- Mar 20, 2019
-
-
James Robinson authored
This adds a top-level BUILD.gn file and defines it as the root (by removing the root_build override in //.gn) and adds forwarding rules to the top-level target definitions in //build/gn/BUILD.gn. Also moves the file generation steps to the top level BUILD.gn file. The other rules in //build/gn/BUILD.gn should move too, but they are currently referred to by infrastructure configuration files. They can move after this lands and rolls to GI and the infra configs are updated. Change-Id: I2043cc507944158cfe691db884d9fc79c2e80cc4
-
Robert Lord authored
This is the first of two or three changes that implement TextField for input methods making changes to legacy InputMethodEditorClient text fields. Once this proxy is complete, the next steps will be: - Migrate latin-ime and any other input methods to use this new TextField interface, their edits still proxied to legacy text fields. - Bring up a new default input method for physical keyboards, replacing the functionality previously hard-coded in Ime. - Allow Flutter and other clients to implement a TextField and expose it to this service, connecting to any input methods. - Migrate all legacy text fields to the new TextField API. - Shut down the proxy when no InputMethodEditorClient users remain. In this change: - Added conversion from legacy TextInputState into new TextFieldState - Added functions for creating and accessing TextPoints - ImeService now also serves a discoverable TextInputContext service - Ime now also optionally serves a TextField service for input methods, although it currently replies to all requests with BAD_REQUEST. - Added support for running the standard TextField integration test suite on the TextField proxy implementation in Ime, although since that implementation is still a work in progress, this test is marked as `#[ignore]` for now. Change-Id: Iec302a734c25b6e4daae8d467f9b0545f8361da1
-
George Kulakowski authored
I tested this by `fx format-code`ing in libc and in the zircon public directory. Change-Id: I611bbecbedb29968247ceebda1b517499d10345d
-
Nathaniel Manista authored
It's not bad content; it's just not content that we can handle right now and it's also content that we don't need. Test: Full build in CLion. Bug: BLD-393 Change-Id: If1b23351754243056b52efbdf79617256292521f
-
Xo Wang authored
Some host binary is generating //default.profraw. It probably shouldn't be littering, but git should also ignore these files. Test: None Change-Id: I0fa0ce47a7671513a30c81cbb6dc1e8a4f7b2189
-
Taylor Cramer authored
Change-Id: Ibbe4515cfe9fdb63dcadb2eb1e7d9b148e3242c3
-
George Kulakowski authored
Bug: ZX-3357 #comment numeric limits Change-Id: Id66b37d37b6c715e2874f474ff8f481e553a7970
-
Xo Wang authored
Test: fx build Change-Id: I66d53aa427e6c1fda1273c41a574c6bfa1387b12
-
Sam Hansen authored
Change-Id: Iae0e7d2251e59bf7ad622bd87ddd41e863b2f611
-
Nick Pollard authored
What was previously bt-host-integration-tests no longer exercises solely the bt-host subsystem, so the name is now a misnomer. We rename the binary to bt-integration-tests to more accurately describe it. In future, we may split the tests into multiple binaries if there is a need for it. Also add a README.md for the integration tests and link it from the top level bluetooth README.md, to fill in a documentation gap. Change-Id: Id4e3db89b37034ba39c75a056b3e980703c9a734
-
Dan Johnson authored
Also: * Update Display format to return new ?hash=$MERKLE format. * Add test to check FuchsiaPkgURI -> string -> FuchsiaPkgUri round-trip for all the existing "valid" URI cases. * Update the-book/package_uri.md to describe the new format. Bug: PKG-560 #done Test: astro still boots. Change-Id: I0a2c5201e4ae6ef97e5aeab7d77fa125a551d0bf
-
Yu Shan authored
Since the rust tool chain is updated to 1.35.0, we no longer need the allow deprecated hack. Bug: SEC-226 Tests: local build Change-Id: I42eafc60a37512caffdd423cef9539077e9335cc
-
Arman Uguray authored
- Renamed btlib -> bt - Made usage of "::bt" vs "bt" consistent - Made namespace declarations consistent s.t. namespace bt::foo:bar becomes: namespace bt { namespace foo { namespace bar { Bug: BT-757 #done Test: bt-host-unittests and bt-host-integration-tests Change-Id: I45eee4a070e18df90a63390ba1c869a87f15ce48
-
John Bauman authored
