srd: Performance improvements for various PDs.

Ignore/skip identical samples in most (low-level) PDs, as we're usually
(but not necessarily always) only interested in pin changes.

This yields a significant performance improvement for the PDs.

The mechanism was already used in the 'i2s', 'jtag', and 'lpc' PDs, but not
yet in all supported low-level decoders. The following PDs now also use
this mechanism: 'dcf77', 'i2c', 'spi', 'uart', and 'usb_signalling'.

Thanks Lars-Peter Clausen <lars@metafoo.de> for bringing this to our
attention.

1 files changed, 7 insertions, 1 deletions

diff --git a/decoders/dcf77/dcf77.py b/decoders/dcf77/dcf77.py
index 4da9bc4..c4132e9 100644
--- a/decoders/dcf77/dcf77.py
+++ b/decoders/dcf77/dcf77.py
@@ -50,6 +50,7 @@ class Decoder(srd.Decoder):
 
     def __init__(self, **kwargs):
         self.state = 'WAIT FOR RISING EDGE'
+        self.oldpins = None
         self.oldval = None
         self.oldpon = None
         self.samplenum = 0
@@ -197,7 +198,12 @@ class Decoder(srd.Decoder):
             raise Exception('Invalid DCF77 bit: %d' % c)
 
     def decode(self, ss, es, data):
-        for (self.samplenum, (val, pon)) in data:
+        for (self.samplenum, pins) in data:
+
+            # Ignore identical samples early on (for performance reasons).
+            if self.oldpins == pins:
+                continue
+            self.oldpins, (val, pon) = pins, pins
 
             # Always remember the old PON state.
             if self.oldpon != pon: