- May 01, 2019
-
-
David Worsham authored
Tested: No behavior changed; modular_tests, started story in sysUI SCN-1018 #comment SCN-1291 #comment Change-Id: I8509501a15094d7d381557c186a9d0463488a956
-
Christopher Tam authored
Standardize comments such that doc comments ("///") are always associated with some declaration. Otherwise, comments should be normal ("//"). Test: Compilation, as this does not change anything that is not comment-related. Change-Id: I930321b3030f649fe923f90ead7e02271980e99b
-
Pat Erley authored
There is at least 1 case where we're getting audio glitches due to the vmo having stale data in it. To potentially make us resilient to these sorts of glitches, lag out copy out of the vmo behind the current position in the vmo by 1 notification. Test: Ran e2e test locally and it passed. Change-Id: I11063ee1797d6ca3263c97977b8c6dd1b5f4edb1
-
Justin Mattson authored
TEST=verify components still launch! CF-705 #done Change-Id: Ib298702bf485018054ad42dcfa264f59d83e7e14
-
Ankur Mittal authored
Copied code from flutter_runner, fixed documentation and added tests. Once this lands, we should delete original code. Change-Id: Ic67e1d8576764a495d81486bf1db9d075f90f4ec
-
Roland McGrath authored
The C library needs weak stub definitions of the ubsan runtime entry points that instrumented code in the C library itself can call. None of these will be called at runtime unless there is a bug, so the stubs all just trap. Generalize the magic for including the sanitizer runtime libraries in the filesystem manifest from asan to also handle the ubsan case. Add back the ubsan-based variants so they get noop-test variants built. These variants of the test now pass, where previously they would crash. Bug: ZX-3986 #comment ubsan variants working Test: /boot/sys/test/noop-test.ubsan; /boot/sys/test/noop-test.asan-ubsan Change-Id: Ic18e2f8cceb50cbc76f4a4e99961c44b84155381
-
Roland McGrath authored
The mini-process library builds into itself a function that it copies into the child "mini" process, which has no other setup or runtime. So this code must be compiled with no instrumentation, even when the library itself is instrumented. Test: Build with variants=["profile"] and core-tests-{arm64,x64}.zbi now passes tests Change-Id: Ie8fd8eef39ec7b9ad94148b9f840ac0b28ad960f
-
Christopher R. Johnson authored
This change does the following: - iquery now uses libiquery for discovery and parsing locations - libinspect-discovery is removed Next follow-ups: - Factor out formatting to libiquery - Change iquery to use libiquery for formatting CF-714: #progress Change-Id: Ib594f5861f5ba03aaccbad8686f558936ab43fa0
-
Alex Min authored
This changelist changes basemgr to now check for a runtime login mode override from SetUI service and not use auto-login guest mode if it conflicts with this value. The basemgr also persists the compile time configuration to SetUi service if no specification has been made yet. This change also contains a test for fxr/276756. SU-167 #done TEST: fx run-test modular_login_tests Change-Id: I7ae9f9ee6dcf973567b10f1574380d42c2b480a6
-
Yifei Teng authored
To prevent users from doing silly things with them. FIDL-586 #done TEST: No behavior change Change-Id: Ica57761910f7e77ca67c04870e56fbf2f67f706c
-
Ruchira Ravoori authored
Paths have changed. The test has to change accordingly Test: device-enumeration test has no block errors. Other parts of the test fail. That will be addressed in another changeset. Change-Id: I501265e446da86555282c2a4301c526db4b9a18f
-
Leo Lung authored
CF-162 #comment Change-Id: I5121b370f8e7d8ca53f0461929012ada04b6a0f5
-
Mike Voydanoff authored
TEST: builds Change-Id: I9d02a91554a7f1031a543212ca5e2a7bbd551a32
-
Payam Moradshahi authored
In order to properly shutdown and resume the display system on mediatek, the LCD power control needs to occur at specific times during shutdown sequence. Therefore, it was not possible to simply enable the LCD power from the board driver anymore Test: Tested on two versions of Cleo: One with display initialized from the bootloader and the other without any display initialization. Also tested on qemu for unit testing. On Target: Test: display-test --mediatek Result: Manually observed display screen rendering images while the test was running On QEMU: Test: runtests -t dsi-host-test Result: All tests passed Change-Id: I690378cafe59b8ace0451fb328b45b702693a560
-
Sean Klein authored
Change-Id: I84caec9804228c47920b29fb6efd2044201ae20b
-
Tim Detwiler authored
This change sets up a simple sanity test that demonstrates the ability to simulate a guest system interacting over vsock. To accomplish this we introduce a new testing library for faking out the launching of guest systems; providing an API to simulate guest interactions instead of launching real virtual machines. Test: fx run-test linux_runner_tests Change-Id: Ifc63ddff4d57200cac17a3922a487526f332df7c
-
Cristián Donoso authored
This is basically an extension of the breakpoint request to change the BreakpointType on the AddOrChangeBreakpoint debug ipc call. All handling is transparent to the client. Updated the debug logging to show better information. TEST=Added a unit test to verify the setup is doing the expected thing. Manual testing on the console. It works! Change-Id: I4e85e9cba3e2fff9ecc5dadeab4a49014b325fef
-
Joshua Seaton authored
Else, runtests will try to run it like a test by default (as the bringup builders are now trying to do). Bug: INTK-1014 #comment Change-Id: I28b9269ec4257fb2faff9826872a9f29bdf15b30
-
Ankur Mittal authored
DX-387 #progress Change-Id: I3c0c78625452d0ddc9162ae38ada18e67ceb477d
