This change creates a concatenation of the ids.txt produced by the build
that that produced by the .build-id archives downloaded from CIPD - and
passes that to dump_breakpad_symbols.

Tested locally.

Bug: INTK-884 #comment
Bug: INTK-969 #comment
Change-Id: I9d96f8098239ca6a1c828be0a44dbe36a0c3b14a

Hooks up function name searching for autocomplete for locations ("break"
command, etc.). This also completes class and namespace names.

The behavior isn't perfect but this is a good first step we can iterate
from.

Adds some additional test infrastructure to be able to mock targets, and
generalize the existing process mocks.

Adds a name to the FoundName class since currently namespaces and
templates have no way to return the matched name when there is more than
one. Previously we were only doing exact matches for these kinds so a
separate name was not needed.

Change-Id: I44e88165ff66a7b9a6766ff5979da286af6d1bbf

Accommodate "--packages" -> "--with-base", and "ARCH" ->
"PRODUCT.BOARD".  Remove now not nearly as relevant "--flutter_profile"
suggestion.

Test: Documentation only change
Change-Id: I832c8ae78f25715e67ffbe941a8263d115ea8f64

This reverts commit 07bc2f67.

Reason for revert: Broke Chromium tests, specifically because of changes in dlopen_vmo.
Bug: ZX-3943 #comment

Original change's description:
> [fdio] Lift zx_vmo_replace_as_executable out of fdio_get_vmo_*
> 
> Today, anything calling any of these methods is getting a VMO that is
> marked executable.  Not every caller actually needs an executable VMO;
> generally only things that try to launch other programs need this.
> The fdio_get_vmo_* calls are documented as returning a read-only,
> non-executable VMO; however, as far as I can tell, they have never
> actually done so.
> 
> To whittle down the overuse of executable VMOs, we can push the "mark
> things executable" responsiblity up to various callers.  After some
> digging, I believe this changeset covers the universe of launchers and
> loaders that need executable VMOs in the Fuchsia tree.
> 
> Tests: paved and booted astro successfully
> Bug: SEC-252 #comment Lift zx_vmo_replace_as_executable out of fdio_get_vmo_*
> 
> Change-Id: Ieb6e0c5c12dff152b1cd57cd736bb25df5989332

TBR=raggi@google.com,zarvox@google.com,ojh@google.com

Change-Id: I4362a7ad288e5fa3b826bd367844d8187d863f23
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: SEC-252 #comment Lift zx_vmo_replace_as_executable out of fdio_get_vmo_*

It seems stable enough and lacks a few specific bugs from ngunwind.

Change-Id: I7f9d4ac05e60185c65eee80436a8602c8a10ce40

ZX-3917 #comment

Change-Id: Ieddef8095a297fab9ed0176d3a4606fd0c2ce168

We're going to move tutorials under tutorial/. Bigger changs coming,
this is a quick fix to deal with linking issues.

Change-Id: I0d0a48f7c78691fe03042f666ae3ffdbd42a85b1

This used to be used by infra recipes, but is no longer.

Change-Id: If4328102ce6ae5c746546edff1c762bee05881e3

Test: Compiled
Change-Id: I445b87227d5aab2148c61a6ed23bbfff3b18a038

This was removed in 95a893db.

Change-Id: I05c263957e4f990ad53ec0bf98ac5e8df88ec006

This test demonstrates a couple of tricky behaviors:
- join_promises using a move-only result.
- join_promises combined with a continuation that may
return an error.

Change-Id: I12bfaab47ef0c2b8013ed2caa8d30356cca279cc

Add common vital stats for common fs operations and use them in
minfs. Some of these stats will be used in blobfs.

Tested: Added unit test for storage-metrics library and
ran minfs unit tests which verifies some of the stats.
storage-metrics output should look like

$ ls > /data/x
$ ls /data/
d  2       60 .
-  1      229 x
$ ls > /data/y
$ rm /data/y
$ rm /data/y
$ storage-metrics --enable /data
Metrics for: /data
Filesystem Metrics for: /data
General IO metrics

create.success.total_calls:         15
create.success.total_time_spent:    252906636
create.success.maximum_latency:     112202778
create.success.minimum_latency:     1515318
create.success.bytes_transferred:   0

read.success.total_calls:         259
read.success.total_time_spent:    2065087249
read.success.maximum_latency:     111109852
read.success.minimum_latency:     13983
read.success.bytes_transferred:   0

....

unlink.success.total_calls:         1
unlink.success.total_time_spent:    25167756
unlink.success.maximum_latency:     25167756
unlink.success.minimum_latency:     25167756
unlink.success.bytes_transferred:   0

....

lookup.success.total_calls:         494
lookup.success.total_time_spent:    7935389481
lookup.success.maximum_latency:     165046359
lookup.success.minimum_latency:     47616
lookup.success.bytes_transferred:   0

....

create.failure.total_calls:         0
create.failure.total_time_spent:    0
create.failure.maximum_latency:     0
create.failure.minimum_latency:     0
create.failure.bytes_transferred:   0

....

lookup.failure.total_calls:         12
lookup.failure.total_time_spent:    189320416
lookup.failure.maximum_latency:     62403873
lookup.failure.minimum_latency:     125763
lookup.failure.bytes_transferred:   0

....

