Support logging to syslog

[fabled]

It would be beneficial if logging to syslog instead of (or in addition to) files would be supported.

[alexpevzner]

OK, accepted. Thanks a lot, @fabled, for your patience and collaboration!

[fabled]

OK, accepted. Thanks a lot, @fabled, for your patience and collaboration!

Thanks!

I'm actually trying to debug an issue where a HP printer randomly drops the last page of a job. Usually in a duplex printout the last half of the page is empty. The printer errors the job, but it seems that on IPP level (based on Cups backend debug logs) everything seems OK. So I'm trying to figure out if something weird happens on lower USB levels. Or if its just the printer firmware broken.

[fabled]

The printer errors the job, but it seems that on IPP level (based on Cups backend debug logs) everything seems OK. So I'm trying to figure out if something weird happens on lower USB levels.

To clarify. Even if we send the same exact print data, it usually works ok, but randomly this issue will happen.

The exact data sent to printer is not identical because Go http will rework the chunked transfer-encoding chunks. Any idea if it would be easily possibly to change so that the HTTP body chunks sent are bit identical?

[alexpevzner]

Changing this is not easy, and I actually depend on a fact that chunks sent to the printer are independent from the chunks received from the client, because it allows me to always finish the HTTP transaction correctly, even if client drops the connection in the middle and without need to buffer the entire client's data before forwarding to the printer.

However, data actually sent to the printer (at the real USB level) can be logged by enabling the trace-usb log level.

[alexpevzner]

You may play a little bit with the ipp-usb predecessor, the ippusbxd program (https://github.com/OpenPrinting/ippusbxd/).

Unlike ipp-usb, this program just relays incoming TCP connection to the USB, without any modifications or interpretation.