In our logs, we have lots of lines like the following:

Thu Jun  5 21:56:16 2014: ---- begin
Thu Jun  5 21:56:16 2014: Loading document...
Thu Jun  5 21:56:16 2014: Applying style sheets...
Thu Jun  5 21:56:16 2014: Preparing document...
Thu Jun  5 21:56:16 2014: Converting document...
Thu Jun  5 21:56:16 2014: finished: success
Thu Jun  5 21:56:16 2014: ---- end

...which are entirely useless to us, because the logfile is shared by many Prince processes, and we have no way to tell which process is responsible for which "begin/end" block. A process ID on each line would actually make such a combined log useful.

Thanks!

Is it that the messages are interleaved, and the process ID would enable you to deinterleave them? Or do you record the process ID elsewhere, so it allows you to correlate error messages with specific jobs? I'm wondering if there are also other ways in which we could make the logging more convenient.

The latter. For instance, sometimes we get Prince jobs hanging for long periods of time, and they need to be manually killed. Trying to find the relevant process' log messages is... involved.

Do you have the input document that causes Prince to hang? That is definitely something we would like to investigate. If the process is hanging indefinitely, then I suspect that there might not be useful log messages to find.

Further investigation suggests that it's trying to fetch an image from a broken remote server that's leaving the connection open and just not returning any data. Without an explicit --http-timeout set, Prince appears to block forever on the HTTP request. (What's up with that? Shouldn't there be a reasonable default timeout, instead of none at all?)

Good point, we will change the default to 60 seconds.

This default has been changed in Prince 10. The old (bad?) behaviour can be obtained by passing --no-http-timeout on the command-line.