[OS X TeX] novice query concerning international characters

[cfrees at imapmail.org]

Just to follow-up: it turned out the problem I thought was associated
with o circumflex was actually due to stray invisible characters. It
just happened that every paragraph which contained such a stray also
contained and o circumflex and vice-versa. I finally solved it after
examining the files in Terminal and comparing character codes with those
in my minimal test file. Made me wish TeXShop had a (souped-up)
equivalent of WordPerfect's make-everything-visible mode!

Thanks again,
cfr

On Thu 9th Oct, 2008 at 22:47, cfrees at imapmail.org seems to have written:

> Thanks very much. I know I tried that at one point but didn't really
> know what should work so tried a bunch of things. My minimal test now
> works, as you say. I apparently didn't try the right combination with a
> sufficiently minimal test.
>
> Now I know what should work, I should be able to figure out why my
> non-minimal document is still protesting. :)
>
> Much appreciated,
> cfr
>
> On Thu 9th Oct, 2008 at 23:13, Maarten Sneep seems to have written:
>
>> Hi,
>> 
>> This is actually how I discovered the solution, so this is something you 
>> may want to refer to for unrelated problems.
>> 
>> On 9 okt 2008, at 22:52, Dr. Clea F. Rees wrote:
>>
>>> 	\documentclass[11pt,welsh]{article}
>>> 	\usepackage{babel}
>>> 	\usepackage[utf8]{inputenc}
>>> with or without
>>> 	\usepackage{lmodern}
>>> produces the same results.
>>> 
>>> I use the Welsh keyboard for input which makes it easy to type
>>> characters with circumflexes. That is, the 7 vowels: a, e, i, o, u, w,
>>> y. But using a different keyboard (e.g. US) produces the same results.
>>> 
>>> Things work fine for a, e, i and u. Not so fine for the others. w and y
>>> are not typeset at all if typed with a circumflex. (Presumbably because
>>> OT1, T1 are missing slots for w and y circumflex.) I also get console 
>>> errors about o circumflex and I'm not sure why:
>>> 	! Package inputenc Error: Keyboard character used is undefined
>>> 	(inputenc)                in inputencoding `utf8'.
>>> 	See the inputenc package documentation for explanation.
>>> The output looks fine for o with a circumflex, but every occurrence
>>> triggers this error.
>> 
>> First check: Latin Modern has the characters ? (LATIN SMALL LETTER Y WITH 
>> CIRCUMFLEX, unicode 0177, utf-8 C5B7) and ? (LATIN SMALL LETTER W WITH 
>> CIRCUMFLEX, unicode 0175, utf-8 C5B5) in all shape variants. Sorry for 
>> shouting, this is how the character panel names them.
>> 
>> So it should be possible to use these characters, but perhaps the mapping 
>> is incorrect. This is not something I know a lot about. Possible solutions:
>> 
>> - use XeTeX, or more precisely: xelatex. This path is unicode/utf-8 from 
>> start to finish, so there is not a lot that can mess things up.
>> 
>> - Access the characters directly. In trying to figure out how to do this, I 
>> used 'texdoc inputenc'. This opens the documentation for inputenc, but this 
>> only lists 8-bit encodings, not utf8. Then I recalled that the ucs package 
>> is a possibility. 'texdoc ucs' This opens a text file in the terminal, and 
>> in it I find:
>> 
>> Question: When I try to activate options in \usepackage[...]{ucs},
>> LaTeX complains about an option clash.
>> 
>> Answer: ucs.sty probably already got loaded via
>> \usepackage[utf8x]{inputenc}. Try loading ucs.sty first or set the
>> options with \SetUnicodeOption.
>> 
>> Then I recalled that loading the utf8 encoding through ucs, requires 
>> \usepackage[utf8x]{inputenc}.
>> 
>> Minimal test file:
>> 
>> \documentclass[11pt,welsh]{article}
>> \usepackage{babel}
>> \usepackage[utf8x]{inputenc}
>> \usepackage{lmodern}
>> \begin{document}
>> ??
>> \end{document}
>> 
>> This gives the desired results.
>> 
>> Maarten
>