[Git][xorg-team/lib/libei][debian-unstable] 210 commits: meson.build: bump to use gnu11 over gnu99

Jeremy Bícha pushed to branch debian-unstable at X Strike Force / lib / libei

Commits:

36f16411

by Peter Hutterer at 2023-06-09T10:27:55+10:00

meson.build: bump to use gnu11 over gnu99

static asserts require C11

7beaef95
by Jeremy Bícha at 2023-08-11T08:16:11-04:00
```
New upstream version 0.99.2
```
b24c5585
by Jeremy Bícha at 2023-08-11T08:51:43-04:00
```
New upstream version 1.0.0
```

7ddd70e9

by Peter Hutterer at 2023-08-29T11:39:20+10:00

Add ei_device_get_region_at() to obtain a region for a point

Helper function so clients don't need to loop through the regions to
find the matching one.

b7ab63c3

by Peter Hutterer at 2023-08-30T09:11:08+10:00

util: fix iobuf_take_fd invalid memmove for multiple fds

Wrong calculation resulted in memmoving only the first 32 bytes (i.e. 8
fds) instead of the first 32 fds, resulting in an infinite loop when
cleaning up an iobuf with more than 8 fds.

2fbd2298

by Peter Hutterer at 2023-08-30T09:11:08+10:00

util: fix iobuf_append_fd OOB when we have too many fds

Cannot happen in deploymentts since we never have more than one anyway.

Fixes #43

76652350

by Peter Hutterer at 2023-08-30T09:18:26+10:00

Add a mapping_id to the regions

This allows a caller to match up a region with other data, e.g. in the
remote desktop case the same mapping_id can be assigned to the pipewire
stream that represents that output.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>

e03c047b

by Peter Hutterer at 2023-08-30T09:36:40+10:00

proto: require Python 3.9

ei-scanner relies on some 3.9 features and since that has been out for
almost 3 years now, let's make it a requirement.

Fixes #39

7115e9c4

by Peter Hutterer at 2023-08-30T09:46:36+10:00

test: rework the oeffis dbus tests to be pytest-compatible

DBusMock is unittest based and the documentation points users to that
approach. That approach is limiting however because we can't use all
pytest features (see [1]). Luckily, the parent class in dbusmock doesn't
really do much so we can emulate the functionality ourselves - all we
need to do is call the same setUp/tearDowns and be done with it.

This means we can move the dbus-monitor and mainloop handling into
fixtures too which makes the code a fair bit nicer to read.

[1] https://docs.pytest.org/en/7.1.x/how-to/unittest.html#pytest-features-in-unittest-testcase-subclasses

4f2fd161
by Peter Hutterer at 2023-08-31T12:41:13+10:00
```
test: close the sockets on exit
```

8e95c7d8

by Peter Hutterer at 2023-08-31T12:41:13+10:00

util: add a strv_from_mem() helper function

Generates a strv with the buffer as hex string.

8f911d5e

by Peter Hutterer at 2023-08-31T12:41:13+10:00

test: log the buffers after reading from them

Should help debug some issues with the encoding

f235052f

by Peter Hutterer at 2023-08-31T12:41:13+10:00

test: ensure munit debug messages are visible on failure

abe85e05

by Peter Hutterer at 2023-08-31T13:20:05+10:00

meson: build libutil with hidden symbols visibility

We were leaking some of the utility functions, let's not do that.

This is technically an ABI break but if you're relying on libei to
export those functions...well, don't.

1bac0c28

by Peter Hutterer at 2023-08-31T13:30:36+10:00

CI: update the commit for the last ABI break

abe85e051 stopped the symbol leaks which will look like an ABI break.

90040096

by Peter Hutterer at 2023-08-31T13:37:13+10:00

test: rename a variable to avoid clashes

Future changes will use slen further below because it's the best name,
so let's rename this one since it's less impactful.

397ee6d7

by Peter Hutterer at 2023-08-31T13:37:13+10:00

brei: avoid pointer casts for 32/64 bit values

Copy those onto the buffer with a memcpy rather than via a
(const char *) cast to make sure we work on s390x too.

a223ce86

by Peter Hutterer at 2023-08-31T13:37:13+10:00

util: switch iobuf to uint8_t to avoid sign issues

char is signed or unsigned, so let's avoid this by using uint8_t.

f9c6ea30

by Peter Hutterer at 2023-08-31T13:37:13+10:00

brei: copy protocol strings out of the protocol buffers

With our strings being a 4-byte header followed by the string itself,
they're virtually guaranteed not to be 8-byte aligned. This causes an
issue on some architectures so we need to copy the string out before we
access it.

Since strings are the only protocol type with that extra buffer, let's
hack this in with the minimum effort approach - a null-terminated
char * pointer array that's filled with the strings as they appear in
on the wire. The brei_arg->s points to one of those strings as needed.

This means we can drop the brei_string struct, thanks to pointer alignment
issues this struct doesn't work on s390x, so let's drop it, it no longer
serves any meaningful value.

Fixes #41

931632ef

by Peter Hutterer at 2023-08-31T13:37:13+10:00

test: use munit_assert_uint64 for 64-bit values

Otherwise we only take the first 4 bytes which can fail on BE if we're
comparing an array directly (like we do for the string test).

16c7b157

by Peter Hutterer at 2023-08-31T13:42:25+10:00

test: fix 64-bit pointer access for the object ids

memcpy this out so we can compare them.

8713ef2d
by Peter Hutterer at 2023-08-31T14:06:49+10:00
```
libei 1.1.0rc1
```

ac16ba77

by Peter Hutterer at 2023-09-01T12:06:38+10:00

tools: start ei-debug-events for the fd we get from the oeffis demo tool

Once we get the fd, fork off ei-debug-events for that fd so we can debug
what the EIS implementation actually sends us.

c179b3da
by Peter Hutterer at 2023-09-01T12:18:20+10:00
```
tools: fix an indentation issue
```

d9d46305

by Peter Hutterer at 2023-09-01T12:18:20+10:00

tools/debug-events: print the region mapping id

This changes the region print format from a string to a dict

dbd0e971
by Jeremy Bícha at 2023-09-01T11:56:00-04:00
```
New upstream version 1.0.901
```

0820e29b

by Jason Gerecke at 2023-09-06T21:16:08+00:00

Correct documentation for ei_touch_(get|set)_user_data

82cdbc91
by Peter Hutterer at 2023-09-07T15:07:54+10:00
```
libei 1.1.0
```

4929aeae

by Alban Browaeys at 2023-09-14T02:01:04+02:00

Fix duplicate negation in log message

The test is meant for "not emulating", not "not not emulating".

273b5a4e
by Jeremy Bícha at 2023-09-23T12:44:47-04:00
```
New upstream version 1.1.0
```

381c6bea

by Peter Hutterer at 2023-10-05T23:00:20+00:00

ei: improve a debug message, the number we print is the serial

8f6c3558
by Peter Hutterer at 2023-10-11T00:41:13+00:00
```
util: fix clobbered errno in sink_add_source
```

95d1107c

by Peter Hutterer at 2023-10-11T00:41:13+00:00

util: allow sources to have write notifications

In case our outbound buffers are full we need we need to be able to have a
notification when we're able to write again.

There are two approaches to do this: one is to duplicate the source
(and dup() the fd) and use that for write notifications only. The other
approach is to toggle the source's EPOLLOUT flag on demand, thus
sharing the source dispatch. I picked the latter for simplicity.

69e973e6

by Peter Hutterer at 2023-10-11T00:41:13+00:00

ei: queue unsent messages for later delivery if our buffer is full

If our write buffer is full, enqueue the events instead and try to flush
them out later when we can write again.

