init
Some checks failed
Docker. / Ubuntu (push) Has been cancelled
User-agent updater. / User-agent (push) Failing after 15s
Lock Threads / lock (push) Failing after 10s
Waiting for answer. / waiting-for-answer (push) Failing after 22s
Needs user action. / needs-user-action (push) Failing after 8s
Can't reproduce. / cant-reproduce (push) Failing after 8s
Close stale issues and PRs / stale (push) Has been cancelled

This commit is contained in:
allhaileris
2026-02-16 15:50:16 +03:00
commit afb81b8278
13816 changed files with 3689732 additions and 0 deletions

View File

@@ -0,0 +1,4 @@
*.sug text eol=lf
*.good text eol=lf
*.wrong text eol=lf
*.morph text eol=lf

View File

@@ -0,0 +1,3 @@
testSubDir
*.log
*.trs

View File

@@ -0,0 +1,4 @@
# capitalized ngram suggestion test data for
# Sf.net Bug ID 1463589, reported by Frederik Fouvry.
MAXNGRAMSUGS 1
LANG de

View File

@@ -0,0 +1,2 @@
1
Kühlschrank

View File

@@ -0,0 +1,5 @@
Kühlschrank
Kühlschrank
Kühlschrank
Kühlschrank
Kühlschrank

View File

@@ -0,0 +1,5 @@
kuhlschrank
kuehlschrank
kühlschrank
Kuhlschrank
Kuehlschrank

View File

@@ -0,0 +1,5 @@
# capitalized ngram suggestion test data (Unicode version) for
# Sf.net Bug ID 1463589, reported by Frederik Fouvry.
SET UTF-8
MAXNGRAMSUGS 1
LANG de

View File

@@ -0,0 +1,2 @@
1
Kühlschrank

View File

@@ -0,0 +1,5 @@
Kühlschrank
Kühlschrank
Kühlschrank
Kühlschrank
Kühlschrank

View File

@@ -0,0 +1,5 @@
kuhlschrank
kuehlschrank
kühlschrank
Kuhlschrank
Kuehlschrank

View File

@@ -0,0 +1,20 @@
# fix homonym handling for German dictionary project,
# reported by Björn Jacke (sf.net Bug ID 1592880).
SET ISO8859-1
SFX N Y 1
SFX N 0 n .
SFX S Y 1
SFX S 0 s .
SFX P Y 1
SFX P 0 en .
SFX Q Y 2
SFX Q 0 e .
SFX Q 0 en .
COMPOUNDEND z
COMPOUNDPERMITFLAG c
ONLYINCOMPOUND o

View File

@@ -0,0 +1,4 @@
3
weg/Qoz
weg/P
wege

View File

@@ -0,0 +1,3 @@
weg
wege
wegen

View File

@@ -0,0 +1,10 @@
# fix NEEDAFFIX homonym suggestion.
# Sf.net Bug ID 1695964, reported by Björn Jacke.
TRY esianrtolcdugmphbyfvkwESIANRTOLCDUGMPHBYFVKW
MAXNGRAMSUGS 0
NEEDAFFIX h
SFX S Y 1
SFX S 0 s .
SFX e Y 1
SFX e 0 e .

View File

@@ -0,0 +1,3 @@
2
Mull/he
Mull/S

View File

@@ -0,0 +1,3 @@
Mull
Mulle
Mulls

View File

@@ -0,0 +1,3 @@
Mall
Malle
Malls

View File

@@ -0,0 +1,13 @@
# test COMPOUNDRULE bug reported by Björn Jacke
SET ISO8859-1
TRY esijanrtolcdugmphbyfvkwqxz
SFX A Y 5
SFX A 0 e .
SFX A 0 er .
SFX A 0 en .
SFX A 0 em .
SFX A 0 es .
COMPOUNDRULE 1
COMPOUNDRULE vw

View File

@@ -0,0 +1,4 @@
3
arbeits/v
scheu/Aw
farbig/A

View File

@@ -0,0 +1,3 @@
arbeitsfarbig
arbeitsfarbige
arbeitsfarbiger

View File

@@ -0,0 +1,5 @@
# Crashtest for bug 1748408. Parsing an alias file with
# more than (4096 / sizeof(void*)) elements would crash when trying to
# pre-allocate the array of pointers.
# The Hungarian dictionary that caught this had ~1500 entries.
AF 2000

View File

@@ -0,0 +1 @@
dummy

View File

View File

@@ -0,0 +1,3 @@
# Crashtest for bug 1748408.
AF 2000
INVALID something else

View File

@@ -0,0 +1 @@
dummy

View File

View File

@@ -0,0 +1 @@
AM 2000

View File

@@ -0,0 +1 @@
dummy

View File

View File

@@ -0,0 +1,3 @@
# Crashtest for bug 1748408.
AF 2000
INVALID something else

View File

@@ -0,0 +1 @@
dummy

View File

View File

@@ -0,0 +1,6 @@
SET UTF-8
IGNORE ٌٍَُِّْـ
PFX x N 1
PFX x أ ت أ[^ي]

