sequencer: use struct strvec to store merge strategy options

The sequencer stores the merge strategy options in an array of strings
which allocated with ALLOC_GROW(). Using "struct strvec" avoids manually
managing the memory of that array and simplifies the code.

Aside from memory allocation the changes to the sequencer are largely
mechanical, changing xopts_nr to xopts.nr and xopts[i] to xopts.v[i]. A
new option parsing macro OPT_STRVEC() is also added to collect the
strategy options.  Hopefully this can be used to simplify the code in
builtin/merge.c in the future.

Note that there is a change of behavior to "git cherry-pick" and "git
revert" as passing “--no-strategy-option” will now clear any previous
strategy options whereas before this change it did nothing.

Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>

sequencer.h

@@ -2,6 +2,7 @@
 #define SEQUENCER_H
 
 #include "strbuf.h"
+#include "strvec.h"
 #include "wt-status.h"
 
 struct commit;
@@ -60,8 +61,7 @@ struct replay_opts {
 	/* Merge strategy */
 	char *default_strategy;  /* from config options */
 	char *strategy;
-	char **xopts;
-	size_t xopts_nr, xopts_alloc;
+	struct strvec xopts;
 
 	/* Reflog */
 	char *reflog_action;
@@ -80,7 +80,12 @@ struct replay_opts {
 	/* Private use */
 	const char *reflog_message;
 };
-#define REPLAY_OPTS_INIT { .edit = -1, .action = -1, .current_fixups = STRBUF_INIT }
+#define REPLAY_OPTS_INIT {			\
+	.edit = -1,				\
+	.action = -1,				\
+	.current_fixups = STRBUF_INIT,		\
+	.xopts = STRVEC_INIT,			\
+}
 
 /*
  * Note that ordering matters in this enum. Not only must it match the mapping