[mythtv] [PATCH] Multi Artist patch for MythMusic - I'd call it RC1 by now...

[Colin Guthrie]

Hi Thor,

thor wrote:
> This is great progress. Some thoughts though:

Thanks!

> You seem to have commented out existing db updates in dbcheck.cpp and then 
> used the same version number as the updates you commented out for your own 
> updates. So ... anyone running cvs (current) does not get the updates 
> applied, which results in the table being older than the code expects, which 
> results in music scanning that finds no tracks available.

Yeah, I wasn't sure why that code was there as the version number was 
less than it! I now see you iTunes stuff will use it in the future, so 
see the reasoning. I don't know how the whole Schema version stuff works 
with CVS but I take it that there has to be a little bit of manual DB 
tweaking when working on CVS as I guess the schema should only really 
change on actual public releases (e.g. 0.16)?

Am I right in saying I should probably just integrate my shema changes 
with yours and keep the version number the same?

It does still seem odd that the schema changes for an upcoming patch are 
already in dbcheck.cpp CVS. Would this not normally be part of the patch?

> The CTRL-Right thing to move around the ripping table is a bit of a mess, 
> although mitigated by your swap button.

Yeah, some FreeDB entrys get things the wrong way round hense the swap 
button. I don't know how to make the table navigation better, but I 
reckon there will be a way to get it to work with just left+right keys. 
Up/down work afterall... So I can try and see if I can get this better.

> The main shortcoming of the code as it stands, as you've already pointed out, 
> is that people who rip to mp3's won't have the albumartist information 
> preserved in the file itself. Would be awfully nice to fix that, although I 
> know it means a lot of work for getting decent id3v2 tag code in there.

I can do this. Although I don't like MP3 format for quality/patent 
reasons, it shouldn't take too long to replace this stuff. I think the 
id3lib stuff is a dependancy anyway, so no additional packages should be 
needed.

> On a more general thought, the dbcheck code you commented out is stuff I'm 
> actively working on for mfd/mfe.

Oops ;)

> These columns/metadata fields are exactly 
> the same ones that iTunes uses internally. More importantly, they are also 
> the same fields that the daap protocol (e.g. iTunes music sharing) uses. The 
> mfd uses (well, will use) daap to automagically share music content between 
> myth boxes. It's important, therefore, to try and keep daap compatability in 
> mind when thinking about new metadata fields.
> 
> One of these columns is a bool called compilation. That should almost 
> certainly be used by any album-artist type code (and, obviously, set to true 
> if the album is a compilation). That way, any display and/or wire-protocol 
> logic can do a bool check independent of whether albumartist is or is not 
> set/equal to track artist, etc.
> 
> Thoughts?

Yeah, that sounds like a very sensible idea. The compilation bool can be 
incorporated into the stuff I'm doing.

How about I look at your Schema changes and try and piggy back onto 
those fields some more and especially use the compilation flag more. 
Would it make more sense if I used the "composer" field to indicate the 
Album Artist rather than my own field? In ID3v2 there is a separate 
"Composer" field, TCOM which could be checked too and used in preference 
to the TPE4 field... It may be better to keep this separate tho' as it 
is different, just quite close...

What do you think the best approach is with regard to adding the 
album_artist field or just hijaking the composer field? I think I'm 
slightly favouring the separate field method when thinking about it.

I guess in the Classical music genre, the composer and arranger are very 
separate. Most of the compilaions I have, they are arranged by someone, 
but the generic "Various Artists" or "Soundtrack" album artists I tend 
to use, don't really count as "arrangers" (but I guess nothing can ever 
be logically the correct name to use in every circumstance!!)

Also, I take it your patch probably changes a lot of the same files ;) 
This may make it a litte awkward to apply both! If there is anything I 
can do to make things simpler then let me know (e.g. a separate branch 
or separate CVS or something?)

Col.


-- 

+------------------------+
|     Colin Guthrie      |
+------------------------+
|   myth at colin.guthr.ie  |
| http://colin.guthr.ie/ |
+------------------------+