Merge lp://qastaging/~mir-team/mir/public-cookie-api into lp://qastaging/mir

Should the mir_input_event_get_cookie use MirCookie* or just a uint8_t* (or char*?)

FAILED: Continuous integration, rev:3239
http://jenkins.qa.ubuntu.com/job/mir-ci/5989/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5500
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4407
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5456/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/313/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/313/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5453/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/5989/rebuild

FAILED: Continuous integration, rev:3239
http://jenkins.qa.ubuntu.com/job/mir-ci/5991/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5502
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4409
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5458/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/315/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/315/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5455/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/5991/rebuild

FAILED: Continuous integration, rev:3241
http://jenkins.qa.ubuntu.com/job/mir-ci/5996/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5508
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4415
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5464/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/320/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/320/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5461/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/5996/rebuild

FAILED: Continuous integration, rev:3240
http://jenkins.qa.ubuntu.com/job/mir-ci/5995/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5507
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4414
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5463/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/319/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/319/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5460/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/5995/rebuild

FAILED: Continuous integration, rev:3241
http://jenkins.qa.ubuntu.com/job/mir-ci/5997/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5509
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4416
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5465/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/321/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/321/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5462/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/5997/rebuild

FAILED: Continuous integration, rev:3242
http://jenkins.qa.ubuntu.com/job/mir-ci/5998/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5510
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4417
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5466/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/322/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/322/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5463/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/5998/rebuild

FAILED: Continuous integration, rev:3243
http://jenkins.qa.ubuntu.com/job/mir-ci/5999/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5511
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4418
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5467/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/323/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/323/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5464/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/5999/rebuild

FAILED: Continuous integration, rev:3243
https://mir-jenkins.ubuntu.com/job/mir-ci/31/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/31/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/31/rebuild

> Should the mir_input_event_get_cookie use MirCookie* or just a uint8_t* (or
> char*?)

my vote is MirCookie*

After ROAF and I talking for a bit we settled on void* since we are send a stream of *unknown* bytes. Since not just the mac is being sent, so is the timestamp

FAILED: Continuous integration, rev:3244
https://mir-jenkins.ubuntu.com/job/mir-ci/45/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/45/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/45/rebuild

FAILED: Continuous integration, rev:3244
http://jenkins.qa.ubuntu.com/job/mir-ci/6010/
Executed test runs:
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5523/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4430/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5479/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/334/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/334/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5476/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6010/rebuild

The Story Of The Void*:
Brandon and I had a *long* back and forth about client API wrt MirCookie, ranging from having a MirCookie type that can be converted into a MirBlob, to essentially duplicating MirBlob's API against a MirCookie, to a non-opaque struct MirCookie { size_t length, void* data };

Neither of us were particularly happy with any of them.

The void* API here is (a) transparent - it's clear to everyone that they've just got a blob of data, (b) quick to implement, and (c) is minimally constraining for future, better, implementations with better APIs.

General comment: I'm pretty sure you don't need reinterpret_cast<> to go between void* and MirCookie*; static_cast<MirCookie*> should be sufficient. (Wheras you *do* need to reinterpret_cast between uint8_t* and uint64_t*)

42 + * \params[in] ev The input event
43 + * \params[in] cookie An allocated void* with exactly cookie_size bytes
44 + * \params[in] size The size of the MirCookie
45 + */
46 +void mir_input_event_copy_cookie(MirInputEvent const* ev, void* cookie, size_t size);

Two things: do we need to pass size in here? We're only going to assert that it's exactly mir_input_event_get_cookie_size(), right?

And: If we do keep the size, you need to update the comment; no MirCookie here ☺.

Also a couple of inline comments.

FAILED: Continuous integration, rev:3245
https://mir-jenkins.ubuntu.com/job/mir-ci/46/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/46/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/46/rebuild

FAILED: Continuous integration, rev:3245
http://jenkins.qa.ubuntu.com/job/mir-ci/6011/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5524
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4431
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5480/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/335/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/335/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5477/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6011/rebuild

FAILED: Continuous integration, rev:3246
https://mir-jenkins.ubuntu.com/job/mir-ci/47/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/47/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/47/rebuild

FAILED: Continuous integration, rev:3246
http://jenkins.qa.ubuntu.com/job/mir-ci/6012/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5526
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4433
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5482/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/336/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/336/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5479/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6012/rebuild

not too flummoxed about use of void*, especially if you & Chris gave it more thought than I had behind my comment. Some other things that I'm not too flummoxed about (esp given that (as I understand) this is needed in short order)

+ auto const* pev = mir_input_event_get_pointer_event(ev);
no need for * (a few other places)

+ // FIXME Moveing to 160bits soon
typo

+ void* cookie = malloc(size);
would be a bit better to use a RAII type in allocation for this test (and would eliminate the need for a destructor too)

768: std::mutex mutex;
A bit strange to have the mutex be public, and require locking it before accessing some members. Would be better encapsulated to have functions on the object to do that.

