GitLab

Add hashDestroy function pointer as a destroy handler
in fts initialize code that had been removed by previous commit

Change-Id: I0195e64da1cb470b5146f05bfe880ca62fb6d114

Change-Id: I4c01f3f32b0dbe450206f3e8cbf3d4eedfdb6562

Change-Id: I514750dd33ff5bd433d4ee167a2401d829d2fbe9

* changes:
  Fix bugs of database corruption following IO error on systems supporting atomic-write
  Fix bugs of sqlite that returns SIGBUS on disk full in WAL mode

Merge "Change android fix for fts backward compatibility patch of sqlite3.c for newer sqlite version"

SQLite on Android uses AUTO_VACUUM option.
When the disk space is critically low and atomic-write option is enabled,
SQLite writes transactions that modify a single database page on disk
without creating a journal file.

Due to a bug, if an IO or disk full error occurs while transferring
the contents to disk, the single page that was modified in the cache
is not being rolled back - cache corruption.
As a result, applications get a database corruption exception
for next database transactions.

This patch changes createFile() to close pReal in case of IO error.
Therefore, cache is being rolled back.

This patch is originated from www.sqlite.org.
(URL: http://www.sqlite.org/src/tktview?name=df678d738a)

Change-Id: Ieecda65d8458cb591bd4d89d8b423a4479f50ea8

Attempts to prepare a query on a WAL database when the disk space is
critically low result in the process killed with SIGBUS.
The crash happens in walIndexWriteHdr invoked from walIndexRecover.

Some Providers that using WAL mode like as SettingsProvider and MediaProvider
get failed in case of disk full with SIGBUS.
This patch changes unixShmMap() to call fallocate() instead of ftruncate.
In case of disk full, fallocate() will return SQLITE_FULL.

This patch is originated from www.sqlite.org.
(URL: http://www.sqlite.org/src/info/5eaa61ea18)
To simplify error status, returned error code is changed
from SQLITE_IOERR_SHMSIZE to SQLITE_FULL.

[written by Yongil Jang <yi.jang@lge.com>]

Change-Id: Idf3cbdab1ed4a4f2bc1b2b93ab6d8900edc7ee05

From earlier version of Android, patch of FTS3 backwards compatibility is kept until now.
But, this patch doesn't work well on newer version of sqlite3 (3.7.13~) when it was used.
Currently, sqlite3Fts3Init function is called three times to register "fts1", "fts2" and "fts3" with the same module.
It worked very well on earlier version of sqlite but not for newer version.
hashDestroy() function makes an error when it is called after the first destroy of fts3Module.
sqlite3Fts3Init() function initializes "fts4" module and it is called three times to initialize "fts1" and "fts2".

This patch contains changed source code of sqlite3Fts3Init() function to be called once.

[written by Yongil Jang <yi.jang@lge.com>]

Change-Id: I55f522b55f421f73e83508cfc5e08514bd7da925

Bug: 6538393
Change-Id: Iaa20c8bc01c53da08ff18cb1efae80c2be0b0584

Change-Id: Id09ca60b4bcadc232819945f11c98ce7819e1ada

* commit '2d758ba4':
  Enable the powersafe overwrite assumption.

This change improves performance on certain kinds of transactions
such as UPDATEs by allowing SQLite to make stronger assumptions
about data integrity on the disk.

The powersafe overwrite assumption demands that when a byte is
written to a file and a power event occurs then the only bytes
that may have been modified are the ones that were being written.

In absense of the powersafe overwrite assumption, SQLite assumes
that a write of any range of bytes within a disk sector may
potentially cause the rest of the sector to be corrupted.  SQLite
needs to do extra work to compensate for this problem, which adds
overhead to certain transactions.