-
Dale Sather authored
This CL modes the mdns implementation from garnet/bin/mdns to src/connectivity/network/mdns. It also removes some stray debugging log messages that made it into a previous CL. TEST: no behavior change Change-Id: Ief61b42e2183b5bf9b38f262e4fabd39143ab33d
-
David Worsham authored
Tested: No behavior changed; modular_tests, Started story in SysUI SCN-1018 #comment SCN-1291 #comment Change-Id: I916d99e8610852b488a4dfa0bad1ecd2fc03ac30
-
Mark Seaborn authored
Before this change, fs-bench takes 3.3 minutes. The largest instance of the PathWalk test (testing paths containing 1000 components) takes a big chunk of that, at 2.0 minutes. Removing that instance reduces fs-bench's overall running time on the NUC perf bot to 1.3 minutes. Bug: PT-142 Test: garnet-x64-perf-dawson_canyon trybot + check time taken shown in bot log Change-Id: I377a16cf14c56df0fc233c88ec03ecd6a5458be5
-
David Worsham authored
Snapshots themselves are broken, so I cannot test them right now. This CL is not expected to change any bhevior though. A follow-on task will be to fix snapshots and create an integration test. Tested: No behavior changed; modular_tests, started story in sysUI SCN-1018 #comment SCN-1291 #comment Change-Id: I25bf52cf289db946b3abac991ed60b992d97c368
-
Taylor Cramer authored
Test: `fx run-test fuchsia_component_test_outer` Change-Id: Ia3c404c7b4eed12448d1785e74f2946f89a21c91
-
Arman Uguray authored
Bug: BT-89 Test: bt-host-unittests --gtest_filter=*ATT_* Change-Id: I191a895df28771abc2072dbc57c43103466d29a2
-
John Bauman authored
Disabled by default since it can only run on cleo, which we don't have an environment for. Change-Id: I6bcf9e5115106686a55a281e52f2c286e76d063f
-
Craig Tiller authored
Add a fuzzer that ensures that we eventually can form a connection between packet nubs even in the presence of dropped packets. Fix all the bugs it found. The next CL in this series will re-enable tests that caught the fact that there is a problem with this code. Change-Id: I6ea83192eddba9571bfe209bc73c813b14499225
-
Derek Gonyeo authored
Change-Id: I78cb8033c9cc0d3f0bd719f38bd28efe545e7ca5
-
Doug Evans authored
PT-148 #done Change-Id: Ib62de6a9f99aa1c242a23ef190ae61c0c68ac65e
-
Doug Evans authored
FLK-193 #comment patch Change-Id: I5697367313b9d18a23caba0a22c30144ca7659b5
-
Ricardo Vargas authored
- Now there's support for parsing basic NDM/FTL structures. - The NDM list of bad blocks is used by other parts of the tool, for example to avoid reading bad blocks. - A --wear command displays basic wear count information. ZX-3506 #comment in-progress Change-Id: I878e4dd91cf8e5f0a90de6cf3cdc40a8c0febe30
-
Erick Tryzelaar authored
go-tuf G1 replaces the root metadata .ID() method with a .IDs() method, since keys can contain multiple key ids. This patch temporarily comments out a section of the amber daemon test's TestOpenRepository to allow for a soft migration of the go-tuf library. Bug: PKG-687 #comment temporarily disabling a test to ease migration Change-Id: I19f3c7ae8f83bf63ce53de51f337066698bde068
-
Elsbeth Weeks authored
After discussion - changes made by this cl include: - New fuschia-camera-common fidl interface - New ControlV2 protocol that uses common and will replace Control - CreateStreamV2 in sdk that will replace CreateStream. Transition plan: 1) add V2 versions of some protocols and structs, 2) change all instances that use the camera fidl interfaces to use the new versions 3) remove sdk/.../camera.fidl, and all the structs from manager.fidl and fuchsia-hardware-camera/camera_intf.fidl (since they are in fuchsia-camera-common) 4) change V2 names back Change-Id: I93e1560ce2c4694c66daaf193cc3438b6a3c8ada
-
Laura Peskin authored
This is a temporary change to unblock a CL in the cobalt repo which changes the way the EventAggregator handles numeric values of 0. For now the tests will stop logging 0 values. Change-Id: I0da3ac88b173349c4070a4f71b2ef2e5bb07e270
-
Ian McKellar authored
This doesn't remove the fidl_compatibility_test_bin component yet, the test in topaz still depends on that but I have a CL to update it once this has landed in GI. Change-Id: I9b83357ca86f213c09ebbd595e1caa1ea1ef7bfa
-
Cristián Donoso authored
This CL hooks up an experimental feature that is not activated for normal debugger use. The main idea is that the debugger will store backtraces for relevant exceptions (breakpoints, general exceptions, etc.), so that they can be reviewed later. This CL attemps to be minimal about this, but surfacing a feature to the console requires a fair amount of moving parts (settings and verbs). TEST=Unit test injecting an exception and verifying that the cache stored the stack trace correctly. Change-Id: I2fab74ac517d7e31de73cce8048efe6dcefd1cac
-
Jeremy Manson authored
When walking the wire representation of a pointer type, we bumped the marker twice as far as we were supposed to go. This meant that pointer types only worked if they were at the end of the params. Also, added a convenience method to print markers. Change-Id: If5fb409a4d238588b6710ad8456611360cfff0d4
-
Braden Kell authored
Function 0 is shared between all drivers using a particular card. Currently drivers read and write to function 0 directly, but eventually function 0 will only be accessible through one of the methods added here. Bug: ZX-3879 Test: runtests -t sdmmc-test Change-Id: If1dfff0eba05a4b1d6ff255b4800ab4843b7ea78
-
docs-roller authored
Test: CQ Change-Id: I2fb8dc1a01f193ec0b906ab3fe49fe5a012bb567
-
Taylor Cramer authored
Change-Id: Ic753a3ebc4ab1fbc2103134514b72c6e912587ca
-