> not too flummoxed about use of void*, especially if you & Chris gave it more
> thought than I had behind my comment. Some other things that I'm not too
> flummoxed about (esp given that (as I understand) this is needed in short
> order)
>
> + auto const* pev = mir_input_event_get_pointer_event(ev);
> no need for * (a few other places)

I was mainly doing that so you can see its a pointer (I like seeing types sometimes), but yeah auto captures that :). Fixed.

>
> + // FIXME Moveing to 160bits soon
> typo
>

Opps. Fixed.

> + void* cookie = malloc(size);
> would be a bit better to use a RAII type in allocation for this test (and
> would eliminate the need for a destructor too)
Yes! I attempted it but ran into issue when I was using the MirCookie* since the size was not known (should have just written my own destructor for the shared_ptr to cast and delete[].) Fixed.

>
> 768: std::mutex mutex;
> A bit strange to have the mutex be public, and require locking it before
> accessing some members. Would be better encapsulated to have functions on the
> object to do that.

Agreed, and it looks much cleaner :). Fixed.

There's stuff I'm concerned about, but I'm out of time today to review properly - will get to it first thing tomorrow.

(pushing coming in a few :)

void* is an incomplete type, soo to put it into a shared/unique ptr fails on a static_assert. I suppose for testing i could just cast it to a uint8_t, since its in bytes... or just leave it with a malloc/free :)

FAILED: Continuous integration, rev:3247
https://mir-jenkins.ubuntu.com/job/mir-ci/58/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/58/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/58/rebuild

FAILED: Continuous integration, rev:3249
https://mir-jenkins.ubuntu.com/job/mir-ci/59/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/59/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/59/rebuild

FAILED: Continuous integration, rev:3248
http://jenkins.qa.ubuntu.com/job/mir-ci/6024/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5540
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4447
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5496/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/348/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/348/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5493/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6024/rebuild

Bunch of locking issues still in the tests; marked inline.

PASSED: Continuous integration, rev:3250
https://mir-jenkins.ubuntu.com/job/mir-ci/60/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/60/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/60/rebuild

PASSED: Continuous integration, rev:3251
https://mir-jenkins.ubuntu.com/job/mir-ci/61/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/61/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/61/rebuild

Different APIs we have considered:
https://pastebin.canonical.com/147720/

PASSED: Continuous integration, rev:3250
http://jenkins.qa.ubuntu.com/job/mir-ci/6025/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5541
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4448
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5497
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-xenial-touch/281
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/349
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/349/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/349
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/349/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5494
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5494/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/7964
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26690
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/277
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/277/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-xenial-touch/135
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26691

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6025/rebuild

566 + void handle_input_event(MirInputEvent const* iev)
567 + {
568 + std::lock_guard<std::mutex> lk(mutex);
569 + if (mir_input_event_has_cookie(iev))
570 + {
571 + auto const size = mir_input_event_get_cookie_size(iev);
572 + std::vector<uint8_t> cookie(size);
573 +
574 + mir_input_event_copy_cookie(iev, cookie.data());
575 + out_cookies.push_back(cookie);
576 + }
577 +
578 + event_count++;
579 + }
580 +
581 + size_t get_event_count() const
582 + {
583 + std::lock_guard<std::mutex> lk(mutex);
584 + return event_count;
585 + }
586 +
587 + size_t cookie_size() const
588 + {
589 + std::lock_guard<std::mutex> lk(mutex);
590 + return out_cookies.size();
591 + }
592 +
593 + std::vector<uint8_t> back_cookie() const
594 + {
595 + std::lock_guard<std::mutex> lk(mutex);
596 + return out_cookies.back();
597 + }
598 +
599 + bool cookies_empty() const
600 + {
601 + std::lock_guard<std::mutex> lk(mutex);
602 + return out_cookies.empty();
603 + }

While not blocking, this seems to be a bit of a faff for the testcase vs just directly locking the mutex when necessary :).