Change-Id: Ia90699c11040eaa58ad7430b018600c89207a507

We were previously moving the task RefPtr into a smaller scope, so when
we left that scope the task could die leaving a dangling pointer.

ZX-3940

Test: runtests -t exception-test
Change-Id: Ic46910609b13dcd3872b3b8707c5a3f669a5c481

fidlgen seems to only be installed under host_x64 by virtue of being in
the SDK. This patch hooks up the installation action for both fidlgen
and fidlgen_llcpp under //bundles:tools, such that they are always
copied to the right location.

TEST: Make sure the regen.sh script for goldens works under the fx set
core.x64 configuration.

Change-Id: Ie3935bccf4cc53bc97d0d5b8ca5643c9425ad08c

Uses the cmdline parser from Zircon. This is already used by zxdb and is
better suited to our needs than the fxl one.

This fixes a bug with the android unwind flag being broken. The
validated name didn't match the string that was read to enable.

Changes the unwinder flag to specify either "ng" or "aosp" which allows
us to change the defaults back and forth more easily.

Add the --unwind flag to the "fx debug" command.

Change-Id: I65783f125f4af81e1105373f4945a9745a270850

Due to the ordering of vn lookups in fdio_ns_unbind, it is not possible
to unbind a directory at /* when more than one directory at that level
exists.

For example:
bind "/svc"
bind "/pkg"
bind "/test"
unbind "/test"

causes a debug assertion failure.

This change reorders the lookup of the next vnode in the chain to select
the correct node under the root for removal.

Change-Id: Icab5cf37229c93900606cbe2d8e64375c9b3818a

After this CL, recovery is able to draw to an rgb565 framebuffer (which
is what QEMU uses). However, we are still drawing four byte pixels
rather than two byte pixels. The net result is that the recovery UI is
usable, just the wrong color.

Change-Id: I70beddb2327a2986cf9c72e3c6cacb78a80b54e8

Abigen backend uses /// instead of shortdesc for top description.

ZX-3882 #comment [prebuilt] Roll banjo to Apr18

Test: CQ
Change-Id: I5bcde81a7974add21c3c995c56a04e5360c6905f

While it is not impossible to understand, I think a comment helps.
I also believe that moving the code responsible for constructing default
value into the corresponding match branch makes the intent a bit easier
to see as well.

Change-Id: I1e6dbe457130aa1971577bfa3412d1952433fc8c

This CL adds mdns support for multiple config files. The config files
can be used to suppress host name probing (for faster startup) and to
publish service instances. The latter is used to publish the device
as a "_fuchsia._udp." instance, which was previously hard-coded.

TEST: avahi-browse -t _fuchsia._udp.

Change-Id: I2e3657eefdf75a2626a6a73127fd171182c84d10

This adds the ability to add and remove repository configs
inside the resolver and through pkgctl. These settings are
not persisted across reboots.

Test: Added tests to confirm adding/removing works.
Bug: PKG-598 #comment support modification, persistence will be done later.

Change-Id: I7a2d6a94596ac70f779d339eba0da93fefdf9720

Change-Id: I8b5b75bdfd3a39eecea3ec6239f3003ad1941ad3

This CL implements a fidl table validation macro which to
enforce assumptions at type-level. A validated type can
be constructed at fidl endpoints and used internally.

Change-Id: I75644048a0f78c0392b900042eef5323019884e8

With the removal of the DRM FIDL files from fuchsia.media, the
fuchsia.media library no longer requires a dependency on fuchsia.io.
This also adds the fuchsia.io dependency to fuchsia.modular.internal,
which was transitively using it through dependencies on fuchsia.media.

MTWN-271 #done

Change-Id: Ibc6d8614e2c25996cb269a6eec24f11353f75105

Implement control.Forget by clearing the in-memory bonding data of
peers, updating the stashed bonding data, and calling
fuchsia.bluetooth.host.Forget on each adapter.

BT-182
Test: In bt-gap-unittests, store::stash::tests::rm_peer

Change-Id: I9520bf77d1cae949b28574fc99bba8d2b5df6760

FSHOST: Export a PseudoDirectory of services.