Fixes #46

cefef7b1

by Peter Hutterer at 2023-10-11T00:41:13+00:00

Add a static assertion that EAGAIN == EWOULDBLOCK

Not sure what platforms this isn't the case on and I'm even less sure we
need to care so let's fail the build where this is the case and work it
out if someone complains.

fd2999ff

by Peter Hutterer at 2023-10-11T00:49:25+00:00

tools: add a --interval option to the demo server and client

Makes testing a few things easier.

d4bf8840

by Peter Hutterer at 2023-10-11T00:58:23+00:00

util: block SIGALRM during posix_fallocate

posix_fallocate() may be interrupted (EINTR), e.g. by SIGALRM in the X
server's smart scheduler.  On slow systems and for large allocations
this means we cannot ever succeed so let's block SIGALRM like everyone
else does.

See e.g.
https://gitlab.freedesktop.org/xorg/xserver/-/commit/4cfee398726adf89db4b632ade7d6cab8b78282e
and the same code is in libwayland, libdecor, etc.

b8884770

by Peter Hutterer at 2023-10-11T12:25:53+10:00

CI: add a separate job for the documentation build

The pages job is only run once we push to main, so any issues with the
documentation build won't be picked up until it's too late. Split those
up and build the doc in a separate job, the pages job simply takes the
artifacts from that job and publishes them.

43b383eb

by Peter Hutterer at 2023-10-11T12:42:47+10:00

doc/protocol: fix some hugo warnings

WARN 2023/10/11 12:41:19 "_index.md": WARNING you must call the ref /
relref shortcode with '% %' instead of '< >' to work correctly for the
anchor target attribute

59097177

by Peter Hutterer at 2023-10-11T13:09:31+10:00

doc: use a specific revision of the hugo template

Upstream commit 1250bf30a8c2e3c1bfac285489bc63f2c395b641 somehow breaks
things. Doesn't affect hugo v0.111 (F39) but it does break on F38 on
v0.101.

Let's not waste more time debugging this, hardcode the template and in a
few weeks time we can switch to F39 as the build host.

3f768ecb
by Peter Hutterer at 2023-10-11T13:31:47+10:00
```
CI: add missing dependencies line
```

38244b8c

by Peter Hutterer at 2023-10-18T09:44:01+10:00

ei: if sending a message fails, don't send any more

Any error will cause ei_disconnect() which in turn unwinds our context.
Each object removed will cause an ei_send_message() to notify the server
but that will only cause error messages - our socket is already gone.

Fix this by removing the source and checking if the fd is still valid
before we write to it.

Fixes #47

5dcdc696

by Peter Hutterer at 2023-10-19T14:18:57+10:00

ei: remove the source if we get disconnected

If we get disconnected there's no point sending anything anymore, the
server may have already closed the connection.

Otherwise this produces a confusing warning  in the log:

 18:37:31 | INFO | Disconnected by EIS
      ... | WARN | failed to send message: Broken pipe

See https://gitlab.freedesktop.org/libinput/libei/-/merge_requests/262#note_2132619

2df570b2

by Peter Hutterer at 2023-10-24T18:22:48+10:00

tools: add an --iterations argument to ei-demo-client

Limits the number of poll() iterations before triggering a shutdown.
This allows us to use this tool for shutdown testing too - previously
it would just get killed and never performed a proper unwind.

cdbd1d9e

by Peter Hutterer at 2023-10-24T18:25:25+10:00

util: replace most of SIGALRM blocks with EINTR checks

SIGALRM *should* not really be an issue for us, most of our IO shouldn't
take long (posix_fallocate is the exception) so we can just retry it
instead of blocking someone's use of SIGALRM.