Otherwise, I'm OK, pending possible API changes :)

PASSED: Continuous integration, rev:3252
https://mir-jenkins.ubuntu.com/job/mir-ci/63/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/63/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/63/rebuild

PASSED: Continuous integration, rev:3253
https://mir-jenkins.ubuntu.com/job/mir-ci/64/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/64/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/64/rebuild

PASSED: Continuous integration, rev:3252
http://jenkins.qa.ubuntu.com/job/mir-ci/6027/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5543
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4450
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5499
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-xenial-touch/283
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/351
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/351/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/351
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/351/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5496
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5496/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/7966
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26694
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/279
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/279/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-xenial-touch/137
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26695

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6027/rebuild

Oh, whoops! Missed one thing, sorry.

This would appear to be a libmircookie ABI break, so should be accompanied by a SONAME bump.

FAILED: Continuous integration, rev:3254
https://mir-jenkins.ubuntu.com/job/mir-ci/65/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/65/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/65/rebuild

PASSED: Continuous integration, rev:3253
http://jenkins.qa.ubuntu.com/job/mir-ci/6028/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5544
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4451
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5500
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-xenial-touch/284
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/352
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/352/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/352
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/352/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5497
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5497/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/7967
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26696
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/280
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/280/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-xenial-touch/138
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26697

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6028/rebuild

FAILED: Continuous integration, rev:3254
http://jenkins.qa.ubuntu.com/job/mir-ci/6029/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5545
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4452/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5501/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/353/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/353/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5498/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6029/rebuild

+ typedef struct MirCookie MirCookie;

Not needed (at the moment at least), and also may conflict with mir::cookie::MirCookie.

- virtual MirCookie timestamp_to_cookie(uint64_t const& timestamp) = 0;

A *Cookie*Factory that doesn't create cookies seems strange.

Needs discussion/fixing

PASSED: Continuous integration, rev:3255
https://mir-jenkins.ubuntu.com/job/mir-ci/66/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/66/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/66/rebuild

+// FIXME Remove me when we are no longer returning 8 byte MACS
+// We assert that what we return is 8 bytes, so lets test for that for now!
+TEST(MirCookieFactory, assert_8_bytes_for_mac)
+{
+ std::vector<uint8_t> secret{ 0xaa, 0xbb, 0xcc, 0xdd, 0xee, 0xff, 0xde, 0x01 };
+ auto factory = mir::cookie::CookieFactory::create_from_secret(secret);
+ uint64_t mock_timestamp{0x322322322332};
+ auto mac = factory->timestamp_to_mac(mock_timestamp);
+
+ EXPECT_EQ(mac.size(), sizeof(uint64_t));
+}

sizeof(uint64_t) is not a synonym for 8 - although platforms with 16bit characters are of little interest to us.

Nits to fix and questions:

CookieFactory looks less and less like a "factory" interface. Would "CookieAuthority" be a better name?

~~~~

+ virtual bool attest_timestamp(MirCookie const* cookie) = 0;

This is the only place in the interface MirCookie is mentioned. Should we simply lose this function and require the user to call the other overload?

~~~~

+void mir_input_event_copy_cookie(MirInputEvent const* ev, void* cookie);

I'm always cautions of APIs that take a target buffer and no buffer size - while not bulletproof the latter is at least a reminder to the user to obtain the size.

~~~~

+void mir_surface_raise_with_cookie(MirSurface* surface, void const* cookie);

By using void* we loose the type safety that we have with other Mir APIs. Can you provide some justification for this design choice? E.g. will this function be primarily for clients that obtained the cookie from a non-Mir API?

~~~~

- * Copyright © 2015 Canonical Ltd.
+ * Copyright © 2016 Canonical Ltd.

Don't remove the 2015 copyright claim. (I know these headers are legally unnecessary and PITA, but let's try to do it right.)

~~~~

+ /* No mac == no size! */
+ return 0;

It should be a precondition that there's a MAC. I.e. the function should begin with something like:

    mir::require(mir_input_event_has_cookie(ev));

~~~~

+#include "mir/cookie.h"
#include "mir/cookie_factory.h"

The first include of the source file should be the corresponding header (to ensure it compiles by itself).

PASSED: Continuous integration, rev:3255
http://jenkins.qa.ubuntu.com/job/mir-ci/6030/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5547
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4454
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5503
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-xenial-touch/285
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/354
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/354/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/354
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/354/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5500
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5500/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/7969
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26699
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/281
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/281/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-xenial-touch/139
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26703

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6030/rebuild

> +// FIXME Remove me when we are no longer returning 8 byte MACS
> +// We assert that what we return is 8 bytes, so lets test for that for now!
> +TEST(MirCookieFactory, assert_8_bytes_for_mac)
> +{
> + std::vector<uint8_t> secret{ 0xaa, 0xbb, 0xcc, 0xdd, 0xee, 0xff, 0xde,
> 0x01 };
> + auto factory = mir::cookie::CookieFactory::create_from_secret(secret);
> + uint64_t mock_timestamp{0x322322322332};
> + auto mac = factory->timestamp_to_mac(mock_timestamp);
> +
> + EXPECT_EQ(mac.size(), sizeof(uint64_t));
> +}
>
> sizeof(uint64_t) is not a synonym for 8 - although platforms with 16bit
> characters are of little interest to us.

