- Moved the thread check from AbstractCursor to SQLiteCursor.
- Show warning only once per database
Bug 2633883
Change-Id: I915bfba8715a469c45fb68ba1282231279f2a50c
revert part of the CL: Ia561135e974a44ad3e3774ecb23c6a3d0fc38176
and add it to ContentProvider.query() to force query execution
in worker thread, instead of having the main thread incur the cost
of query execution and potentially get ANR
Change-Id: I0ea8a170bd954a421f4ad825f8090319a83a5b2b
1. when an entry is pushed out of cache, it should be released if it is not
in use by any thread. This didn't have to be done when cache was NOT LRU
because the object either got into the cache while the caller had a reference
to it or it didn't. if it didn't get into cache (because cache is full),
the caller's close() released the object. But in LRU cache, an object
could get pushed out of cache due to LRU policy and if it is NOT in use
by any thread at the time it was pushed out of cache, then it
got GC'ed - which caused warnings to be pronted in the log.
2. also delete some unused methods in SQLiteDatabase.
Change-Id: I7831952647d3a057342bcc8ac186a6a26eb58f33
bug:1648729
print warning if a requery is attemped on main thread
bug:2459293
browser death because of NoSuchElementException. don't use iterator
to march thru the ArrayList. use old style for() loop.
bug:1609474
don't allow downgrades of databases
other stuff
use LRUcache to maintain statement-cache in SQLiteDatabase
Change-Id: I3a42022162f4dfcc75f7d0cfcad8f20f1d193249
let the user specify an interface impl class to specify the actions
to take when db corruption is detected.
this class is specified when the database is opened/created.
Change-Id: I84eb57208c8fedfa7235805b0ec58165efdc1560
1. we should let apps set their statement-cache size. right n ow it is 250
is the statement-cache size for all apps and that is wasting
a lot of memory. each prepared statement is averaging about 1k-5K,
depending upon the complexity of sql and schema.
mnake default 25 and let apps increase the size, if they need to.
2. in "adb bugreport" info, print stats on the statement-cache hits/missies
and size (in number of statement cached). this will help us understand
how statement-cache is being used
Change-Id: Ic53a2842ef75823757492778158fcb8d5b4a28bb
The issue is that our code often "executes" a query
on a background thread but iterates over the cursor
on the UI thread. Since we actually do the fetch
on moveToFirst or moveToNext, the query is in reality
often run on the UI thread and causes an ANR.
Change-Id: Ia561135e974a44ad3e3774ecb23c6a3d0fc38176
This also makes the 500ms logging threshold (over which is always
logged, and under which is sub-sampled) configurable via a
SystemProperty, which is mostly useful for interactive debugging when
lock contention is suspected, but could also be useful in the future
as a quick way to adjust this threshold for dogfooders, without code
changes.
Change-Id: I769069d8d870331d89a4aa3239ba50db806fe4d4
and database.close() should NOT set mPath to null.
a few other minor changes included in this CL
1. if it is memory database, no need to delete any file
2. if 2 threads are sharing the same connection, and if corruption occurs
on it, one thread closes the db and deletes db - while the other
thread is still using it. this can cause SQLITE_MISUSE error.
to prevent this, every method in SQLiteDatabse should make sure
db is open before exec'ing sql statements. bug:2531172
Change-Id: I4cb5ab8539f46d7f8b26c3f830d799adf46444b6
this will let us know which cursor is being finalized.
if the heapworker gets wedbed due to the close bug:2536922
then we will never see the warning and never know which cursor
caused this problem
Change-Id: I595cc1014ff7a1114079b33665c092c6a2f3a907
if insert statement doesn't succeed, last inserted rowid
shoudl return -1 - instead of returning rowid of the last
successful insert that may have occurred years before this most
recent insert statement failure.
Change-Id: Ia517292afd58fdb600da900e0ee01fe051d6e618
When query() uses bulkQuery() and we know we're going to need some
metadata right afterwards (number of rows and column index of _id, if
present), just asked for it in the initial binder transaction instead
of immediately fetching it again.
Also, this defers loading column names until the client asks for them.
This gets down the simpler (and very common) use cases of
ContentProvider.query() down to 3 binder calls:
QUERY_TRANSACTION to android.content.ContentProvider$Transport
GET_CURSOR_WINDOW_TRANSACTION to android.database.CursorToBulkCursorAdaptor
CLOSE_TRANSACTION to android.database.CursorToBulkCursorAdaptor
More can still be done, but this is a good bite-sized first piece.
Change-Id: I7ad45949f53e0097ff18c2478d659f0f36929693
a bug in maintaining the cache caused these warnings. when the cache
is full, caching code in SQLiteDatabase dropped an entry from the cache
to accommodate the new one. and if the just-dropped one is not in use
that object got GC'ed and caused a finalizer warning. Calendar is one app
that didn't use ? for bindargs (sometimes) and noticed this bug in that app
Fix is to not add the new enry to cache if the cache is already full.
that will cause the app's close() to release the entry.
another common case where this finalizer warning occurs is when unittests run.
if the test does not close the database in tearDown(), it will cause
database object and the compiled sql statement cache within the database
obj get GC'ed which cause finalizer warnings.
finalizer shoudl not be called ever. add a warning to say that.
adeprecate a few members in SQLiteProgram.java. they should not
have had protected access level. shoudl be package.