View File

@@ -0,0 +1,3 @@
2
أرى/x
أيار/x

View File

@@ -0,0 +1,3 @@
أرى
أيار
ترى

View File

@@ -0,0 +1 @@
تيار

View File

@@ -0,0 +1,5 @@
# test words with three parts
CHECKCOMPOUNDPATTERN 1
CHECKCOMPOUNDPATTERN le fi
COMPOUNDFLAG c

View File

@@ -0,0 +1,4 @@
3
first/c
middle/c
last/c

View File

@@ -0,0 +1 @@
firstmiddlelast

View File

@@ -0,0 +1 @@
lastmiddlefirst

View File

@@ -0,0 +1,4 @@
CHECKCOMPOUNDPATTERN 1
CHECKCOMPOUNDPATTERN /a /b
COMPOUNDFLAG c

View File

@@ -0,0 +1,4 @@
3
foo/ac
bar/c
baz/bc

View File

@@ -0,0 +1,5 @@
foobar
barfoo
bazfoo
barbaz
bazbar

View File

@@ -0,0 +1 @@
foobaz

View File

@@ -0,0 +1,6 @@
COMPOUNDRULE 1
COMPOUNDRULE ab
COMPOUNDBEGIN A
COMPOUNDEND B

View File

@@ -0,0 +1,4 @@
3
foo/aA
bar/b
baz/B

View File

@@ -0,0 +1,2 @@
foobar
foobaz

View File

@@ -0,0 +1,8 @@
# check bad capitalisation of Dutch letter IJ.
TRY i
FORBIDDENWORD *
PFX i N 1
PFX i ij IJ ij
REP 1
REP ij IJ

