Merge branch 'ag/sequencer-reduce-rewriting-todo'

The scripted version of "git rebase -i" wrote and rewrote the todo
list many times during a single step of its operation, and the
recent C-rewrite made a faithful conversion of the logic to C.  The
implementation has been updated to carry necessary information
around in-core to avoid rewriting the same file over and over
unnecessarily.

* ag/sequencer-reduce-rewriting-todo:
  rebase--interactive: move transform_todo_file()
  sequencer: use edit_todo_list() in complete_action()
  rebase-interactive: rewrite edit_todo_list() to handle the initial edit
  rebase-interactive: append_todo_help() changes
  rebase-interactive: use todo_list_write_to_file() in edit_todo_list()
  sequencer: refactor skip_unnecessary_picks() to work on a todo_list
  rebase--interactive: move rearrange_squash_in_todo_file()
  rebase--interactive: move sequencer_add_exec_commands()
  sequencer: change complete_action() to use the refactored functions
  sequencer: make sequencer_make_script() write its script to a strbuf
  sequencer: refactor rearrange_squash() to work on a todo_list
  sequencer: refactor sequencer_add_exec_commands() to work on a todo_list
  sequencer: refactor check_todo_list() to work on a todo_list
  sequencer: introduce todo_list_write_to_file()
  sequencer: refactor transform_todos() to work on a todo_list
  sequencer: remove the 'arg' field from todo_item
  sequencer: make the todo_list structure public
  sequencer: changes in parse_insn_buffer()

builtin/rebase--interactive.c

@@ -14,6 +14,103 @@ static GIT_PATH_FUNC(path_state_dir, "rebase-merge/")
 static GIT_PATH_FUNC(path_squash_onto, "rebase-merge/squash-onto")
 static GIT_PATH_FUNC(path_interactive, "rebase-merge/interactive")
 
