*** WARNING ***
A new version of the spellchecker has just been committed:
this document is out of date, but can still be used if you really
need to know what's goin' on here.
I'll update it ASAP: promise!
*** WARNING ***


The following files should help translators to find typos in po files:

check_dit.sh 
check_all.sh
lang2dict.txt
msgstr_extract.awk

In order to be able to use all the scripts above you need the following packages:
gnuplot
aspell
aspell-bin
aspell-${DICT}

where ${DICT} is the aspell-suffix (see the second column of lang2dict.txt)


Run the following command in order to spellcheck the languages listed
in lang2dict.txt:

	 sh check_all.sh <D-I repository>


This will create a subdirectory "check_${TODAY}" containing various
files organized as described later in this document.

check_dit.sh 
	Usage: check_dit <lang> <aspell-suffix> <D-I repository> <out dir>

	is the "heart" of the project: it spell checks one single language of
	installer and writes the results inside an output dir.

	      <out dir>
		|
		|----nozip
		|     |
		|     ${LANG}.po
		|
		|----zip
		|     |	
		|     ${LANG}_all.txt
		|     ${LANG}_unkn_wl.txt
		|
		graph.png
		stats.txt

	Optional word lists are expected to be found in the "wls" subdir

check_all.sh
	Usage: check_all.sh

	this is the top-level scripts that loops through the languages and
	creates a png graph and a table in text format containing statistics
	for all the languages checked.

lang2dict.txt
	contains the language code to aspell-suffix conversion:
	it is used for example in these cases
	"pt_BR" --> "pt-br"
	"uk" --> "ukr"
	
msgstr_extract.awk
	Usage: find . -name ${LANG}.po |\
	xargs awk -f ~/bin/msgstr_extract > [outputfile]

	script originally sent to me by Frans Pop and slightly modified
	in order to work with all languages

More details about the ideas behind this scripts can be found
here:

http://d-i.alioth.debian.org/spellcheck/

Please send your comments to 

Davide Viti (zinosat@tiscali.it)