Hmm that's wrong, but I bet it's fixed in 7.5 embedded, where I completely rewrote the way covers work.

Can I run 7.5 embedded on a "normal" Linux System? Wil this be backported into the main branch?

Yeah embedded can be run everywhere.

Not on my Ubuntu server 8.04 LTS, see the forum thread, probably OT here :)

J?rg - what exact query are you using?

status - 1 tags:uBJjKlax subscribe:42

Thanks - confirmed:

- connect to the server on the CLI
- create a status listener using the command in comment 6
- play some new music using BMF

-> the tracks.cover column instead of an id now has the path to the cover file

It's broken in embedded?  tracks.cover is supposed to contain a full path.  tracks.coverid has the id for display.  The tag for coverid is 'c'.

Andy,

a) why change these things with every release? In 7.4 "coverart" contains an id, why not add "coverartpath" or something instead of changing the meaning of one field and introducing a new one having the old meaning of the first one?

b) the biggest issue is that the path is invalid. coverart contains an absolute path on the machine the server runs on which is not even URL escaped; I did not find a way to retrieve a cover using that path.

Because artwork is fundamentally broken in current releases, the main issue is the use of track IDs which can change after rescan and this causes cache issues for browsers, SP, iPeng, etc.  Now the id is a hex checksum based on the URL, filesize, and some other stuff thus fixing the cache problems.

Ah. This is good.
But I still don't understand why not to use that in "coverart" instead of "coverartid"?

MY big issue here is that my code for displaying a cover already today looks like this:
	if (item) {
		lArtist.text = (temp) ? temp : [iPengAppAppDelegate globalDictString:@"REMOTE_STREAMING" andDefault:@"Remote-Stream"];
		int ca = 0;
		if ([(temp = [item objectForKey:@"coverart"]) respondsToSelector:@selector(intValue)])
			ca = [temp intValue];
		if (ca)
			path = [NSString stringWithFormat:PATH_ARTWORK_COVER(picSize), [item objectForKey:@"id"]];
		else if ([item objectForKey:@"artwork_track_id"])
			path = [NSString stringWithFormat:PATH_ARTWORK_COVER(picSize), [item objectForKey:@"artwork_track_id"]];
		else if ([item objectForKey:@"artwork_url"])
			path = [item objectForKey:@"artwork_url"];
		else if ([item objectForKey:@"remote"]) {
			artwork.image = [UIImage imageNamed:@"radio.png"];
			mirrorImage.image = nil;
			return;
		} else {
			artwork.image = [[UIImage imageNamed:@"cover.JPG"] stretchableImageWithLeftCapWidth:0.0 topCapHeight:0.0];
			mirrorImage.image = nil;
			return;
		}
	} else {
		lArtist.text = nil;
		artwork.image = [[UIImage imageNamed:@"coverDark.png"] stretchableImageWithLeftCapWidth:0.0 topCapHeight:0.0];
		mirrorImage.image = nil;
		if (player.artwork)
			player.artwork = nil;
		return;
	}

Now for 7.5 I have to add a few more exceptions and special cases. But I can't drop any of the old ones because I have to support other server versions as well.

What is so difficult about using new fields for new information and keeping old ones.
I don't care what _type_ the coverart field is, it may as well be a url, but I want to be able to retieve it from the server in the same way I could retrieve it before, which is not the case with something containing unescaped spaces and slashes.

Andy - what needs to be fixed here? I understand it you think it's a "works as designed" - why a P1 then?

Because _in addition_ to the design change in the case above the URL is also wrong.

 == Auto-comment from SVN commit #30145 to the  repo by agrundman ==
 == https://svn.slimdevices.com/?view=revision&revision=30145 ==

Fixed bug 15204, coverArtExists method was returning the value of the cover column instead of 1

The problem here was just that the 'j' tag was using coverArtExists() which was returning the actual cover column value.  This bug wasn't just related to BMF, but any track that contained a cover.jpg image.

BTW, don't worry about embedded, cover art works great with iPeng under embedded.  The changes I made are internal and don't affect the CLI.

Thanks for fixing this :)
I'll give it a try with the next firmware on the touch.
Still can't get embedded to run on my main server, although I do now have another one to try it out on.

This bug has been marked fixed in a released version of Squeezebox Server or the accompanying firmware or mysqueezebox.com release.

If you are still seeing this issue, please let us know!