Fail gracefully on incorrectly formatted header records

[sobolevnrm]

There are many changes in PDB2PQR since 2.1.1; we've just updated the document with a summary of these changes.

The 3nir.pdb "works" with the latest version (3.1.0, released last week) of PDB2PQR: the program fails gracefully when it encounters a gap in the backbone.

The 3m24_final_tot.pdb fails with 3.1.0 because of CISPEP and CRYST1 records that don't comply with the PDB format. However, it does not fail very gracefully--and probably shouldn't fail at all since this information isn't essential for the PDB2PQR functionality. I'll create a new issue to fix this.

Originally posted by @sobolevnrm in #164 (comment)