Sigh ... 

Almost everything you told us, in flamebait and now more calmly, is
already known.

You could have found this out yourself with a little light reading.

Slimserver is written in perl. Whether you are a perl fan or not,
that's the language it's written in. Warts and all.

perl did not acquire threads until recently, and not all
implementations for all machines work equally well, so the developers
have stayed away from multi-threading until very recently. The entire
slimserver has it's own internal dispatcher so that it can service
music requests and display requests and handle remote control button
presses while scanning and dishing out web pages at the web interface.

[edit] the two threads you see in windows - one is a perl windows
implementation thread that does almost nothing - I think it waits for
sigint or something. It is not a perl second thread. Slimserver is at
the moment single threaded. In 6.5 they're getting their toe in the
water with a second thread. We'll see how that goes.

It is true that (some) file systems have a way of notifying you about
file updates, but the notification method is not uniform across file
systems. Moreover, Slimserver has to work (and does, to a remarkable
extent, it does succeed) across multiple systems and across networks.
So the file store may be on a linux box, but slimserver may be running
on windows. Or vice versa. The problem is complex in it's generality
and not, as I understand it, handled at all well by perl. So the
application would have to get down and dirty in system details, not a
pleasant task.

There is code in slimserver, going back to the early days of release 6,
to catch and detect most common scanning loops. These loops are more
subtle than you might think, because some music file formats spread
themselves across multiple files. For instance, one file format records
the entire album as a monolith and then generates a second file called a
cue sheet. The idea is to prevent gaps between tracks if you don't want
them. But the 2 file format wasn't as well thought out as it might have
been (by the designers of the format, not Slim) and it is possible to
have various out-of-sync problems.

But, basically, what you described as happening shouldn't. 

Which is why we asked to see your logs, to try to get to the bottom of
what's happened, to improve the product so it won't happen again.

You wrote:
> any program running on a desktop computer (in the background) that
> consumes 99% cpu (slimserver does this when working properly too) 
No it doesn't.

If it's consuming 99% of the cpu, for long enough for you to casually
notice, it's not running properly at all.

I run slimserver 24/7 and have done so for about a year now. It runs on
my main computer that also runs AutoCAD Mechanical Desktop (when you
look up pig in the dictionary, that's what you find). Slim may use the
CPU for 10 seconds at a time, in response to a query, but that's it.

And when scanning, once in a blue moon, it is rather more busy for 15
minutes to half an hour or so, but even then you have plenty of
computer left over, at least I do on my 1.8GHz machine.

So *if* that's not your experience, and *if* you actually want to get
to the bottom of it rather than complaining and
*shouting-but-not-for-help in a very loud voice*, turn on the debugging
options and show us what your log shows.

Or not.

Your choice.


-- 
Michaelwagner
------------------------------------------------------------------------
Michaelwagner's Profile: http://forums.slimdevices.com/member.php?userid=428
View this thread: http://forums.slimdevices.com/showthread.php?t=26350

_______________________________________________
discuss mailing list
[email protected]
http://lists.slimdevices.com/lists/listinfo/discuss

Reply via email to