gnuaspell / aspell Goto Github PK

Pawel Salek <pawsa@sf> created a bug report on 2001-09-25 09:11:52 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/27)

There is a apparent crash reported at

http://bugzilla.gnome.org/show_bug.cgi?id=61119

Please update bugzilla when this problem is resolved
(or even earlier, when possible).

#​4 0x40051d60 in pthread_sighandler (signo=11, ctx={gs
= 0, __gsh = 0, fs =
0, __fsh = 0,
+ es = 43, __esh = 0, ds = 43, __dsh = 0, edi =
136750600, esi =
136750600,
+ ebp = 3221213240, esp = 3221213200, ebx =
1074123416, edx = 0, ecx = 0,
+ eax = 136750600, trapno = 14, err = 4, eip =
1074082326, cs = 35, __csh
= 0,
+ eflags = 66118, esp_at_signal = 3221213200, ss =
43, __ssh = 0, fpstate
= 0xbfffcf90,
+ oldmask = 2147483648, cr2 = 76}) at signals.c:97
+#​5 <signal handler called>
+#​6 __pthread_unlock (lock=0x826a608) at
../linuxthreads/sysdeps/i386/i686/pt-machine.h:54
+#​7 0x4004fbbc in __pthread_mutex_unlock
(mutex=0x826a5f8) at mutex.c:163
+#​8 0x4005266e in __funlockfile (stream=0xbfffd144) at
lockfile.c:47
+#​9 0x408de199 in istream & operator>><char,
string_char_traits<char>,
__default_alloc_template<true, 0> > () from
/usr/lib/libstdc++-libc6.2-2.so.3
+#​10 0x40b6c1e7 in autil::split () from
/usr/lib/libaspell.so.10
+#​11 0x40b3ac55 in aspell::Language::setup () from
/usr/lib/libaspell.so.10
+#​12 0x40b44686 in aspell::DataSet::set_check_lang () from
/usr/lib/libaspell.so.10
+#​13 0x40b4b292 in
aspell_default_readonly_ws::ReadOnlyWS::load ()
+ from /usr/lib/libaspell.so.10
+#​14 0x40b47aab in aspell::add_data_set () from
/usr/lib/libaspell.so.10
+#​15 0x40b68992 in aspell::MultiWS::load () from
/usr/lib/libaspell.so.10
+#​16 0x40b47aab in aspell::add_data_set () from
/usr/lib/libaspell.so.10
+#​17 0x40b5881d in aspell::Manager::real_setup () from
/usr/lib/libaspell.so.10
+#​18 0x40b585a0 in aspell::Manager::setup () from
/usr/lib/libaspell.so.10
+#​19 0x40add5df in pspell_aspell::PA_Manager::PA_Manager ()
+ from /usr/lib/libpspell_aspell.so.2
+#​20 0x40adead1 in
libpspell_aspell_LTX_new_pspell_manager_class ()
+ from /usr/lib/libpspell_aspell.so.2
+#​21 0x40027d4a in new_pspell_manager_class () from
/usr/lib/libpspell.so.4
+#​22 0x40029d59 in new_pspell_manager () from
/usr/lib/libpspell.so.4
+#​23 0x8098be6 in balsa_spell_check_start ()

Michael Opdenacker <zumbi@sf> created a bug report on 2002-07-17 17:48:13 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/37)

Hello,

Here is the problem I had building aspell-.33.7.1 on
Red Hat GNU/Linux 7.1:

g++ -DHAVE_CONFIG_H -I. -I. -I.. -I/usr/local/include
-g -O2 -c file_util.cc
-fPIC -DPIC -o .libs/file_util.o
file_util.cc: In function `string
autil::add_possible_dir (const string
&, const string &)':
file_util.cc:33: `need_dir' undeclared (first use this
function)
file_util.cc:33: (Each undeclared identifier is
reported only once for
each function it appears in.)
make[2]: *** [file_util.lo] Error 1

I got rid of this error by installing pspell-.12.2 first.
(pspell was not installed before).

If pspell is a prerequisite for aspell, it seems you should
add a check for this tool in the configure.in file.

What do you think?

Cheers,

Michael.

Anonymous created a bug report on 2001-04-23 06:22:40 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/21)

gmake[2]: Entering directory
`/usr/local/aspell-.33.5/util'
/bin/sh ../libtool --mode=compile c++ -DHAVE_CONFIG_H
-I. -I. -I.. -I/usr/local/include -g -O2 -c
file_util.cc
c++ -DHAVE_CONFIG_H -I. -I. -I.. -I/usr/local/include
-g -O2 -c file_util.cc -DPIC -o .libs/file_util.o
In file included from file_util.cc:4:
simple_fstream.hh:45: macro `fileno' used without args
file_util.cc:41: macro `fileno' used without args
file_util.cc:50: macro `fileno' used without args
file_util.cc:73: macro `fileno' used without args
file_util.cc:92: macro `fileno' used without args
gmake[2]: *** [file_util.lo] Error 1
gmake[2]: Leaving directory
`/usr/local/aspell-.33.5/util'
gmake[1]: *** [all-recursive] Error 1
gmake[1]: Leaving directory `/usr/local/aspell-.33.5'
gmake: *** [all-recursive-am] Error 2

