When retrieving a remote index for replay, validate the cdxj syntax

[machawk1]

Use https://github.com/oduwsdl/ORS/wiki/CDXJ as a basis, except there is a newer version that has a different symbol (# ?) at the start of the metadata line. Does a library or explicit BNF exist for this @ibnesayeed ? This might work better as its own Py library to be utilized elsewhere, too.

[machawk1]

I'm guessing one does not exist given the query for "cdxj" in pypi directs the user to ipwb.

[machawk1]

Or rather, the functionality of pywb could be reused since, y'know, we are already using that very module.

[machawk1]

On a second look, it does not appear that pywb has any sort of CDXJ validation. There is a parse_cdxj() that might throw an exception we can catch.

[machawk1]

Started on this and wrote tests in branch issue-86 but broke something where even sample cdxj won't validate. Will fix here before merging.

[ibnesayeed]

No, we don't have any (E)BNF for CDXJ yet, but we will.