Re: [Demexp-dev] Re: Before merging ketty's branch

[David MENTRE]

Hi all,

Frederic Lehobey <address@hidden> writes:

> On Fri, Oct 13, 2006 at 11:48:53PM +0200, ketty . wrote:
>> On 10/13/06, David MENTRE <address@hidden> wrote:
>> I am still not able to use »« in the sources without breaking the book. It
>> seams to me like latex has some serious issues[1] :(
>> 
>> [1] Recent debian thread:
>> http://lists.debian.org/debian-tetex-maint/2006/03/msg00455.html
>
>   Read http://www.joachim-breitner.de/blog/archives/185-latex-ucs.html

I confirm: I have 'latex-ucs' package on my debian sarge machine put I
can't compile the source with '«' and '»', even if all seems encoded in
UTF-8.

More details:

After applying patch:
"""
diff -r 67feb2c7de75 web/pages.ml.nw
--- a/web/pages.ml.nw   Fri Oct 13 23:14:36 2006 +0200
+++ b/web/pages.ml.nw   Sat Oct 14 20:19:18 2006 +0200
@@ -207,7 +207,7 @@ let fill_question_details q_id srv =
     let str = 
       Array.fold_left 
         (fun str r_id ->
-          str ^ ">>" ^ response_desc r_id ^ "<<" ^ " &nbsp; ") "" 
+          str ^ "»" ^ response_desc r_id ^ "«" ^ " &nbsp; ") "" 
         q_info.q_info_elected_responses in
     Var.Q.winning_responses#set str;
     if Var.logged_in#get then ( (* get own vote *)
"""

I compile and it breaks:
"""
! LaTeX Error: Command \guillemotright unavailable in encoding OT1.

See the LaTeX manual or LaTeX Companion for explanation.
Type  H <return>  for immediate help.
 ...                                              
                                                  
l.210           str ^ "^^c2^^bb
                               " ^ response_desc r_id ^ "^^c2^^ab" ^ " &nbsp...

"""

Even if the generated file is in UTF-8:
 $ file --mime web/pages.ml.tex
 web/pages.ml.tex: text/plain; charset=utf-8


Ok, after some search, the fix is to use the babel, ucs and utf8x latex
packages:

diff -r 67feb2c7de75 doc/demexp-book.tex
--- a/doc/demexp-book.tex       Fri Oct 13 23:14:36 2006 +0200
+++ b/doc/demexp-book.tex       Sat Oct 14 21:13:30 2006 +0200
@@ -7,7 +7,9 @@
 \usepackage{graphicx}
 \usepackage{verbatim}
 \usepackage{fancyvrb}
-\usepackage[utf8]{inputenc}
+\usepackage{ucs}
+\usepackage[utf8x]{inputenc}
+\usepackage[english]{babel}
 \usepackage[bookmarks=true]{hyperref} %% should be the last package included


But in fact, it does not work for other random UTF-8 chars, like: ¶ŋĸłµ¢”

So you've been warned, this is a temporary fix. For other unicode
characters, one must load specific Unicode table but I don't know how to
do that.

Interesting files (on my debian sarge):

 * /usr/share/doc/latex-ucs-doc/*.ps.gz (<rant>Those !!&$$@ debian
                                         developpers have once again
                                         made a separate package for the
                                         doc!!  Why?? </rant>)

   The docs to understand.

 * /usr/share/doc/latex-ucs/ltxmacrs.txt 

   Gives for a missing \SYMBOL the table to load (but how?? The
   languages.ps.gz doc only speaks about languages for {babel} package).


I've pushed the patch, let me know if it still works on your machine
ketty.


Best wishes,
d.
-- 
GPG/PGP key: A3AD7A2A David MENTRE <address@hidden>
 5996 CC46 4612 9CA4 3562  D7AC 6C67 9E96 A3AD 7A2A