View File

@@ -0,0 +1,3 @@
1
ijs/i
Ijs/*

View File

@@ -0,0 +1,2 @@
ijs
IJs

View File

@@ -0,0 +1 @@
IJs, ijs

View File

@@ -0,0 +1 @@
Ijs

View File

@@ -0,0 +1,666 @@
## Process this file with automake to create Makefile.in
SUBDIRS = suggestiontest
TEST_EXTENSIONS = .dic
AM_TESTS_ENVIRONMENT = export HUNSPELL=$(top_builddir)/src/tools/hunspell; \
export ANALYZE=$(top_builddir)/src/tools/analyze; \
export LIBTOOL=$(top_builddir)/libtool;
DIC_LOG_COMPILER = $(top_srcdir)/tests/test.sh
TESTS = \
affixes.dic \
condition.dic \
condition_utf.dic \
base.dic \
base_utf.dic \
allcaps.dic \
allcaps_utf.dic \
allcaps2.dic \
allcaps3.dic \
keepcase.dic \
dotless_i.dic \
i58202.dic \
map.dic \
rep.dic \
sug.dic \
sug2.dic \
wordpair.dic \
sugutf.dic \
phone.dic \
ph.dic \
ph2.dic \
flag.dic \
flaglong.dic \
flagnum.dic \
flagutf8.dic \
slash.dic \
forbiddenword.dic \
nosuggest.dic \
alias.dic \
alias2.dic \
alias3.dic \
breakdefault.dic \
break.dic \
needaffix.dic \
needaffix2.dic \
needaffix3.dic \
needaffix4.dic \
needaffix5.dic \
circumfix.dic \
fogemorpheme.dic \
onlyincompound.dic \
complexprefixes.dic \
complexprefixes2.dic \
complexprefixesutf.dic \
conditionalprefix.dic \
zeroaffix.dic \
utf8.dic \
utf8_bom.dic \
utf8_bom2.dic \
utf8_nonbmp.test \
compoundflag.dic \
compoundrule.dic \
compoundrule2.dic \
compoundrule3.dic \
compoundrule4.dic \
compoundrule5.dic \
compoundrule6.dic \
compoundrule7.dic \
compoundrule8.dic \
compoundaffix.dic \
compoundaffix2.dic \
compoundaffix3.dic \
compoundforbid.dic \
checkcompounddup.dic \
checkcompoundtriple.dic \
simplifiedtriple.dic \
checkcompoundrep.dic \
checkcompoundrep2.dic \
checkcompoundcase2.dic \
checkcompoundcaseutf.dic \
checkcompoundpattern.dic \
checkcompoundpattern2.dic \
checkcompoundpattern3.dic \
checkcompoundpattern4.dic \
utfcompound.dic \
checksharps.dic \
checksharpsutf.dic \
germancompounding.dic \
germancompoundingold.dic \
i35725.dic \
i53643.dic \
i54633.dic \
i54980.dic \
maputf.dic \
reputf.dic \
ignore.dic \
ignoreutf.dic \
right_to_left_mark.dic \
1592880.dic \
1695964.dic \
1463589.dic \
1463589_utf.dic \
IJ.dic \
i68568.dic \
i68568utf.dic \
1706659.dic \
1748408-1.dic \
1748408-2.dic \
1748408-3.dic \
1748408-4.dic \
digits_in_words.dic \
colons_in_words.dic \
ngram_utf_fix.dic \
morph.dic \
1975530.dic \
fullstrip.dic \
iconv.dic \
iconv2.dic \
oconv.dic \
oconv2.dic \
encoding.dic \
korean.dic \
opentaal_forbiddenword1.dic \
opentaal_forbiddenword2.dic \
opentaal_keepcase.dic \
arabic.dic \
2970240.dic \
2970242.dic \
breakoff.dic \
opentaal_cpdpat.dic \
opentaal_cpdpat2.dic \
2999225.dic \
onlyincompound2.dic \
forceucase.dic \
nepali.dic \
hu.dic \
warn.dic \
timelimit.dic \
ignoresug.dic \
limit-multiple-compounding.dic
clean-local:
-rm -rf testSubDir
EXTRA_DIST = \
test.sh \
affixes.aff \
affixes.dic \
affixes.good \
condition.aff \
condition.dic \
condition.good \
condition.wrong \
condition_utf.aff \
condition_utf.dic \
condition_utf.good \
condition_utf.wrong \
base.aff \
base.dic \
base.good \
base.sug \
base.wrong \
base_utf.aff \
base_utf.dic \
base_utf.good \
base_utf.sug \
base_utf.wrong \
allcaps.aff \
allcaps.dic \
allcaps.good \
allcaps.sug \
allcaps.wrong \
allcaps2.aff \
allcaps2.dic \
allcaps2.good \
allcaps2.sug \
allcaps2.wrong \
allcaps3.aff \
allcaps3.dic \
allcaps3.good \
allcaps3.wrong \
allcaps_utf.aff \
allcaps_utf.dic \
allcaps_utf.good \
allcaps_utf.sug \
allcaps_utf.wrong \
keepcase.aff \
keepcase.dic \
keepcase.good \
keepcase.sug \
keepcase.wrong \
dotless_i.aff \
dotless_i.dic \
dotless_i.good \
dotless_i.wrong \
map.aff \
map.dic \
map.sug \
map.wrong \
rep.aff \
rep.dic \
rep.sug \
rep.wrong \
sug.aff \
sug.dic \
sug.sug \
sug.wrong \
sug2.aff \
sug2.dic \
sug2.sug \
sug2.wrong \
wordpair.aff \
wordpair.dic \
wordpair.wrong \
sugutf.aff \
sugutf.dic \
sugutf.sug \
sugutf.wrong \
phone.aff \
phone.dic \
phone.sug \
phone.wrong \
ph.aff \
ph.dic \
ph.sug \
ph.wrong \
ph2.aff \
ph2.dic \
ph2.sug \
ph2.wrong \
alias.aff \
alias.dic \
alias.good \
alias2.aff \
alias2.dic \
alias2.good \
alias2.morph \
alias3.aff \
alias3.dic \
alias3.good \
alias3.morph \
break.aff \
break.dic \
break.good \
break.wrong \
breakdefault.aff \
breakdefault.dic \
breakdefault.good \
breakdefault.sug \
breakdefault.wrong \
circumfix.aff \
circumfix.dic \
circumfix.good \
circumfix.morph \
circumfix.wrong \
fogemorpheme.aff \
fogemorpheme.dic \
fogemorpheme.good \
fogemorpheme.wrong \
onlyincompound.aff \
onlyincompound.dic \
onlyincompound.good \
onlyincompound.sug \
onlyincompound.wrong \
forbiddenword.aff \
forbiddenword.dic \
forbiddenword.good \
forbiddenword.wrong \
nosuggest.aff \
nosuggest.dic \
nosuggest.good \
nosuggest.sug \
nosuggest.wrong \
germancompounding.aff \
germancompounding.dic \
germancompounding.good \
germancompounding.wrong \
germancompoundingold.aff \
germancompoundingold.dic \
germancompoundingold.good \
germancompoundingold.wrong \
needaffix2.aff \
needaffix2.dic \
needaffix2.good \
needaffix2.morph \
needaffix3.aff \
needaffix3.dic \
needaffix3.good \
needaffix3.wrong \
needaffix4.aff \
needaffix4.dic \
needaffix4.good \
needaffix5.aff \
needaffix5.dic \
needaffix5.good \
needaffix5.wrong \
needaffix.aff \
needaffix.dic \
needaffix.good \
needaffix.wrong \
zeroaffix.aff \
zeroaffix.dic \
zeroaffix.good \
zeroaffix.morph \
utf8.aff \
utf8.dic \
utf8.good \
utf8_bom.aff \
utf8_bom.dic \
utf8_bom.good \
utf8_bom2.aff \
utf8_bom2.dic \
utf8_bom2.good \
utf8_nonbmp.aff \
utf8_nonbmp.dic \
utf8_nonbmp.good \
utf8_nonbmp.sug \
utf8_nonbmp.wrong \
utfcompound.aff \
utfcompound.dic \
utfcompound.good \
utfcompound.wrong \
compoundflag.aff \
compoundflag.dic \
compoundflag.good \
compoundflag.wrong \
compoundrule.aff \
compoundrule.dic \
compoundrule.good \
compoundrule.wrong \
compoundrule2.aff \
compoundrule2.dic \
compoundrule2.good \
compoundrule2.wrong \
compoundrule3.aff \
compoundrule3.dic \
compoundrule3.good \
compoundrule3.wrong \
compoundrule4.aff \
compoundrule4.dic \
compoundrule4.good \
compoundrule4.wrong \
compoundrule5.aff \
compoundrule5.dic \
compoundrule5.good \
compoundrule5.morph \
compoundrule5.wrong \
compoundrule6.aff \
compoundrule6.dic \
compoundrule6.good \
compoundrule6.wrong \
compoundrule7.aff \
compoundrule7.dic \
compoundrule7.good \
compoundrule7.wrong \
compoundrule8.aff \
compoundrule8.dic \
compoundrule8.good \
compoundrule8.wrong \
compoundaffix.aff \
compoundaffix.dic \
compoundaffix.good \
compoundaffix.wrong \
compoundaffix2.aff \
compoundaffix2.dic \
compoundaffix2.good \
compoundaffix3.aff \
compoundaffix3.dic \
compoundaffix3.good \
compoundaffix3.wrong \
compoundforbid.aff \
compoundforbid.dic \
compoundforbid.good \
compoundforbid.wrong \
checkcompounddup.aff \
checkcompounddup.dic \
checkcompounddup.good \
checkcompounddup.wrong \
checkcompoundcase.aff \
checkcompoundcase.dic \
checkcompoundcase.good \
checkcompoundcase.wrong \
checkcompoundcase2.aff \
checkcompoundcase2.dic \
checkcompoundcase2.good \
checkcompoundcase2.wrong \
checkcompoundcaseutf.aff \
checkcompoundcaseutf.dic \
checkcompoundcaseutf.good \
checkcompoundcaseutf.wrong \
checkcompoundrep.aff \
checkcompoundrep.dic \
checkcompoundrep.good \
checkcompoundrep.wrong \
checkcompoundrep2.aff \
checkcompoundrep2.dic \
checkcompoundrep2.good \
checkcompoundtriple.aff \
checkcompoundtriple.dic \
checkcompoundtriple.good \
checkcompoundtriple.wrong \
simplifiedtriple.aff \
simplifiedtriple.dic \
simplifiedtriple.good \
simplifiedtriple.wrong \
checkcompoundpattern.aff \
checkcompoundpattern.dic \
checkcompoundpattern.good \
checkcompoundpattern.wrong \
checkcompoundpattern2.aff \
checkcompoundpattern2.dic \
checkcompoundpattern2.good \
checkcompoundpattern2.wrong \
checkcompoundpattern3.aff \
checkcompoundpattern3.dic \
checkcompoundpattern3.good \
checkcompoundpattern3.wrong \
checkcompoundpattern4.aff \
checkcompoundpattern4.dic \
checkcompoundpattern4.good \
checkcompoundpattern4.wrong \
checksharps.aff \
checksharps.dic \
checksharps.good \
checksharps.sug \
checksharps.wrong \
checksharpsutf.aff \
checksharpsutf.dic \
checksharpsutf.good \
checksharpsutf.sug \
checksharpsutf.wrong \
conditionalprefix.aff \
conditionalprefix.dic \
conditionalprefix.good \
conditionalprefix.morph \
conditionalprefix.wrong \
flaglong.aff \
flaglong.dic \
flaglong.good \
flagnum.aff \
flagnum.dic \
flagnum.good \
flag.aff \
flag.dic \
flag.good \
flagutf8.aff \
flagutf8.dic \
flagutf8.good \
complexprefixes.aff \
complexprefixes.dic \
complexprefixes.good \
complexprefixes.wrong \
complexprefixes2.aff \
complexprefixes2.dic \
complexprefixes2.good \
complexprefixesutf.aff \
complexprefixesutf.dic \
complexprefixesutf.good \
complexprefixesutf.wrong \
i35725.aff \
i35725.dic \
i35725.good \
i35725.sug \
i35725.wrong \
i53643.aff \
i53643.dic \
i53643.good \
i53643.wrong \
i54633.aff \
i54633.dic \
i54633.good \
i54633.sug \
i54633.wrong \
i54980.aff \
i54980.dic \
i54980.good \
i58202.aff \
i58202.dic \
i58202.good \
i58202.sug \
i58202.wrong \
maputf.aff \
maputf.dic \
maputf.sug \
maputf.wrong \
reputf.aff \
reputf.dic \
reputf.sug \
reputf.wrong \
slash.aff \
slash.dic \
slash.good \
ignore.aff \
ignore.dic \
ignore.good \
ignoreutf.aff \
ignoreutf.dic \
ignoreutf.good \
right_to_left_mark.aff \
right_to_left_mark.dic \
right_to_left_mark.good \
1592880.aff \
1592880.dic \
1592880.good \
1695964.aff \
1695964.dic \
1695964.sug \
1695964.wrong \
1463589.aff \
1463589.dic \
1463589.sug \
1463589.wrong \
1463589_utf.aff \
1463589_utf.dic \
1463589_utf.sug \
1463589_utf.wrong \
IJ.aff \
IJ.dic \
IJ.good \
IJ.sug \
IJ.wrong \
i68568.aff \
i68568.dic \
i68568.wrong \
i68568utf.aff \
i68568utf.dic \
i68568utf.wrong \
1706659.aff \
1706659.dic \
1706659.wrong \
1748408-1.aff \
1748408-1.dic \
1748408-1.good \
1748408-2.aff \
1748408-2.dic \
1748408-2.good \
1748408-3.aff \
1748408-3.dic \
1748408-3.good \
1748408-4.aff \
1748408-4.dic \
1748408-4.good \
digits_in_words.aff \
digits_in_words.dic \
digits_in_words.wrong \
colons_in_words.aff \
colons_in_words.dic \
ngram_utf_fix.aff \
ngram_utf_fix.dic \
ngram_utf_fix.good \
ngram_utf_fix.sug \
ngram_utf_fix.wrong \
morph.aff \
morph.dic \
morph.good \
morph.morph \
1975530.aff \
1975530.dic \
1975530.good \
1975530.wrong \
fullstrip.aff \
fullstrip.dic \
fullstrip.good \
iconv.aff \
iconv.dic \
iconv.good \
iconv2.aff \
iconv2.dic \
iconv2.good \
oconv.aff \
oconv.dic \
oconv.good \
oconv.sug \
oconv.wrong \
oconv2.aff \
oconv2.dic \
oconv2.good \
oconv2.morph \
oconv2.root \
oconv2.test \
encoding.aff \
encoding.dic \
encoding.good \
opentaal_forbiddenword1.aff \
opentaal_forbiddenword1.dic \
opentaal_forbiddenword1.good \
opentaal_forbiddenword1.sug \
opentaal_forbiddenword1.wrong \
opentaal_forbiddenword2.aff \
opentaal_forbiddenword2.dic \
opentaal_forbiddenword2.good \
opentaal_forbiddenword2.sug \
opentaal_forbiddenword2.wrong \
opentaal_forbiddenword2.aff \
opentaal_forbiddenword2.dic \
opentaal_forbiddenword2.good \
opentaal_forbiddenword2.sug \
opentaal_forbiddenword2.wrong \
opentaal_keepcase.aff \
opentaal_keepcase.dic \
opentaal_keepcase.good \
opentaal_keepcase.sug \
opentaal_keepcase.wrong \
arabic.aff \
arabic.dic \
arabic.wrong \
2970240.aff \
2970240.dic \
2970240.good \
2970240.wrong \
2970242.aff \
2970242.dic \
2970242.good \
2970242.wrong \
breakoff.aff \
breakoff.dic \
breakoff.good \
breakoff.wrong \
opentaal_cpdpat.aff \
opentaal_cpdpat.dic \
opentaal_cpdpat.good \
opentaal_cpdpat.wrong \
opentaal_cpdpat2.aff \
opentaal_cpdpat2.dic \
opentaal_cpdpat2.good \
opentaal_cpdpat2.wrong \
2999225.aff \
2999225.dic \
2999225.good \
korean.aff \
korean.dic \
korean.good \
korean.wrong \
onlyincompound2.aff \
onlyincompound2.dic \
onlyincompound2.good \
onlyincompound2.wrong \
forceucase.aff \
forceucase.dic \
forceucase.good \
forceucase.sug \
forceucase.wrong \
nepali.aff \
nepali.dic \
nepali.good \
hu.aff \
hu.dic \
hu.good \
warn.aff \
warn.dic \
warn.good \
timelimit.aff \
timelimit.dic \
timelimit.good \
timelimit.wrong \
ignoresug.aff \
ignoresug.dic \
ignoresug.good \
ignoresug.morph \
limit-multiple-compounding.aff \
limit-multiple-compounding.dic \
limit-multiple-compounding.good \
limit-multiple-compounding.wrong \
\
utf8_nonbmp.test

View File

@@ -0,0 +1,7 @@
# simple example for affix compression (see Hunspell(4))
PFX A Y 1
PFX A 0 re .
SFX B Y 2
SFX B 0 ed [^y]
SFX B y ied y

View File

@@ -0,0 +1,4 @@
3
hello
try/B
work/AB

View File

@@ -0,0 +1,7 @@
hello
try
tried
work
worked
rework
reworked

View File

@@ -0,0 +1,12 @@
# aliases for flag vectors (AF)
# AB -> 1
# A -> 2
AF 2
AF AB
AF A
SFX A Y 1
SFX A 0 x .
SFX B Y 1
SFX B 0 y/2 .

View File

@@ -0,0 +1,2 @@
1
foo/1

View File

@@ -0,0 +1,4 @@
foo
foox
fooy
fooyx

View File

@@ -0,0 +1,17 @@
# aliases for flag vectors (AF) and morphological descriptions (AM)
# AB -> 1
# A -> 2
AF 2
AF AB
AF A
AM 3
AM is:affix_x
AM ds:affix_y
AM po:noun xx:other_data
SFX A Y 1
SFX A 0 x . 1
SFX B Y 1
SFX B 0 y/2 . 2

View File

@@ -0,0 +1,2 @@
1
foo/1 3

View File

@@ -0,0 +1,4 @@
foo
foox
fooy
fooyx

View File

@@ -0,0 +1,12 @@
> foo
analyze(foo) = st:foo po:noun xx:other_data
stem(foo) = foo
> foox
analyze(foox) = st:foo po:noun xx:other_data is:affix_x
stem(foox) = foo
> fooy
analyze(fooy) = st:foo po:noun xx:other_data ds:affix_y
stem(fooy) = fooy
> fooyx
analyze(fooyx) = st:foo po:noun xx:other_data ds:affix_y is:affix_x
stem(fooyx) = fooy

View File

@@ -0,0 +1,18 @@
# morph. aliases with complex prefixes
COMPLEXPREFIXES
WORDCHARS _
AM 4
AM affix_1/
AM affix_2/
AM /suffix_1
AM [stem_1]
PFX A Y 1
PFX A 0 tek . 1
PFX B Y 1
PFX B 0 met/A . 2
SFX C Y 1
SFX C 0 _test_ . 3

View File

@@ -0,0 +1,2 @@
1
ouro/BC 4

View File

@@ -0,0 +1,4 @@
ouro
metouro
tekmetouro
ouro_test_

View File

@@ -0,0 +1,8 @@
> ouro
analyze(ouro) = [stem_1] ouro:ts
> metouro
analyze(metouro) = affix_2/ ouro:ts [stem_1]
> tekmetouro
analyze(tekmetouro) = affix_1/ affix_2/ ouro:ts [stem_1]
> ouro_test_
analyze(ouro_test_) = [stem_1] ouro:ts /suffix_1

View File

@@ -0,0 +1,8 @@
# check uppercase forms of allcaps word + affix and words with mixed casing
WORDCHARS '.
SFX S N 1
SFX S 0 's .
PFX L N 1
PFX L 0 L' .

View File

@@ -0,0 +1,4 @@
2
OpenOffice.org
UNICEF/S
Afrique/L

View File

@@ -0,0 +1,6 @@
OpenOffice.org
OPENOFFICE.ORG
UNICEF's
UNICEF'S
L'AFRIQUE
L'Afrique

View File

@@ -0,0 +1,3 @@
OpenOffice.org
UNICEF
UNICEF's

View File

@@ -0,0 +1,3 @@
Openoffice.org
Unicef
Unicef's

View File

@@ -0,0 +1,6 @@
# forbidden all caps words are case sensitive
# iPod -> ipodos ("iPodic" in Hungarian)
FORBIDDENWORD *
SFX s N 1
SFX s 0 os .

View File

@@ -0,0 +1,4 @@
3
iPod/s
iPodos/*
ipodos

View File

@@ -0,0 +1,4 @@
iPod
IPOD
ipodos
IPODOS

View File

@@ -0,0 +1,2 @@
iPod
ipodos

View File

@@ -0,0 +1,2 @@
ipod
iPodos

View File

@@ -0,0 +1,10 @@
# homonym support
WORDCHARS '
SFX s N 1
SFX s 0 s .
SFX S N 1
SFX S 0 's .

View File

@@ -0,0 +1,7 @@
4
UNESCO/S
Unesco/S
Nasa/S
NASA/S
ACTS
act/s

View File

@@ -0,0 +1,13 @@
UNESCO
Unesco
UNESCO's
Unesco's
UNESCO'S
NASA
Nasa
NASA's
Nasa's
NASA'S
ACTS
acts
Acts

View File

@@ -0,0 +1,4 @@
unesco
unesco's
nasa
nasa's

View File

@@ -0,0 +1,6 @@
SET UTF-8
WORDCHARS '.
SFX S N 1
SFX S 0 's .

View File

@@ -0,0 +1,3 @@
2
OpenOffice.org
UNICEF/S

View File

@@ -0,0 +1,4 @@
OpenOffice.org
OPENOFFICE.ORG
UNICEF's
UNICEF'S

View File

@@ -0,0 +1,3 @@
OpenOffice.org
UNICEF
UNICEF's

View File

@@ -0,0 +1,3 @@
Openoffice.org
Unicef
Unicef's

View File

@@ -0,0 +1,6 @@
SET UTF-8
TRY أ
IGNORE ٌٍَُِّْ
PFX Aa Y 1
PFX Aa 0 0/X0 أ[^ي]

View File

@@ -0,0 +1,2 @@
1
ب

View File

@@ -0,0 +1 @@
ـ

View File

@@ -0,0 +1,192 @@
# OpenOffice.org's en_US.aff file
SET ISO8859-1
TRY esianrtolcdugmphbyfvkwz'
WORDCHARS .'
PFX A Y 1
PFX A 0 re .
PFX I Y 1
PFX I 0 in .
PFX U Y 1
PFX U 0 un .
PFX C Y 1
PFX C 0 de .
PFX E Y 1
PFX E 0 dis .
PFX F Y 1
PFX F 0 con .
PFX K Y 1
PFX K 0 pro .
SFX V N 2
SFX V e ive e
SFX V 0 ive [^e]
SFX N Y 3
SFX N e ion e
SFX N y ication y
SFX N 0 en [^ey]
SFX X Y 3
SFX X e ions e
SFX X y ications y
SFX X 0 ens [^ey]
SFX H N 2
SFX H y ieth y
SFX H 0 th [^y]
SFX Y Y 1
SFX Y 0 ly .
SFX G Y 2
SFX G e ing e
SFX G 0 ing [^e]
SFX J Y 2
SFX J e ings e
SFX J 0 ings [^e]
SFX D Y 4
SFX D 0 d e
SFX D y ied [^aeiou]y
SFX D 0 ed [^ey]
SFX D 0 ed [aeiou]y
SFX T N 4
SFX T 0 st e
SFX T y iest [^aeiou]y
SFX T 0 est [aeiou]y
SFX T 0 est [^ey]
SFX R Y 4
SFX R 0 r e
SFX R y ier [^aeiou]y
SFX R 0 er [aeiou]y
SFX R 0 er [^ey]
SFX Z Y 4
SFX Z 0 rs e
SFX Z y iers [^aeiou]y
SFX Z 0 ers [aeiou]y
SFX Z 0 ers [^ey]
SFX S Y 4
SFX S y ies [^aeiou]y
SFX S 0 s [aeiou]y
SFX S 0 es [sxzh]
SFX S 0 s [^sxzhy]
SFX P Y 3
SFX P y iness [^aeiou]y
SFX P 0 ness [aeiou]y
SFX P 0 ness [^y]
SFX M Y 1
SFX M 0 's .
SFX B Y 3
SFX B 0 able [^aeiou]
SFX B 0 able ee
SFX B e able [^aeiou]e
SFX L Y 1
SFX L 0 ment .
REP 88
REP a ei
REP ei a
REP a ey
REP ey a
REP ai ie
REP ie ai
REP are air
REP are ear
REP are eir
REP air are
REP air ere
REP ere air
REP ere ear
REP ere eir
REP ear are
REP ear air
REP ear ere
REP eir are
REP eir ere
REP ch te
REP te ch
REP ch ti
REP ti ch
REP ch tu
REP tu ch
REP ch s
REP s ch
REP ch k
REP k ch
REP f ph
REP ph f
REP gh f
REP f gh
REP i igh
REP igh i
REP i uy
REP uy i
REP i ee
REP ee i
REP j di
REP di j
REP j gg
REP gg j
REP j ge
REP ge j
REP s ti
REP ti s
REP s ci
REP ci s
REP k cc
REP cc k
REP k qu
REP qu k
REP kw qu
REP o eau
REP eau o
REP o ew
REP ew o
REP oo ew
REP ew oo
REP ew ui
REP ui ew
REP oo ui
REP ui oo
REP ew u
REP u ew
REP oo u
REP u oo
REP u oe
REP oe u
REP u ieu
REP ieu u
REP ue ew
REP ew ue
REP uff ough
REP oo ieu
REP ieu oo
REP ier ear
REP ear ier
REP ear air
REP air ear
REP w qu
REP qu w
REP z ss
REP ss z
REP shun tion
REP shun sion
REP shun cion

View File

@@ -0,0 +1,29 @@
28
created/U
create/XKVNGADS
imply/GNSDX
natural/PUY
like/USPBY
convey/BDGS
look/GZRDS
text
hello
said
sawyer
NASA
rotten
day
tomorrow
seven
FAQ/SM
can't
doesn't
etc
won't
lip
text
horrifying
speech
suggest
uncreate/V
Hunspell

View File

@@ -0,0 +1,28 @@
created
uncreate
uncreated
imply
implied
unnatural
conveyed
sawyer
NASA
FAQs
can't
doesn't
won't
Created
Hello
HELLO
NASA
etc.
etc
HELLO
lip.
text.
NASA.
Text.
TEXT.
Hunspell.
HUNSPELL.
HUNSPELL...

View File

@@ -0,0 +1,11 @@
looked, look
text
hello
said
rotten day, rotten-day, rotten
tomorrow
seven
NASA
horrifying
speech
suggest

View File

@@ -0,0 +1,11 @@
loooked
texxt
hlelo
seid
rottenday
tomorow
seeeven
Nasa
horrorfying
peech
sugesst

View File

@@ -0,0 +1,198 @@
# OpenOffice.orgs en_US.aff file
# with Unicode apostrophe:
SET UTF-8
TRY esianrtolcdugmphbyfvkwzESIANRTOLCDUGMPHBYFVKWZ'
MAXNGRAMSUGS 1
WORDCHARS .'
PFX A Y 1
PFX A 0 re .
PFX I Y 1
PFX I 0 in .
PFX U Y 1
PFX U 0 un .
PFX C Y 1
PFX C 0 de .
PFX E Y 1
PFX E 0 dis .
PFX F Y 1
PFX F 0 con .
PFX K Y 1
PFX K 0 pro .
SFX V N 2
SFX V e ive e
SFX V 0 ive [^e]
SFX N Y 3
SFX N e ion e
SFX N y ication y
SFX N 0 en [^ey]
SFX X Y 3
SFX X e ions e
SFX X y ications y
SFX X 0 ens [^ey]
SFX H N 2
SFX H y ieth y
SFX H 0 th [^y]
SFX Y Y 1
SFX Y 0 ly .
SFX G Y 2
SFX G e ing e
SFX G 0 ing [^e]
SFX J Y 2
SFX J e ings e
SFX J 0 ings [^e]
SFX D Y 4
SFX D 0 d e
SFX D y ied [^aeiou]y
SFX D 0 ed [^ey]
SFX D 0 ed [aeiou]y
SFX T N 4
SFX T 0 st e
SFX T y iest [^aeiou]y
SFX T 0 est [aeiou]y
SFX T 0 est [^ey]
SFX R Y 4
SFX R 0 r e
SFX R y ier [^aeiou]y
SFX R 0 er [aeiou]y
SFX R 0 er [^ey]
SFX Z Y 4
SFX Z 0 rs e
SFX Z y iers [^aeiou]y
SFX Z 0 ers [aeiou]y
SFX Z 0 ers [^ey]
SFX S Y 4
SFX S y ies [^aeiou]y
SFX S 0 s [aeiou]y
SFX S 0 es [sxzh]
SFX S 0 s [^sxzhy]
SFX P Y 3
SFX P y iness [^aeiou]y
SFX P 0 ness [aeiou]y
SFX P 0 ness [^y]
SFX M Y 1
SFX M 0 's .
SFX B Y 3
SFX B 0 able [^aeiou]
SFX B 0 able ee
SFX B e able [^aeiou]e
SFX L Y 1
SFX L 0 ment .
REP 88
REP a ei
REP ei a
REP a ey
REP ey a
REP ai ie
REP ie ai
REP are air
REP are ear
REP are eir
REP air are
REP air ere
REP ere air
REP ere ear
REP ere eir
REP ear are
REP ear air
REP ear ere
REP eir are
REP eir ere
REP ch te
REP te ch
REP ch ti
REP ti ch
REP ch tu
REP tu ch
REP ch s
REP s ch
REP ch k
REP k ch
REP f ph
REP ph f
REP gh f
REP f gh
REP i igh
REP igh i
REP i uy
REP uy i
REP i ee
REP ee i
REP j di
REP di j
REP j gg
REP gg j
REP j ge
REP ge j
REP s ti
REP ti s
REP s ci
REP ci s
REP k cc
REP cc k
REP k qu
REP qu k
REP kw qu
REP o eau
REP eau o
REP o ew
REP ew o
REP oo ew
REP ew oo
REP ew ui
REP ui ew
REP oo ui
REP ui oo
REP ew u
REP u ew
REP oo u
REP u oo
REP u oe
REP oe u
REP u ieu
REP ieu u
REP ue ew
REP ew ue
REP uff ough
REP oo ieu
REP ieu oo
REP ier ear
REP ear ier
REP ear air
REP air ear
REP w qu
REP qu w
REP z ss
REP ss z
REP shun tion
REP shun sion
REP shun cion
McDonaldssá/w
McDonaldsszá/g3) st:McDonald’s po:noun_prs is:TRANS
McDonaldsszal/g4) st:McDonald’s po:noun_prs is:INSTR
McDonaldssal/w

View File

@@ -0,0 +1,30 @@
28
created/U
create/XKVNGADS
imply/GNSDX
natural/PUY
like/USPBY
convey/BDGS
look/GZRDS
text
hello
said
sawyer
NASA
rotten
day
tomorrow
seven
FAQ/SM
cant
doesnt
etc
wont
lip
text
horrifying
speech
suggest
uncreate/V
Hunspell
İzmir

View File

@@ -0,0 +1,33 @@
created
uncreate
uncreated
imply
implied
unnatural
conveyed
sawyer
NASA
FAQs
cant
doesnt
wont
Created
Hello
HELLO
NASA
etc.
etc
HELLO
lip.
text.
NASA.
Text.
TEXT.
Hunspell.
HUNSPELL.
İzmir
İZMİR
İzmir.
İZMİR.
Imply
IMPLY

View File

@@ -0,0 +1,13 @@
looked, look
text
hello
said
rotten day, rotten-day, rotten
tomorrow
seven
NASA
horrifying
speech
suggest
Imply
IMPLY

View File

@@ -0,0 +1,13 @@
loooked
texxt
hlelo
seid
rottenday
tomorow
seeeven
Nasa
horrorfying
peech
sugesst
İmply
İMPLY

View File

@@ -0,0 +1,10 @@
# word break points test, recursive break at dash and n-dash
SET UTF-8
BREAK 2
BREAK -
BREAK
WORDCHARS -
FORBIDDENWORD !

Some files were not shown because too many files have changed in this diff Show More