| ActorsChild.cpp |
Constants
**************************************************************************** |
119837 |
| ActorsChild.h |
|
25042 |
| ActorsParent.cpp |
|
814524 |
| ActorsParent.h |
|
2281 |
| crashtests |
|
5 |
| FileInfo.cpp |
static |
4499 |
| FileInfo.h |
|
2056 |
| FileManager.h |
|
2994 |
| FileSnapshot.cpp |
aActorDestroyedNormally |
12324 |
| FileSnapshot.h |
|
5089 |
| IDBCursor.cpp |
|
25971 |
| IDBCursor.h |
|
5371 |
| IDBDatabase.cpp |
|
36625 |
| IDBDatabase.h |
|
7664 |
| IDBEvents.cpp |
|
2928 |
| IDBEvents.h |
|
3452 |
| IDBFactory.cpp |
|
27529 |
| IDBFactory.h |
|
7095 |
| IDBFileHandle.cpp |
aWrapAsDOMRequest |
18494 |
| IDBFileHandle.h |
|
6997 |
| IDBFileRequest.cpp |
|
3311 |
| IDBFileRequest.h |
|
2455 |
| IDBIndex.cpp |
|
17527 |
| IDBIndex.h |
aKeysOnly |
6231 |
| IDBKeyRange.cpp |
|
10002 |
| IDBKeyRange.h |
|
4692 |
| IDBMutableFile.cpp |
|
6386 |
| IDBMutableFile.h |
|
2964 |
| IDBObjectStore.cpp |
|
76711 |
| IDBObjectStore.h |
|
11144 |
| IDBRequest.cpp |
|
15262 |
| IDBRequest.h |
|
6423 |
| IDBResult.h |
|
5594 |
| IDBTransaction.cpp |
|
30671 |
| IDBTransaction.h |
|
8967 |
| IndexedDatabase.h |
|
3283 |
| IndexedDatabaseInlines.h |
|
3630 |
| IndexedDatabaseManager.cpp |
|
29775 |
| IndexedDatabaseManager.h |
|
5148 |
| Key.cpp |
Here's how we encode keys:
Basic strategy is the following
Numbers: 0x10 n n n n n n n n ("n"s are encoded 64bit float)
Dates: 0x20 n n n n n n n n ("n"s are encoded 64bit float)
Strings: 0x30 s s s ... 0 ("s"s are encoded unicode bytes)
Binaries: 0x40 s s s ... 0 ("s"s are encoded unicode bytes)
Arrays: 0x50 i i i ... 0 ("i"s are encoded array items)
When encoding floats, 64bit IEEE 754 are almost sortable, except that
positive sort lower than negative, and negative sort descending. So we use
the following encoding:
value < 0 ?
(-to64bitInt(value)) :
(to64bitInt(value) | 0x8000000000000000)
When encoding strings, we use variable-size encoding per the following table
Chars 0 - 7E are encoded as 0xxxxxxx with 1 added
Chars 7F - (3FFF+7F) are encoded as 10xxxxxx xxxxxxxx with 7F
subtracted
Chars (3FFF+80) - FFFF are encoded as 11xxxxxx xxxxxxxx xx000000
This ensures that the first byte is never encoded as 0, which means that the
string terminator (per basic-strategy table) sorts before any character.
The reason that (3FFF+80) - FFFF is encoded "shifted up" 6 bits is to maximize
the chance that the last character is 0. See below for why.
When encoding binaries, the algorithm is the same to how strings are encoded.
Since each octet in binary is in the range of [0-255], it'll take 1 to 2
encoded unicode bytes.
When encoding Arrays, we use an additional trick. Rather than adding a byte
containing the value 0x50 to indicate type, we instead add 0x50 to the next
byte. This is usually the byte containing the type of the first item in the
array. So simple examples are
["foo"] 0x80 s s s 0 0 // 0x80 is 0x30 + 0x50
[1, 2] 0x60 n n n n n n n n 1 n n n n n n n n 0 // 0x60 is 0x10 + 0x50
Whe do this iteratively if the first item in the array is also an array
[["foo"]] 0xA0 s s s 0 0 0
However, to avoid overflow in the byte, we only do this 3 times. If the first
item in an array is an array, and that array also has an array as first item,
we simply write out the total value accumulated so far and then follow the
"normal" rules.
[[["foo"]]] 0xF0 0x30 s s s 0 0 0 0
There is another edge case that can happen though, which is that the array
doesn't have a first item to which we can add 0x50 to the type. Instead the
next byte would normally be the array terminator (per basic-strategy table)
so we simply add the 0x50 there.
[[]] 0xA0 0 // 0xA0 is 0x50 + 0x50 + 0
[] 0x50 // 0x50 is 0x50 + 0
[[], "foo"] 0xA0 0x30 s s s 0 0 // 0xA0 is 0x50 + 0x50 + 0
Note that the max-3-times rule kicks in before we get a chance to add to the
array terminator
[[[]]] 0xF0 0 0 0 // 0xF0 is 0x50 + 0x50 + 0x50
As a final optimization we do a post-encoding step which drops all 0s at the
end of the encoded buffer.
"foo" // 0x30 s s s
1 // 0x10 bf f0
["a", "b"] // 0x80 s 0 0x30 s
[1, 2] // 0x60 bf f0 0 0 0 0 0 0 0x10 c0
[[]] // 0x80
|
28633 |
| Key.h |
|
12205 |
| KeyPath.cpp |
|
16529 |
| KeyPath.h |
|
2968 |
| moz.build |
|
2850 |
| nsIIDBPermissionsRequest.idl |
Interface for IDB permission requests. This is passed as the
subject for the permission request observer service notifications.
|
856 |
| PBackgroundIDBCursor.ipdl |
|
2083 |
| PBackgroundIDBDatabase.ipdl |
|
2300 |
| PBackgroundIDBDatabaseFile.ipdl |
|
481 |
| PBackgroundIDBDatabaseRequest.ipdl |
|
710 |
| PBackgroundIDBFactory.ipdl |
|
1429 |
| PBackgroundIDBFactoryRequest.ipdl |
|
981 |
| PBackgroundIDBRequest.ipdl |
|
3642 |
| PBackgroundIDBSharedTypes.ipdlh |
|
5918 |
| PBackgroundIDBTransaction.ipdl |
|
976 |
| PBackgroundIDBVersionChangeTransaction.ipdl |
|
1711 |
| PBackgroundIndexedDBUtils.ipdl |
|
947 |
| PermissionRequestBase.cpp |
aExpireTime |
6558 |
| PermissionRequestBase.h |
|
2225 |
| ProfilerHelpers.h |
|
14552 |
| ReportInternalError.cpp |
no IDB in private window |
1069 |
| ReportInternalError.h |
Don't evaluate |res| more than once |
2310 |
| ScriptErrorHelper.cpp |
aSourceLine |
6626 |
| ScriptErrorHelper.h |
nsIScriptError::xxxFlag |
1295 |
| SerializationHelpers.h |
|
2930 |
| test |
|
186 |