scalar: implement `scalar diagnose`

Over the course of Scalar's development, it became obvious that there is a need for a command that can gather all kinds of useful information that can help identify the most typical problems with large worktrees/repositories. The `diagnose` command is the culmination of this hard-won knowledge: it gathers the installed hooks, the config, a couple statistics describing the data shape, among other pieces of information, and then wraps everything up in a tidy, neat `.zip` archive. Note: originally, Scalar was implemented in C# using the .NET API, where we had the luxury of a comprehensive standard library that includes basic functionality such as writing a `.zip` file. In the C version, we lack such a commodity. Rather than introducing a dependency on, say, libzip, we slightly abuse Git's `archive` machinery: we write out a `.zip` of the empty try, augmented by a couple files that are added via the `--add-file*` options. We are careful trying not to modify the current repository in any way lest the very circumstances that required `scalar diagnose` to be run are changed by the `diagnose` run itself. Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Johannes Schindelin committed May 28, 2022 at 16:11 UTC aa5c79a33156c2f086ca3a149c11f1434d10f5ce
3 files changed +170
contrib/scalar/scalar.c
+144
@@ -11,6 +11,7 @@
11 #include "dir.h"
12 #include "packfile.h"
13 #include "help.h"
14 +#include "archive.h"
15
16 /*
17 * Remove the deepest subdirectory in the provided path string. Path must not
@@ -260,6 +261,47 @@ static int unregister_dir(void)
261 return res;
262 }
263
264 +static int add_directory_to_archiver(struct strvec *archiver_args,
265 + const char *path, int recurse)
266 +{
267 + int at_root = !*path;
268 + DIR *dir = opendir(at_root ? "." : path);
269 + struct dirent *e;
270 + struct strbuf buf = STRBUF_INIT;
271 + size_t len;
272 + int res = 0;
273 +
274 + if (!dir)
275 + return error_errno(_("could not open directory '%s'"), path);
276 +
277 + if (!at_root)
278 + strbuf_addf(&buf, "%s/", path);
279 + len = buf.len;
280 + strvec_pushf(archiver_args, "--prefix=%s", buf.buf);
281 +
282 + while (!res && (e = readdir(dir))) {
283 + if (!strcmp(".", e->d_name) || !strcmp("..", e->d_name))
284 + continue;
285 +
286 + strbuf_setlen(&buf, len);
287 + strbuf_addstr(&buf, e->d_name);
288 +
289 + if (e->d_type == DT_REG)
290 + strvec_pushf(archiver_args, "--add-file=%s", buf.buf);
291 + else if (e->d_type != DT_DIR)
292 + warning(_("skipping '%s', which is neither file nor "
293 + "directory"), buf.buf);
294 + else if (recurse &&
295 + add_directory_to_archiver(archiver_args,
296 + buf.buf, recurse) < 0)
297 + res = -1;
298 + }
299 +
300 + closedir(dir);
301 + strbuf_release(&buf);
302 + return res;
303 +}
304 +
305 /* printf-style interface, expects `<key>=<value>` argument */
306 static int set_config(const char *fmt, ...)
307 {
@@ -500,6 +542,107 @@ cleanup:
542 return res;
543 }
544
545 +static int cmd_diagnose(int argc, const char **argv)
546 +{
547 + struct option options[] = {
548 + OPT_END(),
549 + };
550 + const char * const usage[] = {
551 + N_("scalar diagnose [<enlistment>]"),
552 + NULL
553 + };
554 + struct strbuf zip_path = STRBUF_INIT;
555 + struct strvec archiver_args = STRVEC_INIT;
556 + char **argv_copy = NULL;
557 + int stdout_fd = -1, archiver_fd = -1;
558 + time_t now = time(NULL);
559 + struct tm tm;
560 + struct strbuf path = STRBUF_INIT, buf = STRBUF_INIT;
561 + int res = 0;
562 +
563 + argc = parse_options(argc, argv, NULL, options,
564 + usage, 0);
565 +
566 + setup_enlistment_directory(argc, argv, usage, options, &zip_path);
567 +
568 + strbuf_addstr(&zip_path, "/.scalarDiagnostics/scalar_");
569 + strbuf_addftime(&zip_path,
570 + "%Y%m%d_%H%M%S", localtime_r(&now, &tm), 0, 0);
571 + strbuf_addstr(&zip_path, ".zip");
572 + switch (safe_create_leading_directories(zip_path.buf)) {
573 + case SCLD_EXISTS:
574 + case SCLD_OK:
575 + break;
576 + default:
577 + error_errno(_("could not create directory for '%s'"),
578 + zip_path.buf);
579 + goto diagnose_cleanup;
580 + }
581 + stdout_fd = dup(1);
582 + if (stdout_fd < 0) {
