Installing and Using Unicode

From VistApedia
Revision as of 15:06, 4 February 2011 by Ivaldes1 (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

No automated installer for this yet. Here are Sam's instructions:

I have done this like 10-20 times. Here is how I generally proceed. I
rather exclusively use Debian Systems everywhere I go (I fell into a
habit), so my instructions are geared towards them. We ran Unicode
GT.M on CentOS here; and it wasn't much different from what I recall.

Add these to your environment file:

export gtm_chset="UTF-8" (try to run it now--it will complain about libicuio)
export gtm_icu_version="4.2" (or whichever one you have) (try now:
will still complain about libicuio)
Use strace against mumps -direct, and compare where GT.M looks with
locate's output. Make the symlink. (try now. It should work. But no
routines will.)

Change gtm_dist to use the utf8 compiled GT.M system routines
export gtm_dist=<gtmpath>/utf8
Delete all of the VISTA object files. They will be regenerated as
needed by VISTA.

If you have any characters outside of ASCII in the database (a clean
one should have none), they are illegal UTF-8 and need to be
converted. Run the globals through iconv to convert them. That needs a
tiny bit of programming.

You need to restart xinetd. Sometimes it gives me trouble and I end up
running it in the foreground (xinetd -d) to examine what it is
complaining about.

Now you have a UTF-8 Mumps database. Medsphere's OpenVista RPC (with
some modifications to send UTF-8 instead of ASCII) and my port of the
BMX 2 broker will communicate fine with Mumps in all of the world's
languages.

Sam