Skip to contentSkip to conversation

seq: arecordmidi: Add num-events option

PATCH In review [email protected] 4 participants 11 messages Jan 29, 2019 – Feb 11, 2019

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 seq: arecordmidi: Add num-events option

From: Prashant Malani <[email protected]>

Add a command line option to automatically exit after recording a fixed number of MIDI events. This allows a program using arecordmidi to expect a MIDI file to be written automatically when the specified number of events have been received, instead of having to send a SIGINT or SIGTERM programmatically.

It also avoids the need to have the arecordmidi process running in the background, and then constantly stat the output file to check if any bytes have been written to it (this makes for less predictable and longer-running tests).

This functionality finds use in Chrome OS functional testing, since having to send SIGTERM/SIGINT programmatically and then wait for the output file adds unpredictability and delay to the tests.

The addition of this command-line option should (hopefully) not break any existing usage.

seq/aplaymidi/arecordmidi.c +20 -3
index 19dbb7d..ca25bfa 100644
@@ -690,7 +690,8 @@ static void help(const char *argv0)
 		"  -t,--ticks=ticks           resolution in ticks per beat or frame\n"
 		"  -s,--split-channels        create a track for each channel\n"
 		"  -m,--metronome=client:port play a metronome signal\n"
-		"  -i,--timesig=nn:dd         time signature\n",
+		"  -i,--timesig=nn:dd         time signature\n"
+		"  -n,--num-events=events     fixed number of events to record, then exit\n",
 		argv0);
 }
 
@@ -706,7 +707,7 @@ static void sighandler(int sig)
 
 int main(int argc, char *argv[])
 {
-	static const char short_options[] = "hVlp:b:f:t:sdm:i:";
+	static const char short_options[] = "hVlp:b:f:t:sdm:i:n:";
 	static const struct option long_options[] = {
 		{"help", 0, NULL, 'h'},
 		{"version", 0, NULL, 'V'},
@@ -719,6 +720,7 @@ int main(int argc, char *argv[])
 		{"dump", 0, NULL, 'd'},
 		{"metronome", 1, NULL, 'm'},
 		{"timesig", 1, NULL, 'i'},
+		{"num-events", 1, NULL, 'n'},
 		{ }
 	};
 
@@ -727,6 +729,9 @@ int main(int argc, char *argv[])
 	struct pollfd *pfds;
 	int npfds;
 	int c, err;
+	/* If |num_events| isn't specified, leave it at 0. */
+	int num_events = 0;
+	int events_received = 0;
 
 	init_seq();
 
@@ -775,6 +780,11 @@ int main(int argc, char *argv[])
 		case 'i':
 			time_signature(optarg);
 			break;
+		case 'n':
+			num_events = atoi(optarg);
+			if (num_events <= 0)
+				fatal("num_events must be greater than 0");
+			break;
 		default:
 			help(argv[0]);
 			return 1;
@@ -864,13 +874,20 @@ int main(int argc, char *argv[])
 			err = snd_seq_event_input(seq, &event);
 			if (err < 0)
 				break;
-			if (event)
+			if (event) {
 				record_event(event);
+				events_received++;
+			}
 		} while (err > 0);
 		if (stop)
 			break;
+		if (num_events && (events_received == num_events))
+			break;
 	}
 
+	if (num_events && events_received < num_events)
+		fputs("Warning: Received signal before num_events\n", stdout);
+
 	finish_tracks();
 	write_file();
 
Signature · 1 line
2.20.1.495.gaa96b0ce6b-goog
#1 lore raw

Hi,

seq/aplaymidi/arecordmidi.c int main(int argc, char *argv[]) · line 784

Quoted 6 lines · On 2019/01/29 10:44, [email protected] wrote:
seq/aplaymidi/arecordmidi.c +5 -0
@@ -775,6 +780,11 @@ int main(int argc, char *argv[])
+		case 'n':
+			num_events = atoi(optarg);
+			if (num_events <= 0)
+				fatal("num_events must be greater than 0");
+			break;

We cannot distinguish two cases; 0 is given and including non-numeraical characters. Please use strtol() or so and check return value and errno for safe.

