On Mon, 8 Oct 2007, Stefan Kost wrote:
hi,
I took another attempt at spell checking sources using
@(#) International Ispell Version 3.1.20 (but really Aspell 0.60.4)
I run
aspell -c --mode=ccpp -p ./.aspell_dict.pws xxx.c
and I create the personal dictionary from the tags file and
file-name basenames. This is to automatically have a sensible
wordlist. When trying to use the list I get error like this:
Error: ./.aspell_dict.pws: The word "adapter_test" is invalid. The
character '_' (U+5F) may not appear in the middle of a word.
Error: ./.aspell_dict.pws: The word "appendix-win32" is invalid.
The character '3' (U+33) may not appear in the middle of a word.
Error: ./.aspell_dict.pws: The word "appendix-win32" is invalid.
The character '3' (U+33) must be followed by an alphabetic character.
I already started to muck with my /usr/lib/aspell/en.dat and adding
special ' -*- _ **- - -*- 1 -** 2 -** 3 -**
and so on. This helps with allowing to get rid of the *in the
middle' warnings but not of the latter (must be followed by an
alphabetic character).
With Aspell You can only have one "special" character in a row. Known
Limitation. I'm sure I have written about it before.
The personal dictionary starts with:
personal_ws-1.1 en 701
What about having some 'dummy' language for this purpose? I mean,
when using generated wordlists, I don't want any affix handling and
the like. Just adding these workds as valid words.
That could work. But if you start making numbers letters than every
number will be marked as a misspelling is that really what you want?
While speaking of personal dictionaries, could aspell please save
those sorted? If keeping them in the source repository it would be
handy to have minimal diff.
There is a patch to this effect:
http://sourceforge.net/tracker/index.php?func=detail&aid=1333573&group_id=245&atid=300245