bsl / send-sds Goto Github PK

Hi, I see there's no license file in the repo, so if I fork it to make any improvements I'm violating your copyright, which I assume is not your intention.

See https://help.github.com/articles/licensing-a-repository/ for more explanation, and https://help.github.com/articles/adding-a-license-to-a-repository/ for the easiest way to add one to a Github repo.

In the method send_file, when get_response returns WAIT -- the packet number is increased regardless and the next packet is sent. The unit is then one packet behind and the ACK that is received from it is then never handled correctly by the utility.

I ran into two issues with my E-my ESi-32 on firmware 2.10, both which seem to confuse the utility and break sending of samples. Of course the problem lies with the hardware unit, but it can be solved in software.

The issues that I observed are:

• When initiating a MIDI Sample Dump, it sends a Time Signature packet that confuses the utility. The length of this Time Signature packet is variable, yet always larger than SDS_RESPONSE_LENGTH, so that breaks the utility's logic.
• The unit sends corrupted SysEx packets (I observe multiple Start-Of-SysEx bytes, like F0...F0...F7). This issue can be mitigated by dropped what was read so far upon detecting a new F0 (SOX) while waiting for an F0 (EOX)

The utility blocks for no apparent reason when the MIDI device is in use by another process. The culprit seems to be line 30 in midi.c:
r = snd_rawmidi_open(&handle_in, &handle_out, device, 0);

If SND_RAWMIDI_NONBLOCK is set as the flags, then the utility properly exits with a human readable error:
r = snd_rawmidi_open(&handle_in, &handle_out, device, SND_RAWMIDI_NONBLOCK );