`git for-each-repo` stops on non-existent repo (Windows?)

[derrickstolee]

`git for-each-repo` stops on non-existent repo (Windows?)

[dscho]

Sketch of the functional change I am proposing:

diff --git a/builtin/for-each-repo.c b/builtin/for-each-repo.c
index 28186b30f548..35ab7f34ca15 100644
--- a/builtin/for-each-repo.c
+++ b/builtin/for-each-repo.c
@@ -32,6 +32,7 @@ static int run_command_on_repo(const char *path, int argc, const char ** argv)
 int cmd_for_each_repo(int argc, const char **argv, const char *prefix)
 {
 	static const char *config_key = NULL;
+	int stop_on_error = 0;
 	int i, result = 0;
 	const struct string_list *values;
 	int err;
@@ -39,6 +40,8 @@ int cmd_for_each_repo(int argc, const char **argv, const char *prefix)
 	const struct option options[] = {
 		OPT_STRING(0, "config", &config_key, N_("config"),
 			   N_("config key storing a list of repository paths")),
+		OPT_BOOL(0, "stop-on-error", &stop_on_error,
+			 N_("stop at the first repository where the operation failed")),
 		OPT_END()
 	};
 
@@ -55,8 +58,9 @@ int cmd_for_each_repo(int argc, const char **argv, const char *prefix)
 	else if (err)
 		return 0;
 
-	for (i = 0; !result && i < values->nr; i++)
-		result = run_command_on_repo(values->items[i].string, argc, argv);
+	for (i = 0; (!stop_on_error || !result) && i < values->nr; i++)
+		if (run_command_on_repo(values->items[i].string, argc, argv))
+			result = 1;
 
 	return result;
 }

[dscho]

The first non-existent repo stopped the loop.

This is indeed the case 😞:

https://github.com/git/git/blob/v2.43.0/builtin/for-each-repo.c#L58-L59

It would probably make sense to retain the ability to stop on the first failure, but make that an option, by default continuing (but of course still exiting with an error if the operation failed in at least one repository).

[derrickstolee]

Sketch of the functional change I am proposing:

That would be the backwards-compatible fix. The other approach would be to change behavior by attempting all repos and returning success only if every repo succeeded. But Git doesn't typically do backwards-incompatible things. Who knows who depends on this early termination behavior?

Please also update the background maintenance schedule to include --stop-on-error on its use of for-each-repo to auto-heal users when they reconfigure background maintenance.