Hmm should I just not assert this then? Since really all im assert is that our mac == sizeof(uint64_t). So possibly I should just change the name.

> Nits to fix and questions:
>
> CookieFactory looks less and less like a "factory" interface. Would
> "CookieAuthority" be a better name?
>
> ~~~~

Agreed.

>
> + virtual bool attest_timestamp(MirCookie const* cookie) = 0;
>
> This is the only place in the interface MirCookie is mentioned. Should we
> simply lose this function and require the user to call the other overload?
>
> ~~~~

The reason for keeping this, is theres no way to de construct a MirCookie in the public. So once the content hub gets a mir cookie it wont know how to *attest* it.

>
> +void mir_input_event_copy_cookie(MirInputEvent const* ev, void* cookie);
>
> I'm always cautions of APIs that take a target buffer and no buffer size -
> while not bulletproof the latter is at least a reminder to the user to obtain
> the size.
>
> ~~~~

Thats sounds quite reasonable to me.
>
> +void mir_surface_raise_with_cookie(MirSurface* surface, void const* cookie);
>
> By using void* we loose the type safety that we have with other Mir APIs. Can
> you provide some justification for this design choice? E.g. will this
> function be primarily for clients that obtained the cookie from a non-Mir API?
>
> ~~~~

I agree, and will be switching to a MirCookie* for both the copy and these functions since void* and MirCookie* will both be incomplete types.

>
> - * Copyright © 2015 Canonical Ltd.
> + * Copyright © 2016 Canonical Ltd.
>
> Don't remove the 2015 copyright claim. (I know these headers are legally
> unnecessary and PITA, but let's try to do it right.)
>
> ~~~~

Opps. Yeah Ill fix those!

>
> + /* No mac == no size! */
> + return 0;
>
> It should be a precondition that there's a MAC. I.e. the function should begin
> with something like:
>
> mir::require(mir_input_event_has_cookie(ev));
>
> ~~~~

This is true, if we have a function that tells you if an event has a cookie we better require it to actually have one when attempting to construct it!.

>
> +#include "mir/cookie.h"
> #include "mir/cookie_factory.h"
>
> The first include of the source file should be the corresponding header (to
> ensure it compiles by itself).

Will fix!

> + typedef struct MirCookie MirCookie;
>
> Not needed (at the moment at least), and also may conflict with
> mir::cookie::MirCookie.

Opps thought I removed all thoses... Had an idea at one point that I think made it into a file :)
>
> - virtual MirCookie timestamp_to_cookie(uint64_t const& timestamp) = 0;
>
> A *Cookie*Factory that doesn't create cookies seems strange.
>
> Needs discussion/fixing

Right it doesnt create any MirCookies but it creates what a cookie is. With alans comment, changing the name to:
CookieAuthority

> > sizeof(uint64_t) is not a synonym for 8 - although platforms with 16bit
> > characters are of little interest to us.
>
> Hmm should I just not assert this then? Since really all im assert is that our
> mac == sizeof(uint64_t). So possibly I should just change the name.

On reflection, if unit8_t exists we're safe to assume 8 bit chars. (And if it doesn't we don't compile.)

> > + virtual bool attest_timestamp(MirCookie const* cookie) = 0;
> >
> > This is the only place in the interface MirCookie is mentioned. Should we
> > simply lose this function and require the user to call the other overload?
> >
> > ~~~~
>
> The reason for keeping this, is theres no way to de construct a MirCookie in
> the public. So once the content hub gets a mir cookie it wont know how to
> *attest* it.

are you conflating ::MirCookie and mir::cookie::MirCookie?

> > > + virtual bool attest_timestamp(MirCookie const* cookie) = 0;
> > >
> > > This is the only place in the interface MirCookie is mentioned. Should we
> > > simply lose this function and require the user to call the other overload?
> > >
> > > ~~~~
> >
> > The reason for keeping this, is theres no way to de construct a MirCookie in
> > the public. So once the content hub gets a mir cookie it wont know how to
> > *attest* it.
>
> are you conflating ::MirCookie and mir::cookie::MirCookie?

