Instructions for importing a new release of sqlite from sqlite.org.

First, you need to be on Linux.

Find the release you want at:
  http://www.sqlite.org/cvstrac/timeline

Search for "Milestone", and find the appropriate release.  Click
through, and snag the "Date" for use in DATE line below.
Unfortunately, the actual displayed date string on that page will not
work, you have to make it YYYY/MM/DD.  [TODO(shess) Determine if the
fact that sqlite.org thinks it's running under UTC is relevant.]

DATE='2007/01/24 09:54:56'

# Get to the third_party/sqlite directory in your Chromium client.
cd ../third_party/sqlite

# Run the super awesome automatic merge tool (requires kdiff3).
# NOTE(shess): The following (which runs google_generate_preprocessed.sh) will
# produce different output on grhat versus goobuntu; I've been running it on
# goobuntu.
./google_update_sqlite.sh "$DATE"

# Resolve any conflicts.  Figure out if we've got everything we should
# have (see below), or if we can omit any changes we no longer need.

# TODO(shess) If we don't want the CVS dirs.
# find sqlite_vendor -name CVS -execdir rm -rf {} + -prune

# Find a sucker.  Send review.
# TODO(shess) Describe an appropriate comment style.  Seems like it
# should include the DATE line, and the sqlite version number.


--------------------------------------------

Why all this convolution?  The problem being solved is that we want to
harvest changes from the sqlite CVS tree, and merge them with any
local changes we make.  In CVS, you would do this using a "vendor
import", which is essentially a branch dedictated to the vendor
version which is merged with local changes.

third_party/sqlite_vendor/... is the CVS checkout for a particular
build from sqlite.org.  third_party/sqlite_google/... is the local
version, with our local modifications.  So we update the sqlite_vendor
tree, then use perforce to downintegrate changes into our
locally-modified tree.  The downintegrate will call out any
conflicting changes, but will otherwise just merge things together.
Basically, sqlite_vendor is a gateway between CVS and perforce.

Scott Hess <shess@google.com>, April 9, 2007.

--------------------------------------------

How to run the SQLite tests for the Gears version of SQLite on Linux.

cd ../third_party/sqlite_google/
mkdir build
cd build
make -f ../Makefile.linux-gcc testfixture
make -f ../Makefile.linux-gcc test > /tmp/test.log
egrep -v 'Ok$' /tmp/test.log
# When run on a locally-mounted disk, my output ends with:
# 0 errors out of 57887 tests

Scott Hess <shess@google.com>, December 11, 2007

--------------------------------------------

As of September 12, 2008, these are our changes from sqlite_vendor:

 - fts2.c disables fts2_tokenizer().
 - sqlite3Poison() in src/btree.c.
 - BEGIN defaults to BEGIN IMMEDIATE in parse.y.
 - Tweak to SQLITE_EXTENSION_INIT* in sqlite3ext.h.
   - That implied a change in src/test_autoext.c for testing.
 - Added fts.test and fts1.test in tests, modified quick.test.
 - src/os_symbian.cc.
 - Modifications to Makefile.linux-gcc and main.mk for compiling
   SQLite tests.
 - Compile warning fixed in func.c (check if this is still needed)
 - Fixed a typo bug in fts2_icu.c: "strlen(nInput)" (filed upstream as
   http://www.sqlite.org/cvstrac/tktview?tn=3543)

Changes from Chrome:
 - I marked all changes I made with "evanm", so you can find them with
   "grep evanm *".
 - Most files include sqlite3ext.h with SQLITE_CORE #defined, but two don't:
   fts2_tokenizer.c and icu.c.  Without this #define, the calls in
   fts2_tokenizer.c try to go through some pointer to the sqlite API instead
   of calling the functions directly (to work as a loadable module), but then
   crash (because the other files never initialize that loadable module
   support).  As a hack I #defined it in these files, but it'd be nice to
   figure out what really ought to happen here (perhaps this file is new and
   hasn't been tested to verify it works right).  Update: Seems this is an
   issue we get because we're using fts2 instead of fts3.
 - shell_icu.c is a Chrome-specific file used to load our ICU data.  shell.c
   has been modifed to call into shell_icu.c.
 - fts2_icu.c has a critical bug. U8_NEXT is used over a UTF-16 string. It's
   rep$ by U16_NEXT (jungshik)
 - Added a new function sqlite3Preload we use to prime the database cache. It
   allows much faster performance by reading the file in one contiguous
   operation rather than bringing it in organically, which involves a lot of
   seeking.