Ed Robitaille <erobitaille@sf> created a bug report on 2000-11-12 23:23:24 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/16)

Ran strace;
aspell loads files while in phoenetic search strace shoes SigKill

Using latest pspell/dict and aspell
compiled on SuSE 6.3 using egcs-2.91.6

There was only 1 warning listed during make
a pointer was so large it was made into an unsigned pointer
this occured in the 'rope' section

Anonymous created a bug report on 2001-03-11 22:54:57 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/20)

Aspell 32.6 and pspell 11.2 does a stack dump when
running in cygwin
environment. This happens with cygwin.dll 1.1.8-1 and
cygwin.dll
1.1.8-2. It is reported not to occur with later
versions of the
cygwin.dll, but aspell does appear to have a problem.

I have tried the later versions of aspell 33 and pspell
12 and had the same
problem. This version of pspell by the way will not
build on Solaris 2.6, 2.7
or 2.8 platforms and this has been reported.

SEQUOIA-> aspell -c --dont-backup README
0 [main] aspell 63 handle_exceptions: Exception:
STATUS_ACCESS_VIOLATION
11562 [main] aspell 63 stackdump: Dumping stack trace
to aspell.exe.stackdump
Segmentation fault (core dumped)

When I run aspell within gdb below is the stack at time
of the segment
violation:
(gdb) run -c --dont-backup README
Starting program: /usr/local/bin/aspell.exe -c
--dont-backup README

Program received signal SIGSEGV, Segmentation fault.
autil::VectorHashTable<aspell_default_readonly_ws::ReadOnlyWS::WordLookupParms>:
:FindIterator::FindIterator (this=0x2d3eb34,
ht=0xa022af0, k=@0x2d3eb60)
at inc/language.hh:81
81 SpecialChar special(char c) const {return
special_[to_uchar(c)];}

Below is the stack dump file:
Exception: STATUS_ACCESS_VIOLATION at eip=004C3F6F
eax=0A01FE60 ebx=6AA37C61 ecx=0A01FE58 edx=0A01FE64
esi=0A010388 edi=02D3EC84
ebp=02D3E94C esp=02D3EBEC
program=c:\cygwin\usr\local\bin\aspell.exe
cs=001B ds=0023 es=0023 fs=003B gs=0000 ss=0023
Stack trace:
Frame Function Args
02D3E94C 004C3F6F (00000000, 00000000, 00000000,
00000000)
02D3EC44 004C3F6F (02D3EC84, 0A01FE58, 02D3ECB0,
00000000)
02D3ECA4 0042B57C (0A01FE28, 0A010388, 0A01E3DC,
004DB538)
02D3ED04 00435A82 (02D3FBA4, 0A010388, 02D3ED84,
00449729)
02D3ED64 00435B1A (02D3FBA4, 0A010388, 0A01038B,
00000008)
02D3EF54 004DD43A (02D3FA04, 02D3F13B, 02D3FC14,
00408AEE)
02D3FC14 0040CD4F (02D30001, 004011F3, 00000000,
00000005)
02D3FEB4 00402387 (00000004, 1A020CA8, 0A010008,
FFFFFFFE)
02D3FF10 61003859 (00000000, 00000000, 77F763D3,
77F763DB)
02D3FF40 61003A3D (00401260, 00000000, 00000000,
00000000)
02D3FF60 61003A7C (00000000, 00000000, 00000000,
00000000)
02D3FF90 004812CF (00401260, 00000000, 00000246,
8011748B)
02D3FFC0 0040103D (00000000, 00000000, 7FFDF000,
7FFDF000)
02D3FFF0 77F1B9EA (00401000, 00000000, 000000B0,
00000100)
End of stack trace

sven sanwald <blabla_1@sf> created a bug report on 2001-08-02 06:44:05 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/25)

Hello,

I tried to compile the new aspell-.33.6.3 with the
aspell-.33-fix1.diff
patch. But there is something wrong.
Where is the mistake?
What can I do?

