Rom

[fixed] more bugs with String (bytes) columns in rom for Python 3.x

[fixed] bug with String (bytes) columns in rom for Python 3.x

[fixed] an issue with the docs not generating 20+ notes and warnings from the
source code
[changed] moved on_delete action from OneToMany to ManyToOne, where it belongs
[added] OneToOne column with on_delete functionality, thanks to github user
https://github.com/jmmills for the feature request
[changed] on_delete reference scanning now based ManyToOne/OneToOne
relationships, ensuring that references are now always handled as part of
on_delete processing
[added] missing 'String' column for Python 3.x, which is really just a bytes
column that shares its name with the 2.x String column
[updated] and clarified documentation relating to column arguments and for
String, Text, ManyToOne, OneToMany, and OneToOne columns
[added] on_delete 'set null' and 'set default' options for ManyToOne and
OneToOne columns
[added] OneToMany columns can be used to reference the reverse of a OneToOne
relationship column

[updated] Copyright year, development status, Python version compatability,
and stopped installing the test package as part of installation.
[fixed] Tests for Python 2.6 .
[fixed] Issue with prefix/suffix indexes on string columns using the new
SIMPLE and CASE_INSENSITIVE keygen functions

[fixed] a double-delete issue with setting a column/attribute to None when the
column/attribute was already empty/None. Thanks to Github user
https://github.com/hakib for the bug report.
[added] keyword-only argument to .get_by() for explicitly fetching numeric
ranges when both a numeric and unique index are specified on a column.
Thanks to Github user https://github.com/hakib for the use-case where the
ambiguity was discovered.
[added] more documentation for optional arguments passed to .get_by()
[changed] Indexes on String/Text columns will start to warn about a
non-specified keygen function in 0.30.0, and will raise an exception in
0.31.0 and later when one isn't specified. This is to ensure that users
are getting the index on their data that is expected. The existing
full-text indexing keygen for string/text will get aliases of
rom.FULL_TEXT, rom.columns.FULL_TEXT, and rom.util.FULL_TEXT .
[added] rom.SIMPLE and rom.CASE_INSENSITIVE basic order keygens for String and
Text column indexes, as well as similarly named aliases in rom.columns and
rom.util, like rom.FULL_TEXT has.

[added] Integer columns can now be unique. Thanks to Github user
https://github.com/egalitarian for the feature request for this obviously
missing feature.