arcctgx / tagreaders Goto Github PK

Something like:

> csvscrob -t now <files ...>

Tracks shorter than 30 seconds are not supposed to be scrobbled, and they are ignored. But current time should advance by the length of this track. This is not the case now:

> csvscrob -t 0:00:00 12_you_suffer.ogg 01_continue.ogg 
track too short (4s), skipping.
"Lull", "Continue", "Continue", "2017-04-30 00:04:00", "", "3720"

The 2nd track should be scrobbled at 00:04:04.

last.fm uses the following logic to determine if the track will be scrobbled:

• if track is less than 8 minutes long, then scrobble past half of track
• if track is 8 minutes long or longer, scrobble past 4 minutes

But this logic is used only to determine whether the track will be scrobbled. If the track is scrobbled, the timestamp of playback start is used.

Currently csvscrob uses time at half of the track (or at 4 minutes into the track if it's long) as timestamp, which is incorrect. Since we assume all given tracks must be scrobbled, we shouldn't care about the logic described above.

Double quotes embedded in the titles should be escaped by doubling them, e.g.:
Theme From "To Kill a Dead Man" becomes Theme From ""To Kill a Dead Man""

Perhaps UTF-8 output should remain the default, but there could be a command line switch for disabling it.

> csvscrob -t "2017-04-29 23:58:45" 01_stream_endless.ogg 
"Lull", "Stream Endless", "Dreamt About Dreaming", "2017-04-30 01:02:45", "", "709

This is correctly scrobbled after 4 minutes, but for some reason one hour is added to the time.