Sincerely

Sven Sanwald

make all-recursive
make[1]: Entering directory
`/drv2/src/OFFICE/Spell/Aspell/aspell-.33.6.3'
Making all in util
make[2]: Entering directory
`/drv2/src/OFFICE/Spell/Aspell/aspell-.33.6.3/util'
make[2]: Nothing to be done for `all'.
make[2]: Leaving directory
`/drv2/src/OFFICE/Spell/Aspell/aspell-.33.6.3/util'
Making all in filter
make[2]: Entering directory
`/drv2/src/OFFICE/Spell/Aspell/aspell-.33.6.3/filter'
make[2]: Nothing to be done for `all'.
make[2]: Leaving directory
`/drv2/src/OFFICE/Spell/Aspell/aspell-.33.6.3/filter'
Making all in lib
make[2]: Entering directory
`/drv2/src/OFFICE/Spell/Aspell/aspell-.33.6.3/lib'
Making all in inc
make[3]: Entering directory
`/drv2/src/OFFICE/Spell/Aspell/aspell-.33.6.3/lib/inc'
make[3]: Nothing to be done for `all'.
make[3]: Leaving directory
`/drv2/src/OFFICE/Spell/Aspell/aspell-.33.6.3/lib/inc'
make[3]: Entering directory
`/drv2/src/OFFICE/Spell/Aspell/aspell-.33.6.3/lib'
/bin/sh ../libtool --mode=compile c++ -DHAVE_CONFIG_H
-I. -I. -I.. -I./inc -I../ -I../util -I../filter
-I/usr/local/include -g -O2 -c config.cc
c++ -DHAVE_CONFIG_H -I. -I. -I.. -I./inc -I../
-I../util -I../filter -I/usr/local/include -g -O2 -c
config.cc -fPIC -DPIC -o .libs/config.o
config.cc:156: `CONF_DIR' was not declared in this
scope
config.cc:158: `DATA_DIR' was not declared in this
scope
config.cc:159: `DICT_DIR' was not declared in this
scope
config.cc:175: `DATA_DIR' was not declared in this
scope
make[3]: *** [config.lo] Error 1
make[3]: Leaving directory
`/drv2/src/OFFICE/Spell/Aspell/aspell-.33.6.3/lib'
make[2]: *** [all-recursive] Error 1
make[2]: Leaving directory
`/drv2/src/OFFICE/Spell/Aspell/aspell-.33.6.3/lib'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory
`/drv2/src/OFFICE/Spell/Aspell/aspell-.33.6.3'
make: *** [all-recursive-am] Error 2

Domenico Andreoli <cavok@sf> created a patch on 2001-11-06 15:22:23 UTC
(Orig. from https://sourceforge.net/p/aspell/patches/6)

aspell supports a keyboard layout file to find
characters next to each other. This allows it to locate
words that may have been misspelled because you hit the
wrong key.

The package does not ship with a DVORAK keymap; one is
attached to this message. It would be great to have
this included in the package...

slava <slavushka@sf> created a bug report on 2001-08-26 01:53:28 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/26)

aspell -a
+

aborts. Isn't it suppose to switch to TeX mode? This
is what ispell does.
This prevents from using aspell in XEmacs with
ispell.el in a (la)tex mode.

aspell -v
@(#) International Ispell Version 3.1.20 (but really
Aspell .33.7.1 alpha)

Anonymous created a bug report on 2000-12-18 23:49:27 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/19)

When Aspell .32.6 is used with Fly-Spell the memory size of the Aspell process increases very quickly indicating a possible memory leak.

martin.moessner <martin098@sf> created a feature request on 2001-11-20 23:43:40 UTC
(Orig. from https://sourceforge.net/p/aspell/feature-requests/3)

for the german word für aspell does not recognize the
tex encodings f"ur (using the german style) or f&quot;ur.
this problem arises for the german umlauts
ä = "a = &quot;a Ä = "A = &quot;A
ö = "o = &quot;o Ö = "O = &quot;O
ü = "u = &quot;u Ü = "U = &quot;U
and the sharp s
ß = "s = &quot;ss
but also in other languages such as in the french
language.

Joe Feise <jfeise@sf> created a bug report on 2000-04-20 19:43:51 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/7)

The Windows binary crashes on NT4 when trying to spell-check a LaTeX document. It works ok on a plain text file. I am running XEmacs 21.1.
Norton CrashGuard reports this:
aspell caused fault #c0000005 in at address 001b:004d4fc6

To date, 4 fatal errors have been recorded in this program. This particular error was recorded 4 times.

Reported By:
CrashGuard v2.0.1

Report Date:
4/20/00 12:40:57

WindowTitle:
c:\aspell\aspell.exe

Last Message:
MSG(00000000, 0000, 00000000, 00000000)

Program:
aspell
( - 0)

Registers:
EAX=024ef440 CS=001b
EIP=004d4fc6 EFLGS=00000206
EBX=024ef4f0 SS=0023
ESP=024ef328 EBP=024ef4c0
ECX=024ef450 DS=0023
ESI=024ef450 FS=0038
EDX=00453f9b ES=0023
EDI=00411484 GS=0000

Bytes at CS:EIP:
89 42 04 89 ca 83 c4 f8 6a ff 6a 00 8d 4d e0 51

Stack dump:
034fbaf0 004011fb 00453bfe 77fa80c0 004eb1ab 77f0d2a6 004eb320 00000000 00000000 004eb320 004eb320 024ef4e0 034f5800 004eb320 00000000 77641621

Anonymous created a bug report on 2000-01-18 12:55:23 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/1)

When checking an all capitalized version of an proper none such as KEVIN aspell thinks KEVIN is wrong but also suggestion comes up with KEVIN as the first suggestion.

Anonymous created a bug report on 2000-07-25 21:55:49 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/12)

System SuSE 6.3
gnome 1.2
pspell-.11.0.1

When compiling aspell-0.32 I get the following error
Making all in data
make[2]: Entering directory `/app/work/aspell-.32/data'
cd ../; bd=`pwd`; cd scowl/final;
cat english-*.10 english-*.20 english-*.35 english-*.50
special-*.35 special-*.50
| $bd/src/aspell --local-data-dir=$bd/data/
create master $bd/data/english-med-only
make[2]: *** [english-med-only] Error 134
make[2]: Leaving directory `/app/work/aspell-.32/data'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/app/work/aspell-.32'
make: *** [all-recursive-am] Error 2

This also happened with release 0.30/0.31, whereas 0.29 run without problems.