Doing it here mostly means that the rest of libei doesn't have to care
about EINTR, it'll just be handled transparently (just like before with
the SIGALRM blocks).

a3c80109

by Peter Hutterer at 2023-10-24T18:25:25+10:00

util: don't check for EINTR on close()

See https://gitlab.freedesktop.org/libinput/libei/-/merge_requests/261#note_2131802

19c949a4

by Peter Hutterer at 2023-10-24T18:25:25+10:00

util: retry posix_fallocate() if it fails

We block SIGALRM since d4bf8840a48854a822951707764b452bacf3dd56 so any
remaining EINTR is something we should probably retry with.

6758b997

by Peter Hutterer at 2023-10-24T18:25:25+10:00

util: add helper macros for the ELF section handling

f558eeda

by Peter Hutterer at 2023-10-24T18:25:25+10:00

util: add a global setup option to our munit wrapper

This allows us to handle/manipulate argv and/or pass a userdata to the
tests. Same ELF section approach as we already have with the MUNIT_TEST
macro but this time it's for a global setup.

Note that for the generated __start and __stop section variables we have
to have at least one entry in that section which we do with one
hardcoded (and ignored) one.

05d79f6e

by Peter Hutterer at 2023-10-24T18:25:25+10:00

test: set up a timer to trigger SIGALRM

Xwayland uses a timer for the scheduler which means any of our syscalls
can trigger EINTR. Let's make sure we may catch bugs related to that by
setting up our test suite to hammer us with timers.

Can't guarantee this will trigger all bugs but over time it may help or
at least ensure that the low-hanging fruit are all fixed.

57e7cd79

by Peter Hutterer at 2023-11-08T16:04:33+10:00

doc: drop the git depth from the hugo theme

We need a specific SHA so let's not risk that one being outside that
depth that we clone to.

f9c83ed9

by Peter Hutterer at 2023-11-08T23:24:45+00:00

test: expand log messages to accommodate for 4-digit line numbers

35832d9f

by Peter Hutterer at 2023-11-08T23:24:45+00:00

test: make peck_dispatch_until_stable() print more useful debug info

44e7f2cb
by Peter Hutterer at 2023-11-08T23:24:45+00:00
```
test: fix a test missing frame events
```

d31b5a1c

by Peter Hutterer at 2023-11-08T23:24:45+00:00

eis: only queue a seat bind event if the caps change

This triggered an internal bug message in test_ei_seat_bind_unbind_immediately.

If a client unbinds all capabilities this triggers an EIS_SEAT_BIND with
zero caps. peck would call eis_seat_remove() which calls eis_seat_drop().
That queued another EIS_EVENT_SEAT_BIND with zero capabilities, leading
peck to call eis_seat_remove() again on an already removed seat.

34c83370
by Peter Hutterer at 2023-11-08T23:24:45+00:00
```
eis: fix a typo
```

fae41aac

by Peter Hutterer at 2023-11-08T23:24:45+00:00

test: hack a peck_debug() function

To make it easier to print something debuggy from a test.

This is the MVP, it always requires an argument after the format string
but it'll do for now.

7999483a

by Peter Hutterer at 2023-11-09T10:15:38+10:00

test: make bug logs fatal by default

We don't want to paper over bugs in the implementation, so let's make
any ei/eis error message with a bug in it fatal by default.

This needs to be disabled where we test for known-buggy client/EIS
behavior.

4e634aa7

by Peter Hutterer at 2023-11-09T12:47:06+10:00

test: add the ability to add offsets to ei_now()

Build a separate libei-eierpecken.so that is identical to libei.so but
allows adding an offset to ei_now() for the eierpecken tests. That
offset is added to the return value of ei_now(), removing the real-time
dependency of the tests.

In other words, we can call peck_ei_add_time_offset(peck, s2us(5)) to
add 5 seconds to the time and continue the test as if that time has
elapsed.

924341e1

by Peter Hutterer at 2023-11-09T12:47:06+10:00

ei: keep a cache of defunct objects to avoid spamming the log

Because of the asynchronous protocol, we may get this interaction

   server sends A->destroyed()
       client sends A->foo()
       client receives A->destroyed()
   server receives A->foo()
   server sends invalid_object_id(A)
       client receives invalid_object_id(A)

Previously we dropped the object after destroyed() and were thus
guaranteed to warn about the invalid object id for that same object
later.

