One of the manuals I maintain in my day job is in groff, and (to my
surprise) I had to have it translated. The translators managed to not
muck up things *too* badly -- there were a few lines that got run
together in error and a couple macro and cross-ref names that got
translated, but they know to watch for that next time now.
Since I'm not maintaining the translations, I ran the files through
preconv to fix the accented glyphs permanently (some of the files
were Latin-1 and others UTF-8, which gave preconv heartburn when run
using groff -k). Indexing, of course, was interesting. Here's a patch
to the Bentley/Kernighan indexing scripts (CSTR 128) that increases
preconv-awareness:
--- gen.key.old 2009-03-31 08:30:39.000000000 -0400
+++ gen.key 2009-03-31 12:35:41.000000000 -0400
@@ -21,6 +21,16 @@
# Remove some troff commands
gsub(/\\f\(..|\\f.|\\s[+-][0-9]|\\s[0-9][0-9]?/, "", $2)
+ # de-accent Latin characters (assuming preconv for now)
+ gsub(/\\\[u00E0\]/, "a", $2);
+ gsub(/\\\[u00E1\]/, "a", $2);
+ gsub(/\\\[u00E2\]/, "a", $2);
+ gsub(/\\\[u00E8\]/, "e", $2);
+ gsub(/\\\[u00E9\]/, "e", $2);
+ gsub(/\\\[u00ED\]/, "i", $2);
+ gsub(/\\\[u00F3\]/, "o", $2);
+ gsub(/\\\[u00F1\]/, "n", $2);
+
# underscore -> 0, so "foo_gorp" sorts before "food"
gsub(/_/, "0", $2)
--- format.old 2006-05-23 10:30:49.000000000 -0400
+++ format 2009-03-31 11:09:05.000000000 -0400
@@ -35,6 +35,9 @@
else
$0 = $2
+ # preserve Unicode glyphs (\[u....] - preconv awareness)
+ gsub( /(\[u....\])/, "%\&%", $0 );
+
if ($0 ~ /%/) {
quoted = 1
The gen.key patch probably needs to be filled out a little more, but
this was enough to get my project properly sorted (pun not intended).
Aside from the few glitches, most of which are a matter of
familiarity (or lack thereof), the project went reasonably smoothly.
I'm sure the hyphenation is a little wonky... there's a hyphen.fr
file (which I assume could be used for proper French hyphenation, if
installed properly). Is there anywhere I could grab a Spanish
hyphenation file?
-- Larry