Thanks,
Klaus Dahlke
[email protected]

Ralf Schlatterbeck <schlatterbeck@sf> created a bug report on 2001-12-08 22:39:36 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/31)

The personal wordlist file (specified with --personal
option) will create an empty file via NFS if the file
did not yet exist, it will ignore the contents of an
already existing file.

This is with Linux running kernel 2.2.19 on a debian
potato system.

Anonymous created a bug report on 2001-05-04 04:59:58 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/22)

I just got the latest pspell (0.12.1) and aspell
(0.33.5). I
compiled on Solaris 8. No problems (except I had to
disable
ltdl on pspell and build the modules manually) in
compiling
and installing. But when I run
aspell -c /etc/hosts
I get
ld.so.1: aspell: fatal: relocation error: file aspell:
symbol newterm32: referenced symbol not found

ldd `which aspell ` gives me:

libaspell.so.10 =>
/u/visitors2/rfinkel/local/lib/libaspell.so.10
libm.so.1 => /usr/lib/libm.so.1
libpspell.so.4 =>
/u/visitors2/rfinkel/local/lib/libpspell.so.4
libpspell-modules.so.1 =>
/u/visitors2/rfinkel/local/lib/libpspell-modules.so.1
libpspell-impl.so.4 =>
/u/visitors2/rfinkel/local/lib/libpspell-impl.so.4
libcurses.so.1 =>
/usr/ucblib/libcurses.so.1
libc.so.1 => /usr/lib/libc.so.1
libpspell_aspell.so.2 =>
/u/visitors2/rfinkel/local/lib/libpspell_aspell.so.2
libtermcap.so.1 =>
/usr/ucblib/libtermcap.so.1
libucb.so.1 => /usr/ucblib/libucb.so.1
libdl.so.1 => /usr/lib/libdl.so.1
libelf.so.1 => /usr/lib/libelf.so.1
/usr/platform/SUNW,Ultra-5_10/lib/libc_psr.so.1

Kevin Atkinson <kevina@sf> created a patch on 2000-09-21 05:28:23 UTC
(Orig. from https://sourceforge.net/p/aspell/patches/2)

Chris Herborth <chrish_texar@sf> created a bug report on 2001-11-09 19:03:29 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/30)

The --add-sgml-check and --rem-sgml-check docs are
misleading:

--add|rem-sgml-check=<str> sgml tags to always check.

That suggests that you can make it check or ignore
SGML elements (often called "tags")... in reality, it
seems to be for checking or ignoring tag attributes
(such as the "alt" attribute in the "img" element of
HTML).

So, changing "tags" to "attributes" in the docs and
help message will fix this problem.

It would be _extremely_ useful to be able to skip
spell checking on the contents of certain elements;
for example, if you've got code in <pre> blocks,
maybe --add-sgml-skip=pre would totally skip that hunk
of the document while spellchecking...

Henning Niss <hniss@sf> created a bug report on 2000-11-24 14:46:23 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/17)

Checking a TeX file (perhaps other files as well) that ends with a comment
without a newline produces a segmentation fault.

Try Aspell with the following file:

A short test file.
%%% Local Variables:
%%% mode: latex
%%% TeX-master: t
%%% End:

(where there is no newline after End:)

Local fix:
--- filter/tex-t.hh.orig Sun Mar 26 05:03:47 2000
+++ filter/tex-t.hh Fri Nov 24 15:34:45 2000
@@ -61,6 +61,7 @@
// deal with comments
if (c == '%' && !prev_backslash) in_comment = true;
if (in_comment && c == '\n') in_comment = false;
+ if (in_comment && c == '\0') in_comment = false;
if (in_comment) return !parms->check_comments;

if (top.in_what == Name) {

Anonymous created a bug report on 2000-02-13 00:33:41 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/2)

In certain rare cases, where < rules and follow-up rules come together the algorithm produces wrong phonetics.

Erik Dorfman <eryq@sf> created a bug report on 2000-08-21 20:04:50 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/13)

Aspell: 32.5
OS: Solaris [SunOS 5.7]
Compiler: g++ 2.95.2

Dumps core during "make" at the following command:

Entering directory `/export/home/cs/depot/src/aspell-.32.5/data'
cd ../; bd=`pwd`; cd scowl/final;
cat english-*.60 english-*.70
| $bd/src/aspell --local-data-dir=$bd/data/
create master $bd/data/english-lrg-only

The rest of the make (before and after this) goes ok.

Rocky Bernstein <rockyb@sf> created a patch on 2001-08-05 17:46:40 UTC
(Orig. from https://sourceforge.net/p/aspell/patches/4)

Got compiler error in bulding aspell-.33.7 on Solaris 8
using g++ version 3.0:

/bin/sh ../libtool --mode=compile g++ -DHAVE_CONFIG_H
-I. -I. -I.. -I./inc -I../ -I../util -I../filter
-I/opt/platform/include -I/opt/platform/include
-I/opt/platform/X11/include -g -O2 -c manager.cc
In file included from ../util/vector_hash-t.hh:18,
from readonly_ws.cc:77:
../util/vector_hash.hh:244: friend declaration requires
class-key, i.e. `friend
class autil::VectorHashTable<Parms>'
../util/vector_hash.hh:258: friend declaration requires
class-key, i.e. `friend
class autil::VectorHashTable<Parms>::FindIterator'

The problems seem to be coming from util/vector_hash.hh
these lines:
friend HashTable;

friend FindIterator;

Comparing with aspell-.33.6.3 I see that the word
"class" was deleted between the versions.

