Mailing List Archive

Mispeelings going wrong
The Dutch misspellings page seems to have gone crazy, showing for many words
not only the pages with mis-spelled words, but also the pages where the
word or part of it is spelled correctly. For example, the second misspelling on
http://nl.wikipedia.org/w/wiki.phtml?title=Speciaal:Maintenance&subfunction=mispeelings
is 'adfundum', being a mis-spelling of 'ad fundum'. There are 18 pages there,
but those are actually pages with 'ad' rather than 'ad fundum'. What is going
on? Is there a technical error or is there something wrong with our
[[Wikipedia:Veel voorkomende spelfouten]]?

Andre Engels
Re: Mispeelings going wrong [ In reply to ]
On mer, 2003-01-15 at 07:30, Andre Engels wrote:
> The Dutch misspellings page seems to have gone crazy, showing for many words
> not only the pages with mis-spelled words, but also the pages where the
> word or part of it is spelled correctly. For example, the second misspelling on
> http://nl.wikipedia.org/w/wiki.phtml?title=Speciaal:Maintenance&subfunction=mispeelings
> is 'adfundum', being a mis-spelling of 'ad fundum'. There are 18 pages there,
> but those are actually pages with 'ad' rather than 'ad fundum'. What is going
> on? Is there a technical error or is there something wrong with our
> [[Wikipedia:Veel voorkomende spelfouten]]?

Some of the entries used tabs instead of spaces to separate the
misspelled word from the comment/correction. This code was hard-coded to
check for spaces:

$x = explode ( " " , $x ) ;
$x = array_shift ( $x ) ;

so it considered "adfundum\t(ad" to be its search word.

I've replaced it with this:

$x = preg_replace( '/^(\S+).*$/', '$1', $x );

It should now accept any whitespace as a delimiter instead of just
spaces.

-- brion vibber (brion @ pobox.com)
Re: Mispeelings going wrong [ In reply to ]
Is it possibel to exclude the language links?

example;

http://nl.wikipedia.org/w/wiki.phtml?title=Speciaal:Maintenance&subfunction
=mispeelings

The error "electron (elektron)"

The article http://nl.wikipedia.org/wiki/Elektron

The only "electron" is [[en:Electron]]



Giskart
Re: Re: Mispeelings going wrong [ In reply to ]
Giskart wrote:
> Is it possibel to exclude the language links?

Only by excluding them from searches altogether (bad!) or perhaps
loading and checking the text of every match manually (ugh).

-- brion vibber (brion @ pobox.com)