+static int add_exec_commands(struct string_list *commands)
+{
+	const char *todo_file = rebase_path_todo();
+	struct todo_list todo_list = TODO_LIST_INIT;
+	int res;
+
+	if (strbuf_read_file(&todo_list.buf, todo_file, 0) < 0)
+		return error_errno(_("could not read '%s'."), todo_file);
+
+	if (todo_list_parse_insn_buffer(the_repository, todo_list.buf.buf,
+					&todo_list)) {
+		todo_list_release(&todo_list);
+		return error(_("unusable todo list: '%s'"), todo_file);
+	}
+
+	todo_list_add_exec_commands(&todo_list, commands);
+	res = todo_list_write_to_file(the_repository, &todo_list,
+				      todo_file, NULL, NULL, -1, 0);
+	todo_list_release(&todo_list);
+
+	if (res)
+		return error_errno(_("could not write '%s'."), todo_file);
+	return 0;
+}
+
+static int rearrange_squash_in_todo_file(void)
+{
+	const char *todo_file = rebase_path_todo();
+	struct todo_list todo_list = TODO_LIST_INIT;
+	int res = 0;
+
+	if (strbuf_read_file(&todo_list.buf, todo_file, 0) < 0)
+		return error_errno(_("could not read '%s'."), todo_file);
+	if (todo_list_parse_insn_buffer(the_repository, todo_list.buf.buf,
+					&todo_list)) {
+		todo_list_release(&todo_list);
+		return error(_("unusable todo list: '%s'"), todo_file);
+	}
+
+	res = todo_list_rearrange_squash(&todo_list);
+	if (!res)
+		res = todo_list_write_to_file(the_repository, &todo_list,
+					      todo_file, NULL, NULL, -1, 0);
+
+	todo_list_release(&todo_list);
+
+	if (res)
+		return error_errno(_("could not write '%s'."), todo_file);
+	return 0;
+}
+
+static int transform_todo_file(unsigned flags)
+{
+	const char *todo_file = rebase_path_todo();
+	struct todo_list todo_list = TODO_LIST_INIT;
+	int res;
+
+	if (strbuf_read_file(&todo_list.buf, todo_file, 0) < 0)
+		return error_errno(_("could not read '%s'."), todo_file);
+
+	if (todo_list_parse_insn_buffer(the_repository, todo_list.buf.buf,
+					&todo_list)) {
+		todo_list_release(&todo_list);
+		return error(_("unusable todo list: '%s'"), todo_file);
+	}
+
+	res = todo_list_write_to_file(the_repository, &todo_list, todo_file,
+				      NULL, NULL, -1, flags);
+	todo_list_release(&todo_list);
+
+	if (res)
+		return error_errno(_("could not write '%s'."), todo_file);
+	return 0;
+}
+
+static int edit_todo_file(unsigned flags)
+{
+	const char *todo_file = rebase_path_todo();
+	struct todo_list todo_list = TODO_LIST_INIT,
+		new_todo = TODO_LIST_INIT;
+	int res = 0;
+
+	if (strbuf_read_file(&todo_list.buf, todo_file, 0) < 0)
+		return error_errno(_("could not read '%s'."), todo_file);
+
+	strbuf_stripspace(&todo_list.buf, 1);
+	res = edit_todo_list(the_repository, &todo_list, &new_todo, NULL, NULL, flags);
+	if (!res && todo_list_write_to_file(the_repository, &new_todo, todo_file,
+					    NULL, NULL, -1, flags & ~(TODO_LIST_SHORTEN_IDS)))
+		res = error_errno(_("could not write '%s'"), todo_file);
+
+	todo_list_release(&todo_list);
+	todo_list_release(&new_todo);
+
+	return res;
+}
+
 static int get_revision_ranges(const char *upstream, const char *onto,
 			       const char **head_hash,
 			       char **revisions, char **shortrevisions)
@@ -66,13 +163,13 @@ static int do_interactive_rebase(struct replay_opts *opts, unsigned flags,
 				 const char *onto, const char *onto_name,
 				 const char *squash_onto, const char *head_name,
 				 const char *restrict_revision, char *raw_strategies,
-				 const char *cmd, unsigned autosquash)
+				 struct string_list *commands, unsigned autosquash)
 {
 	int ret;
 	const char *head_hash = NULL;
 	char *revisions = NULL, *shortrevisions = NULL;
 	struct argv_array make_script_args = ARGV_ARRAY_INIT;
-	FILE *todo_list;
+	struct todo_list todo_list = TODO_LIST_INIT;
 
 	if (prepare_branch_to_be_rebased(opts, switch_to))
 		return -1;
@@ -94,34 +191,29 @@ static int do_interactive_rebase(struct replay_opts *opts, unsigned flags,
 	if (!upstream && squash_onto)
 		write_file(path_squash_onto(), "%s\n", squash_onto);
 
-	todo_list = fopen(rebase_path_todo(), "w");
-	if (!todo_list) {
-		free(revisions);
-		free(shortrevisions);
-
-		return error_errno(_("could not open %s"), rebase_path_todo());
-	}
-
 	argv_array_pushl(&make_script_args, "", revisions, NULL);
 	if (restrict_revision)
 		argv_array_push(&make_script_args, restrict_revision);
 
-	ret = sequencer_make_script(the_repository, todo_list,
+	ret = sequencer_make_script(the_repository, &todo_list.buf,
 				    make_script_args.argc, make_script_args.argv,
 				    flags);
-	fclose(todo_list);
 
 	if (ret)
 		error(_("could not generate todo list"));
 	else {
 		discard_cache();
-		ret = complete_action(the_repository, opts, flags,
-				      shortrevisions, onto_name, onto,
-				      head_hash, cmd, autosquash);
+		if (todo_list_parse_insn_buffer(the_repository, todo_list.buf.buf,
+						&todo_list))
+			BUG("unusable todo list");
+
+		ret = complete_action(the_repository, opts, flags, shortrevisions, onto_name,
+				      onto, head_hash, commands, autosquash, &todo_list);
 	}
 
