to interpolate
139 +into the resulting output. For each outputing line, the following
140 +names can be used:
141 +
142 +objectmode::
143 + The mode of the object.
144 +objecttype::
145 + The type of the object (`blob` or `tree`).
146 +objectname::
147 + The name of the object.
148 +objectsize[:padded]::
149 + The size of the object ("-" if it's a tree).
150 + It also supports a padded format of size with "%(size:padded)".
151 +path::
152 + The pathname of the object.
153 +
154 GIT
155 ---
156 Part of the linkgit:git[1] suite
builtin/ls-tree.c
+150 -3
@@ -23,6 +23,8 @@ static int ls_options;
23 static struct pathspec pathspec;
24 static int chomp_prefix;
25 static const char *ls_tree_prefix;
26 +static const char *format;
27 +
28 struct show_tree_data {
29 unsigned mode;
30 enum object_type type;
@@ -37,10 +39,77 @@ static const char * const ls_tree_usage[] = {
39 };
40
41 static enum ls_tree_cmdmode {
40 - MODE_LONG = 1,
42 + MODE_DEFAULT = 0,
43 + MODE_LONG,
44 MODE_NAME_ONLY,
45 } cmdmode;
46
47 +static void expand_objectsize(struct strbuf *line, const struct object_id *oid,
48 + const enum object_type type, unsigned int padded)
49 +{
50 + if (type == OBJ_BLOB) {
51 + unsigned long size;
52 + if (oid_object_info(the_repository, oid, &size) < 0)
53 + die(_("could not get object info about '%s'"),
54 + oid_to_hex(oid));
55 + if (padded)
56 + strbuf_addf(line, "%7"PRIuMAX, (uintmax_t)size);
57 + else
58 + strbuf_addf(line, "%"PRIuMAX, (uintmax_t)size);
59 + } else if (padded) {
60 + strbuf_addf(line, "%7s", "-");
61 + } else {
62 + strbuf_addstr(line, "-");
63 + }
64 +}
65 +
66 +static size_t expand_show_tree(struct strbuf *sb, const char *start,
67 + void *context)
68 +{
69 + struct show_tree_data *data = context;
70 + const char *end;
71 + const char *p;
72 + unsigned int errlen;
73 + size_t len = strbuf_expand_literal_cb(sb, start, NULL);
74 +
75 + if (len)
76 + return len;
77 + if (*start != '(')
78 + die(_("bad ls-tree format: element '%s' does not start with '('"), start);
79 +
80 + end = strchr(start + 1, ')');
81 + if (!end)
82 + die(_("bad ls-tree format: element '%s' does not end in ')'"), start);
83 +
84 + len = end - start + 1;
85 + if (skip_prefix(start, "(objectmode)", &p)) {
86 + strbuf_addf(sb, "%06o", data->mode);
87 + } else if (skip_prefix(start, "(objecttype)", &p)) {
88 + strbuf_addstr(sb, type_name(data->type));
89 + } else if (skip_prefix(start, "(objectsize:padded)", &p)) {
90 + expand_objectsize(sb, data->oid, data->type, 1);
91 + } else if (skip_prefix(start, "(objectsize)", &p)) {
92 + expand_objectsize(sb, data->oid, data->type, 0);
93 + } else if (skip_prefix(start, "(objectname)", &p)) {
94 + strbuf_add_unique_abbrev(sb, data->oid, abbrev);
95 + } else if (skip_prefix(start, "(path)", &p)) {
96 + const char *name = data->base->buf;
97 + const char *prefix = chomp_prefix ? ls_tree_prefix : NULL;
98 + struct strbuf quoted = STRBUF_INIT;
99 + struct strbuf sbuf = STRBUF_INIT;
100 + strbuf_addstr(data->base, data->pathname);
101 + name = relative_path(data->base->buf, prefix, &sbuf);
102 + quote_c_style(name, &quoted, NULL, 0);
103 + strbuf_addbuf(sb, &quoted);
104 + strbuf_release(&sbuf);
105 + strbuf_release(&quoted);
106 + } else {
107 + errlen = (unsigned long)len;
108 + die(_("bad ls-tree format: %%%.*s"), errlen, start);
109 + }
110 + return len;
111 +}
112 +
113 static int show_recursive(const char *base, size_t baselen, const char *pathname)