Well only mir::cookie::MirCookie should be used? The attest_timestamp is in the mir::cookie namespace. Not sure if they are different and being combined? I Could be wrong :)

Though there will be an issue... if i move from a void* to a MirCookie* in the public API... not sure how to combine a MirCookie* into a mir::cookie::MirCookie*? Just cast it? I would like a MirCookie* vs void* :)

FAILED: Continuous integration, rev:3258
https://mir-jenkins.ubuntu.com/job/mir-ci/69/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/69/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/69/rebuild

Some concerns addressed. Abstaining until I have time to review again.

FAILED: Continuous integration, rev:3258
http://jenkins.qa.ubuntu.com/job/mir-ci/6032/
Executed test runs:
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5553/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4460/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5509/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/356/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/356/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5506/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6032/rebuild

PASSED: Continuous integration, rev:3259
https://mir-jenkins.ubuntu.com/job/mir-ci/71/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/71/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/71/rebuild

PASSED: Continuous integration, rev:3259
http://jenkins.qa.ubuntu.com/job/mir-ci/6034/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5555
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4462
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5511
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-xenial-touch/287
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/358
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/358/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/358
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/358/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5508
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5508/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/7973
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26718
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/283
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/283/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-xenial-touch/140
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26720

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6034/rebuild

+/*
+ *
+ * \params[in] ev The input event
+ * \params[in] cookie An allocated void* with exactly cookie_size bytes
+ * \params[in] size The size of the allocated void*
+ */
+void mir_input_event_copy_cookie(MirInputEvent const* ev, MirCookie* cookie, size_t size);

When comments and code disagree, both are likely wrong.

While I'm not sure I fully understand the intended usage I suspect that have two scenarios:

1. Getting a MirCookie from an event.
2. Converting a MirCookie to and from a chunk of memory.

In neither of these does it make sense for a user to supply allocate and supply a pointer to an incomplete type.

How about:

MirCookie* mir_input_event_get_cookie(MirInputEvent const* ev);

and either:

MirCookie* mir_blob_to_mir_cookie(MirBlob* blob);
MirBlob* mir_blob_from_mir_cookie(MirCookie* cookie);

Or:

size_t mir_cookie_get_buffer_size(MirCookie* cookie);
void mir_cookie_to_buffer(MirCookie* cookie, void* buffer, size_t size);
MirCookie* mir_cookie_from_buffer(void const* buffer, size_t size);

[Just comment]

It (still) seems odd that CookieAuthority deals with both MirCookies and with pairs of timestamp and MAC. Surely a /cookie/ authority should deal with Cookies? Would it not make more sense as:

auto make_cookie(Timestamp t) -> unique_ptr<MirCookie>

auto authenticated_timestamp(MirCookie const& c) -> Timestamp

Having clients aware of the MAC seems like a leaky abstraction

FAILED: Continuous integration, rev:3261
https://mir-jenkins.ubuntu.com/job/mir-ci/89/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/89/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/89/rebuild

FAILED: Continuous integration, rev:3262
https://mir-jenkins.ubuntu.com/job/mir-ci/91/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/90/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/91/rebuild

FAILED: Continuous integration, rev:3263
https://mir-jenkins.ubuntu.com/job/mir-ci/92/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/92/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/92/rebuild

PASSED: Continuous integration, rev:3264
https://mir-jenkins.ubuntu.com/job/mir-ci/93/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/93/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/93/rebuild

FAILED: Continuous integration, rev:3261
http://jenkins.qa.ubuntu.com/job/mir-ci/6064/
Executed test runs:
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5596/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4503/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5552/console
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/388
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/388/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/388
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/388/artifact/work/output/*zip*/output.zip
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5549/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6064/rebuild

FAILED: Continuous integration, rev:3264
http://jenkins.qa.ubuntu.com/job/mir-ci/6066/
Executed test runs:
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5599/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4506/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5555/console
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/390
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/390/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/390
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/390/artifact/work/output/*zip*/output.zip
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5552/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6066/rebuild

FAILED: Continuous integration, rev:3264
http://jenkins.qa.ubuntu.com/job/mir-ci/6068/
Executed test runs:
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5602/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4509/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5558/console
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/392
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/392/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/392
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/392/artifact/work/output/*zip*/output.zip
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5555/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6068/rebuild

PASSED: Continuous integration, rev:3265
https://mir-jenkins.ubuntu.com/job/mir-ci/95/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/95/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/95/rebuild

FAILED: Continuous integration, rev:3265
http://jenkins.qa.ubuntu.com/job/mir-ci/6069/
Executed test runs:
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5603/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4510/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5559/console
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/393
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/393/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/393
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/393/artifact/work/output/*zip*/output.zip
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5556/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6069/rebuild

PASSED: Continuous integration, rev:3266
https://mir-jenkins.ubuntu.com/job/mir-ci/96/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/96/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/96/rebuild

FAILED: Continuous integration, rev:3266
http://jenkins.qa.ubuntu.com/job/mir-ci/6070/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5605
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4512/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5561/console
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/394
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/394/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/394
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/394/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5558
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5558/artifact/work/output/*zip*/output.zip
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/8011/console
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26821

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6070/rebuild

394 +/* Returns an allocated MirCookie that you must release with mir_cookie_release
394 + *
395 + * \params[in] ev The input event
396 + * \return An allocated and ref'ed cookie
397 + */

There's no need to say “allocated” or “ref'ed” here.