For your information:
http://git.alsa-project.org/?p=alsa-utils.git;a=blob;f=axfer/main.c;hb=HEAD#l44

Regards

Takashi Sakamoto

#2 lore raw

Hi Takashi,

Quoted 14 lines · Takashi Sakamoto

Hi,

seq/aplaymidi/arecordmidi.c int main(int argc, char *argv[]) · line 784

Quoted 6 lines · On 2019/01/29 10:44, [email protected] wrote:
seq/aplaymidi/arecordmidi.c +5 -0
@@ -775,6 +780,11 @@ int main(int argc, char *argv[])
+             case 'n':
+                     num_events = atoi(optarg);
+                     if (num_events <= 0)
+                             fatal("num_events must be greater than 0");
+                     break;

We cannot distinguish two cases; 0 is given and including non-numeraical characters. Please use strtol() or so and check return value and errno for safe.

Quoted 3 lines

For your information:

http://git.alsa-project.org/?p=alsa-utils.git;a=blob;f=axfer/main.c;hb=HEAD#l44

Can I duplicate this in arecordmidi.c ? I'm loath to introduce a header, and then refactor axfer/main.c.

Quoted 7 lines

Regards

Takashi Sakamoto

#3 lore raw

Hi,

Quoted 2 lines · Prashant Malani

Can I duplicate this in arecordmidi.c ? I'm loath to introduce a header, and then refactor axfer/main.c.

I don't mind it.

When parsing string with non-numeric characters, the behaviour of atoi() is undefined still in C11.

$ cd alsa-utils.git
$ git grep atoi | grep optarg | wc
     39     158    1919

Mmm... Many parts are written without enough care of the undefined behaviour...

Thanks

Takashi Sakamoto

#4 lore raw

PATCH 1/1 seq: arecordmidi: Add num-events option

From: Prashant Malani <[email protected]>

Add a command line option to automatically exit after recording a fixed number of MIDI events. This allows a program using arecordmidi to expect a MIDI file to be written automatically when the specified number of events have been received, instead of having to send a SIGINT or SIGTERM programmatically.

It also avoids the need to have the arecordmidi process running in the background, and then constantly stat the output file to check if any bytes have been written to it (this makes for less predictable and longer-running tests).

This functionality finds use in Chrome OS functional testing, since having to send SIGTERM/SIGINT programmatically and then wait for the output file adds unpredictability and delay to the tests.

The addition of this command-line option should (hopefully) not break any existing usage.

seq/aplaymidi/arecordmidi.c +44 -3
index 19dbb7d..c0d0569 100644
@@ -86,6 +86,25 @@ static int ts_num = 4; /* time signature: numerator */
 static int ts_div = 4; /* time signature: denominator */
 static int ts_dd = 2; /* time signature: denominator as a power of two */
 
+/* Parse a decimal number from a command line argument. */
+static long arg_parse_decimal_num(const char *str, int *err)
+{
+	long val;
+	char *endptr;
+
+	errno = 0;
+	val = strtol(str, &endptr, 0);
+	if (errno > 0) {
+		*err = -errno;
+		return 0;
+	}
+	if (*endptr != '\0') {
+		*err = -EINVAL;
+		return 0;
+	}
+
+	return val;
+}
 
 /* prints an error message to stderr, and dies */
 static void fatal(const char *msg, ...)
@@ -690,7 +709,8 @@ static void help(const char *argv0)
 		"  -t,--ticks=ticks           resolution in ticks per beat or frame\n"
 		"  -s,--split-channels        create a track for each channel\n"
 		"  -m,--metronome=client:port play a metronome signal\n"
-		"  -i,--timesig=nn:dd         time signature\n",
+		"  -i,--timesig=nn:dd         time signature\n"
+		"  -n,--num-events=events     fixed number of events to record, then exit\n",
 		argv0);
 }
 
@@ -706,7 +726,7 @@ static void sighandler(int sig)
 
 int main(int argc, char *argv[])
 {
-	static const char short_options[] = "hVlp:b:f:t:sdm:i:";
+	static const char short_options[] = "hVlp:b:f:t:sdm:i:n:";
 	static const struct option long_options[] = {
 		{"help", 0, NULL, 'h'},
 		{"version", 0, NULL, 'V'},
@@ -719,6 +739,7 @@ int main(int argc, char *argv[])
 		{"dump", 0, NULL, 'd'},
 		{"metronome", 1, NULL, 'm'},
 		{"timesig", 1, NULL, 'i'},
+		{"num-events", 1, NULL, 'n'},
 		{ }
 	};
 
@@ -727,6 +748,9 @@ int main(int argc, char *argv[])
 	struct pollfd *pfds;
 	int npfds;
 	int c, err;
+	/* If |num_events| isn't specified, leave it at 0. */
+	long num_events = 0;
+	long events_received = 0;
 
 	init_seq();
 
@@ -775,6 +799,16 @@ int main(int argc, char *argv[])
 		case 'i':
 			time_signature(optarg);
 			break;
+		case 'n':
+			err = 0;
+			num_events = arg_parse_decimal_num(optarg, &err);
+			if (err != 0) {
+				fatal("Couldn't parse num_events argument: %s\n",
+					strerror(-err));
+			}
+			if (num_events <= 0)
+				fatal("num_events must be greater than 0");
+			break;
 		default:
 			help(argv[0]);
 			return 1;
@@ -864,13 +898,20 @@ int main(int argc, char *argv[])
 			err = snd_seq_event_input(seq, &event);
 			if (err < 0)
 				break;
-			if (event)
+			if (event) {
 				record_event(event);
+				events_received++;
+			}
 		} while (err > 0);
 		if (stop)
 			break;
+		if (num_events && (events_received == num_events))
+			break;
 	}
 
+	if (num_events && events_received < num_events)
+		fputs("Warning: Received signal before num_events\n", stdout);
+
 	finish_tracks();
 	write_file();
 
Signature · 1 line
2.20.1.495.gaa96b0ce6b-goog
#5 lore raw
In reply to Takashi Sakamoto
Quoted 2 lines · Takashi Sakamoto

When parsing string with non-numeric characters, the behaviour of atoi() is undefined still in C11.

A completely non-numeric string is specified to return zero.

The behaviour is undefined only if the value cannot be represented, i.e., if it overflows.

Regards,
Clemens

#6 lore raw

Interesting. Thanks for pointing that out, Clemens.

In any case, I've sent both versions (atoi v/s strtol), so whichever one is preferred can be used (would be slightly inclined towards atoi, since it's used elsewhere in the file, and because of Clemens' comment; additionally, there is less code duplication with the atoi version).

Best regards,

Quoted 13 lines · Clemens Ladisch
Quoted 2 lines · Takashi Sakamoto

When parsing string with non-numeric characters, the behaviour of atoi() is undefined still in C11.

A completely non-numeric string is specified to return zero.

The behaviour is undefined only if the value cannot be represented, i.e., if it overflows.

Regards,
Clemens

#7 lore raw

Friendly ping. Sounds like Clemens mentioned atoi() should be OK (given his explanation regarding non-numeric strings, along with consistency with the rest of the utility).
Are there any other concerns with this patchset?

Thanks once again for taking the time to review.

Best regards,

Quoted 26 lines · Prashant Malani

Interesting. Thanks for pointing that out, Clemens.

In any case, I've sent both versions (atoi v/s strtol), so whichever one is preferred can be used (would be slightly inclined towards atoi, since it's used elsewhere in the file, and because of Clemens' comment; additionally, there is less code duplication with the atoi version).

Best regards,

Quoted 13 lines · Clemens Ladisch
Quoted 2 lines · Takashi Sakamoto

When parsing string with non-numeric characters, the behaviour of atoi() is undefined still in C11.

A completely non-numeric string is specified to return zero.

The behaviour is undefined only if the value cannot be represented, i.e., if it overflows.