 	free(revisions);
 	free(shortrevisions);
+	todo_list_release(&todo_list);
 	argv_array_clear(&make_script_args);
 
 	return ret;
@@ -140,6 +232,7 @@ int cmd_rebase__interactive(int argc, const char **argv, const char *prefix)
 	const char *onto = NULL, *onto_name = NULL, *restrict_revision = NULL,
 		*squash_onto = NULL, *upstream = NULL, *head_name = NULL,
 		*switch_to = NULL, *cmd = NULL;
+	struct string_list commands = STRING_LIST_INIT_DUP;
 	char *raw_strategies = NULL;
 	enum {
 		NONE = 0, CONTINUE, SKIP, EDIT_TODO, SHOW_CURRENT_PATCH,
@@ -224,6 +317,14 @@ int cmd_rebase__interactive(int argc, const char **argv, const char *prefix)
 		warning(_("--[no-]rebase-cousins has no effect without "
 			  "--rebase-merges"));
 
+	if (cmd && *cmd) {
+		string_list_split(&commands, cmd, '\n', -1);
+
+		/* rebase.c adds a new line to cmd after every command,
+		 * so here the last command is always empty */
+		string_list_remove_empty_items(&commands, 0);
+	}
+
 	switch (command) {
 	case NONE:
 		if (!onto && !upstream)
@@ -231,7 +332,7 @@ int cmd_rebase__interactive(int argc, const char **argv, const char *prefix)
 
 		ret = do_interactive_rebase(&opts, flags, switch_to, upstream, onto,
 					    onto_name, squash_onto, head_name, restrict_revision,
-					    raw_strategies, cmd, autosquash);
+					    raw_strategies, &commands, autosquash);
 		break;
 	case SKIP: {
 		struct string_list merge_rr = STRING_LIST_INIT_DUP;
@@ -243,7 +344,7 @@ int cmd_rebase__interactive(int argc, const char **argv, const char *prefix)
 		break;
 	}
 	case EDIT_TODO:
-		ret = edit_todo_list(the_repository, flags);
+		ret = edit_todo_file(flags);
 		break;
 	case SHOW_CURRENT_PATCH: {
 		struct child_process cmd = CHILD_PROCESS_INIT;
@@ -256,20 +357,21 @@ int cmd_rebase__interactive(int argc, const char **argv, const char *prefix)
 	}
 	case SHORTEN_OIDS:
 	case EXPAND_OIDS:
-		ret = transform_todos(the_repository, flags);
+		ret = transform_todo_file(flags);
 		break;
 	case CHECK_TODO_LIST:
-		ret = check_todo_list(the_repository);
+		ret = check_todo_list_from_file(the_repository);
 		break;
 	case REARRANGE_SQUASH:
-		ret = rearrange_squash(the_repository);
+		ret = rearrange_squash_in_todo_file();
 		break;
 	case ADD_EXEC:
-		ret = sequencer_add_exec_commands(the_repository, cmd);
+		ret = add_exec_commands(&commands);
 		break;
 	default:
 		BUG("invalid command '%d'", command);
 	}
 
+	string_list_clear(&commands, 0);
 	return !!ret;
 }

rebase-interactive.c

@@ -1,10 +1,35 @@
 #include "cache.h"
 #include "commit.h"
-#include "rebase-interactive.h"
 #include "sequencer.h"
+#include "rebase-interactive.h"
 #include "strbuf.h"
+#include "commit-slab.h"
+#include "config.h"
+
+enum missing_commit_check_level {
+	MISSING_COMMIT_CHECK_IGNORE = 0,
+	MISSING_COMMIT_CHECK_WARN,
+	MISSING_COMMIT_CHECK_ERROR
+};
+
+static enum missing_commit_check_level get_missing_commit_check_level(void)
+{
+	const char *value;
+
+	if (git_config_get_value("rebase.missingcommitscheck", &value) ||
+			!strcasecmp("ignore", value))
+		return MISSING_COMMIT_CHECK_IGNORE;
+	if (!strcasecmp("warn", value))
+		return MISSING_COMMIT_CHECK_WARN;
+	if (!strcasecmp("error", value))
+		return MISSING_COMMIT_CHECK_ERROR;
+	warning(_("unrecognized setting %s for option "
+		  "rebase.missingCommitsCheck. Ignoring."), value);
+	return MISSING_COMMIT_CHECK_IGNORE;
+}
 