Struct members are added to specify whether memory should be allocated in the RAM domain, and a function is added to check what coherency domain the memory is in. At the moment the new struct members are ignored. That will be changed in a follow-up patch, after drivers are changed to initialize them. Change-Id: I34f9a7215c8d4881baac0f15baeb6535478a2c04
-
Robert Hahn authored
WLAN-996 #done Test=Verified auto-connect behavior across device restarts. Change-Id: I52161792063deb2175c8d0ae939441d171e093f6
-
Alain Vongsouvanh authored
The cloud TUF repos are not set up yet, but those can be useful through a local TUF repo. Bug: DX-980 Change-Id: Ia3314b91ed7b63a7ca62aa3889f992c5852d43e7
-
John Bauman authored
The PowerVR firmware is a MIPS ELF binary, which the rest of the code was trying (and failing) to handle. Change-Id: I8b5bed778dac90678d02d626e9a5d4b2284ec640
-
Drew Fisher authored
This interface isn't exercised until the next patch in the series, where we remove zxcrypt's self-unsealing and make the paver and block-watcher explicitly unseal zxcrypt devices. Change-Id: I89300d24ed6ba37801e8e953925e5678efc4cd71
-
Przemek Pietrzkiewicz authored
This patch adds a UI toggle underneath each replica view and flips each toggle when clicked. In a subsequent CL, we will plumb this down to the in-memory cloud provider to control synchronization state. LE-663 Test: tiles_ctl add fuchsia-pkg://fuchsia.com/voila#meta/voila.cmx Change-Id: Ib49d018ef56b3881779827afbba462e7b6aa265b
-
Isma Cortes Vazquez authored
Also try to improve the error messages that get logged. TESTED=FUCHSIA_IPV4_ADDR=<wrong ip> ./out/x64/mytest then watched it retry hopelessly. The test had to be modified to increase its default timeout. Change-Id: I5b3a7b9d791d7b546cbca831a82c6d5207a64372
-
Ruchira Ravoori authored
This changeset adds ZX_PROTOCOL_POWER that operates on a particular power domain. This has the same semantics of GPIO for now. A subsystem will request for a power domain in its board driver. The device driver will get the power protocol on a power domain, using the index of that domain in the array of power domains requested in their board driver. Test: Used the power protocol apis in mt8167-touch and was able to enable/ read status of a domain correctly. Change-Id: I47a8809264e89e3abe466956baf079ac53dbdb90
-
Casey Dahlin authored
This is a much better place for this logic. While we're here, don't show symbol repos in sym-stat if we don't actually find a .build-id folder. Change-Id: Ia84e63a74847c931a00ca0154616a3242ca7eba8
-
John Bauman authored
Some clients need to negotiate what level of cache coherency that is used with a buffer. For example, the Intel display hardware isn't cache-coherent with anything else, while the GPU is cache coherent with the CPU, and can be configured to disable caching for some textures. It's much more efficient to only require that data be in RAM in these cases, rather than require extra cache flushing. Add a flag so that clients can advertise RAM domain support, and some simple heuristics for when to use it. CPU coherency is still the default; advertising support for CPU coherency is temporarily added to the display drivers, even though it's not correct, until consumers that require it are moved off. Change-Id: Ie4a7295d6c08223193a1c2e04a399a91112e34a9
-
Tamir Duberstein authored
Change-Id: Ide99124c5630b1d4334a4a1136fe946a716e4dc8
-
Scott Graham authored
This adds a new safe math library based on Chromium's base/numerics. This library is dependency-free and header-only, and provides well-defined semantics for safely and performantly handling a variety of numeric operations, including most common arithmetic operations and conversions. One common use is to range check and terminate on failure. e.g.: int some_value = ...; size_t buffer_size = safemath::checked_cast<size_t>(some_value); will result in program termination if some_value is out of range for buffer_size. There are many other usage patterns described in README.md. If you land here in git log and wish to update this library, see zircon/third_party/ulib/safemath/do-update.sh! ZX-3415 #comment [safemath] Add third_party/ulib/safemath OSRB-156 #comment [safemath] Add third_party/ulib/safemath Test: CQ Change-Id: I243835521195f1a308e01b00c162b0152f85fb75
-
Mohan Srinivasan authored
Separate ACPI code from the x86 board driver. In future, we can move the ACPI namespace walk code and drivers into a separate library. Test: Boot NUC, Pixelbook and make sure ACPI devices are initialized. Change-Id: I0b8a5de90adeb70777d74947aeb64eff25666cf9
