jdb stacktrace regex doesn't account for thousands separator

[nemrod]

Presumably because of a locale setting, my jdb outputs in this format:
[1] com.example.Class.Method (Class.java:1,361)
Note that the line number has a thousands separator in the form of a comma. This causes the regex to not match and vebugger doesn't recognise the breakpoint being hit. I've successfully changed it to accept everything between the colon and the parenthesis, and then substitute() comma with nothing. I could do a PR with what I have, but it seems to me there should be a better way; either forcing jdb to output it in a predictable way (doing locale stuff), or if vim has a better way to parse numbers according to locale, or something.