Adding "class" back in seems to make everything work
for me.

Below you should find a patch for this simple change.

Ian Haywood <ihaywood@sf> created a bug report on 2001-06-21 11:37:13 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/23)

The British varient dictionary that comes with aspell
has some problems.. It is not really complete, with a
number of American words in the "common" english-*-only
file.

The VarCon word list on wordlist.sourceforge.net has
most of the
variants needed, could you remove all these words from
english-lrg-only, and put the variants in
american-lrg-only and british-lrg-only files.

Thanks

Xianglong Yuan <yuanx@sf> created a bug report on 2001-06-22 20:07:33 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/24)

dvfdsfsd

Joe Feise <jfeise@sf> created a bug report on 2000-04-20 21:41:29 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/8)

The Windows binary crashes on NT4 when trying to spell-check a LaTeX document. It works ok on a plain text file. I am running XEmacs 21.1.
Norton CrashGuard reports this:
aspell caused fault #c0000005 in at address 001b:004d4fc6

To date, 4 fatal errors have been recorded in this program. This particular error was recorded 4 times.

Reported By:
CrashGuard v2.0.1

Report Date:
4/20/00 12:40:57

WindowTitle:
c:\aspell\aspell.exe

Last Message:
MSG(00000000, 0000, 00000000, 00000000)

Program:
aspell
( - 0)

Registers:
EAX=024ef440 CS=001b
EIP=004d4fc6 EFLGS=00000206
EBX=024ef4f0 SS=0023
ESP=024ef328 EBP=024ef4c0
ECX=024ef450 DS=0023
ESI=024ef450 FS=0038
EDX=00453f9b ES=0023
EDI=00411484 GS=0000

Bytes at CS:EIP:
89 42 04 89 ca 83 c4 f8 6a ff 6a 00 8d 4d e0 51

Stack dump:
034fbaf0 004011fb 00453bfe 77fa80c0 004eb1ab 77f0d2a6 004eb320 00000000 00000000 004eb320 004eb320 024ef4e0 034f5800 004eb320 00000000 77641621

Anonymous created a bug report on 2000-02-22 10:17:00 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/5)

Igor Markov wrote:
> aspell complained about .html --- while it has HTML, .html
> will often happen in URLs. In this case the URL consisted
> of good words, except for html at the end.

Kevin Atkinson wrote:
>Aspell normally ignores url and the like a url contains :// or two or
>more dots. Also ignores words with @ in the middle. Won't ignore
>jdk.jdk/jfkjfk but a future version will.
>
>Please submit it not ignoring jdk.djk/djkdj as aon the Bug Tracker so I
>can keep track of it.

Benjamin Lerman <quisar@sf> created a bug report on 2000-07-12 21:52:33 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/11)

When I use aspell with the -l option, with the french or the american disctionnary, it begin to take all the memory of the computer (192 Mo RAM, 60Mo Swap) and then crash the computer.

My Linux is a 2.2.16 kernel, glibc 2.1

Hal Schectman <schectman@sf> created a bug report on 2002-03-19 18:48:00 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/34)

Doing the following (using the attached file):

aspell -H -a < aspell.htm

Will cause aspell to hang. It writes its signature (@
(#) International Ispell Version 3.1.20 (but really
Aspell .31 alpha)) and then nothing. Removing the -H
avoids the problem.

I reduced the aspell.htm file quite a bit and the
result is attached. Interesting things about its
contents:

1. The text to be spell checked is all on one line. We
do this to make tracking the location of spelling
errors easier. If it is sent to aspell formatted on
multiple lines, aspell is much happier.
2. The offending piece seems to be the initial
comment: "<!-- a's -->". If you remove the single
quote, aspell is happy. I believe the single quote is
confusing aspell and it gets lost trying to process
the rest of the text. There's something about the rest
of the text, though, possibly its length, that is part
of the equation.

Thanks very much!

Hal Schectman
Atomz Corporation

Vitaliy Vorobyov <deathsoft@sf> created a patch on 2002-03-12 20:34:53 UTC
(Orig. from https://sourceforge.net/p/aspell/patches/7)

Dictionary compressor decompressor has a small bug. The
pipes stdin and stdout are opened in text (translated
mode), when some symbols like CR,LF,CTRL+Z are translated.
In this case end of input file situation may occur earlier
then it realy have to occur. This bug is occured when
decompressing russian dictionary coded in koi8r (only
25% of file is decompressed). To resolve this problem
some code must be added to file compress.cc:
after all "#include <>" directives

#ifdef _WIN32
#include <io.h>
#include <fcntl.h>
#endif
at the begin of main function:
#ifdef _WIN32
_setmode( _fileno( stdin ), _O_BINARY );
_setmode( _fileno( stdout ), _O_BINARY );
#endif

line:
cout << cur << '\n';
must be replaced by:
cout << cur
#ifdef _WIN32
<< '\r'
#endif
<< '\n';

This code switch pipes mode from text to binary
(untranslated) and adds WIN32 like line ends.

