> head I think; I just checked it out using cvs. Syan, this is fixed. For how and why this is safe see the comments at the top of sql/gmBlobViews.sql.
BTW, just doing a redo-v2.sh from a HEAD branch won't do you much good. I mean, it will indeed setup a v2 database usable with a 0.2.x client (where x < 3) but the same CVS tree does not contain a corresponding client. If you want to setup a 0.2 client you should checkout the latest revision of the rel-0-2-patches branch and go from there. The redo-v2.sh script in the HEAD branch is intended to setup the baseline database ready for transformation into v3. The procedure is as follows: To setup a *new* v3 database do this: - from HEAD branch - run redo-v2.sh - run update-v2_v3.sh To transform an *existing* v2 database into v3 do this: - verify (via checksum) that you have a v2 database - from HEAD branch - run update-v2_v3.sh Later on this procedure won't change, it'll only ever get extended: - make/have v2 - update-v2_v3.sh - update-v3_v4.sh - ... That way we can guarantuee a safe and maintainable upgrade path. Due to using versioned database names (gnumed_v2, gnumed_v3, gnumed_vX) we can also guarantuee entirely non-desctructive, always-fallback upgrades. Karsten -- GPG key ID E4071346 @ wwwkeys.pgp.net E167 67FD A291 2BEA 73BD 4537 78B9 A9F9 E407 1346 _______________________________________________ Gnumed-devel mailing list [email protected] http://lists.gnu.org/mailman/listinfo/gnumed-devel