583 + res = error_errno(_("could not duplicate stdout"));
584 + goto diagnose_cleanup;
585 + }
586 +
587 + archiver_fd = xopen(zip_path.buf, O_CREAT | O_WRONLY | O_TRUNC, 0666);
588 + if (archiver_fd < 0 || dup2(archiver_fd, 1) < 0) {
589 + res = error_errno(_("could not redirect output"));
590 + goto diagnose_cleanup;
591 + }
592 +
593 + init_zip_archiver();
594 + strvec_pushl(&archiver_args, "scalar-diagnose", "--format=zip", NULL);
595 +
596 + strbuf_reset(&buf);
597 + strbuf_addstr(&buf, "Collecting diagnostic info\n\n");
598 + get_version_info(&buf, 1);
599 +
600 + strbuf_addf(&buf, "Enlistment root: %s\n", the_repository->worktree);
601 + write_or_die(stdout_fd, buf.buf, buf.len);
602 + strvec_pushf(&archiver_args,
603 + "--add-virtual-file=diagnostics.log:%.*s",
604 + (int)buf.len, buf.buf);
605 +
606 + if ((res = add_directory_to_archiver(&archiver_args, ".git", 0)) ||
607 + (res = add_directory_to_archiver(&archiver_args, ".git/hooks", 0)) ||
608 + (res = add_directory_to_archiver(&archiver_args, ".git/info", 0)) ||
609 + (res = add_directory_to_archiver(&archiver_args, ".git/logs", 1)) ||
610 + (res = add_directory_to_archiver(&archiver_args, ".git/objects/info", 0)))
611 + goto diagnose_cleanup;
612 +
613 + strvec_pushl(&archiver_args, "--prefix=",
614 + oid_to_hex(the_hash_algo->empty_tree), "--", NULL);
615 +
616 + /* `write_archive()` modifies the `argv` passed to it. Let it. */
617 + argv_copy = xmemdupz(archiver_args.v,
618 + sizeof(char *) * archiver_args.nr);
619 + res = write_archive(archiver_args.nr, (const char **)argv_copy, NULL,
620 + the_repository, NULL, 0);
621 + if (res) {
622 + error(_("failed to write archive"));
623 + goto diagnose_cleanup;
624 + }
625 +
626 + if (!res)
627 + fprintf(stderr, "\n"
628 + "Diagnostics complete.\n"
629 + "All of the gathered info is captured in '%s'\n",
630 + zip_path.buf);
631 +
632 +diagnose_cleanup:
633 + if (archiver_fd >= 0) {
634 + close(1);
635 + dup2(stdout_fd, 1);
636 + }
637 + free(argv_copy);
638 + strvec_clear(&archiver_args);
639 + strbuf_release(&zip_path);
640 + strbuf_release(&path);
641 + strbuf_release(&buf);
642 +
643 + return res;
644 +}
645 +
646 static int cmd_list(int argc, const char **argv)
647 {
648 if (argc != 1)
@@ -801,6 +944,7 @@ static struct {
944 { "reconfigure", cmd_reconfigure },
945 { "delete", cmd_delete },
946 { "version", cmd_version },
947 + { "diagnose", cmd_diagnose },
948 { NULL, NULL},
949 };
950
contrib/scalar/scalar.txt
+12
@@ -14,6 +14,7 @@ scalar register [<enlistment>]
14 scalar unregister [<enlistment>]
15 scalar run ( all | config | commit-graph | fetch | loose-objects | pack-files ) [<enlistment>]
16 scalar reconfigure [ --all | <enlistment> ]
17 +scalar diagnose [<enlistment>]
18 scalar delete <enlistment>
19
20 DESCRIPTION
@@ -139,6 +140,17 @@ reconfigure the enlistment.
140 With the `--all` option, all enlistments currently registered with Scalar
141 will be reconfigured. Use this option after each Scalar upgrade.
142
143 +Diagnose
144 +~~~~~~~~
145 +
146 +diagnose [<enlistment>]::
147 + When reporting issues with Scalar, it is often helpful to provide the
148 + information gathered by this command, including logs and certain
149 + statistics describing the data shape of the current enlistment.
150 ++
151 +The output of this command is a `.zip` file that is written into
152 +a directory adjacent to the worktree in the `src` directory.
153 +
154 Delete
155 ~~~~~~
156
contrib/scalar/t/t9099-scalar.sh
+14
@@ -98,4 +98,18 @@ test_expect_success '`scalar [...] <dir>` errors out when dir is missing' '
98 grep "cloned. does not exist" err
99 '
100
101 +SQ="'"
102 +test_expect_success UNZIP 'scalar diagnose' '
103 + scalar clone "file://$(pwd)" cloned --single-branch &&
104 + scalar diagnose cloned >out 2>err &&
105 + sed -n "s/.*$SQ\\(.*\\.zip\\)$SQ.*/\\1/p" <err >zip_path &&
106 + zip_path=$(cat zip_path) &&
107 + test -n "$zip_path" &&
108 + unzip -v "$zip_path" &&
109 + folder=${zip_path%.zip} &&
110 + test_path_is_missing "$folder" &&
111 + unzip -p "$zip_path" diagnostics.log >out &&
112 + test_file_not_empty out
113 +'
114 +
115 test_done