headers
David Stone | 7 Mar 2009 18:36
Picon

xindy support for Welsh language

I've been trying to write a Welsh language module for xindy, but am
unsure what to do, even after reading the source and documentation
(though perhaps I haven't looked in the right place).

Welsh uses the Latin alphabet, but has digraphs ch, dd, ff, ng, ll,
ph, rh, th, which occur as separate letters just after c, d, f, g, l,
p, r, t respectively.  This is rather like Spanish ch, ll with the
traditional spelling.  Also, vowels including w and y can take
accents: acute, grave, circumflex, diaeresis (some combinations are
rare; circumflex is common); accents are ignored when sorting.  Some
other letters are only used in borrowed words.  See
e.g. http://en.wikipedia.org/wiki/Welsh_alphabet

Because of combinations like w-circumflex, it seemed simplest to use
utf8.

I've been trying to use make-rules.pl, and using traditional Spanish
as a model, spanish/traditional-utf8.pl.in, and have created
welsh/utf8.pl.in locally.  But I do not understand why $alphabet seems
to have a fixed number of elements, nor how to change it to what I
require.  The comments suggest that you have to have an element for
unused letters, which is left as [] if you do not require it.  But
there is no element for several of the Welsh digraphs.  Should I be
setting  <at> letter_group_names?  Is there some other array which I need
to set?

Eventually I want to use the result with texindy, by making the .xdy
file and then specifying it with -M.

If the result works, it can of course be included with standard
(Continue reading)

Permalink | Reply | 8 comments |
headers
Joachim Schrod | 9 Mar 2009 21:13
Picon
Favicon

Re: Accents lost, even with LC2 companion and "xindy revisited" conference example

Denis Bitouzé writes:

Hi,

to keep you posted:
> 
> I'm using the xindy version available with TeXLive 2008, on a linux
> box:
> 
> [bitouze <at> drums divers]$ xindy -V
> xindy release: unknown
> xindy script version: 1.08
> xindy kernel version: 2.3
> xindy run time engine: i686-linux-gnu, version 2.2
> CLISP version 2.44.1 (2008-02-23) (built on tug.org [130.225.2.178])
>     architecture: I686
> 
> I tried to test the french example given in the LaTeX Companion
> version 2, and also in the conference:

I'm analyzing the changes between the last version where it still
worked and the first version where it didn't. (That happened when we
both restructured our distribution to use autoconf and merged lots of
outstanding patches; so it's quite some changes.)

What basically happens is that the LICR encoding normalization doesn't
work as expected, xindy modules are probably loaded in the wrong
order. (That was also changed at that time.)

It is definitively a system error and not a usage error.
(Continue reading)

Permalink | Reply | 6 comments |
headers
Denis Bitouzé | 9 Mar 2009 21:23
Picon

Re: Accents lost, even with LC2 companion and "xindy revisited" conference example

Le lundi 09/03/09 à 21h13,
Joachim Schrod <jschrod <at> acm.org> a écrit :

> I'm analyzing the changes between the last version where it still
> worked and the first version where it didn't. (That happened when we
> both restructured our distribution to use autoconf and merged lots of
> outstanding patches; so it's quite some changes.)
> 
> What basically happens is that the LICR encoding normalization doesn't
> work as expected, xindy modules are probably loaded in the wrong
> order. (That was also changed at that time.)
> 
> It is definitively a system error and not a usage error.

OK, thanks for your answer. So, could we expect a new version soon? :)
--

-- 
Denis

------------------------------------------------------------------------------
Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
-OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise
-Strategies to boost innovation and cut costs with open source participation
-Receive a $600 discount off the registration fee with the source code: SFAD
http://p.sf.net/sfu/XcvMzF8H
Permalink | Reply | 1 comment |
headers
Joachim Schrod | 9 Mar 2009 21:33
Picon
Favicon

Re: Accents lost, even with LC2 companion and "xindy revisited" conference example

Denis Bitouzé writes:
> Le lundi 09/03/09 à 21h13,
> Joachim Schrod <jschrod <at> acm.org> a écrit :
> 
> > I'm analyzing the changes between the last version where it still
> > worked and the first version where it didn't. (That happened when we
> > both restructured our distribution to use autoconf and merged lots of
> > outstanding patches; so it's quite some changes.)
> > 
> > What basically happens is that the LICR encoding normalization doesn't
> > work as expected, xindy modules are probably loaded in the wrong
> > order. (That was also changed at that time.)
> > 
> > It is definitively a system error and not a usage error.
> 
> OK, thanks for your answer. So, could we expect a new version soon? :)

I hope so, I'm working on it.

