| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
We now append the singly linked list at the end of the queue.
Previously the list was added after the first element of the
unenqueued list.
Bug: 17381967
Change-Id: I6d4eb1053e5a2babd559e410b42344a04490e7b5
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
We left this out when manually auditing the list of classes
that don't throw and silently use the default (like Calendar)
vs those that do. See bugs 16938922 and 15849709.
bug: 17394277
Change-Id: I6e1dae3c9251dfd48221914351a504f24f4bdf5d
|
|/
|
|
|
|
|
|
|
|
| |
The calendars will use System.currentTimeMillis if we don't
explicitly call setTimeMillis or setTime. We don't particularly
care about time here, we just need to check that all Locale
derived fields are equal.
bug: 17377252
Change-Id: Ibd5e74c237e5520580f1f44b4e98b021000639fa
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The problem was as follows:
The GC finishes and enqueues all cleared references on the
ReferenceQueue.unenqueued list. Then someone calls
finalizeAllEnqueued which enqueues a finalizer reference directly on
the finalizer reference queue. However the recently cleared references
had not yet been processed by the ReferenceQueueDaemon resulting
in the sentinal being finalized before the objects which were freed
in the GC.
The fix is enqueuing the sentinel finalizer reference on the unenqueued
queue.
(cherry picked from commit 2cc3d41ff2dfd673fa1ab4ef2a46525933d523ec)
Change-Id: I0598199ef56aa8bd9e1fa40c39d50b2785dc5c10
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Changed SocketHandler to re-introduce a "socket != 0"
check unintentionally removed by commit
232c91dce1760b180155256d1896d0d80375ff6d.
Some failures are the result of the test code being newer
than the libcore code. Android took its last libcore update
from apache-harmony some time after r820767 but before r929953
and its test-code update at r1097236. This explains a few
things.
Updated FileHandler to deal with null system properties.
This is an equivalent change to
http://svn.apache.org/viewvc?view=revision&revision=935099
Updated XMLFormatter to handle basic XML escaping. This is
an equivalent change to
http://svn.apache.org/viewvc?view=revision&revision=929953
Unsuppress tests that are fixed by this.
Bug: 13882147
Change-Id: I36321b488da6d4e08fbb0d28d0e366f9b06e6371
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Based on looking at IANA and wikipedia.
Bug: 16978217
(cherry picked from commit 93235f904ce83fe138c90a90c66d0e78f9f405b8)
Change-Id: I272436facd3c2f5dbc8c455fc340293b83f98453
|
|\ \ |
|
| |/
| |
| |
| |
| |
| |
| |
| | |
Replace {@link} references to hidden methods with {@code}.
(cherry picked from commit f73b1089c7236a7268ccce10546cc340a36f8c4b)
Change-Id: Iff245542979c2bbcea8b6c4b4632547c6aac4da4
|
|/
|
|
|
|
|
| |
(cherry-pick of f6cf9efb212e572dcd2e902ca461af6323793dbf.)
Bug: 17203955
Change-Id: I0999fc0ff295986b92e31568d96e321b9e7ffb2c
|
|
|
|
|
|
|
|
|
|
|
| |
Framework".
Using a non-existent TLD is now a Bad Idea.
(cherry-pick of a7d651e95e38974fa0de19247e8341df98a0b1b1.)
Bug: 17146413
Change-Id: I6a68b4609d59a6d56ba780d03ffcf24a2759db48
|
|\ |
|
| |
| |
| |
| |
| |
| | |
Fix a missing case from one constructor by delegating to another.
Change-Id: I8b85c97586dde2d72c9ac21216c594aa810d761b
|
|/
|
|
|
|
|
| |
(cherry-pick of ebe417d7808ac5bce5bc14d36d6af4c31218326d.)
Bug: 16938600
Change-Id: I7334c1ea57ea0593b3e0bd7943f8a5208c704881
|
|
|
|
|
|
|
|
|
|
| |
Runtime.freeMemory is the largest allocation that's
possible *without a heap expansion*. That's a poor proxy
for "can i allocate this byte array" and leads to flaky
CTS tests.
bug: 17021619
Change-Id: I108138bf2402e899a40a0b3081c083f733e7c999
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Because the tests cases are executed in the same runtime and the class
linker may cache generated optimized files, init tests need to use their
own optimized directories (so that they can reliable test the number of
generated files).