394 +/* Returns the MirCookie associated with this input event.
394 + *
395 + * \params[in] ev The input event
396 + * \return A reference to the MirCookie associated with this input event.
397 + * This must be released with a call to mir_cookie_release().
398 + */

Likewise:
415 +/* Turns the user copyed buffer back into a MirCookie, which must be
416 + * released with mir_cookie_release
417 + *
418 + * \params[in] buffer The buffer used to get a MirCookie back
419 + * \return A newely allocated MirCookie
420 + */
421 +MirCookie const* mir_cookie_from_buffer(void const* buffer);

should maybe be:
415 +/* Create a MirCookie from a serialised representation.
417 + *
418 + * \params[in] buffer The buffer containing a serialised MirCookie.
418 + * The buffer may be freed immediately after this call.
419 + * \return A reference to a MirCookie.
418 + * This must be released with a call to mir_cookie_release().
420 + */
421 +MirCookie const* mir_cookie_from_buffer(void const* buffer);

This can never return NULL, right?

Actually: we should probably pass in the length of the buffer here and return NULL if it doesn't match what we expect. Clients will get this passed in from other sources, and it'll courteous to tell them if it's easy to detect that it's wrong.

443 +* \param [in] cookie The void* from the input event that you want to raise the window from.
Ba baw!

1143 + mir_surface_raise_with_cookie;
1144 + mir_input_event_has_cookie;
1145 + mir_cookie_get_size;
1146 + mir_input_event_get_cookie;
1147 + mir_cookie_copy_to_buffer;
1148 + mir_cookie_from_buffer;
1149 + mir_cookie_release;

These should be in MIR_CLIENT_unreleased, not MIR_CLIENT_DETAIL (which is pseudo-private implementation details).

+ if (raw_cookie.size() < cookie_size_from_format(mir::cookie::Format::HMAC_SHA_1_8))

Should be !=, right?

Why did you move mir::SecurityCheckFailed?

2029 + auto const& cookie_ptr = cookie_authority->unmarshall_cookie(cookie_bytes);

I'm surprised this passed valgrind? unmarshall_cookie() returns a std::unique_ptr<>, that you take a *reference* to, and then the std::unique_ptr<> gets *destroyed* at the end of the statement.

Likewise in a bunch of other places. Why doesn't this blow up?

A: http://herbsutter.com/2008/01/01/gotw-88-a-candidate-for-the-most-important-const/. I wonder why people think C++ is needlessly arcane‽.

PASSED: Continuous integration, rev:3267
https://mir-jenkins.ubuntu.com/job/mir-ci/104/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/104/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/104/rebuild

PASSED: Continuous integration, rev:3267
https://mir-jenkins.ubuntu.com/job/mir-ci/106/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/106/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/106/rebuild

First pass - cosmetic stuff

+/* Returns an allocated MirCookie that you must release with mir_cookie_release
+ *
+ * \params[in] ev The input event
+ * \return An allocated and ref'ed cookie
+ */
+MirCookie const* mir_input_event_get_cookie(MirInputEvent const* ev);

1. The /precondition/ that the event has a cookie is missing.

2. "An allocated and ref'ed cookie" is implementation detail the user isn't interested in.

~~~~

+ * \return The size needed to allocate a buffer
+ */
+size_t mir_cookie_get_size(MirCookie const* cookie);

While allocation is likely what the client will do, surely this is just the size of buffer needed by mir_cookie_copy_to_buffer().

~~~~

+ * \params[in] buffer The allocated buffer to copy the MirCookie into
+ * \params[in] size The size of the allocated buffer
+ */
+void mir_cookie_copy_to_buffer(MirCookie const* cookie, void* buffer, size_t size);

%s/allocated //

~~~~

+MirCookie const* mir_cookie_from_buffer(void const* buffer);

I think it would be useful validation for the buffer size to be supplied and checked. While not bulletproof the it is a reminder to the user that size matters.