     Joachim

------------------------------------------------------------------------------
Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
-OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise
-Strategies to boost innovation and cut costs with open source participation
-Receive a $600 discount off the registration fee with the source code: SFAD
http://p.sf.net/sfu/XcvMzF8H
Permalink | Reply | 0 comments |
headers
Joachim Schrod | 9 Mar 2009 21:46
Picon
Favicon

Re: xindy support for Welsh language

David Stone writes:
> I've been trying to write a Welsh language module for xindy, but am
> unsure what to do, even after reading the source and documentation
> (though perhaps I haven't looked in the right place).

No, that's more because we don't have good documentation. :-(

> Welsh uses the Latin alphabet, but has digraphs ch, dd, ff, ng, ll,
> ph, rh, th, which occur as separate letters just after c, d, f, g, l,
> p, r, t respectively.  This is rather like Spanish ch, ll with the
> traditional spelling.  Also, vowels including w and y can take
> accents: acute, grave, circumflex, diaeresis (some combinations are
> rare; circumflex is common); accents are ignored when sorting.  Some
> other letters are only used in borrowed words.  See
> e.g. http://en.wikipedia.org/wiki/Welsh_alphabet
> 
> Because of combinations like w-circumflex, it seemed simplest to use
> utf8.
> 
> I've been trying to use make-rules.pl, and using traditional Spanish
> as a model, spanish/traditional-utf8.pl.in, and have created
> welsh/utf8.pl.in locally.  But I do not understand why $alphabet seems
> to have a fixed number of elements, nor how to change it to what I
> require.  The comments suggest that you have to have an element for
> unused letters, which is left as [] if you do not require it.  But
> there is no element for several of the Welsh digraphs.

Have you discovered test1/test1.pl? It explains the $alphabet stuff:

# Every line describes one letter of the alphabet (in all its variants).
(Continue reading)

Permalink | Reply | 7 comments |
headers
David Stone | 10 Mar 2009 21:31

Re: xindy support for Welsh language

Joachim Schrod wrote:
> David Stone writes:

>> ... I've been trying to use make-rules.pl, and using traditional Spanish
>> as a model, spanish/traditional-utf8.pl.in, and have created
>> welsh/utf8.pl.in locally.  But I do not understand why $alphabet seems
>> to have a fixed number of elements, nor how to change it to what I
>> require.  The comments suggest that you have to have an element for
>> unused letters, which is left as [] if you do not require it.  But
>> there is no element for several of the Welsh digraphs.
> 
> Have you discovered test1/test1.pl? It explains the $alphabet stuff:
> 
> [some explanations snipped]
>
> # There can be empty arrays [] which are called slots. They are used for
> # mixing alphabets of different languages.

yes, this is the part I didn't understand.  It seemed that there was a
fixed super-list of letters, and that I should use the ones I wanted
and leave the others as empty slots.  What I didn't understand is what
to do when I had letters which didn't have a slot already there. like
ff.  If I inserted a new letter ff after f, then when I tried using the
resulting .xdy then the letters heading each letter's section in the
resulting index did not match the words under the letter.  So
I guessed that somewhere else there is a master list of all the
letters, which I had to update to match the extra element of
$alphabet that I had inserted.  But my guess may be wrong.

> Does this information help you further?
(Continue reading)

Permalink | Reply | 0 comments |
headers
Thomas Henlich | 11 Mar 2009 15:16
Picon
Favicon

Re: xindy support for Welsh language

Am 09.03.2009, 21:46 Uhr, schrieb Joachim Schrod <jschrod <at> acm.org>:

> So the number of elements in $alphabet (characters or diacritic
> variants) is not fixed, but there are limits.

