Compiling LMS

[dolodobendan]

Hello Michael, I had a look at library.db again. Do I get this right:

Every browsable menu level (album, year, genre) has its own table (with the exception of the multipurposed contributor. Why?)

And another question:

The Custom Scan tables are in persist.db in one huge table. Loading the Custom Browse menus sometimes(?) is very slow, I might even get timeouts (what would speed it up: a faster disk, more RAM, a faster CPU?). Is this because the table is not split into "entry points" like the native tags but is just one large list?

[mherger]

> Every browsable menu level (album, year, genre) has its own table (with
> the exception of the multipurposed contributor. Why?)

Every entity and relation has its own table. Some of those entities
obviously then are used to create the menus. But it's not "hey, we want
a new menu, we need to change the database schema".

> The Custom Scan tables are in persist.db in one huge table. Loading the

This would be a question for Erland.

--

Michael

[dolodobendan]

Every entity and relation has its own table. Some of those entities
obviously then are used to create the menus. But it's not "hey, we want
a new menu, we need to change the database schema".

Perfect, thanks!

This would be a question for Erland.

I meant in general, but ok.

[mherger]

> The Custom Scan tables are in persist.db in one huge table. Loading the
> Custom Browse menus sometimes(?) is very slow, I might even get timeouts
> (what would speed it up: a faster disk, more RAM, a faster CPU?). Is
> this because the table is not split into "entry points" like the native
> tags but is just one large list?

Oh, I guess I didn't read to the end - stopped after reading "Custom
Scan" twice :-D.

It's hard to tell what the bottleneck is there. It could be crazy
complex queries, lack of indices, or resources. You'd have to run some
analysis to see what's going on.

(Database analysis isn't easy. In my previous job we once flew in a DB
specialist from IBM, because an application didn't run on the 250k USD
iSeries server, while it performed nicely on the dev's cheap Windows
machine. After two days of looking into it he came to the conclusion of
"can't you do less queries?")

--

Michael

[dolodobendan]

Oh, I guess I didn't read to the end - stopped after reading "Custom
Scan" twice :-D.

Don't we all.

It's hard to tell what the bottleneck is there. It could be crazy
complex queries, lack of indices, or resources. You'd have to run some
analysis to see what's going on.

(Database analysis isn't easy. In my previous job we once flew in a DB
specialist from IBM, because an application didn't run on the 250k USD
iSeries server, while it performed nicely on the dev's cheap Windows
machine. After two days of looking into it he came to the conclusion of
"can't you do less queries?")

Yeah, I don't wanna (or could) do that. Thanks!