- Create an fshost Registry, which exports two services:
  "/svc/fuchsia.fshost.Filesystems":
    A collection of mounted filesystems.
  "/svc/fuchsia.fshost.Registry":
    A registration server, where filesystems can request to
    be served by "fuchsia.fshost.Filesystems".
- Delete "FshostConnections", opting to store references to other
filesystems directly in the fshost namespace instead.
- Refactor fshost to avoid access to global variables.
- Add unit tests to fshost.

DEVCOORDINATOR: Plumb the fshost services to svchost.

- Supply a handle to both svchost and fshost, such that the fshost
export directory can also be exported under "/svc".

SVCHOST: Export the fshost services.

- Create a ServiceProxy Vnode, which acts as a service endpoint, but
may also forward directory requests to the underlying connection.
This enables entries in "/svc" to implement the directory protocol,
in addition to any other protocols they may speak.
- Permit the fs::Service vnode from banning directory flags. Services
may implement the directory protocol, so these should not be disallowed.

Change-Id: Ieac9323a9211bd986ad1d19bfe2754197f67bac2

done #MTWN-270

TEST: no behavior change; fx run-test mediasession_tests
Change-Id: I32a83e2cc2cc689adca3ee37c66aae88f0b4a695

Create a new HidDevice class to parse
touchscreens. This only parses a specific
touchscreen, but it should be able to handle
any given device that supports Linux
Multitouch Evdev commands.

At the moment it has hardcoded support for
5 fingers. It also only supports X and Y
positions and does not support pressure or
finger size.

DX-1364 #done

Change-Id: Iecff5155c330375294e64e811982de10b83029ab

This seperates out the device specific from the virtio
code so more devices can be created. At the moment we
only support Keyboards.

TEST: Manually tested that keyboard still works
      in QEMU.

DX-1364 #comment

Change-Id: Id3d8e0db9fa019b2ac876876ff3344b470d2b5b9

Also:
- marking terminal nodes as final classes
- removing unnecessary virtual destructors on non-polymorphic classes

Change-Id: Icdec84e8b48b5138e1f92d90bbe487e1e14ae3bc

BUG: NET-2097 #done

Parse the following IPv6 Extension Headers:
- Hop-by-Hop Options
  - The only options being parsed right now are Pad1 & PadN
- Destination Options
  - The only options being parsed right now are Pad1 & PadN
- Routing
  - Supporting parsing of routing headers with routing type 0
- Fragment

This change does not include parsing of the following IPv6 Extension
Headers (they will be in a seperate change so this change doesn't get
any bigger):
- Authentication
- Encapsulating Security Payload

Change-Id: I2273bea90e55c0f2e6917623d2394174c2e02b52

Use carefully selected z values so that no overlapping elements
have the same z. This is something that Carnelian will eventually
handle for apps, but can’t at the moment.

Also fix .cargo/config so fargo can work again.

Change-Id: I2545398cfe1b3e26eedd36d8129dc19edc6f06fa

Depends on relevant CLs in other repos to land first.

Change-Id: Icf654d3a531f3716ce5093288a724b011c92251a

This adds a directory tree that can be used to identify packages that are
part way through being cached, and provides blob installation endpoints for
their blobs.

Change-Id: I54d0425401e0dcf4e96e2990da4a22b7b19173a8

Change-Id: Ib85730783f9c83fa2c74c11f68fa1a8916bbccc7

This isn't hooked up to anything yet, but we'll need it if we want to
re-scan for symbols when a new server becomes available.

Change-Id: I7628ba5faa231c6401ff72c04a4de80a13965dd1

Change-Id: I5b7b8d23ddc1dcdf7a8b0102ac42237bf841da1a

We've found that the mean running times of performance tests tend to
vary between launches of the same process and between reboots.

We want to calculate the mean across process launches and reboots, so
we need to sample accordingly.  This change implements sampling across
process relaunches.  Sampling across reboots will be done later.

 * Before: We launch the perftest process once, and (for each test
   case) do 1000 test runs in that process.  We treat this as a sample
   of size 1000 when calculating confidence intervals.

 * After: We launch the perftest process 30 times, and (for each test
   case) do 100 test runs in each process.  We treat this as a sample
   of size 30 when calculating confidence intervals.

Bug: IN-646
Test: "python perfcompare_test.py" + perfcompare trybot
Change-Id: Ic0176dcc394e87015302bbe3d8588a68408fca05

Change-Id: I114884b2146902218b00b8b975a4d916596d0d15

C++ doesn't generate default move constructors, so we manually write
one.

Change-Id: I3a34e52edd0040eed92ff1a78f6b5ff8333d1202