bug#7732: ispell-buffer always gives bogus (0) and (1) choices

bug#7732: ispell-buffer always gives bogus (0) and (1) choices

[jidanni]

Why can't i/flyspell get the order correctly?

$ cat file
transfered
extention
$ ispell file

 0: extension
 1: extent ion
 2: extent-ion

Whereas ispell-buffer always has these bogus (0) and (1) choices:

$ LANG=C LC_ALL=C HOME=/tmp emacs -nw file
M-x ispell-buffer

Affix rules generate and capitalize this word as shown below:
        transfer+ed
Use option `i' to accept this spelling and put it in your private dictionary.(0) transfer ed
(1) transfer-ed  (2) transferred  (3) transferee  (4) transferees  (5) transformed  (6) transversed
(7) transferee's

(0) extent ion  (1) extent-ion  (2) extension  (3) extenuation  (4) extensions  (5) extinction
(6) extortion  (7) extending  (8) extension's  (9) extensional  (:) extenuation's

By the way, why does
$ spell file #and ispell too
not say anything about 'transfered', but in emacs ispell does?
P.S.,
$ dict transfered
No definitions found for "transfered", perhaps you mean:
gcide:  Transferred  Transferee
stardic:  transferred  transferee
xdict:  transferred  transferee
$ apt-show-versions emacs-snapshot
emacs-snapshot/sid uptodate 1:20101212-2
ii  iamerican             3.1.20.0-7            An American English dictionary for ispell
ii  ibritish              3.1.20.0-7            A British English dictionary for ispell
ii  myspell-en-us         1:3.2.1-2             English_american dictionary for myspell
ii  wamerican-huge        6-3                   American English dictionary words for /usr/share/dict

bug#7732: ispell-buffer always gives bogus (0) and (1) choices

[Eli Zaretskii]

> From: jidanni@jidanni.org
> Date: Sat, 25 Dec 2010 09:55:36 +0800
> Cc: rfrancoise@debian.org
> 
> Why can't i/flyspell get the order correctly?
> 
> $ cat file
> transfered
> extention
> $ ispell file
> 
>  0: extension
>  1: extent ion
>  2: extent-ion
> 
> Whereas ispell-buffer always has these bogus (0) and (1) choices:
> 
> $ LANG=C LC_ALL=C HOME=/tmp emacs -nw file
> M-x ispell-buffer
> 
> Affix rules generate and capitalize this word as shown below:
>         transfer+ed
> Use option `i' to accept this spelling and put it in your private dictionary.(0) transfer ed
> (1) transfer-ed  (2) transferred  (3) transferee  (4) transferees  (5) transformed  (6) transversed
> (7) transferee's
> 
> (0) extent ion  (1) extent-ion  (2) extension  (3) extenuation  (4) extensions  (5) extinction
> (6) extortion  (7) extending  (8) extension's  (9) extensional  (:) extenuation's

I cannot reproduce that.  I get the same choices as you show from
ispell outside Emacs, and in the same order.

Are you sure ispell-buffer doesn't use aspell or hunspell instead of
the real ispell?  What are your values of ispell-really-aspell and
ispell-really-hunspell?  In my case, I know that's really ispell that
ispell.el invokes, because I don't have the other two installed.

> By the way, why does
> $ spell file #and ispell too
> not say anything about 'transfered', but in emacs ispell does?

I don't understand the question: you just showed above that ispell
_does_ say something about 'transfered'.

> $ dict transfered
> No definitions found for "transfered", perhaps you mean:
  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Exactly.

bug#7732: ispell-buffer always gives bogus (0) and (1) choices

[jidanni]

>>>>> "EZ" == Eli Zaretskii <eliz@gnu.org> writes:

EZ> Are you sure ispell-buffer doesn't use aspell or hunspell instead of
EZ> the real ispell?  What are your values of ispell-really-aspell and
EZ> ispell-really-hunspell?  In my case, I know that's really ispell that
EZ> ispell.el invokes, because I don't have the other two installed.

ispell-really-hunspell nil
ispell-really-aspell "0.60.6"

>> By the way, why does
>> $ spell file #and ispell too
>> not say anything about 'transfered', but in emacs ispell does?

EZ> I don't understand the question: you just showed above that ispell
EZ> _does_ say something about 'transfered'.

Only inside emacs. Outside we see

$ echo transfered  |/usr/bin/spell # http://en.wiktionary.org/wiki/transfered #misspelling
$ echo transferred |/usr/bin/spell
$ echo transferrred|/usr/bin/spell
transferrred

$ cat t
transfered
$ ispell t
$

bug#7732: ispell-buffer always gives bogus (0) and (1) choices

[Eli Zaretskii]

> From: jidanni@jidanni.org
> Cc: 7732-done@debbugs.gnu.org, rfrancoise@debian.org
> Date: Sun, 26 Dec 2010 08:57:03 +0800
> 
> >>>>> "EZ" == Eli Zaretskii <eliz@gnu.org> writes:
> 
> EZ> Are you sure ispell-buffer doesn't use aspell or hunspell instead of
> EZ> the real ispell?  What are your values of ispell-really-aspell and
> EZ> ispell-really-hunspell?  In my case, I know that's really ispell that
> EZ> ispell.el invokes, because I don't have the other two installed.
> 
> ispell-really-hunspell nil
> ispell-really-aspell "0.60.6"

That's it: you are actually using aspell inside Emacs.  So I'm closing
this bug.

> $ echo transfered  |/usr/bin/spell # http://en.wiktionary.org/wiki/transfered #misspelling
> $ echo transferred |/usr/bin/spell
> $ echo transferrred|/usr/bin/spell
> transferrred
> 
> $ cat t
> transfered
> $ ispell t

Only the last one is of (casual) interest to this issue, and I cannot
reproduce it: for me, ispell highlights "transfered" as a misspelling
and shows exactly the same choices as inside Emacs.  Maybe you should
do a "type ispell", perhaps there are more surprises there.

bug#7732: Bug#608023: ispell-buffer always gives bogus (0) and (1) choices

[jidanni]

Package: aspell
Version: 0.60.6-4
X-debbugs-Cc: 7732@debbugs.gnu.org, rfrancoise@debian.org

>>>>> "EZ" == Eli Zaretskii <eliz@gnu.org> writes:
EZ> That's it: you are actually using aspell inside Emacs.  So I'm closing
EZ> this bug.

I see. http://debbugs.gnu.org/7732 must be an aspell bug!

EZ> do a "type ispell", perhaps there are more surprises there.
$ type ispell
ispell is hashed (/usr/bin/ispell)