I have quite an opaque problem.
I have a regular JTable that I have filled with values coming from a text file, a report. Some of these fields contain double numbers, formatted in the scientific way (e.g. 1.25E-7).
My methodology is to parse these strings to a double and fill the table column with the result. On failure to parse the number, I insert a "Double.NaN"
Next I have a table sorter. As the code is now, if you click on the column header containing the double numbers, it will sort the scientific notation values correctly under Linux, Mac, and WindowsXP 64bits but NOT under WindowsXP 32bits. In that OS, it seems to sort the rows by its
string value, not as a scientific notation double (eg. 1.25E-3 before 2.56E-5).
What the 'ellll am I doing wrong!?
Here is the row sorter code: