updated for version 7.0103v7.0103

5 files changed, 81 insertions, 2 deletions

diff --git a/runtime/spell/README.txt b/runtime/spell/README.txt
index 8ea5272f0..44dedb7af 100644
--- a/runtime/spell/README.txt
+++ b/runtime/spell/README.txt
@@ -2,7 +2,7 @@ The spell files included here are in Vim's special format.  You can't edit
 them.  See ":help spell" for more information.
 
 
-Copyright
+COPYRIGHT
 
 The files used as input for the spell files come from the OpenOffice.org spell
 files.  Most of them go under the LGPL or a similar license.
@@ -10,3 +10,78 @@ files.  Most of them go under the LGPL or a similar license.
 Copyright notices for specific languages are in README_??.txt.  Note that the
 files for different regions are merged, both to save space and to make it
 possible to highlight words for another region different from bad words.
+
+
+GENERATING .SPL FILES
+
+This involves downloading the files from the OpenOffice.org server, applying a
+patch and running Vim to generate the .spl file.  To do this all in one go use
+the Aap program (www.a-a-p.org).  It's simple to install, it only requires
+Python.
+
+You can also do it manually:
+1. Fetch the right spell file from:
+   http://ftp.services.openoffice.org/pub/OpenOffice.org/contrib/dictionaries
+
+2. Unzip the archive:
+	unzip LL_RR.zip
+
+3. Apply the patch:
+	patch < LL_RR.diff
+
+4. If the language has multiple regions do the above for each region.  E.g.,
+   for English there are five regions: US, CA, AU, NZ and GB.
+
+5. Run Vim and execute ":mkspell".  Make sure you do this with the correct
+   locale, that influences the upper/lower case letters and word characters.
+   On Unix it's something like:
+   	env LANG=en_US.UTF-8 vim
+	mkspell! en en_US en_AU en_CA en_GB en_NZ
+
+6. Repeat step 5 for other locales.  For English you could generate a spell
+   file for latin1, utf-8 and ASCII.  ASCII only makes sense for languages
+   that have very few words with non-ASCII letters.
+
+Now you understand why I prefer using the Aap recipe :-).
+
+
+MAINTAINING A LANGUAGE
+
+Every language should have a maintainer.  His tasks are to track the changes
+in the OpenOffice.org spell files and make updated patches.  Words that
+haven't been added/removed from the OpenOffice lists can also be handled by
+the patches.
+
+It is important to keep the version of the .dic and .aff files that you
+started with.  When OpenOffice brings out new versions of these files you can
+find out what changed and take over these changes in your patch.  When there
+are very many changes you can do it the other way around: re-apply the changes
+for Vim to the new versions of the .dic and .aff files.
+
+This procedure should work well:
+
+1. Obtain the zip archive with the .aff and .dic files.  Unpack it as
+   explained above and copy (don't rename!) the .aff and .dic files to
+   .orig.aff and .orig.dic.  Using the Aap recipe should work, it will make
+   the copies for you.
+
+2. Tweak the .aff and .dic files to generate the perfect .spl file.  Don't
+   change too much, the OpenOffice people are not stupid.  However, you may
+   want to remove obvious mistakes.  And remove single-letter words that
+   aren't really words, they mess up the suggestions (English has this
+   problem).
+
+3. Make the diff file.  "aap diff" will do this for you.  If a diff would be
+   too big you might consider writing a Vim script to do systematic changes.
+   Do check that someone else can reproduce building the spell file.  Send the
+   result to Bram for inclusion in the distribution.  Bram will generate the
+   .spl file and upload it to the ftp server (if he can't generate it you will
+   have to send him the .spl file too).
+
+4. When OpenOffice makes a new zip file available you need to update the
+   patch.  "aap check" should do most of the work for you: if there are
+   changes the .new.dic and .new.aff files will appear.  You can now figure
+   out the differences with .orig.dic and .orig.aff, adjust the .dic and .aff
+   files and finally move the .new.dic to .orig.dic and .new.aff to .orig.aff.
+
+5. Repeat step 4. regularly.
diff --git a/runtime/spell/en.ascii.spl b/runtime/spell/en.ascii.spl
index 1b51921ed..045e07ba7 100644
--- a/runtime/spell/en.ascii.spl
+++ b/runtime/spell/en.ascii.spl
diff --git a/runtime/spell/en.latin1.spl b/runtime/spell/en.latin1.spl
index b759d2f16..f92ba01c7 100644
--- a/runtime/spell/en.latin1.spl
+++ b/runtime/spell/en.latin1.spl
diff --git a/runtime/spell/en.utf-8.spl b/runtime/spell/en.utf-8.spl
index 420d3af06..83aea8d05 100644
--- a/runtime/spell/en.utf-8.spl
+++ b/runtime/spell/en.utf-8.spl
diff --git a/runtime/spell/he/main.aap b/runtime/spell/he/main.aap
index 413c5e2c3..9adb109e8 100644
--- a/runtime/spell/he/main.aap
+++ b/runtime/spell/he/main.aap
@@ -9,12 +9,16 @@
 SPELLDIR = ..
 FILES    = he_IL.aff he_IL.dic
 
-all: $(SPELLDIR)/he.utf-8.spl ../README_he.txt
+all: $(SPELLDIR)/he.utf-8.spl $(SPELLDIR)/he.iso-8859-8.spl ../README_he.txt
 
 $(SPELLDIR)/he.utf-8.spl : $(VIM) $(FILES)
         :sys env LANG=he_IL.UTF-8
 		$(VIM) -e -c "mkspell! $(SPELLDIR)/he he_IL" -c q
 
+$(SPELLDIR)/he.iso-8859-8.spl : $(VIM) $(FILES)
+        :sys $(VIM) -e -c "set enc=iso-8859-8"
+                        -c "mkspell! $(SPELLDIR)/he he_IL" -c q
+
 ../README_he.txt : README_he_IL.txt
         :copy $source $target