[Refdb-users] ANN: refdb-0.9.4-pre4 available for testing

[Markus H. <mar...@mh...>]

Hi all,

there's a new prerelease available at:

http://refdb.sourceforge.net/pre/refdb-latest.tar.gz

We're getting closer to 0.9.4, but there were some substantial changes
compared to the previous prerelease which make another pre seem like a
good idea.

This release requires the brand-new libdbi-0.7.2 and
libdbi-drivers-0.7.1 versions released last weekend. refdbd gained an
additional dependency as it now requires libiconv.

Besides fixing a couple of bugs and memory leaks, this prerelease
improves the character encoding issue. As Marc has found out, the
previous releases worked ok for risx data only if you happened to use
UTF-8 anyway. In all other cases RefDB still worked, but you might
have gotten an incorrect sorting order in the bibliography in border
cases.

The character encoding stuff now works like this (this will go into
the manual eventually):

- you select a suitable character encoding for your
  database. PostgreSQL allows to select an encoding per database, so
  you can use a suitable one by running createdb -E encoding
  dbname. MySQL allows only one encoding per server instance. Set this
  on the mysqld command line or in the config file. SQLite allows only
  two encodings as a compile-time option.

- RIS input can use any encoding that your local libiconv
  understands. If it is not the default (UTF-8), you have to specify
  the input encoding using the addref/updateref -E switch (or set
  fromencoding in refdbcrc accordingly)

- risx and xnote input can use one of these encodings: UTF-8, UTF-16,
  ISO-8859-1, US-ASCII. The encoding must be declared in the
  processing instruction (the first line of the risx file) like this:
  <?xml version="1.0"? encoding="UTF-16">
  unless it is the default (UTF-8).
  If you have to use a different encoding, use the iconv command line
  utility to convert the data to one of the acceptable encodings
  before sending the data to refdbd.

- getref and getnote send back the data in the encoding used by the
  database unless you request a different encoding with the -E
  switch. You can request any output encoding that your local libiconv
  supports. The same applies to the refdbib tool.

- the refdbc:whichdb shows the character encoding of the current
  database.

regards,
Markus


-- 
Markus Hoenicka
mar...@ca...
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de