~~~~

+* \param [in] cookie The void* from the input event that you want to raise the window from.
+*/
+void mir_surface_raise_with_cookie(MirSurface* surface, MirCookie const* cookie);

The comment and the code disagree

~~~~

+ HMAC_SHA_1_8

Is there a reason for the non standard naming style? https://unity.ubuntu.com/mir/cppguide/index.html?showone=Enumerator_Names#Enumerator_Names

PASSED: Continuous integration, rev:3270
https://mir-jenkins.ubuntu.com/job/mir-ci/110/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/110/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/110/rebuild

From what I've been hearing, I think the terms "cookie" and "marshalling" are over-abstracting the situation. It's better to use more concrete nouns and verbs so the reader more easily understands what's going on. But this is not a review.

FAILED: Continuous integration, rev:3267
http://jenkins.qa.ubuntu.com/job/mir-ci/6080/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5618
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4525/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5574/console
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/404
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/404/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/404
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/404/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5571
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5571/artifact/work/output/*zip*/output.zip
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/8017/console
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26854

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6080/rebuild

FAILED: Continuous integration, rev:3270
http://jenkins.qa.ubuntu.com/job/mir-ci/6084/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5623
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4530/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5579/console
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/408
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/408/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/408
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/408/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5576
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5576/artifact/work/output/*zip*/output.zip
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/8021/console
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26867

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6084/rebuild

FAILED: Continuous integration, rev:3270
http://jenkins.qa.ubuntu.com/job/mir-ci/6088/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5627
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4534/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5583/console
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/412
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/412/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/412
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/412/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5580
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5580/artifact/work/output/*zip*/output.zip
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/8024/console
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26873

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6088/rebuild

PASSED: Continuous integration, rev:3271
https://mir-jenkins.ubuntu.com/job/mir-ci/113/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/113/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/113/rebuild

FAILED: Continuous integration, rev:3271
http://jenkins.qa.ubuntu.com/job/mir-ci/6091/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5630
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4537
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5586/console
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/415
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/415/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/415
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/415/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5583
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5583/artifact/work/output/*zip*/output.zip
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/8026/console
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26876

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6091/rebuild

FAILED: Continuous integration, rev:3271
http://jenkins.qa.ubuntu.com/job/mir-ci/6092/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5631
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4538
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5587/console
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/416
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/416/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/416
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/416/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5584
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5584/artifact/work/output/*zip*/output.zip
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/8027/console
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26878

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6092/rebuild

PASSED: Continuous integration, rev:3272
https://mir-jenkins.ubuntu.com/job/mir-ci/114/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/114/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/114/rebuild

PASSED: Continuous integration, rev:3273
https://mir-jenkins.ubuntu.com/job/mir-ci/116/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/116/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/116/rebuild

PASSED: Continuous integration, rev:3272
http://jenkins.qa.ubuntu.com/job/mir-ci/6093/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5633
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4540
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5589
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-xenial-touch/311
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/417
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/417/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/417
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/417/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5586
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5586/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/8029
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26885
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/307
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/307/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-xenial-touch/163
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26892

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6093/rebuild

Ok, I think. Two inline nonblocking doc comments.

Mostly cosmetic things.

mir::cookie::Array => mir::cookie::Blob

Comment suggestions
---
+
386 +/* Tells if the MirInputEvent has a MirCookie
387 + *
388 + * \params[in] ev The input event
389 + * \return True if the input event has a MirCookie
390 + */

change to:

+
386 +/* Query if an input event contains a cookie
387 + *
388 + * \params[in] ev The input event
389 + * \return True if the input event contains a cookie
390 + */

---

---
/* Returns the MirCookie associated with this input event.