Bug: 16806146
(cherry picked from commit 58c130c1eb99cc72998150d813b3e31896e3bfaf)
Change-Id: I8111cc7ea640f3bc94821fcd4380acfc078c746f
|
|
|
|
|
| |
Bug: 17049821
Change-Id: I309ecfad7e38cf664518a92d6812ffbd90060b79
|
|\ |
|
| |
| |
| |
| |
| | |
Bug: 16162465
Change-Id: Iea4d39e1bb74be7ad9041aa3dbb5809d53f63faf
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Originally, we called getsockoptLinger unconditionally, using ENOTSOCK
to recognize the non-socket case, but the cost of throwing an exception
is so high (both on dalvik and art) that we worked around that by doing
an fstat first. This was mostly fine, but fstat calls onReadFromDisk,
which isn't true for the socket case (which causes upset for threads
that are allowed to touch the network but not allowed to touch disk,
which Chrome uses, even though it's not immediately obvious to me why
that makes any sense).
(cherry-pick of d636777d624428dfc51d1181f47602f389eb4cce.)
Bug: 4972558
Bug: 12133112
Change-Id: Ic0b89d1f181e436b4f43c1c04c1f706883ab84c0
|
|/
|
|
|
|
|
|
|
|
|
|
|
| |
We were just parsing them as millisecond values, which is wrong.
".1" fractional second is 100 milliseconds (we were parsing it as 1 ms),
and ".765432" fractional seconds is 765 milliseconds (we were
parsing it as 765432ms.
In most cases, this was hidden by the fact that we were formatting
such values incorrectly as well.
bug: 16969112
Change-Id: Id287684f78691090ac4a6c3029d7909f2c1e1310
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
LocaleData.get was changed to throw for null locales instead
of falling back to the default, so we need to adjust callers
accordingly.
Turns out all other callers would've thrown anyway (for different
reasons) so this change makes those exceptions explicit and
adds tests for them.
bug: 16938922
Change-Id: I3e77435a6b414b002b96a4f8deb44b51591c5a51
|
|
|
|
| |
Change-Id: I4dba2cfe07cf0c5ce28a0ff4540aa68d9c664a2a
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The functions ceil, floor, hypot, IEEEremainder, nextafter, and rint
have all been implemented in Java to save the native call. Benchmarking
numbers to show this performance improvement are included:
Native call (before change):
benchmark=Ceil: 97.40 ns; σ=0.02 ns @ 3 trials
benchmark=Floor: 98.73 ns; σ=0.02 ns @ 3 trials
benchmark=Hypot: 714.62 ns; σ=0.44 ns @ 3 trials
benchmark=IEEEremainder: 180.31 ns; σ=0.08 ns @ 3 trials
benchmark=NextAfterD: 110.47 ns; σ=0.02 ns @ 3 trials
benchmark=NextAfterF: 96.20 ns; σ=0.01 ns @ 3 trials
benchmark=Rint: 92.21 ns; σ=0.02 ns @ 3 trials
Java implementation (after change):
benchmark=Ceil: 47.32 ns; σ=0.01 ns @ 3 trials
benchmark=Floor: 42.03 ns; σ=0.00 ns @ 3 trials
benchmark=Hypot: 69.42 ns; σ=0.37 ns @ 3 trials
benchmark=IEEEremainder: 49.64 ns; σ=0.04 ns @ 3 trials
benchmark=NextAfterD: 63.32 ns; σ=0.17 ns @ 3 trials
benchmark=NextAfterF: 64.67 ns; σ=0.12 ns @ 3 trials
benchmark=Rint: 31.10 ns; σ=0.41 ns @ 10 trials
Change-Id: I9ee926514c0661cdb7a1e6ae3acfe5c31a4b9f20
|
|/
|
|
|
|
|
|
|
|
|
|
|
| |
This was broken by the removal of the pre-computed raw offsets from
the tzdata file. I think that's still the direction we want to go (with
us hopefully using more of icu4j at some point, and eventually relying
solely on the icu time zone data), so this patch adds code to lazily
evaluate all the offsets by instantiating all the time zones.
(cherry-pick of 065d7764ac1dfe74ee94d17ca6c810de37b57d3e.)
Bug: 16947622
Change-Id: I6d1dfe5ee6c99338f9807c3af5b6f04539c256c3
|
|
|
|
|
|
| |
zip archive makes a copy of a prefix and maintains it's lifecycle now.
Change-Id: I55adc4c9eefee96b80038fb7dbb95dbe498d5a73
|
|
|
|
|
|
|
| |
Call EndIteration in IterationHandle destructor. This will free the
memory allocated for cookie in StartIteration.
Change-Id: I12c861c7da2437113c669a0c7f41b5585ed94c72
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Changes cosh, exp, expm1, log, log10, sinh, cbrt,
asin, acos, atan, tanh, log1p
These functions are now implemnted in java to
save the overhead of a jni call
Change-Id: I4d3b70dd1693378af5b04cbbe0fecd95d97c1cff
|
| |
| |
| |
| |
| |
| |
| | |
Yet we were incorrectly upper-casing them inside forLanguageTag.
bug: 16811867
Change-Id: I088161c4c7769908f35b9f2b7d945fb1cd1fd7ac
|
| |
| |
| |
| |
| |
| |
| | |
It seems that this field is never used outside the internal constructor so it
can be a local variable instead.