114 {
115 int i;
@@ -71,6 +140,38 @@ static int show_recursive(const char *base, size_t baselen, const char *pathname
140 return 0;
141 }
142
143 +static int show_tree_fmt(const struct object_id *oid, struct strbuf *base,
144 + const char *pathname, unsigned mode, void *context)
145 +{
146 + size_t baselen;
147 + int recurse = 0;
148 + struct strbuf sb = STRBUF_INIT;
149 + enum object_type type = object_type(mode);
150 +
151 + struct show_tree_data data = {
152 + .mode = mode,
153 + .type = type,
154 + .oid = oid,
155 + .pathname = pathname,
156 + .base = base,
157 + };
158 +
159 + if (type == OBJ_TREE && show_recursive(base->buf, base->len, pathname))
160 + recurse = READ_TREE_RECURSIVE;
161 + if (type == OBJ_TREE && recurse && !(ls_options & LS_SHOW_TREES))
162 + return recurse;
163 + if (type == OBJ_BLOB && (ls_options & LS_TREE_ONLY))
164 + return 0;
165 +
166 + baselen = base->len;
167 + strbuf_expand(&sb, format, expand_show_tree, &data);
168 + strbuf_addch(&sb, line_termination);
169 + fwrite(sb.buf, sb.len, 1, stdout);
170 + strbuf_release(&sb);
171 + strbuf_setlen(base, baselen);
172 + return recurse;
173 +}
174 +
175 static int show_default(struct show_tree_data *data)
176 {
177 size_t baselen = data->base->len;
@@ -145,11 +246,33 @@ static int show_tree(const struct object_id *oid, struct strbuf *base,
246 return recurse;
247 }
248
249 +struct ls_tree_cmdmode_to_fmt {
250 + enum ls_tree_cmdmode mode;
251 + const char *const fmt;
252 +};
253 +
254 +static struct ls_tree_cmdmode_to_fmt ls_tree_cmdmode_format[] = {
255 + {
256 + .mode = MODE_DEFAULT,
257 + .fmt = "%(objectmode) %(objecttype) %(objectname)%x09%(path)",
258 + },
259 + {
260 + .mode = MODE_LONG,
261 + .fmt = "%(objectmode) %(objecttype) %(objectname) %(objectsize:padded)%x09%(path)",
262 + },
263 + {
264 + .mode = MODE_NAME_ONLY, /* And MODE_NAME_STATUS */
265 + .fmt = "%(path)",
266 + },
267 + { 0 },
268 +};
269 +
270 int cmd_ls_tree(int argc, const char **argv, const char *prefix)
271 {
272 struct object_id oid;
273 struct tree *tree;
274 int i, full_tree = 0;
275 + read_tree_fn_t fn = show_tree;
276 const struct option ls_tree_options[] = {
277 OPT_BIT('d', NULL, &ls_options, N_("only show trees"),
278 LS_TREE_ONLY),
@@ -170,6 +293,9 @@ int cmd_ls_tree(int argc, const char **argv, const char *prefix)
293 OPT_BOOL(0, "full-tree", &full_tree,
294 N_("list entire tree; not just current directory "
295 "(implies --full-name)")),
296 + OPT_STRING_F(0, "format", &format, N_("format"),
297 + N_("format to use for the output"),
298 + PARSE_OPT_NONEG),
299 OPT__ABBREV(&abbrev),
300 OPT_END()
301 };
@@ -190,6 +316,10 @@ int cmd_ls_tree(int argc, const char **argv, const char *prefix)
316 ((LS_TREE_ONLY|LS_RECURSIVE) & ls_options))
317 ls_options |= LS_SHOW_TREES;
318
319 + if (format && cmdmode)
320 + usage_msg_opt(
321 + _("--format can't be combined with other format-altering options"),
322 + ls_tree_usage, ls_tree_options);
323 if (argc < 1)
324 usage_with_options(ls_tree_usage, ls_tree_options);
325 if (get_oid(argv[0], &oid))
@@ -211,6 +341,23 @@ int cmd_ls_tree(int argc, const char **argv, const char *prefix)
341 tree = parse_tree_indirect(&oid);
342 if (!tree)
343 die("not a tree object");
214 - return !!read_tree(the_repository, tree,
215 - &pathspec, show_tree, NULL);
344 + /*
345 + * The generic show_tree_fmt() is slower than show_tree(), so
346 + * take the fast path if possible.