-
Brett Wilson authored
Add a mapping for the builtin C/C++ types and the common stdint types. This is currently unused, it will be used in the future as a fallback for type lookup when a type name is not found in the debugged program. This currently handles one-word types only. In a future pass we'll handle things like "short signed int". Converts BaseType's constructor's "size" parameter to uint32_t which matches the type of the size use by the Type class. Change-Id: I2fd49c0a4f6967f68f6553d7564d02e492ff8432
-
David Worsham authored
This is a reland of 69ddd85b Original change's description: > [scenic] Remove Present2 > > Test: present_view tile_view; present_view simplest_embedder > SCN-1289 #done > > Change-Id: I3dee35a0c4f8f4d4b9d342d2f33eb2ce2abc716c Change-Id: I5a9e7332b7c576421fe6a3e9079edee7d5c4f268
-
Casey Dahlin authored
Moving to a state machine has some complexity benefits, and will set us up to implement better tab completion. Change-Id: Iae93b8f600b002fa84caf1bd2170cb674abdb2b3
-
David Worsham authored
It is no longer used anywhere, but give people a month to stop using it just in case. SCN-892 #comment SCN-990 #comment Change-Id: I40dc9aacbdc8b0fc0ed45595033ae612b0fdb8d2
-
Tim Detwiler authored
To initialize the stateful partition, run: (host) $ cd $FUCHSIA_DIR (host) $ fx scp \ garnet/bin/guest/pkg/biscotti_guest/images/x64/stateful.img \ "[$(fx netaddr --fuchsia)]:/data/r/sys/fuchsia.com:biscotti:0#meta:biscotti.cmx" PD-196 #done Test: Initialized stateful.img with the above and booted the VM. Change-Id: I0299dd70f3be5e010a6f9ef7e149376b5495638e
-
Ankur Mittal authored
DX-387 #comment CF-664 #comment Change-Id: I80e0216f7bb57f7f5aa428f87aac096e6aa64d03
-
Ankur Mittal authored
Fix a regression which broke Environmentservices when started on a seperate thread. DX-387 #comment TEST=fx run-test component_cpp_tests Change-Id: I4f490883cf40138696c0cc65cc8ca727fa6533dc
-
Ruchira Ravoori authored
This changes adds implementation for some of the protocol functions. At the moment this includes only ENABLE and DISABLE of power domains. Modifying the audio/thermal and other drivers to use these apis will be done as part of another changeset. Test: Tested that I was able to write to the registers and read back the correct values with https://fuchsia-review.googlesource.com/c/fuchsia/+/264672 on touch driver. Change-Id: I2935ee3b08eeaf32010bddf1d128ba32a5503149
-
Andres Oportus authored
Smc call shows different replies for valid image ids. Even though there is no valid image to be loaded yet, this shows we are communicating with QSEE and getting meaningful errors. Test: enabling TEST_SMC shows images supported and different auth errors Change-Id: Icebb84913a911d91c900a2cb9a46044455669091
-
Felipe Archondo authored
Test: compiles Change-Id: I098ba7552c537ce08e6b8a8318b8897a3a42cabf
-
Christopher Anderson authored
A quick conversion to make MmioConfig compatible with ddk::MmioBuffer for unit tests. MmioViews will be used in the future. Change-Id: I618fb1b468f65e89253c994199ec37d03d3ed0c7
-
Nick Pollard authored
Changes bt-gap so that if an error occurs when adding or removing a host, this does not terminate the whole program. Due to timing issues, it is possible during add_adapter that a bt-host device has been removed before we attempt to initialize it (open a channel, set connectable, enable background scan etc.), which results in an error. We should log this error, but as this is a situation which could occur in normal behavior, we should not terminate the program. Test: This behavior can be seen by adding and quickly removing an hci device before bt-gap has finished initializing it. This is hard to do as a human but can be implemented in a test. This was previously uncovered by the set_active_host test in bt-integration-tests, but the current behavior no longer produces this. Change-Id: Ife27b14dc6b711c9a4b7c9e3cebc5aa21ea742e3
-
Satsuki Ueno authored
Change-Id: I10aa91918147c13e0975f8903c18c7fbe8c7c5ac
-
Adam Barth authored
This is a reland of 0d4b27fe Original change's description: > [fdio] Remove lib/fdio/util.h > > All clients have moved to including the actual headers they need. > > Change-Id: I584d55510f15cc33217cc979482bafb41ca849fc Change-Id: I176f4ca2d36df76a9e237f66c8d695feb65c2f61
-