That's right. If you do not care about mixing languages, you do need to  
worry about the slots or letters that are already there. You can change  
them to whatever your language requires. There is no other  
alphabet-specific stuff in another file, so I can't see what went wrong  
with your letter groups. Can you post an example .pl file, + xindy input  
and output?

>> Should I be setting  <at> letter_group_names? Is there some other array
>> which I need to set?
>
> No, one doesn't need to set  <at> letter_group_names. It is actually not
> used, AFAICS. Thomas -- any comments on that variable, what you
> planned to to with it or what you did?

This is just some leftover from a work version I guess. Ignore it.

Regards, Thomas
--

-- 
      Dipl.-Ing. Thomas Henlich
      Liebigstr. 38, 01187 Dresden
      tel.: +49 [0]351 4712357
      mobil: +49 [0]175 1723884

------------------------------------------------------------------------------
Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
(Continue reading)

Permalink | Reply | 0 comments |
headers
Zdenek Wagner | 16 Mar 2009 11:57
Picon

Definition of enumerable location class

Hello,
I have found inthe xindy Technical Report that it is possible to
define an enumerable location class by a mapping function. However, I
have not found an exact description nor example. I would like to have
page numbers in Devanagari where digits zero to nine are
U+0966...U+096F. Could you point me to some document?

Sincerely,

--

-- 
Zdeněk Wagner
http://hroch486.icpf.cas.cz/wagner/
http://icebearsoft.euweb.cz

------------------------------------------------------------------------------
Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
easily build your RIAs with Flex Builder, the Eclipse(TM)based development
software that enables intelligent coding and step-through debugging.
Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
Permalink | Reply | 4 comments |
headers
Zdenek Wagner | 16 Mar 2009 12:24
Picon

xindy bugs + fixes (xindy.sty, perl scripts)

Hi all,
I tried to contact Andreas Schlechte and Roger Kehr but probably the
e-mail addresses given in the files are old and no longer working. I
hope I find some developers here.

I have just started to use xindy and I found a few bugs both in the
perl script and xindy.sty

1. In the POD documentation of both xindy.pl and texindy.pl the second
option named --version should be --verbose instead.

2. File xindy.pl contains on line 659:
$variant =~ s/din/din5007/ ;
However, if a user requests --language=german-din5007, the variant
would be set incorrectly to dindin5007. The regexp should either be
anchored e.g. as s/din$/din5007/ or even better would be:
$variant = 'din5007' if $variant eq 'din';

3. As reported at
http://xindy.sourceforge.net/html-mail-archive/msg00038.html xindy.sty
adds undesired spaces. This is caused by missing percent signs at the
ends of several lines. I have found these lines and fixed them.

4. You write in the xindy.sty documentation that the backlashes and
quotes are not escaped in a user written macro. The reason of the
problem is that \indexindy has to change categories of several
characters. However, in the definition shown in the manual the
argument is consumed by the user macro and \indexindy gets the tokens
with categories already set. Doubling the backslash usually works
because \\ is defined as a robust macro but may not work in some
(Continue reading)

Permalink | Reply | 0 comments |
headers
Joachim Schrod | 22 Mar 2009 10:45
Picon
Favicon

Re: Accents lost, even with LC2 companion and "xindy revisited" conference example

Joachim Schrod writes:
> Denis Bitouzé writes:
> 
> > 
> > I tried to test the french example given in the LaTeX Companion
> > version 2, and also in the conference:

[the canonical cote example doesn't work any more.]

> What basically happens is that the LICR encoding normalization doesn't
> work as expected, xindy modules are probably loaded in the wrong
> order. (That was also changed at that time.)

Sorry that it took so long. Attached is a new version of the texindy
user command. You should be able to simply use it as a drop in
replacement, the file is located in the TL's bin directory.

Afterwards, the sorting of french accented chars doesn't ignore the
accents any more.

Please tell us if it doesn't work for you.

Best,
	Joachim

--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Joachim Schrod			Email: jschrod <at> acm.org
xindy maintainer		http://www.xindy.org/
Roedermark, Germany
(Continue reading)

Permalink | Reply | 3 comments |

Gmane