Bugfixes and improved error handling

It's a new year and a new version of dullard!

Changelog

• Empty cells that begin a row are no longer skipped (Issue #12)
• Added error handling (fc225b3, 2e4c378)
• Cells formatted as percentages are now read correctly (eabe2db)

Backwards incompatible changes

• Zip reading errors that previously raised Zip::Error now raise Dullard::Error.
• Some invalid XLSX files that previously either raised no errors or miscellaneous runtime errors now raise Dullard::Error as errors are encountered.

Updated dependencies and date support

Small release, but with brand spanking new support for date parsing in spreadsheets.

Thanks @phendryx and @lucke84 for contributing!

Bugfixes, sheet size metadata support, dropping Ruby 1.9 support

This release fixes an issue where worksheets are accessed with the wrong filename, and fixes reading empty cells.

If worksheets have metadata for the number of rows in the sheet, it will be returned with rows.size. This feature breaks compatibility with Ruby < 2.0, but if there is user demand, I will try to backport the library to 1.9.

Thanks to @c0, @chewi, and @pythonicrubyist for contributing!