Regards,
Clemens

#8 lore raw

Friendly ping.

Quoted 38 lines · Prashant Malani

Friendly ping. Sounds like Clemens mentioned atoi() should be OK (given his explanation regarding non-numeric strings, along with consistency with the rest of the utility).
Are there any other concerns with this patchset?

Thanks once again for taking the time to review.

Best regards,

Quoted 26 lines · Prashant Malani

Interesting. Thanks for pointing that out, Clemens.

In any case, I've sent both versions (atoi v/s strtol), so whichever one is preferred can be used (would be slightly inclined towards atoi, since it's used elsewhere in the file, and because of Clemens' comment; additionally, there is less code duplication with the atoi version).

Best regards,

Quoted 13 lines · Clemens Ladisch
Quoted 2 lines · Takashi Sakamoto

When parsing string with non-numeric characters, the behaviour of atoi() is undefined still in C11.

A completely non-numeric string is specified to return zero.

The behaviour is undefined only if the value cannot be represented, i.e., if it overflows.

Regards,
Clemens

#9 lore raw
Quoted 2 lines · Prashant Malani

Friendly ping.

Since no one raised objection, I applied now.

thanks,

Takashi

Quoted 46 lines
Quoted 5 lines · Prashant Malani

Friendly ping. Sounds like Clemens mentioned atoi() should be OK (given his explanation regarding non-numeric strings, along with consistency with the rest of the utility).
Are there any other concerns with this patchset?

Quoted 33 lines

Thanks once again for taking the time to review.

Best regards,

Quoted 26 lines · Prashant Malani

Interesting. Thanks for pointing that out, Clemens.

In any case, I've sent both versions (atoi v/s strtol), so whichever one is preferred can be used (would be slightly inclined towards atoi, since it's used elsewhere in the file, and because of Clemens' comment; additionally, there is less code duplication with the atoi version).

Best regards,

Quoted 13 lines · Clemens Ladisch
Quoted 2 lines · Takashi Sakamoto

When parsing string with non-numeric characters, the behaviour of atoi() is undefined still in C11.

A completely non-numeric string is specified to return zero.

The behaviour is undefined only if the value cannot be represented, i.e., if it overflows.

Regards,
Clemens

_______________________________________________
Alsa-devel mailing list
[email protected]
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel

#10 lore raw

Thanks Takashi!

Quoted 64 lines · On Mon, Feb 11, 2019, 00:15 Takashi Iwai <[email protected]> wrote:
Quoted 2 lines · Prashant Malani

Friendly ping.

Since no one raised objection, I applied now.

thanks,

Takashi

Quoted 5 lines
Quoted 2 lines · Prashant Malani

Friendly ping. Sounds like Clemens mentioned atoi() should be OK (given his explanation regarding non-numeric strings, along with consistency

with

Quoted 5 lines
Quoted 5 lines

the rest of the utility).
Are there any other concerns with this patchset?

Thanks once again for taking the time to review.

Quoted 8 lines
Quoted 8 lines

Best regards,

Quoted 3 lines · Prashant Malani

Interesting. Thanks for pointing that out, Clemens.

In any case, I've sent both versions (atoi v/s strtol), so whichever

one

Quoted 1 line
Quoted 1 line
Quoted 1 line

is preferred can be used (would be slightly inclined towards atoi,

since

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

it's used elsewhere in the file, and because of Clemens' comment; additionally, there is less code duplication with the atoi version).

Best regards,

On Tue, Jan 29, 2019 at 8:19 AM Clemens Ladisch <[email protected]> wrote:

Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 2 lines · Takashi Sakamoto

When parsing string with non-numeric characters, the behaviour of atoi() is undefined still in C11.

A completely non-numeric string is specified to return zero.

The behaviour is undefined only if the value cannot be represented,

i.e.,

Quoted 12 lines
Quoted 7 lines
Quoted 7 lines
Quoted 6 lines

if it overflows.

Regards,
Clemens

_______________________________________________
Alsa-devel mailing list
[email protected]
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel