search

giellalt / bugzilla-dummy

0 stars 0 forks source link

strange compound behavior (Bugzilla Bug 522) #508

Closed albbas closed 16 years ago

albbas commented 17 years ago

This issue was created automatically with bugzilla2github

Bugzilla Bug 522

Date: 2007-09-21T13:15:12+02:00 From: Thomas Omma <> To: Thomas Omma <> CC: pbeinema, sjur.n.moshagen

Last updated: 2007-11-23T14:44:48+01:00

albbas commented 17 years ago

Comment 1929

Date: 2007-09-21 13:15:12 +0200 From: Thomas Omma <>

Davvisámi, public beta 2, 2007-09-19

One example on strange compounding behavior that seems to be rather usual:

4-part compound, not accepted by the speller:

muorralastagieraolmmoš (muorra#lasta#giera#olmmoš)

When compound is divided we get following results:

muorralastagiera - accepted (muorra#lasta#giera)

gieraolmmoš - accepted (giera#olmmoš)

lastagiera - accepted (lasta#giera)

lastagieraolmmoš - NOT accepted (lasta#giera#olmmoš)

albbas commented 17 years ago

Comment 1930

Date: 2007-09-21 13:22:34 +0200 From: Thomas Omma <>

Another example:

4-part compound, not accepted by the speller:

veahkkegoahtesilbageaidnu (veahkke#goahte#silba#geaidnu)

When compound is divided we get following results:

goahtesilbageaidnu - accepted (goahte#silba#geaidnu)

silbageaidnu - accepted (silba#geaidnu)

goahtesilba - accepted (goahte#silba )

veahkkegoahtesilba - NOT accepted (veahkke#goahte#silba )

albbas commented 17 years ago

Comment 1931

Date: 2007-09-21 13:30:43 +0200 From: Sjur Nørstebø Moshagen <>

(In reply to comment #1)

veahkkegoahtesilba - NOT accepted (veahkke#goahte#silba )

veahkkegoahte

is also NOT accepted. This case is thus a problem with veahkke- compounds.

albbas commented 17 years ago

Comment 1932

Date: 2007-09-21 13:39:10 +0200 From: Thomas Omma <>

It is not a a problem with veahkke- compounds, because -goahte is shortened form of -goahti: Veahkkegoahti is recognized.

albbas commented 17 years ago

Comment 1936

Date: 2007-09-24 10:17:55 +0200 From: Thomas Omma <>

here is a 4-word compound that is recognized:

muorralastagierraolmmoš (muorra#lasta#gierra#olmmoš)

albbas commented 17 years ago

Comment 1937

Date: 2007-09-24 11:26:23 +0200 From: Knowledge Concepts/Peter Beinema <>

2007-09-24, Peter Beinema: several of the examples can't be reproduced:

*** "muorralastagieraolmmoš" is recognised, as are all of its possible 1, 2, and 3 part sub-words; goes for Windows as well as for Mac.

*** "veahkkegoahtesilbageaidnu" is NOT recognised. "silbageaidnu" is, "goahti" is (but "goahte" not) and "veahkke" is not. So: not good, but different from findings below.

I suspect this is a lexicon issue. Further investigation implies decompilation of the lexicon and investigation of its contents.

albbas commented 17 years ago

Comment 1944

Date: 2007-09-24 14:41:44 +0200 From: Thomas Omma <>

veahkke (with lowering) is only in compounds. Base-form is veahkki

albbas commented 17 years ago

Comment 1945

Date: 2007-09-24 14:43:21 +0200 From: Sjur Nørstebø Moshagen <>

Peter, can you add what version of the lexicon your results were achieved with? Ie type 'nuvviD', and select the version string, then copy and paste into this bug report.

albbas commented 17 years ago

Comment 1946

Date: 2007-09-25 08:10:17 +0200 From: Knowledge Concepts/Peter Beinema <>

2007-09-25 Peter Beinema: repro of "muorralastagieraolmmoš" FAILED: the word and all combinations of its parts are accepted. Repro results on "veahkkegoahtesilbageaidnu" SUCCEEDED: correspond with Thomas' results exactly (see below). Versions used: .dll: 2007-09-18; .lex: public beta 1, 2007-05 AND public beta 2, 2007-09-08 OS: windows XP+SP2 Office: Office 2003 = office 11

albbas commented 16 years ago

Comment 2250

Date: 2007-11-23 14:44:48 +0100 From: Sjur Nørstebø Moshagen <>

Fixed with BOE change.

Case closed.