GitLab

* changes:
  Suppressing BufferedReader jtreg tests we're better off not passing.

Bug: 2187020

* changes:
  Fix HttpURLConnection's chunked encoding behavior.

A third-party developer who wasn't reading to the end of the stream found that
their next connection would return junk, which turned out to be the tail of
the first response (thanks to connection recycling). Make sure we clean up
a chunked-encoding stream before allowing the connection to be reused.

Enhance our test web server to implement chunked encoding properly, with
configurable chunk sizes, rather than just responding with a single chunk.

Bug: http://code.google.com/p/android/issues/detail?id=2939

* changes:
  Make cacerts.bks available to the simulator too.

This lets you successfully run more tests on a sim-eng build than otherwise,
by bringing us in line with the normal on-device configuration.

Do this work at JNI registration time, as we do for almost everything else.
(I did this to rule out a warning from the dalvikvm deadlock prediction
code, which doesn't like the unusual lock ordering at initialization time,
and although it didn't make any difference to that, I prefer to have a
defined static order of initialization.)

* changes:
  Fix a buffer overrun in OSNetworkSystem.

Given an IPv6 address in a non-standard (Java-specific) form, we used to
overrun an on-stack buffer that was assuming the standard (shorter) form.
Make the buffer large enough for both forms, and reject anything that's
still too large.

Found by Inet6AddressTest.

* changes:
  Jit stress mode: translate everything we can and self verify.

This represents a general clean-up of some existing command-line
options: -Xjitthreshold:num and -Xjitblocking.  The jit threshold
controls how quickly we treat a Dalvik address as a potential trace
head.  Normally this is set around 200 (and the range is 0..255, where
0 is in effect 256 and 1 means begin trace selection on first visit).

-Xjitblocking forces the system to pause execution whenever a translation
request is made and resume when that translation is complete.  Normally
the system make a request but continues execution (to avoid jitter).

Additionally, if WITH_SELF_VERIFICATION is defined, we force blocking
to be true, and set the threshold to 1.  And finally, we treat
threshold==1 as a special case and disable the 2nd-level trace-building
filter - which causes the system to immediately start trace selection.

Our DOM parser didn't support { or š character references,
and didn't merge adjacent text nodes into one (so "a&b" would be
three text nodes rather than one; SAX allows the former, but DOM
guarantees the latter).

This patch fixes both bugs, and adds tests.

Bug: 2607 (and duplicates)

Much of this is spurious whitespace changes, but there's some increased
"relaxation". I deliberately lost the Android-specific change that was
avoiding Runtime, since we do now have Runtime. I've added an Android-specific
change to comment out some System.out logging that's been added upstream.

I'd tell you the upstream revision number, but they're still using CVS, so
there isn't one.

am 216438fc: (-s ours) am 83859891: am 3dc6a811: (-s ours) DO NOT MERGE. CTS test cleanup. Mark tests as KnownFailures and BrokenTests.

Merge commit '216438fc405d78069b1042cc21fc43dc85f860a9'

* commit '216438fc405d78069b1042cc21fc43dc85f860a9':
  DO NOT MERGE. CTS test cleanup. Mark tests as KnownFailures and BrokenTests.

* changes:
  Introduce "just interpret" chainable pseudo-translation.

This is the first step towards enabling translation & self-cosim stress modes.
When trace selection begins, the trace head address is pinned and
remains in a limbo state until the translation is complete.  Previously,
if the trace selected aborted for any reason, the trace head would remain
forever in limbo.  This was not a correctness problem, but caused some
small performance anomolies and made life more difficult for self-cosimulation
mode.

This CL introduces a pseudo-translation that simply routes control to
the interpreter.  When we detect that a trace selection attempt has
failed, the trace head is associated with this fully-chainable
pseudo-translation.  This also has the benefit for self-cosimulation that
we are guaranteed forward progress.

am 83859891: am 3dc6a811: (-s ours) DO NOT MERGE. CTS test cleanup. Mark tests as KnownFailures and BrokenTests.

Merge commit '838598912874c49de45aba3baac644d177920115' into eclair-mr2-plus-aosp

* commit '838598912874c49de45aba3baac644d177920115':
  DO NOT MERGE. CTS test cleanup. Mark tests as KnownFailures and BrokenTests.

Merge commit '3dc6a81182178a8ac9570a7679cf24b6a2665a0b' into eclair-mr2

* commit '3dc6a81182178a8ac9570a7679cf24b6a2665a0b':
  DO NOT MERGE. CTS test cleanup. Mark tests as KnownFailures and BrokenTests.

OpenSSL (in addition to -lssl, which was already listed).

Change-Id: Iddc53ee474dae030755395266a35f472f9b568b5

* changes:
  Our XML serializer permits \0, resulting in malformed documents.

* changes:
  Sanitize text before emitting to XML.

This is the absolute least amount of sanitization that will work.
I'll consider adding more as necessary; currently I'm not too
interested in spending the time to make this perfect.

This failing test demonstrates the problem. It also adds AllTests
plumbing for this test and some missing ones.

* changes:
  Clean up sieb.[ch], including changing a reference to <malloc.h> to instead be the more standard <stdlib.h>.

* changes:
  Remove 'GTE CyberTrust Root', which expired on 2/23/06.

be the more standard <stdlib.h>.

I have no idea what "sieb" means (stands for?), and it looks like these
files may want to be placed in the nearest recycling receptacle, but I'm
not ready to go there today.

Change-Id: I2d4116b742e61018fdbea127070154545ff82dbd

Handle select(2) errors by throwing exceptions from native code (which is
simpler and provides more useful detail messages).

Factor out exception throwing commonality. This is probably ready for JNIHelp.h
now.

Remove sockSelect. Simplify selectWait by observing that it's only ever called
with a "read" fdset.

Remove a little more SOCKERR_* cruft.

Remove a few unused #defines.

* changes:
  More OSNetworkSystem cleanup.

* changes:
  Fix DEADLOCK_DETECTION mode for the Jit.

* changes:
  Remove cruft in the default libdvm.so and add 3 more special targets.

Fix danfuzz's BSD woes by removing uses of non-standard s6_addr32,
replace isMappedAddress with IN6_IS_ADDR_V4MAPPED, and change
convertMappedToIpv4 to reduce duplication in its callers.

Remove the TODO suggesting we consider inet_ntop instead of
getnameinfo; Ulrich Drepper says there's never any reason to
do that: http://people.redhat.com/drepper/userapi-ipv6.html

Remove our last use of sockaddr_in in favor of sockaddr_storage,
so we never have to wonder "is that big enough?" again.

Explain why we convert IPv4-mapped addresses to IPv4 addresses
when there's no obvious reason to do so.

Fix another user of the SOCKERR_* cruft.

Address the FD_ISSET and what-exception-to-throw-on-BSD comments
from danfuzz's last patch.

* changes:
  Remove deadlock-prone thread pool.

This is only test code, so let's not make it any more complicated than it
needs to be.

(The code sometimes locked the Vector before the individual Worker, and
sometimes the other way round.)

The calls and structs provided by the underlying platform for
multicast aren't yet exactly well-established POSIX standards, and the
current Linux code is still not considered to be super-robust, and as
such it isn't a good idea to try to duplicate its functionality for
other platforms. We will revisit the issue once the Linux side has
matured a bit.

Change-Id: I630086e22f6980726fd4523cb60dfc8c9fb9b011

* changes:
  Fix for inline string indexof; added regression tests