 *
 * \pre The input event must have a MirCookie
 * \params[in] ev The input event
 * \return A reference to the MirCookie associated with this input event
 * This must be released with a call to mir_cookie_release()
 */

 change to:

/* Returns the cookie associated with an input event.
 *
 * \pre The input event must have a MirCookie
 * \params[in] ev An input event
 * \return The cookie associated with the given input event
 * The cookie must be released by calling mir_cookie_release
 */
---

/* The size of the buffer needed to serialize this MirCookie
 *
 * \params[in] cookie The MirCookie
 * \return The size needed for a buffer
 */
size_t mir_cookie_get_size(MirCookie const* cookie);

change to:

/* Queries the size needed to serialize a given cookie
 *
 * \params[in] cookie A cookie instance
 * \return The size of the serialized representation of the given cookie
 */
size_t mir_cookie_size(MirCookie const* cookie);
---

---
/* Copy the MirCookie into an allocated buffer
 *
 * \pre The size must be equal to mir_cookie_get_size
 * \params[in] cookie The MirCookie
 * \params[in] buffer The allocated buffer to copy the MirCookie into
 * \params[in] size The size of the allocated buffer
 */

 change to:

/* Serializes a cookie into the given buffer
 *
 * \pre The size must be equal to mir_cookie_get_size
 * \params[in] cookie A cookie instance
 * \params[in] buffer A buffer which is filled with the serialized representation
                      of the given cookie
 * \params[in] size The size of the given buffer
 */
void mir_cookie_copy_to_buffer(MirCookie const* cookie, void* buffer, size_t size);

To be consistent with mir_cookie_from_buffer
mir_cookie_copy_to_buffer => mir_cookie_to_buffer

---
/* Create a MirCookie from a serialised representation
 *
 * \pre The size must be equal to mir_cookie_get_size
 * \params[in] buffer The buffer containing a serialised MirCookie.
 * The buffer may be freed immediately after this call.
 * \return A reference to a MirCookie.
 * This must be released with a call to mir_cookie_release().
 */
MirCookie const* mir_cookie_from_buffer(void const* buffer, size_t size);

change to:

/* Create a cookie from a serialized representation
 *
 * \pre The size must be equal to mir_cookie_get_size
 * \params[in] buffer The buffer containing a serialized cookie.
 * The buffer may be freed immediately after this call.
 * \return A MirCookie instance. The instance must be re...

+#include "mir/cookie_array.h"

Seeing this caused me to notice an existing issue:

It really ought reflect the "namespace == directory" convention used elsewhere: #include "mir/cookie/array.h"

Not sure if we should fix, or just live with it as "mir/cookie_factory.h" is pre-existing.

Oops on the last one I meant:

---
void mir_cookie_copy_to_buffer(MirCookie const* cookie, void* buffer, size_t size)
{
    mir::require(size == mir::cookie::array_size);
    memcpy(buffer, cookie, size);
}
---

With a client side MirCookie class suggested above would just become:
void mir_cookie_copy_to_buffer(MirCookie const* cookie, void* buffer, size_t size)
{
    return cookie->copy_to(buffer, size);
}

and mir_cookie_release turns into:

+void mir_cookie_release(MirCookie const* cookie)
{
    delete cookie;
}

FAILED: Continuous integration, rev:3273
http://jenkins.qa.ubuntu.com/job/mir-ci/6098/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5641
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4548
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5597
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/422
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/422/artifact/work/output/*zip*/output.zip
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/422/console
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5594
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5594/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/8035
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26902

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6098/rebuild

PASSED: Continuous integration, rev:3273
http://jenkins.qa.ubuntu.com/job/mir-ci/6100/
Executed test runs:
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5644
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4551
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5600
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-xenial-touch/318
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/424
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/424/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/424
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/424/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5597
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5597/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-touch/8038
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26908
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/314
        deb: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-xenial-armhf/314/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-runner-xenial-touch/170
    SUCCESS: http://s-jenkins.ubuntu-ci:8080/job/touch-flash-device/26918

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6100/rebuild

FAILED: Continuous integration, rev:3276
https://mir-jenkins.ubuntu.com/job/mir-ci/125/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/125/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/125/rebuild

FAILED: Continuous integration, rev:3276
http://jenkins.qa.ubuntu.com/job/mir-ci/6105/
Executed test runs:
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5655/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4562/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5611/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/429/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/429/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5608/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6105/rebuild

PASSED: Continuous integration, rev:3277
https://mir-jenkins.ubuntu.com/job/mir-ci/126/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/126/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/126/rebuild

FAILED: Continuous integration, rev:3277
http://jenkins.qa.ubuntu.com/job/mir-ci/6106/
Executed test runs:
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-android-vivid-i386-build/5656/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-clang-vivid-amd64-build/4563/console
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-vivid-touch/5612/console
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/430
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-amd64-ci/430/artifact/work/output/*zip*/output.zip
    SUCCESS: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/430
        deb: http://jenkins.qa.ubuntu.com/job/mir-xenial-i386-ci/430/artifact/work/output/*zip*/output.zip
    FAILURE: http://jenkins.qa.ubuntu.com/job/mir-mediumtests-builder-vivid-armhf/5609/console

Click here to trigger a rebuild:
http://s-jenkins.ubuntu-ci:8080/job/mir-ci/6106/rebuild

PASSED: Continuous integration, rev:3278
https://mir-jenkins.ubuntu.com/job/mir-ci/127/
Executed test runs:
    None: https://mir-jenkins.ubuntu.com/job/generic-update-mp/127/console

Click here to trigger a rebuild:
https://mir-jenkins.ubuntu.com/job/mir-ci/127/rebuild