Last modified: 2012-03-09 15:55:31 UTC
The current regex, roughly "end capture after the second dot followed by whitespace produces wildly inaccurate results for sentences with dots in the middle, for example if article title contains dots: https://en.wikipedia.org/w/api.php?action=opensearch&format=xmlfm&search=.s.p.%20v&limit=10 <Item> <Text xml:space="preserve">S. P. Venkatesh</Text> <Description xml:space="preserve">S. P. </Description> <Url xml:space="preserve">https://en.wikipedia.org/wiki/S._P._Venkatesh</Url> </Item> <Item> <Text xml:space="preserve">S. P. Velumani</Text> <Description xml:space="preserve">S. P. </Description> <Url xml:space="preserve">https://en.wikipedia.org/wiki/S._P._Velumani</Url> </Item> It should be something like "first dot followed by whitespace after a certain number of characters".
Fixed (well, improved) in r113475.