-void append_todo_help(unsigned edit_todo, unsigned keep_empty,
+void append_todo_help(unsigned keep_empty, int command_count,
+		      const char *shortrevisions, const char *shortonto,
 		      struct strbuf *buf)
 {
 	const char *msg = _("\nCommands:\n"
@@ -24,6 +49,15 @@ void append_todo_help(unsigned edit_todo, unsigned keep_empty,
 ".       specified). Use -c <commit> to reword the commit message.\n"
 "\n"
 "These lines can be re-ordered; they are executed from top to bottom.\n");
+	unsigned edit_todo = !(shortrevisions && shortonto);
+
+	if (!edit_todo) {
+		strbuf_addch(buf, '\n');
+		strbuf_commented_addf(buf, Q_("Rebase %s onto %s (%d command)",
+					      "Rebase %s onto %s (%d commands)",
+					      command_count),
+				      shortrevisions, shortonto, command_count);
+	}
 
 	strbuf_add_commented_lines(buf, msg, strlen(msg));
 
@@ -53,39 +87,103 @@ void append_todo_help(unsigned edit_todo, unsigned keep_empty,
 	}
 }
 
-int edit_todo_list(struct repository *r, unsigned flags)
+int edit_todo_list(struct repository *r, struct todo_list *todo_list,
+		   struct todo_list *new_todo, const char *shortrevisions,
+		   const char *shortonto, unsigned flags)
 {
-	struct strbuf buf = STRBUF_INIT;
 	const char *todo_file = rebase_path_todo();
+	unsigned initial = shortrevisions && shortonto;
 
-	if (strbuf_read_file(&buf, todo_file, 0) < 0)
-		return error_errno(_("could not read '%s'."), todo_file);
+	/* If the user is editing the todo list, we first try to parse
+	 * it.  If there is an error, we do not return, because the user
+	 * might want to fix it in the first place. */
+	if (!initial)
+		todo_list_parse_insn_buffer(r, todo_list->buf.buf, todo_list);
 
-	strbuf_stripspace(&buf, 1);
-	if (write_message(buf.buf, buf.len, todo_file, 0)) {
-		strbuf_release(&buf);
-		return -1;
-	}
+	if (todo_list_write_to_file(r, todo_list, todo_file, shortrevisions, shortonto,
+				    -1, flags | TODO_LIST_SHORTEN_IDS | TODO_LIST_APPEND_TODO_HELP))
+		return error_errno(_("could not write '%s'"), todo_file);
+
+	if (initial && copy_file(rebase_path_todo_backup(), todo_file, 0666))
+		return error(_("could not copy '%s' to '%s'."), todo_file,
+			     rebase_path_todo_backup());
+
+	if (launch_sequence_editor(todo_file, &new_todo->buf, NULL))
+		return -2;
+
+	strbuf_stripspace(&new_todo->buf, 1);
+	if (initial && new_todo->buf.len == 0)
+		return -3;
 
-	strbuf_release(&buf);
+	/* For the initial edit, the todo list gets parsed in
+	 * complete_action(). */
+	if (!initial)
+		return todo_list_parse_insn_buffer(r, new_todo->buf.buf, new_todo);
 