On flash devices, writing a byte causes the sector to be rewritten
somewhere else on the device.  Eventually the old sector may be
erased and reused.  The powersafe overwrite assumption therefore
holds as long as the flash device never leaves behind (observable)
half-written sectors in case of a power event.  This is a fairly
safe assumption to make.  Some flash devices apparently include
measures to ensure that the last write fully completes when
power is interrupted.  In any case, the window of opportunity
for corruption to occur is very small.

Bug: 6484633
Change-Id: I0b0ef664b3a2a0e85c6099f0130d0affc5562be5

am 38d33f70: am 1eabc582: Merge "GetPhonebookIndex: Fix handling for minor Japanese kana characters"

* commit '38d33f70':
  GetPhonebookIndex: Fix handling for minor Japanese kana characters

* commit '1eabc582':
  GetPhonebookIndex: Fix handling for minor Japanese kana characters

Bug: 6176510
Change-Id: I482e980b3377a9606fc23e72fd25d0c648f718d4

No code changes.  This is just preparing for an upgrade.

Bug: 6176510
Change-Id: I5be7cf74e8d72d1bdb16b09b6bee8330c67ed7fc

* commit '221a8c1c':

am 82446a48: (-s ours) am efe32461: am c69e7a40: Changed default mode for journal files created by sqlite.

* commit '82446a48':
  Changed default mode for journal files created by sqlite.

* commit 'b8b540ae':

* commit 'efe32461':
  Changed default mode for journal files created by sqlite.

* commit 'c69e7a40':
  Changed default mode for journal files created by sqlite.

* commit 'c69e7a40':
  Changed default mode for journal files created by sqlite.

Previous to this, the temporary journal files created by sqlite
were globally readable, leading to the possible leakage of
sensitive data (CVE-2011-3901).

Bug: 6222215
Change-Id: I29a85c9077c622b038e7fb190c3d7833b924df99

Previous to this, the temporary journal files created by sqlite
were globally readable, leading to the possible leakage of
sensitive data (CVE-2011-3901).

Bug: 6222215
Change-Id: I29a85c9077c622b038e7fb190c3d7833b924df99

Linux has a working fdatasync() so we should use it!
This change alone improves performance by around 20% by reducing
the number of synchronous I/O writes of irrelevant metadata
changes.

Change-Id: Iebe0c32a7e018ec0ced2347b7d62b2b7dc5aa376

* commit '4b9bb0e3':

* commit '4a6bed40':

Change-Id: I866cdfbee1fcdf144c850566d682c69e39fa04d7

It seems some people really are depending on the $1 syntax
to bind variables.  Ok.

Bug: 5926203
Change-Id: Iae0cc6a0882b536fb5f3cdad687a6d7b41843d9f

Turn off stuff we don't need.
Enable FTS4.
Enable ICU tokenizer.

Bug: 5530406
Bug: 5465099
Change-Id: Idcc52050a4d048a64bfe8484e419dc8d2b27d6ed

This error code was introduced at some point to help track when
a database could not be closed because a statement was not finalized.
Now that the DB wrappers have been rewritten, it is technically no longer
possible for this to happen, so we can remove this.

Change-Id: Ia3c8167a0af522c852adbbd15c83c3c8f5732b26

Change-Id: I67183b854c34e9436ff604d0fd73b4ec8ee414f2

Ported all original Android-specific patches.

Change-Id: I2c5801a92d783e2c5cba72147d9fb777ad7c8443

See https://android-git.corp.google.com/g/#/c/157220

Bug: 5449033
Change-Id: I5ffb58901712f4db93de474c30833a46ebb4bc7c

See https://android-git.corp.google.com/g/156016

Bug: 5449033
Change-Id: Iad9bddb01ef0d9b28e3969b9f8dfbb2cc2830b30

This adds minor Japanese kana character handling to meet the
unicode collation algorithm.

-Normalize digraphs (yori, koto) into first reading letter.
-Treat dakuten, dot, onbiki, iteration marks as a symbol.
-Add handling for minor small katakana letter.
Signed-off-by: Yutaro Ogasawara <yutaro.ogasawara@gmail.com>