Hello Endre,
I can see three problems in what you are trying to do: the encoding of the
tables, the line endings, and the whole concept of including one .ctb file
in another.
First of all, I would recommend that you use Notepad++ and not the regular
Notepad included with Windows. Notepad++ is able to handle both Unicode
flavours and different line-endings much more gracefully.
When writing or editing tables, as far as I know, you have to use either
7-bit ASCII or UTF-8 Unicode without byte order mark (BOM. However, the
regular Notepad cannot write UTF-8 files without adding BOM. This is a
Microsoft thing.
I dont think that using windows line-endings in the tables will ruin
anything. However, passing the line-endings through Liblouis can produce
various results according to my experience. It really depends on how the
tables in question handles line-endings. Actually, what I do is strip the
line-endings before throwing the string at Liblouis and then adding the
relevant line-ending afterwards.
Including one table into another is primarily used to include e.g. a table
with character definitions into a table containing grade 1 or grade 2 rules.
However, if you include a full set of rules into another full set of rules,
the result will be a mix of the two definitions, generally with the first
definition taking precedence. Also, if one or both of the tables use
multi-pass rules or hyphenation files, the result would probably be quite
unpredictable.
Generally, the table for one language should contain rules on how to deal
with text in other languages or characters outside the native alphabet.
However, these rules may not exist for some tables or be inadequate or
impractical for longer sections in a text book on a foreign language. In
that case, you would probably have to build a mechanism to feet the text
book section by section to liblouis and change tables when it is needed.
I am working on Windows myself, and it does create weird problems dealing
with liblouis sometimes. I used to run Ubuntu in a virtual machine on top of
Windows and use that for the tasks where I had to use Linux. However, I have
recently started to use the Linux subsystem on Windows 10 in stead, and I
have successfully been able to build liblouis in this environment.
So, I cant supply you with a liblouis that is totally independent of
encoding and line-endings, but otherwise, I hope that I have contributed to
helping you solve your problems.
Best regards
Bue
Fra: liblouis-liblouisxml-bounce@xxxxxxxxxxxxx
[mailto:liblouis-liblouisxml-bounce@xxxxxxxxxxxxx] På vegne af Csapó Endre
Sendt: 4. december 2016 08:30
Til: liblouis-liblouisxml@xxxxxxxxxxxxx
Emne: [liblouis-liblouisxml] Need for a correctly working Windows version of
the Liblouis
Hi Everyone on the list,
My name is Endre Csapo, I am from Hungary. I am a blind programmer who
recently have got a hand on a very old (more than 20 years) and very
rugged-looking but almost perfectly working refreshable Braille display.
I have the luck to know at least 4 languages plus my native one and I have a
lot of e-books on those languages.
Among them are Esperanto, russian, bulgarian and the greek. Almost all of
them are using non-latin writing systems and Braille shrift too.
The very best of the language teaching books on greek are written and
published by russians, I have a very nice Esperanto textbook for hungarian
readers, so these books have russian and greek, hungarian and Esperanto at
the same time.
In Unicode it is an easy task to read them on a computer with a screen
reader and a Braille display.
But I would like to give some of them for making Braille hard copy. And the
place capable of paper Braille print can accept only Braille ASCII files,
e.g. .brf or .brl files.
The LibLouis would be an ideal preprocessor for such multilingual books.
But to use it correctly in my case, for example, I must include the
esperanto .ctb into the hungarian, the russian into the greek and so on.
And when I did so, editing the appropriate tables under Windows with
Notepad, the LibLouis and the Pandoc in case of the Sendto shortcut, stopped
to work, it threw a handful of nasty error messages about badly encoded
table files.
As a bonus, the Pandoc also gave an angry reaction to UTF-16 text inputs,
both big- and little-endian, but this is an another story...
I will not! install a Linux on my computer, it is a non-disassemblable
laptop to change hard disks, and I never trusted multiboot systems because
of stability issues.
My request would be this:
Please, someone build me a Windows version of the full Liblouis system which
is absolutely insensitive to the Windows/DOS CR/LF line endings, uses only
this type of EOL exclusively, both in the table files and in the input .txt
and in the output .brl files also.
I ask for this because recently I haven't got any correct compilation
environment for the LibLouis sources. (I am using exclusively onli the
Windows version of the .NET SDK for application development.)
If somebody will do me this favor, please send me a download link to the
.zip file.
It would be also a very good idea to upload or link it into the LibLouis
Download page. Perhaps others also had already stumbled on this annoyance...
Thanks for it in forward.
With best regards:
Endre
