Re: [Gnumed-devel] Data pack logging and debugging imported medications

[Karsten Hilbert]

On Mon, Dec 05, 2011 at 06:53:53AM +0000, Jim Busser wrote:

> Is there any way now to no longer have to include all of the above in GNUmed, 
> so that we can instead require people who wish to use them to load them 
> themselves? How would this work?
> 
> - the bootstrapper script would (effective with database v17) remove them 
> where a database had been bootstrapped from scratch, but would leave them 
> alone where a user was upgrading a (possibly production) database?
>
> - would in production databases need special data packs just to remove 
> unwanted language data which had formed part of earlier GNUmed releases?

That would be the proper approach.

> ===============================================
> 
> Is there programmatic significance to the 'lang = fr'

Not that I remember.

> -- how would it have acquired the language 'fr' despite that my machine is 
> en_CA and FreeDiams provides the Canadian brands in *english* ?

By bootstrapping.

> Is the running of data packs anywhere recorded in GNUmed?

Yes, in the log file.

> I do not seem to see any entries in
> 
>       gm.access_log
> 
> that would correspond to when I did run some data packs.

That's right.

Individual data packs could simply use the

        select gm.log_script_insertion(filename, version);

which would at least record the event in gm.schema_revision.

OTOH, that's rather voluntary on the side of the data pack
creators and it seems useful to "force" this on them
courtesy of GNUmed code *running* the data pack - which I'll
add.

Now, the gm.log_script_insertion() has its own bug which
I'll fix which will properly document the insertion in the
access log.

Karsten
-- 
GPG key ID E4071346 @ gpg-keyserver.de
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346