Fix this by keeping a cache of defunct object IDs that we know about and
ignoring errors for recently dropped objects.

Every 20 ei_dispatch() calls we drop any defunct objects that were
unregistered more than 5 seconds ago.

Fixes #49

38132d6f

by Peter Hutterer at 2023-11-09T12:55:56+10:00

ei: improve debug messages for keymap failures

Let's be more informative about what failed.

Fixes #48

9697aa9c

by Peter Hutterer at 2023-12-05T14:25:10+10:00

pre-commit: autoupdate and add two more useful checkers

We don't have symlinks but let's check for them anyway. And check for
merge conflict leftoers too.

49363168
by Peter Hutterer at 2023-12-06T08:19:00+10:00
```
libei 1.2.0
```
06a8819d
by Jeremy Bícha at 2023-12-05T18:53:33-05:00
```
New upstream version 1.2.0
```

55e5deed

by Peter Hutterer at 2023-12-13T17:55:16+10:00

ci: bump to use F39

And drop the job to test the minimum meson version in the process. F39
ships with python 3.12 which causes this error with meson 0.56.0

proto/meson.build:17:0: ERROR: <ExternalProgram 'python3' -> ['/usr/bin/python3']> is not a valid python or it is missing setuptools

We could put the effort into setting up the CI to fix this, but we
won't.

97d66604

by Peter Hutterer at 2023-12-14T10:41:19+10:00

CI: drop the git depth for the ABI check job

We need to make sure that we can checkout whichever sha was the last ABI
break.

eaed3883

by Peter Hutterer at 2023-12-14T10:44:54+10:00

CI: fix a hardcoded fedora and replace it with the jinja var

870123f5

by Peter Hutterer at 2023-12-14T10:44:54+10:00

CI: move the minium meson version job to debian

See 55e5deedd8d2681e1c87a14f2d92008b89888f47 where this job was removed.

6ce695db

by Peter Hutterer at 2023-12-14T00:58:18+00:00

test: yield the proxy, not the subprocess for the liboeffis fixtures

There's virtually no use for the process, so let's yield the proxy
object instead because we can make use of that.

d2977865

by Peter Hutterer at 2023-12-14T00:58:18+00:00

oeffis: OEFFIS_DEVICE_ALL_DEVICES should translate to "all"

As the portal documentation [1] says:
  Bitmask of what device types to request remote controlling of. Default is all.

The default is only triggered if we do not submit the types at all, the
current behavior of sending a value of 0 means "none". Fix this by
skipping the "types" key if we try to select for all devices.

[1] https://flatpak.github.io/xdg-desktop-portal/docs/doc-org.freedesktop.impl.portal.RemoteDesktop.html

08f1d410
by Peter Hutterer at 2024-02-05T13:33:04+10:00
```
libei 1.2.1
```

93efd3d1

by David Redondo at 2024-03-05T22:47:05+00:00

eis-demo-server: Send a sensible discrete scroll value

So it can be a good example for people and doesn't log about
server bugs.

54ca521d

by Peter Hutterer at 2024-03-12T04:25:33+00:00

proto: highlight the region 'hight' typo in the documentation

Changing this would be an API-breaking change (depending on how
bindings are generated) so we'll have to live with this typo for the
foreseeable future.

Closes #53

a91dac8d
by Jeremy Bícha at 2024-04-01T14:18:51-04:00
```
New upstream version 1.2.1
```

33b4a619

by Matt Turner at 2024-04-08T12:21:09-04:00

test: Raise SIGALRM interval to 50µs

On some platforms, an interval of 5µs is short enough that the test
spends its time almost exclusively processing SIGALRMs and never
progresses otherwise. Raising the interval to 50µs allows the test to
pass in a fraction of a second.