Kurt Lidl <lidl@sf> created a bug report on 2002-01-31 22:16:21 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/32)

Attempting to compile the sources to aspell outside
of the source tree does not work.

Making a subdirectory of the source tree to compile
in doesn't work either.

Repeat by:

1) unpack sources
2) cd aspell-.33.7.1
3) mkdir obj
4) cd obj
5) ../configure --disable-static --with-pic --with-gnu-ld
6) make

Watch it fail, typically like this:

[...]
Making all in filter
/bin/sh ../libtool --mode=compile g++ -DHAVE_CONFIG_H
-I. -I../../filter -I.. -I../util -I/usr/local/include
-g -O2 -c ../../filter/filter.cc
mkdir .libs
g++ -DHAVE_CONFIG_H -I. -I../../filter -I.. -I../util
-I/usr/local/include -g -O2 -c ../../filter/filter.cc
-fPIC -DPIC -o .libs/filter.o
In file included from ../../filter/filter.cc:2:
../../filter/filter.hh:6: clone_ptr.hh: No such file or
directory
In file included from ../../filter/repl.hh:10,
from ../../filter/filter.cc:3:
../../filter/try_hard.hh:3: string_map.hh: No such file
or directory
*** Error code 1

Stop.
*** Error code 1

Stop.
*** Error code 1

Stop.

Pspell, for what it is worth, is broken the same way.

-Kurt

Petr Mladek <pmladek@sf> created a patch on 2001-08-27 14:22:33 UTC
(Orig. from https://sourceforge.net/p/aspell/patches/5)

Hello,

I had problems to compile aspell-.33.7 on axp with
gcc-3.0.1.

Simply fix is in attached patch.

I hope that the fix is good.

With greetings,

Petr Mladek
software developer
---------------------------------------------------------------------
SuSE CR, s.r.o. e-mail:
[email protected]
Drahobejlova 27 tel:+420 2
8309 5373
190 00 Praha 9 fax:+420 2
8309 5374
Ceska republika
http://www.suse.cz

Anonymous created a bug report on 2000-12-07 11:00:37 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/18)

While trying to spell check pam.sgml, aspell dies. It seems to be related to using 'Ignore all' quite a bit to get past things like file paths and special jargon.

backtrace:
Program received signal SIGSEGV, Segmentation fault.
afilter::FilterItrRootClass<_Rope_const_iterator<char, allocator<char> >, afilter::itr_equal<_Rope_const_iterator<char, allocator<char> > > >::next (this=0x8147b28) at ../filter/filter.hh:167
167 else return *ritr;
(gdb) bt
#​0 afilter::FilterItrRootClass<_Rope_const_iterator<char, allocator<char> >, afilter::itr_equal<_Rope_const_iterator<char, allocator<char> > > >::next (this=0x8147b28) at ../filter/filter.hh:167
#​1 0x400f7d46 in afilter::RegionSkipItr<afilter::SgmlSkip<afilter::RegionSkipFilterTypes> >::next ()
from /usr/lib/libaspell.so.8
#​2 0x400ec009 in afilter::ReplItr<afilter::DualRepl<afilter::CharReplBase<afilter::char2uni>, afilter::CharReplBase<afilter::uni2char> >, -1, -1>::next () from /usr/lib/libaspell.so.8
#​3 0x400f5d3d in afilter::BlockSkipItr<afilter::UrlSkip<afilter::BlockSkipFilterTypes> >::next ()
from /usr/lib/libaspell.so.8
#​4 0x4009f07d in aspell::TokenItr::advance () from /usr/lib/libaspell.so.8
#​5 0x805e447 in void aspell::check<_Rope_const_iterator<char, allocator<char> >, afilter::itr_equal<_Rope_const_iterator<char, allocator<char> > >, aspell::DoNothing> (state=@0xbffff8d8, print_status=@0xbffff617)
at ../lib/inc/check.hh:170
#​6 0x80558f4 in check (interactive=true) at aspell.cc:857
#​7 0x804e870 in main (argc=4, argv=0xbffffa64) at aspell.cc:274
#​8 0x401d7c1c in __libc_start_main () from /lib/libc.so.6

This is version 0.32.6. I get similar results with 0.31.

roudkerk <roudkerk@sf> created a bug report on 2002-06-28 09:48:03 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/36)

In the British dict, there are several words ending in
"-wards" which are missing (although the "-ward" form
is included).

Examples: towards, forwards, backwards, upwards,
downwards, onwards, northwards, southwards, westwards,
homewards, inwards, outwards, heavenwards, skywards.

(Strangely eastwards and northeastwards are allowed,
but northwards, westwards... are not.)

In most of these cases the "-ward" form is an
adjective, while the "-wards" form is an adverb (but
Americans usually use the -ward form in both cases).

Ispell seems to accept these words because of its affix
rules (even when using the American dictionary).

Cheers.

