[Abcde-users] questions about cddb files

[Steve McIntyre]

Hi Dan!

On Fri, May 21, 2021 at 08:24:30PM -0500, Dan Hargens wrote:
>I've been using abcde for quite a while, and enjoy using it. Thanks for taking
>the time to make it in the first place. I've used it to rip 1192 CDs and I'm
>still going.

*grin* Cool!

>2 concerns: I've been trying to understand something about how abcde uses the
>cddb files.
>
>1.
>A Long Time Ago the filenames used to be long - 28-characters, the same as the
>"DISCID=" line inside the file. I think maybe it's the MusicBrainz format.
>More recently, they're now the short form (8 chars), the same as the first
>field returned from cd-discid.
>
>Why is this? Was it related to when freedb dropped?

No, not so much. When I added the MusicBrainz support initially, we
would only support one lookup mechanism at a time. I copied the idea
of using the "cddb" ID as a filename for uniqueness, but of course
MusicBrainz uses a different format - the 28-character hash instead of
the 8-character  So I used that.

Then I decided to rationalise the lookup options and allow using all
of them (cddb, MusicBrainz and CD-Text) for lookups on any given
run. At that point I needed to use a unique CD ID that would work with
all the lookup options, and for the sake of convenience I simply
picked the 8-character cddb-style ID. I hope that explains!

>2.
>As part of my process of automating rips, I wrote a script that uses abcde to
>first only retrieve the disc info (cddb), which it then cleans up and changes
>smart-quotes to stupid ones, fixes weird album titles and other things, then
>does the full rip using abcde again (this saves me from having to re-rip it
>because I don't like the defaults or something). Sometimes the filename is NOT
>the first field returned from cd-discid, but something else (but still 8 chars
>long).
>
>Why is this? How do I know what the 'new' filename is that I need to use? To
>clean up the file, I assume the cd-discid is right and use that. That doesn't
>work when the name changes. I suppose I could use the most-recent file in the
>directory, but that assumes I'm not editing other files for a re-rip (maybe I
>needed to manually change the title, or some other such reason), but this feels
>sloppy somehow. Any suggestions on how to handle this?

Oh, ick. That sounds odd, and almost definitely a bug! The major
points for using cddb ID is that it's meant to be (A) unique (ish) and
(B) stable. If you're seeing it change, then point B here is
broken. If you can reproduce this, please file a bug with some debug
logs!

-- 
Steve McIntyre, Cambridge, UK.                                steve at einval.com
Who needs computer imagery when you've got Brian Blessed?