input: vcd: skip BOM at beginning of file

[libsigrok.git] / src / hardware / hp-3457a / protocol.c

diff --git a/src/hardware/hp-3457a/protocol.c b/src/hardware/hp-3457a/protocol.c
index 320d8d0abfe114c657ebbf627809365b8a80c9dc..618a99cd4cc324379310b8a8818bf450ccb0de90 100644 (file)
--- a/src/hardware/hp-3457a/protocol.c
+++ b/src/hardware/hp-3457a/protocol.c
@@ -216,9 +216,17 @@ static int calculate_num_zero_digits(double measurement, double range)
        return zero_digits;
 }
 
+/*
+ * Until the output modules understand double precision data, we need to send
+ * the measurement as floats instead of doubles, hence, the dance with
+ * measurement_workaround double to float conversion.
+ * See bug #779 for details.
+ * The workaround should be removed once the output modules are fixed.
+ */
 static void acq_send_measurement(struct sr_dev_inst *sdi)
 {
        double hires_measurement;
+       float measurement_workaround;
        int zero_digits, num_digits;
        struct sr_datafeed_packet packet;
        struct sr_datafeed_analog analog;
@@ -241,12 +249,13 @@ static void acq_send_measurement(struct sr_dev_inst *sdi)
        packet.payload = &analog;
 
        sr_analog_init(&analog, &encoding, &meaning, &spec, num_digits);
-       encoding.unitsize = sizeof(double);
+       encoding.unitsize = sizeof(float);
 
        meaning.channels = sdi->channels;
 
+       measurement_workaround = hires_measurement;
        analog.num_samples = 1;
-       analog.data = &hires_measurement;
+       analog.data = &measurement_workaround;
 
        meaning.mq = devc->measurement_mq;
        meaning.unit = devc->measurement_unit;