Last modified: 2006-12-18 01:14:03 UTC
Older ISBN may be internally divided by hyphens or spaces. The software recognizes divisions by hyphen correctly but not ISBN divided by spaces. If a space would be added to the string search (e. g. (regex) "ISBN ([0123456789-Xx ]+)" instead of (regex) "ISBN ([0123456789-Xx]+)") this bug could be fixed.
[[ISBN]] says both hyphens and spaces are valid, so I guess this should be fixed. But in the mean time, you could work around it by just using hyphens instead -- it's not as if the choice delimiters actually means anything.
Fixed in r18074.
A hyphen with a space on each side is also sometimes used as a delimiter between the ISBN and the other text on a line. So, a good fix shall also check that a space and a hyphen may not occur next to each other in a link. For example: ISBN 123456789 - 2005 In this case the year may not become a part of the ISBN link.
The example above now makes a link for the ISBN including the year, so it looks for ISBN 1234567892005 on Special:Booksources. See [[User:JePe/temp]]
It could be expanded to (?:[0-9]{3}[-\ ]?)?[0-9]+[-\ ]?[0-9]+[-\ ]?[0-9]+[-\ ]?[0-9xX] but that's quite a bit ickier than [0-9Xx\ \-]+ I don't think it would make a noticeable difference to execution time, though.
Fixed in r18415. This change makes the magic ISBN detection much stricter; it should still match all valid ISBNs, but there might be some breakage somewhere if someone had previously relied on it matching invalid ones as well.
