add crc error message

[devZer0]

example message from devZer0/libsml-testing#13

cat eBZ_DD3_2R06.bin | ./sml_server -

libsml: error: CRC checksum mismatch
libsml: error: CRC checksum mismatch

also see discussion there

[jahir]

libs should only print errors to stderr as a last resort. but as it is already used (occasionally), it seems ok. AFAIK, there is currently no better way to get errors out of libsml... (#93 is WIP)

So if nobody objects, this should be merged.

[r00t-]

this seems to relate to #103

[r00t-]

@devZer0:
good idea (it seems i missed this MR before!), just a few details:

your commit has the username set to root, can you fix this?
(git config --global --add user.email "..." --add user.name="..." git commit --amend --reset-author git push --force)

the formatting isn't ideal, as it now seems the comment "Workaround for Holley DTZ541"
applies to the whole block, and not just the the condition.

also imho the message could be more helpful.

i would write this like:

	if (
		*msg->crc != sml_crc16_calculate(&(buf->buffer[msg_start]), len)) &&
		// Workaround for Holley DTZ541, uses CRC-16/Kermit
		*msg->crc != sml_crc16kermit_calculate(&(buf->buffer[msg_start]), len)
	){
		fprintf(stderr, "libsml: sml_message_parse(): crc mismatch, dropping message\n");
		goto error;
	}

[mbehr1]

I was thinking whether returning e.g. a "fake" sml reading that maps to obis error codes like S.1.1.16 (device status) S.1.1.7 (failed readout attempts) would make it easier processable.

[r00t-]

@mbehr1:

I was thinking whether returning e.g. a "fake" sml reading ...

we shouldn't be doing such a thing.
we simply need to extend the API, if the current one doesn't fit our requirements.
that should even still be backwards compatible for code that does not need the new features - same as for #93

[r00t-]

if @devZer0 doesn't mind or doesn't respond, we can just merge #114 instead.

[devZer0]

thanks for reminder, i'm totally busy at the moment, so i'm fine if you drop my commit or replace with something other/better

[r00t-]

@devZer0: thanks!

closing as superseded by #114