- Introduces unix_read_interruptible() which is like unix_read() except
that it can return EINTR.
- The big idea is that the Windows ReadConsoleInput() API will return an
event on window resize and then we return EINTR from
unix_read_interruptible() just like Unix.
- Only handles horizontal resize since Windows doesn't seem to give an
event for vertical resize when no special screen buffer is used. This
should be sufficient for the primary use case of adb on Windows
(people are not running vi in the first place).
Change-Id: Id8d1710b559834c8098f2d7fbecedf2d0ade4b88
Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
There is at least one app that assumes that you will always get at least
one line of output in stdout from ps.
To fix this, move error output to stdout, and move the check of whether
/proc can be opened until after the ps header is printed.
Bug: 26554285
Change-Id: I6d9342aafd5c6f728735507cdd87a48a8e0373ac
Weaved replaces brillo::Any with base::Value in its client library,
so need to update clients.
Also this removes last D-Bus dependencies from metrics_collector.
BUG: 26558300
Change-Id: I82d571d7e2bd17347daeb1b70f48ec05378894f0
Call addService using a reference to a strong pointer to the service object,
such that the proper IBinder object is ref counted, instead of passing the
object's this pointer.
Bug: None
Change-Id: I72cac08c835a17b8cc263b2dc34864420b7a2841
http://r.android.com/166419 changed `adb shell` behavior to not
allocate a remote PTY for non-interactive commands, but adbd relied on
having a PTY to properly terminate the subprocess.
One impact of this is that when using older versions of adb or passing
the -x flag, `adb screenrecord` wasn't properly terminating and closing
out the video file.
This CL restores the old behavior for legacy shell connections: always
use a PTY, but put it in raw mode if the client is doing local PTY
input/output processing itself.
Bug: http://b/26742824
Change-Id: I9ee630c0ff0d2d6a0db367387af7123deea79676
boot events in the LOCAL_INIT_RC file, bootstat.rc.
This change also creates the /data/misc/bootstat dir during the
filesystem creation block in init.rc.
Bug: 21724738
Change-Id: I2ad6913c0235d5f926c6ef8b3c65d650159c6ed7
Without this setting, arguments passed to a function can be packed in
more than one line, with more than one argument per line.
For example, the following code will be formated as this without the
BinPackArguments:
CallMyFunction(
with_more_than_one, argument_per_line,
event_they_dont_fit_in_a_single_line);
On the other hand, with BinPackArguments: false, that code will be
split in:
CallMyFunction(
with_more_than_one,
argument_per_line,
event_they_dont_fit_in_a_single_line);
or all the arguments in the same line if they all fit.
Bug: None
TEST=None
Change-Id: I48627e66e2a841f89418c0763e7fe6c3e6ea7387
BacktraceOffline is the only code that uses libLLVM. Create
a new shared library, backtrace_offline.so, that references this
so that including libbacktrace.so doesn't wind up pulling in libLLVM.
Change-Id: I90c502d833f6c464c3810dffde52a0c6da241cce
Best practice for services that reside on the system, vendor, and odm
partitions is to have a corresponding init .rc file in the
/{system,vendor,odm}/etc/init directory, which contains its service
entry and any related actions.
Change-Id: I26204f9a0f09ce8069f3791ccd9ee61d164a048f
Mix the contents of /proc/cmdline and /default.prop
into /dev/urandom. /proc/cmdline often contains
androidboot.serialno, a device-specific unique
identifier. Similarly, /default.prop contains the
build fingerprint and timestamp, which vary between
device families.
Change-Id: I8803b38c7089b2a1217b99a7c1808b29a3b138cf
This CL adds two more checks:
* When populating a closed log with the metadata, make sure the
SystemProfileSetter is not null.
* When using the update engine client library, check if the update
engine client object was created correctly. The returning object can
be null if update engine's binder service is not running.
Bug: 26731548
Change-Id: I8a615c87f33e15a163d56ccfa3a932a0908aac5f
Previously GetFeatureSet() on an invalid target would just return an
empty feature set, leading to some invalid assumptions, e.g. if there
isn't exactly one device connected this happens:
$ adb shell -t
error: target doesn't support PTY args -Tt
This CL adds a success/failure return value to GetFeatureSet(), and
also adds an option to print errors to stderr since that's the most
common behavior.
This will cause a slight difference in behavior for install/uninstall.
Previously they would block until the device was available, now they
print an error and quit immediately, which seems to be the more common
behavior for adb functions.
Bug: http://b/26387641
Change-Id: I0ea6ffaec922e04b9946e84f05c3870e5b549fde
This CL moves the network address parsing function from adb to libbase
so that it can be used by fastboot as well as adb.
libbase seemed like the right choice because:
1. It already has some parsing functions (parseint)
2. The net address parsing function uses the libbase string
functions so we have a libbase dependency anyway.
The parsing function has been modified slightly to make the canonical
address optional, and debug logging on success has been removed.
For adb the only functional difference is that parsing a network
address will no longer print the result to the debug log, which seemed
unnecessary.
Bug: http://b/26236380
Change-Id: Ife6df02937225fc66de87884d3572d79c092c522
Yabin Cui