-	transform_todos(r, flags | TODO_LIST_SHORTEN_IDS);
+	return 0;
+}
+
+define_commit_slab(commit_seen, unsigned char);
+/*
+ * Check if the user dropped some commits by mistake
+ * Behaviour determined by rebase.missingCommitsCheck.
+ * Check if there is an unrecognized command or a
+ * bad SHA-1 in a command.
+ */
+int todo_list_check(struct todo_list *old_todo, struct todo_list *new_todo)
+{
+	enum missing_commit_check_level check_level = get_missing_commit_check_level();
+	struct strbuf missing = STRBUF_INIT;
+	int res = 0, i;
+	struct commit_seen commit_seen;
+
+	init_commit_seen(&commit_seen);
 
-	if (strbuf_read_file(&buf, todo_file, 0) < 0)
-		return error_errno(_("could not read '%s'."), todo_file);
+	if (check_level == MISSING_COMMIT_CHECK_IGNORE)
+		goto leave_check;
 
-	append_todo_help(1, 0, &buf);
-	if (write_message(buf.buf, buf.len, todo_file, 0)) {
-		strbuf_release(&buf);
-		return -1;
+	/* Mark the commits in git-rebase-todo as seen */
+	for (i = 0; i < new_todo->nr; i++) {
+		struct commit *commit = new_todo->items[i].commit;
+		if (commit)
+			*commit_seen_at(&commit_seen, commit) = 1;
 	}
 
-	strbuf_release(&buf);
+	/* Find commits in git-rebase-todo.backup yet unseen */
+	for (i = old_todo->nr - 1; i >= 0; i--) {
+		struct todo_item *item = old_todo->items + i;
+		struct commit *commit = item->commit;
+		if (commit && !*commit_seen_at(&commit_seen, commit)) {
+			strbuf_addf(&missing, " - %s %.*s\n",
+				    find_unique_abbrev(&commit->object.oid, DEFAULT_ABBREV),
+				    item->arg_len,
+				    todo_item_get_arg(old_todo, item));
+			*commit_seen_at(&commit_seen, commit) = 1;
+		}
+	}
 
-	if (launch_sequence_editor(todo_file, NULL, NULL))
-		return -1;
+	/* Warn about missing commits */
+	if (!missing.len)
+		goto leave_check;
 
-	transform_todos(r, flags & ~(TODO_LIST_SHORTEN_IDS));
+	if (check_level == MISSING_COMMIT_CHECK_ERROR)
+		res = 1;
 
-	return 0;
+	fprintf(stderr,
+		_("Warning: some commits may have been dropped accidentally.\n"
+		"Dropped commits (newer to older):\n"));
+
+	/* Make the list user-friendly and display */
+	fputs(missing.buf, stderr);
+	strbuf_release(&missing);
+
+	fprintf(stderr, _("To avoid this message, use \"drop\" to "
+		"explicitly remove a commit.\n\n"
+		"Use 'git config rebase.missingCommitsCheck' to change "
+		"the level of warnings.\n"
+		"The possible behaviours are: ignore, warn, error.\n\n"));
+
+leave_check:
+	clear_commit_seen(&commit_seen);
+	return res;
 }

rebase-interactive.h

@@ -3,9 +3,14 @@
 
 struct strbuf;
 struct repository;
+struct todo_list;
 
-void append_todo_help(unsigned edit_todo, unsigned keep_empty,
+void append_todo_help(unsigned keep_empty, int command_count,
+		      const char *shortrevisions, const char *shortonto,
 		      struct strbuf *buf);
-int edit_todo_list(struct repository *r, unsigned flags);
+int edit_todo_list(struct repository *r, struct todo_list *todo_list,
+		   struct todo_list *new_todo, const char *shortrevisions,
+		   const char *shortonto, unsigned flags);
+int todo_list_check(struct todo_list *old_todo, struct todo_list *new_todo);
 
 #endif

sequencer.h

@@ -10,6 +10,7 @@ struct repository;
 const char *git_path_commit_editmsg(void);
 const char *git_path_seq_dir(void);
 const char *rebase_path_todo(void);
+const char *rebase_path_todo_backup(void);
 
 #define APPEND_SIGNOFF_DEDUP (1u << 0)
 