(P.S. Why does the British dictionary insist on
Roumania
instead of Romania? All the dictionaries I can find
say
Romania/Rumania.)

Carsten Krueger <carson@sf> created a bug report on 2000-05-23 10:55:00 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/9)

- Windows NT 4.0
- aspell-.30.1-win32-i386

"aspell --lang=german create master ./german < german.0"

when trying to create a master file from a big wordlist aspell stops after about 30 words! the master file is imcomplete! configuration files seem to be at the right place and of the right content!

Devin Weaver <ktohg@sf> created a patch on 2001-02-21 00:50:34 UTC
(Orig. from https://sourceforge.net/p/aspell/patches/3)

As many can get tripped up on library version numbers and since aspell requires pspell version >= 12; I added a check to the configure script. It might be best to model this over GTK+'s autoconf script. But time prevaling this was a quicker hack.

Greg Schafer <gschafer@sf> created a bug report on 2002-02-01 15:12:29 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/33)

Hi
I'm having a weird problem building aspell version 33.7.1

The weird thing is that it used to build fine on my
previous system. The only differences I can think of
that may be relevant are

Old System New System
glibc-2.2.4 glibc-2.2.5
binutils-2.11.2 binutils-2.11.92.0.12.3

both systems are linux 2.4.17 using patched gcc-2.95.4

Here is the tail of the output:

make[2]: Entering directory
`/home/gws/src/aspell-.33.7.1/data'
cd ../; bd=`pwd`; cd scowl/final;
cat english-*.10 english-*.20 english-*.35 english-*.50
english-*.60 english-*.65
special-*.35 special-*.50
| $bd/src/aspell --lang=english
create master $bd/data/english-med-only
make[2]: *** [english-med-only] Error 134
make[2]: Leaving directory
`/home/gws/src/aspell-.33.7.1/data'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/gws/src/aspell-.33.7.1'
make: *** [all-recursive-am] Error 2

And here is a backtrace. Hope its useful.

#​0 0x401f6611 in kill () from /lib/libc.so.6
#​1 0x401f6418 in raise () from /lib/libc.so.6
#​2 0x401f7abb in abort () from /lib/libc.so.6
#​3 0x4018217b in __default_terminate () from
/usr/lib/libstdc++-libc6.2-2.so.3
#​4 0x4018219c in __terminate () from
/usr/lib/libstdc++-libc6.2-2.so.3
#​5 0x40182dbe in __throw () from
/usr/lib/libstdc++-libc6.2-2.so.3
#​6 0x400b4773 in
autil::ConfigData::throw_file_exception
(this=0xbffff770, file=0x806a038
"/home/gws/.aspell.conf") at config_data.cc:30
#​7 0x400776f1 in aspell::Config::read_in
(this=0xbffff880, override=0x80691c0) at
../util/config_data.hh:163
#​8 0x4007753a in aspell::Config::read_in
(this=0xbffff880, override=0x80691c0) at config.cc:64
#​9 0x08055683 in master () at aspell.cc:1001
#​10 0x0804efd3 in main (argc=5, argv=0xbffff9d4) at
aspell.cc:320
#​11 0x401e5591 in __libc_start_main () from /lib/libc.so.6

Regards
Greg

Adam Lopresto <eimi@sf> created a bug report on 2002-08-30 19:30:38 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/40)

When in -a (ispell emulation) mode, if the last line of
input doesn't end with a newline it is ignored, where
ispell -a will still output spelling errors for it.
Basically, pipe() first checks if the last character is
an EOF, and stops processing if it is. The attached
patch makes it first process the line and then check if
it should quit. There may be a more elegant solution,
but this seems to work.

Bill <ascilum_@sf> created a feature request on 2001-07-15 08:27:25 UTC
(Orig. from https://sourceforge.net/p/aspell/feature-requests/2)

While I can appreciate all the work that has gone into
aspell, the largest reason that I won't install it on
our servers is that most students/faculty/staff use
pine and are inherently familar ispell as their
speller. Why is it that in aspell, 'a' doesn't mean
'accept' it means 'add', 'i' doesn't mean 'insert' it
means 'ignore', and 'q' doesn't quit. This may seem
like a trivial point, but I will have 250 proff's
calling me nasty names if I change this interface. I
doubt that I am alone in this complaint.

Is there any way to compile aspell such that it
defaults to using ispell's single letter commands?
Better still, a command line flag?

Anonymous created a bug report on 2000-11-07 14:39:38 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/15)

The program works fine when
checks spelling on the same drive
as aspell is insatlled

but when checking another drive it dosent
find

E:&gt;aspell -c c:\SERIAL.C
Sorry, I can't read the file "/aspell/dict/english".

and I have installed it on c:\aspell

Best Regards A.B

Melvyn Sopacua <nyvlem@sf> created a bug report on 2002-08-25 10:49:38 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/38)

config.cpp is the only place using LC_MESSAGES, so as a
work-around I've set:
CXXFLAGS=-DLC_MESSAGES=1729
(the value was taken from libintl.h)

I think the permanent fix would be to link with libintl.

In any case: LC_MESSAGES is a POSIX construct and not
supported on BSD/OS (perhaps others).

Len Rose <lenrose@sf> created a bug report on 2002-08-28 07:16:33 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/39)

Hi.. sorry to bother you..

