aspell-user
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Aspell-user] Looking for usage advice


From: Christoph Hintermüller
Subject: Re: [Aspell-user] Looking for usage advice
Date: Tue, 15 Feb 2005 21:37:56 +0100
User-agent: KMail/1.7.1

Hi
Am Dienstag, 8. Februar 2005 21:48 schrieb Grzegorz Adam Hankiewicz:

> Are this filters embedded in the aspell binary or external, like say
> a shell script? If they are embedded, it would be quite difficult
> to implement them as a crossplatform tool (in fact, I haven't even
> looked at the possibility of running aspell under win32).
> 
Depends partially on how the aspell library is built. If static linking is 
used than mainly yes. In case of shared build with loadtime linking mainly 
yes too if built schared with runtime linking go ahead and code your own 
basic filters if you need to. Further in aspell 0.60 fiter modes are not 
hardcoded anymore they are rather a file describing a specific combination of 
aspell baseic filters and the case specific settings of the corresponding 
filter options.

How far this is allready funcitonal in Win32 port you have to ask Gary as he 
is the one in charge of the Windows port. 

> Most of the "foreign" text tends to be delimited in special markers,
> but sometimes we might need to translate it anyway or include both
> languages at the same time, so it cannot be said that <xxx> tags
> will always enclose English or Spanish text.
Hm if you would make that strict as allready stated earlier in this thread a 
twopass scan by aspell would suffice as aspell context filter takes care that 
aspell only touches the visible context and leaves the invisible unchanged.

If you need a hirarchical context switching mechanism you would have to code 
your own context filter or at least extend the existing context filter. The 
later was mainly coded for spellchecking c/c++ comments containing 
javadoc/qtdoc/doxygen comments and c/c++ strings without bothering about the 
remaining code.
> Having the advantage of 
> XML as source, it would be possible for us to use additional tags,
> either in the form <aspell-langcode> or <!-- aspell-langcode -->
> if the former bothers the xml book conversion tools. The advantage
> here would be that the translator knows what can be checked or not,
> and that we can use a custom filter.
Hm the xml approach imho would require a context filter capable of handling 
context hirarchies switched based uppon specific tags within the document. 
Never the less spellchecking two languages within one document in Aspell 
0.60.X is still a twopass process as aspell itself cannot switch 
dictionaries. 
  Also, I tried building aspell from CVS. The first
> time it didn't want to install, and now I updated again and there
> are some parse errors in common/speller.hpp.
Report them either on the mailto:address@hidden mailing list or on 
http://sourceforge.net/tracker/?group_id=245
cu
Xris




reply via email to

[Prev in Thread] Current Thread [Next in Thread]