@@ -66,14 +67,60 @@ struct replay_opts {
 };
 #define REPLAY_OPTS_INIT { .action = -1, .current_fixups = STRBUF_INIT }
 
-enum missing_commit_check_level {
-	MISSING_COMMIT_CHECK_IGNORE = 0,
-	MISSING_COMMIT_CHECK_WARN,
-	MISSING_COMMIT_CHECK_ERROR
+/*
+ * Note that ordering matters in this enum. Not only must it match the mapping
+ * of todo_command_info (in sequencer.c), it is also divided into several
+ * sections that matter.  When adding new commands, make sure you add it in the
+ * right section.
+ */
+enum todo_command {
+	/* commands that handle commits */
+	TODO_PICK = 0,
+	TODO_REVERT,
+	TODO_EDIT,
+	TODO_REWORD,
+	TODO_FIXUP,
+	TODO_SQUASH,
+	/* commands that do something else than handling a single commit */
+	TODO_EXEC,
+	TODO_BREAK,
+	TODO_LABEL,
+	TODO_RESET,
+	TODO_MERGE,
+	/* commands that do nothing but are counted for reporting progress */
+	TODO_NOOP,
+	TODO_DROP,
+	/* comments (not counted for reporting progress) */
+	TODO_COMMENT
 };
 
-int write_message(const void *buf, size_t len, const char *filename,
-		  int append_eol);
+struct todo_item {
+	enum todo_command command;
+	struct commit *commit;
+	unsigned int flags;
+	int arg_len;
+	/* The offset of the command and its argument in the strbuf */
+	size_t offset_in_buf, arg_offset;
+};
+
+struct todo_list {
+	struct strbuf buf;
+	struct todo_item *items;
+	int nr, alloc, current;
+	int done_nr, total_nr;
+	struct stat_data stat;
+};
+
+#define TODO_LIST_INIT { STRBUF_INIT }
+
+int todo_list_parse_insn_buffer(struct repository *r, char *buf,
+				struct todo_list *todo_list);
+int todo_list_write_to_file(struct repository *r, struct todo_list *todo_list,
+			    const char *file, const char *shortrevisions,
+			    const char *shortonto, int num, unsigned flags);
+void todo_list_release(struct todo_list *todo_list);
+const char *todo_item_get_arg(struct todo_list *todo_list,
+			      struct todo_item *item);
 
 /* Call this to setup defaults before parsing command line options */
 void sequencer_init_config(struct replay_opts *opts);
@@ -93,19 +140,19 @@ int sequencer_remove_state(struct replay_opts *opts);
  * commits should be rebased onto the new base, this flag needs to be passed.
  */
 #define TODO_LIST_REBASE_COUSINS (1U << 4)
-int sequencer_make_script(struct repository *repo, FILE *out,
-			  int argc, const char **argv,
-			  unsigned flags);
-
-int sequencer_add_exec_commands(struct repository *r, const char *command);
-int transform_todos(struct repository *r, unsigned flags);
-enum missing_commit_check_level get_missing_commit_check_level(void);
-int check_todo_list(struct repository *r);
+#define TODO_LIST_APPEND_TODO_HELP (1U << 5)
+
+int sequencer_make_script(struct repository *r, struct strbuf *out, int argc,
+			  const char **argv, unsigned flags);
+
+void todo_list_add_exec_commands(struct todo_list *todo_list,
+				 struct string_list *commands);
+int check_todo_list_from_file(struct repository *r);
 int complete_action(struct repository *r, struct replay_opts *opts, unsigned flags,
 		    const char *shortrevisions, const char *onto_name,
-		    const char *onto, const char *orig_head, const char *cmd,
-		    unsigned autosquash);
-int rearrange_squash(struct repository *r);
+		    const char *onto, const char *orig_head, struct string_list *commands,
+		    unsigned autosquash, struct todo_list *todo_list);
+int todo_list_rearrange_squash(struct todo_list *todo_list);
 
 /*
  * Append a signoff to the commit message in "msgbuf". The ignore_footer