Re: [Gnumed-devel] Upgrade from gnumed v1.3.7 to 1.4.

[Karsten Hilbert]

On Sun, Nov 17, 2013 at 02:37:02PM +0100, Hilbert, Sebastian wrote:

> Here is the fingerprint for v18 on PG 8.4 and PG 9.2

> Fingerprinting GNUmed database ...
> 
>                 Name: gnumed_v18
>          Schema hash: a0f9efcabdecfb4ddb6d8c0b69c02092
>                 Size: 56 MB
>             Patients: 14
>             Contacts: 6
>             Episodes: 5
>               Issues: 2
>              Results: 5
>         Vaccinations: 1
>            Documents: 3
>              Objects: 4
>         Version (PG): 8.4.1
>        Encoding (PG): UTF8
>      LC_COLLATE (PG): English, United States
>        LC_CTYPE (PG): English, United States

> Fingerprinting GNUmed database ...
> 
>                 Name: gnumed_v18
>          Schema hash: 499896a73f02668c8b1c7345073d8723
>                 Size: 46 MB
>             Patients: 14
>             Contacts: 6
>             Episodes: 5
>               Issues: 2
>              Results: 5
>         Vaccinations: 1
>            Documents: 3
>              Objects: 4
>         Version (PG): 9.2.3
>        Encoding (PG): UTF8
>      LC_COLLATE (PG): German_Germany.1252
>        LC_CTYPE (PG): German_Germany.1252

Here we can see that the hash is, indeed, different.

Also, LC_COLLATE are different which *might* produce
a different sort order (despite GNUmed sorting on
the HEX representation of the MD5 sum of the object
name, thusly only sorting 0-9 and A-F which *really*
should be stable across LC_COLLATE settings.

A re-run with --with-dump will show the differences
in the clear.

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