347 + */
348 + if (format) {
349 + struct ls_tree_cmdmode_to_fmt *m2f;
350 +
351 + fn = show_tree_fmt;
352 + for (m2f = ls_tree_cmdmode_format; m2f->fmt; m2f++) {
353 + if (strcmp(format, m2f->fmt))
354 + continue;
355 +
356 + cmdmode = m2f->mode;
357 + fn = show_tree;
358 + break;
359 + }
360 + }
361 +
362 + return !!read_tree(the_repository, tree, &pathspec, fn, NULL);
363 }
t/t3104-ls-tree-format.sh new
+67
@@ -0,0 +1,67 @@
1 +#!/bin/sh
2 +
3 +test_description='ls-tree --format'
4 +
5 +TEST_PASSES_SANITIZE_LEAK=true
6 +. ./test-lib.sh
7 +
8 +test_expect_success 'ls-tree --format usage' '
9 + test_expect_code 129 git ls-tree --format=fmt -l HEAD &&
10 + test_expect_code 129 git ls-tree --format=fmt --name-only HEAD &&
11 + test_expect_code 129 git ls-tree --format=fmt --name-status HEAD
12 +'
13 +
14 +test_expect_success 'setup' '
15 + mkdir dir &&
16 + test_commit dir/sub-file &&
17 + test_commit top-file
18 +'
19 +
20 +test_ls_tree_format () {
21 + format=$1 &&
22 + opts=$2 &&
23 + fmtopts=$3 &&
24 + shift 2 &&
25 +
26 + test_expect_success "ls-tree '--format=<$format>' is like options '$opts $fmtopts'" '
27 + git ls-tree $opts -r HEAD >expect &&
28 + git ls-tree --format="$format" -r $fmtopts HEAD >actual &&
29 + test_cmp expect actual
30 + '
31 +
32 + test_expect_success "ls-tree '--format=<$format>' on optimized v.s. non-optimized path" '
33 + git ls-tree --format="$format" -r $fmtopts HEAD >expect &&
34 + git ls-tree --format="> $format" -r $fmtopts HEAD >actual.raw &&
35 + sed "s/^> //" >actual <actual.raw &&
36 + test_cmp expect actual
37 + '
38 +}
39 +
40 +test_ls_tree_format \
41 + "%(objectmode) %(objecttype) %(objectname)%x09%(path)" \
42 + ""
43 +
44 +test_ls_tree_format \
45 + "%(objectmode) %(objecttype) %(objectname) %(objectsize:padded)%x09%(path)" \
46 + "--long"
47 +
48 +test_ls_tree_format \
49 + "%(path)" \
50 + "--name-only"
51 +
52 +test_ls_tree_format \
53 + "%(objectmode) %(objecttype) %(objectname)%x09%(path)" \
54 + "-t" \
55 + "-t"
56 +
57 +test_ls_tree_format \
58 + "%(objectmode) %(objecttype) %(objectname)%x09%(path)" \
59 + "--full-name" \
60 + "--full-name"
61 +
62 +test_ls_tree_format \
63 + "%(objectmode) %(objecttype) %(objectname)%x09%(path)" \
64 + "--full-tree" \
65 + "--full-tree"
66 +
67 +test_done