Bug: https://bugs.gentoo.org/916777
Closes: https://gitlab.freedesktop.org/libinput/libei/-/issues/50

23f05643

by David Redondo at 2024-04-09T06:39:13+00:00

Fix API docs for EIS_EVENT_CLIENT_CONNECT

The functions with event in their name don't exist.

da1fa204

by Peter Hutterer at 2024-07-24T12:29:48+10:00

meson.build: bump the meson version by one

This allows us to drop one version check

9ee399c8

by Peter Hutterer at 2024-07-24T17:39:04+10:00

CI: Lock hugo to v0.111 to allow for distro updates

F40 now ships v0.121 which again breaks something in the relearn theme
so let's lock our version here to one we know works and move on with
life. I can't be bothered to relearn hugo and themes every few months
just for a single static website.

40598203
by Peter Hutterer at 2024-07-24T20:19:44+10:00
```
CI: bump to Fedora 40
```

267716a7

by Peter Hutterer at 2024-07-25T11:11:02+10:00

proto: mark the explanation in disconnect as nullable

This is the only string in the current protocol that is nullable and our
DTD allows for that so it's not even an API break. Yay.

Closes #54

87705ef9

by Peter Hutterer at 2024-07-25T11:11:07+10:00

scanner: handle the allow-null attribute for arguments

Currently unused but it's exposed, so yay...

0f811145

by Peter Hutterer at 2024-07-25T01:17:39+00:00

Fix region check for devices with multiple regions

Use the ei(s)_device_in_region() to check for any region.

Closes #56

a424458f

by Peter Hutterer at 2024-07-25T01:17:39+00:00

Devices without regions pass coordinates as-is

If a device doesn't have a region (e.g. physical device) the answer to
"is this in the region" is always yes.

Closes #56

eb1d5a7e

by Peter Hutterer at 2024-07-29T14:32:40+10:00

test: drop a custom cleanup func in favor of _unref_

c4ac0841

by Peter Hutterer at 2024-07-29T14:48:34+10:00

util/memmap: add a helper wrapper around mmap

Makes this easier to use with our unref functions.

dc153c50
by Peter Hutterer at 2024-07-29T14:48:34+10:00
```
util/memfile: include stddef for size_t
```

6ea468c8

by Peter Hutterer at 2024-07-29T14:48:34+10:00

util/memfile: use MAP_SHARED to create the map

Using MAP_PRIVATE means copy-on-write so our data written into the map
immediately disappears again. This leads to a empty string when sending
a keymap to a client.

a924888f

by Peter Hutterer at 2024-07-29T14:53:35+10:00

ei-demo-client: use mmap to read the keymap

Closes #55

3e2e43e3

by Peter Hutterer at 2024-07-29T14:54:31+10:00

ei-demo-client: use xkb_keymap_new_from_buffer()

This is the more correct approach since we get a sized buffer from the
server and people may use this as reference code in their implementation.
Technically we cannot expect a true zero-terminated string from EIS.

Unfortunately this means we need to work around
libxkbcommon#307 to strip trailing zeroes from the buffer if any exist.

865d7152

by Peter Hutterer at 2024-07-29T05:15:53+00:00

util: replace a strncpy with a memcpy

We know exactly the lengths of everything involved here so let's use
memcpy. This way we don't need the stringop-truncation warning (a pragma
clang doesn't support anyway).

e8a84435

by David Redondo at 2024-07-31T07:05:48+00:00

oeffis: Make ConnectToEIS call async

A blocking call can be  problematic when done from inside Xwayland
to the compositor as the compositor could be doing a blocking X call
at the same time. In this instance we found it's likely to happen
because this call will happen shortly after the user accepted the
This is not a problem for the other calls because the portal API
is async for these via request response signalling.

b67cdba8
by Peter Hutterer at 2024-08-01T11:56:23+10:00
```
CI: avoid duplicate pipelines
```

5331dc0b