Change-Id: I2aa26fb3fd13c571224a9aee14c7eea93e5c9591
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | | |
This reverts commit c4da34ca5792e0c73577f868249fdfe8a11a01f6.
Change-Id: I1b93ecfcb8a867a670af717eb25c7076b7097d5c
|
|\ \ \
| |/ /
|/| | |
|
| | |
| | |
| | |
| | |
| | |
| | | |
The SubjectKeyIdentifierStructure class is no longer in Bouncycastle.
Change-Id: Ie5eb91a14ec38f773ef22754d77651a161fd7122
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
It seems that this field is never used outside the internal constructor so it
can be a local variable instead.
Change-Id: Icd9e009a24ae22da8f5dfd18b9f0ab80913129d4
|
|/ /
| |
| |
| |
| |
| |
| |
| | |
and GenericArrayTypeImpl.
Bug: 14590652
Bug: https://code.google.com/p/android/issues/detail?id=74060
Change-Id: Ib9a30b741a8d5f82d21f37ca8959bf5b2893f8db
|
| |
| |
| |
| |
| |
| |
| | |
(cherry-pick of 77926e913a14b9541df38b6175f795f2ca6c5266.)
Bug: 15715370
Change-Id: I2d99f070c3debea01621a172d8189951a56736b5
|
| |
| |
| |
| |
| |
| |
| |
| | |
data[OUTPUT_OFFSET] from the perspective of native
code must always be the offset from the start of the
buffer passed in to it.
Change-Id: I4a5c9595aebd105c57ce2239dc95cf85d5bd6256
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
- After the new Java forLanguageTag implementation, bogus
locales really do result in "und".
- Fix a regression in unicode extension parsing, where we
would always add a "null" to the keyword set.
- Improve test assertions so that the reason for failure
is obvious.
bug: 15849709
Change-Id: I26fb348b872badb7844587cf290429e15dbb0f0e
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
We don't need a replacement for ICU.getScript(String),
Locale.getScript() can be used instead. The old definition
of the function has been left public for the support library
to use (though we could remove it if we modify the support lib).
Note that addLikelySubtag returns a String that's language tag
like, but uses underscores instead of hyphens to separate
subtags.
bug: 15876704
Change-Id: I6c7d289c4b725a83900e38dd5bbed026afd9ac4e
|
| |
| |
| |
| |
| |
| |
| |
| | |
locale and zoneStrings are not accessed outside of
DateFormatSymbols.
Bug: 16502916
Change-Id: I67907cf6b0ebc2e6a2dd0bfe8728ba6d0487d5a3
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Scanner had no mechanism for recovering buffer space it didn't
need.
Now, if the buffer is more than 50% full of ignorable characters
the remaining characters are shuffled to the beginning to recover
space. For most expected usecases this means that the buffer will
stay 1k and contain up to 512 characters of useful data. A
growable, circular character buffer could have been introduced
to avoid the copy but is not worth the effort.
Previously the buffer would double in size until the data or memory
was exhausted, and each read would also double in size to fill the
empty half of the buffer. This was fine providing the input
could fit in memory.
On top of that the 1k, 2k, 4k, etc. buffer was repeatedly turned
into a String and passed to the (native) matcher, and then the
matcher was told to ignore more than half of it.
As a consequence of keeping the buffer a fixed size (and only
filling 50% of it at a time), this change may cause a performance
regression: for most usecases where delimiters are closer together
than 512 bytes, reads after the first will now usually be 512 bytes
and not the 1k, 2k, 4k, etc. it was previously.
Having fixed the test so it doesn't OOM, the test now
takes 6 minutes to pass on host and so is unsuitable for inclusion
in CTS tests and so is being suppressed.
This change also includes so tidy up changes to the test and the
Scanner class.
The implementation could no doubt be improved but the API makes
it inherently slow. It would be surprising if anybody uses the
Scanner class on Android with so many better alternatives.
Bug: 14865710
Change-Id: I40a7fc0c2bfaf6db4e42108efe417303e76bde24
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | | |
Bug: 7955614
Change-Id: Iecdd939f25579f882c11d2957c2a5667d2f7fd3a
|
|\ \ \ \
| |/ / /
|/| | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Adding locale to the serialized form of
DateFormatSymbols.
Some duplicated code around the generation of
numeric / offset timezone strings (e.g. GMT+08:00) has
been removed. A new hidden method has been added
to TimeZone to create this string.
The timezone string lookup has been moved into
LocaleData and it now has a single path rather
than treating the customZoneStrings path as special.
customZoneStrings has been removed from
DateFormatSymbols because it would potentially
have an incorrect value after serialization and
it is no longer required.
Bug: 16502916
Change-Id: I2b317e34ba4772beb372a75dd08c95113408b9cc
|