2007-09-30 (0.1.30)
- fixed for R 2.6.0+ and devel: CHAR_ELT() now returns const char*, removed
CR in src/Makevars.in

2007-07-11 (0.1.30)
- fix for changes in formals() in latest R-devel (cannot do formals() directly
on primitive functions like log(), has to do formals(args(log)) )
- fixed renameSdf documentation

2007-07-10 (0.1.27-0.1.29)
- created method _make_row_index(), similar to R that creates an index
vector, so that methods no longer have to consider the type of passed index
- added [<-.sqlite.vector, tests for that
- fix builds for windows, *bsd (mac)

2007-02-18 (0.1.26)
- updated to sqlite-3.3.13
- added R function sdfSelect, added tests for that
- fixes & enhancements on _get_factor_levels1, _get_vector_index_typed_result,
etc.

2006-10-03 (17:45 Manila, 0.1.25)
- fixed 0.1.24 warnings (errors in R/All.R)
- used sqlite context on aggregates. somehow, sizeof on struct with long
double worked. have to look out for that. 

2006-09-25 (18:50 Manila)
- inlined _expand_buf
- fixed sql buffer resizing in sdf_get_index
- added sdflm2

2006-09-20 (18:50 Manila)
- new sqlite.vector(vec), creates a 1-column sdf from a vector
- sqlite.matrix : dimnames, [, other experimental stuff
- sort.sqlite.vector sorts to a table to the sdf, no more creating a new
sdf
 
2006-09-06 (09:55 Manila)
- fixed linux configure script (add #include <unistd> in src/SQLite/tool/lemon.c)
- removed warnings, errors in configure.win (sh again, "uninitialized" vars)

2006-09-03 (15:34 Manila)
- moved default data dir to ./.SQLiteDF
- "generalized" sqlite vector, in that any table column can be used as an
sqlite.vector (previously, only columns of table sdf_data)
- fixed sqlite.matrix rownames

2006-08-28 (08:00 Manila)
- added sqlite.matrix and some of its methods
- fixed print.sqlite.data.frame bug
- added msg when an existing workspace.db is loaded

2006-08-27 (05:30 Manila)
- added sed script in configure.win to change chmod to ls

2006-08-24 (20:05 Manila)
- moved quantile.default to version conditional
- removed sqlite.matrix export

2006-08-23 (10:27 Manila)
- added mean.sqlite.vector, fixed summary.sqlite.vector, added
summary.sqlite.data.frame
- is.list.sqlite.vector === FALSE
- fixed version conditionals for generic sort
- fixed sdfImportSQLite, added test case
- fixed _copy_factor_levels2
- checked on linux (R-2.3.0, R-devel)

2006-08-23 (10:27 Manila)
- added [], [n], [,] indexing for sdf. added tests for those
- added Prof. Bates print.sqlite.data.frame with modifications,
print.sqlite.vector
- changed inameSdf to return filename too
- changed some Rprintf to error() and warning()
- added doc for rbindSdf
- fixed bug in sort.sqlite.vector
- fixed bug in sdf_get_index, where a new sdf is added to the workspace twice
(one in _create_sdf_skeleton2 and another in sdf_get_index)
- sort() still does not work for 2.3.x (not generic)

2006-08-23 (00:20 Manila)
- added all.equal.sqlite.vector
- added == for character
- updated test script
- used conditional on NAMESPACE for sort
- R CMD check runs for me in linux (R-devel, R-2.3.0), windows (R-2.3.0)

2006-08-22 (19:45 Manila)
- removed rbind S3, replaced with rbindSdf()
- fixed rbindSdf() for devel
- fixed a couple of sqlite.vector functions that still does "inheritance"
- re-added DBI in depends because it has no namespace and DBI::: does not work
- added head(), tail(). however both do not follow the original's row names yet
- added optional tests for sdfImportDBI, rbindSdf
- added doc for typeSvec, has.typeSvec
- runs for me in both R-devel, R-2.3.1

2006-08-22 (10:45 Manila)
- sqlite.data.frame now does not inherit from data.frame
- sqlite.vector does not inherit from the builtin data types
- added attribute sdf.vector.type to hold the type of sqlite.vector
- commented out summary() tests in tests.R, added tests n classes of sdf's
- added function typeSvec and is.typeSvec to query type of sqlite.vector
- fixed row.names comparison. I guess I put that before checking in and after
  running tests, because ==.sqlite.vector is not implemented, only Ops.sqlite.vector.
- updated version to 0.1.13