Solaris 8 (x86), gcc 2.95.2, Solaris ld

Attaching output from make..appears to be an issue with
cpp I guess.

Ed Richley <richley@sf> created a bug report on 2001-10-04 03:31:52 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/28)

Hi:

I upgraded to RH7.1 on an older alpha workstation
(AS200,
21064 processor) and aspell core dumps. Doesn't work at
all.
emacs can't run it (says something about not returning
a
version number).

Here's the output:

bose$ aspell -c notes.txt
Segmentation fault (core dumped)
bose$

I haven't tried to build it from source, but I wanted
to know if
this is a known problem, or what I should do.

Thanks.

Ed Richley
[email protected]

Anonymous created a bug report on 2000-11-01 18:17:04 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/14)

After seemingly successfully compiling aspell 0.32.5 under Cygwin on Windows NT 4.0, running it with -a produces a stack dump.

$ echo test word | aspell -a
0 [main] aspell 11844 handle_exceptions: Exception: STATUS_ACCESS_VIOLATIO
N
1131 [main] aspell 11844 stackdump: Dumping stack trace to aspell.exe.stackdu
mp
Segmentation fault (core dumped)

-------------------
[email protected]

Richard Russon <flatcap@sf> created a feature request on 2001-03-20 03:14:51 UTC
(Orig. from https://sourceforge.net/p/aspell/feature-requests/1)

ATM aspell hex numbers, eg

0xef0

will get queried:

0*xEF*0 ?

Please can you add an option to ignore hex values
(for we technical documentation writers).

Cheers,
FlatCap (Rich).
[email protected]

Anonymous created a bug report on 2000-02-13 00:37:54 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/3)

Let's assume two word which are both "Capital" in the dictionary like "Mary" and "Stewart" in the English one. If you write "MaryStewart" Aspell suggests "Mary stewart" and "Mary-stewart" and does not do the second word Capitalized.

Ed Avis <epaepa@sf> created a bug report on 2002-06-04 11:59:44 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/35)

A while ago I was working on a big list of
american/british/canadian
spelling conversions (more comprehensive than varcon)
and while trying
to automatically generate part of this list I found lots of
inconsistencies in aspell's English wordlists.

These were a few cases of variant spellings being
listed in the wrong
lists, many cases of possessives being present but not
the base word (eg
american includes Grecism's but not Grecism) and
thousands of cases
where words had posessives that shouldn't (eg
Anglicanize's).

So I've got this big big list of inconsistencies, but
in a format that
would make it easy to read with Perl and correct the
wordlists
automatically. The trouble is that Aspell's wordlists
are made by merging together
several others, and if those are 'buggy' there's not
much to do about
it. For words which are _missing_ you could have an
ad-hoc additional
list to make things consistent; for words which
shouldn't be there at
all it's not clear what to do. Although I think that
ispell's
wordlists are responsible for most of those errors, and
I have already
submitted corrections to the ispell maintainer.

Anyway, for the record, the attached file is a big list
of inconsistencies found in the american, british,
canadian and english wordlists.

Anonymous created a bug report on 2000-02-24 07:58:02 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/6)

When spell checking any files on the powerpc checking any files uses up all the memory at a very fast rate.

Stepan Yar <yarex@sf> created a bug report on 2001-10-28 18:37:53 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/29)

>D:\aspell\aspell>aspell.exe --lang="ru" -l
>дурак
>Sorry, I can't read the
>file "D:/ASPELL/ASPELL/dict/ru.rws".

What is wrong?

Anonymous created a bug report on 2000-02-13 00:43:26 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/4)

All of aspell dictionaries accept words with non-word characters in it often creating confusing results. Aspell should really check it the word only contains legal characters before inserting it into the word list. This also apples for the creation of the main word list.

Anonymous created a patch on 2000-02-24 08:17:48 UTC
(Orig. from https://sourceforge.net/p/aspell/patches/1)

Rémi Vanicat <vanicat@sf> created a bug report on 2000-06-12 20:16:12 UTC
(Orig. from https://sourceforge.net/p/aspell/bugs/10)

Hello,

I'have began to write a some configuration file for aspell and french,
and i do have a probleme :

whith this phonec file :
version 0000000000001

C K
ET$ E
EU E
EN A
E E
Č E
É E
Ę E
H _
O O
Ô O
R R
T T
#end of the phonec file

and this word list:
ente
et
etc
été
eth
eue
ôte
rte
te
tee

i have the following (bad) result :
+serveur1-1:~/prog/aspell/share/aspell$ aspell --lang=ete pipe
@(#) International Ispell Version 3.1.20 (but really Aspell .31 alpha)
ete
& ete 9 0: tee, ente, et, te, eth, eue, etc, rte, ôte

even if
+serveur1-1:~/prog/aspell/share/aspell$ aspell --lang=ete soundslike
été ete tee, ente, et, te, eth, eue, etc, rte, ôte
été ETE
ete ETE
tee, TEE
ente, ATE
et, ET
te, TE
eth, ET
eue, EE
etc, ETK
rte, RTE
ôte OTE

as you see été and ete give the same result whith the algorithm, but it's not in the proposition.