bcd: initial import of binary coded decimal en-/decoder #7311
Conversation
miri64
added
the
Type: new feature
miri64
added
the
CI: ready for build
| static void test_bcd_to_byte__greater_0x99(void) | ||
| { | ||
| TEST_ASSERT_EQUAL_INT(100, bcd_to_byte(0xa0)); | ||
| TEST_ASSERT_EQUAL_INT(100, bcd_to_byte(0xa0)); |
There was a problem hiding this comment.
duplicate test for 100, and shouldn't this be the opposite version of test_bcd_from_byte__greater_99 - meaning both should test same values?!
There was a problem hiding this comment.
Removed the duplicate line, but they don't have the same value, since they are out of range => garbled "input", which is the whole point of this test.
|
(in fact the output of |
|
interesting encoding, but which driver is (seriously) using this - and why 😕 ❓ |
|
ah oversaw your reference ... |
Among others (like the DS1307) the x86 RTC |
|
The mensa-cards at FU also store the current account balance as a BCD, IIRC. |
|
As to why according to my quick research mainly a) legacy reasons (see x86 RTC) and simplified ASCII representation. Since the DS1307 is mainly used by the Arduino community I guess the latter applies for this device. |
|
(not only ASCII, but also e.g. 7-segment displays). |
that sounds reasonable, indeed! Wasn't aware that it might still have a good use-case ... mostly read articles stating it is legacy stuff and "not supported by modern CPUs" bla bla ... |
| */ | ||
| static inline uint8_t bcd_from_byte(uint8_t byte) | ||
| { | ||
| return byte + (6 * (byte / 10)); |
There was a problem hiding this comment.
Though this is (more) elegant, it's not as intuitive as ((byte / 10) << 4) | (byte % 10), which makes the conversion IMHO quite explicit. I don't suggest to change this, but it might be worth a comment (or example)?
There was a problem hiding this comment.
Not only is it more elegant, but also more efficient ;-), because you only have one costly division op (counting mod as a division).
|
please squash, too |
|
Squashed |
This imports a simple BCD en-/decoder. I require that for a driver I'm currently porting.