2006-08-21 (9:30 Manila)
- added symbols SDF_RowNamesSymbol, SDF_DimSymbols, SDF_DimNamesSymbols. SDF
  row names is now attached to SDF_RowNamesSymbol ("sdf.row.names")
- added row.names.sqlite.data.frame
- added src/win/opcode.h
- used ScalarLogical(), ScalarInteger(), ...
- removed ";" in All.R
- removed eval()
- made configure, cleanup executable

2006-08-20 (6:35 Manila)
- configure file fixes from Prof. Ripley
- added native routines registration

2006-08-20 (00:32 Manila)
- added docs in HACKING
- fixed latex error on sqlite.data.frame.Rd
- changed configure to copy libsqlite3.la to SQLite/ (like configure.win)
- finished Ops (%%, %/%), and (round, signif, log)
- fixed bug on operand order for Ops
- added test cases

2006-08-19 (9:39 Manila time)
- added configure files from Prof. Ripley
- added documentations for inameSDF, sdfImportSQLite, sdfImportDBI, sdflm.
- added HACKING notes
- added list of functions known to work in SQLiteDF.package.Rd
- changed sqlite_vector accumulator to long double, fixed the accuracy problem

2006-08-18  Douglas Bates  <bates@R-project.org>
       * R/All.R (with.sqlite.data.frame, summary.sqlite.vector,
       as.logical.sqlite.vector, as.integer.sqlite.vector,
       as.double.sqlite.vector): Modify calling sequence of S3 methods to
       correspond to generic.
       * man/sqlite.data.frame.Rd: add keywords

2006-08-18 (6:05pm Manila time)
- fixed R-devel tests failure
- removed Depends on biglm, DBI. used namespace qualifiers 
- added ChangeLog
- added tests on summary
- added version in DESCRIPTION, starting at 0.1.11

2006-08-17
- fix on USE_SDF1 bug, when a function uses 2 sdf's the 1st may be unloaded 
  because of low use count
- fix on transactions
- started implementing sqlite.matrix
- commented out parts of the test that fails due to "minor technicalities" (precision, NA). will test them again...
- summary works, almost the same as original df except for NA handling

2006-08-11
- added sdflm, as.data.frame.sdf, eval, sdfImportDBI, sdfImportSQLite
- added Requires DBI, biglm in DESCRIPTION
- corrected generic-ized is.list, length (already generic)
- added "begin", "commit" in sdf_create_sdf, rbind, and row ops. absence of
  which was the cause of the slowness
- factored out _prepare_attach2() w/c checks the # of attached sqlite db's
  and detaches as necessary.
- fixed bug on detecting R character type, STRSXP instead of CHARSXP
- added utility _str_tolower(), file name & line number in _sqlite_error
- override stats::quantile env
- change Makevars to link statically

2006-08-01
- added methods is.list and as.list to sdf's. lapply (sapply) now works
- rbind(sdf, df) works, however no "labeled" rbind
- finished Op group (%% & %/% implemented wrongly for now...)
- with.sqlite.data.frame is already working
- sort.sqlite.vector() is working
- renamed USE_SDF to USE_SDF1, restricted it to only use buf1 and below.
  there was something affected by this, which calls USE_SDF twice (initially
  i thought it would only be used on top of funcs, so using all buffers is
  ok. if you use it in the middle of the func, it would overwrite all buffers).
- fixed sdf_get_row_count, _create_svector1 (renamed from *2, added ordered),
  renamed _create_sdf_skeleton2 to skeleton1 because it uses USE_SDF1,
  changed _get_row_count2 to have quoted and non-quoted option
- added tests

2006-07-17
- added documentation
- finished Summary group, cummulative stuffs in Math group for sqlite.vector
- revised tests to use stop()
- started rbind for data frames

2006-07-10
- added windows dll's, Makevar.win. windows dll still has the default 10 max
  attached db
- added USE_SDF, w/c does the juggling of which db to be attached

2006-07-10
- moved libsqlite3.* to inst/sqlite/linux, added inst/sqlite/win to have
  sqlite windows static dll
- change libsqlite3.a to have 31 max attach-ed db's
- most of Math.sqlite.vector is working already
- fixed sdf_attach_sdf
- factored out creation of sdf in _create_sdf_skeleton2
- added dupSdf, renameSdf

2006-07-05
- Added test script
- Fixed compile error for Ubuntu dapper (?), as tested by Mr. Medri. 

2006-07-04
- fixed lots of monstrous bugs introduced by refactoring stuffs
- fixed R CMD check errors on documentation

2006-07-04
- added row.names support for sdf
- -pedantic cleanups
- fixes by Prof. Bates: duplicate include & sqlite3_stmt mis-declaration
- minor fixes

2006-06-27
- added sqlite.vector, its indexers, length and conversion to ordinary vec
- indexers for sqlite.data.frame: [, [[, $
- attachSdf, detachSdf

2006-06-26
1st submit