Re: cutting down the TODO list thread

Here is the next section on data types, proposed to be moved to the "not
worth doing" page. As before, if there are any objections, do speak up.
I'll make the move in a few days.

**Datatypes

- Fix data types where equality comparison is not intuitive, e.g. box
There is likely no way to do this without breaking applications. We
already have a big warning about this in the docs.

- Add IMMUTABLE column attribute
Controversial

*Domains (this entire section would go)

- Allow functions defined as casts to domains to be called during casting
- Allow values to be cast to domain types
- Make domains work better with polymorphic functions
Old and difficult

*Date/Time

- Allow TIMESTAMP WITH TIME ZONE to store the original timezone information
- Have timestamp subtraction not call justify_hours()
Very old

- Allow a comma to denote fractional seconds in ISO-8601-compliant times
(and timestamps)
Apparent lack of interest

*Arrays

- Add function to detect if an array is empty
- Improve handling of NULLs in arrays
Lack of interest

*Money (this entire section would go)

- Add locale-aware MONEY type, and support multiple currencies
- MONEY dumps in a locale-specific format making it difficult to restore to
a system with a different locale
The money datatype seems kind of obsolete anyway, and there doesn't seem
to be demand to improve it.

*Text Search

- Allow dictionaries to change the token that is passed on to later
dictionaries
- Consider a function-based API for '@@' searches
Very old

- Improve text search error messages
One of the gripes has been fixed already, in any case it's very old

- tsearch and tsdicts regression tests fail in Turkish locale on glibc
Bug report that refers to locale behavior from 2009

- Improve handling of dash and plus signs in email address user names, and
perhaps improve URL parsing
Difficult

*XML (difficult section -- plenty of bugs which should be fixed, but also
old and low interest)

- Allow XML arrays to be cast to other data types
Very old

- Allow reliable XML operation non-UTF8 server encodings
Difficult

- Move XSLT from contrib/xml2 to a more reasonable location
Lack of consensus

- Improve the XSLT parameter passing API
Lack of consensus

- xpath_table needs to be implemented/implementable to get rid of
contrib/xml2
- xpath_table is pretty broken anyway
Unclear path forward

- better handling of XPath data types
- Improve handling of PIs and DTDs in xmlconcat()
Zero interest

- Restructure XML and /contrib/xml2 functionality
As discussed in the thread, it's an unrealistically large project

--
John Naylor
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company