[ic] guess_cc_type(), does the 'test way' need to be changed ?

ns linux at post10.tele.dk
Wed Sep 14 05:34:30 EDT 2005


Hi, this is mostly for brainstorming...

I did observe that Mike have changed the guess_cc_type() function to 
cover more cards.

Because we (here in DK) can charge different fees for different CC card, 
like :

- danish issued CC ( fee 1.95 + 0.1 %)
- danish issued visa electron (1.95 + 1.25)
- international issued CC's (1.95 + 5.5%)

Due to this I did contact our national clearing house, and after 'some' 
dialog I got a 5 pages listening of prefixes for different CC numbers. 
allthough a lot of them is really local cc cards (like shell or ikea 
cards), it gives some challenges to pick out the segments for the dk 
issued cards.


Eg. Danish issued Visa Electron is prefixed/segmented as following (and 
can be between 13 & 19 digits long)
402620
405245
405266 - 405277
405832 - 405839
405860
405864
405934
406432 - 406438
406443
417500 – 417536
417538 – 417548
417550 – 417599
417600 - 417603

JBC is : 3528 - 3589 (I can not confirm the prefix 1800 & 2131 for JBC, 
can any of you ?)

and so on (DK issued MC is 38 lines long...)

So, I was thinking of a new way to make this CC tests, like a lookup in 
a table, but due to the use of prefixes like 405833 is also 'coveret' by 
the rule /^4(?:\d{12}|\d{15})$/ , I can not se a 'easy' way to make 
another way to make the test, else than expanging the existing routine..


So any ideas ?

- niels



More information about the interchange-users mailing list