fdo-mirrors/libei
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/libinput/libei.git synced 2026-02-15 10:20:33 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions
146 commits 2 branches 17 tags 6.1 MiB
Commit graph

146 commits

This branch
This branch
All branches
Author SHA1 Message Date
Peter Hutterer
d4e274ee6f libeis: flatten the event hierarchy 
Same as already done for libei. There's relatively little benefit here since
we won't have a lot of different events and any caller will do the switch
based on the event type anyway. So let's just export a single event type and
have everything contained in that.

Since this required rewriting all getters, let's move the lot to a new file
and streamline things a bit that way.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-18 16:05:52 +10:00
Peter Hutterer
4565fb6851 libei: add the hooks for pointer/touch ranges 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-18 14:42:23 +10:00
Peter Hutterer
4b0976c36c test: add helpers for the next event 
Makes debugging easier since we can get the string event type that way.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-18 14:42:23 +10:00
Peter Hutterer
81d37cfde5 test: fix behavior handling for accepting/rejecting clients 
Either of those should not modify the behavior for accepting or rejecting
devices.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-18 14:42:23 +10:00
Peter Hutterer
d65c678239 test: add a helper to wait until a stable state 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-18 14:42:23 +10:00
Peter Hutterer
a5ea7a6dd0 test: better event type logging 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-18 14:42:23 +10:00
Peter Hutterer
c39dc360d4 test: terminate the test list with NULL 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-18 14:42:23 +10:00
Peter Hutterer
14d6c87974 libei: use a #define for cleaning up a device 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-18 14:42:23 +10:00
Peter Hutterer
fd9035dfe6 libei: some more API documentation 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-18 14:42:23 +10:00
Peter Hutterer
537fdfde85 Drop ei_next_event_type() and eis_next_event_type() 
Replaced by ei_peek_event() and eis_next_event_type()

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-18 14:42:23 +10:00
Peter Hutterer
b052d30978 test: fix an unused variable complaint 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:34:36 +10:00
Peter Hutterer
7965e8f4df libei: plug a memleak if we get more than one message 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:17:51 +10:00
Peter Hutterer
515b67c6da tools: use a cleanup function for unlinking the socket 
This way we can't forget to remove the file.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:17:51 +10:00
Peter Hutterer
39b5ba74a4 Switch the various coordinates to doubles 
Original idea of 1/1000 of a pixel was to allow subpixels while sending fixed
width down the wire. Let's not care about that and use doubles instead.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:17:51 +10:00
Peter Hutterer
7d58a76254 libei: rework the wire message parsing for slightly better readability 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:17:51 +10:00
Peter Hutterer
09f6ef8d7a libei: use a macro to define the tests 
Having to define a separate name and test function for each test is annoying.
The unit-tests can leak the tests names anyway, so let's shortcut this with a
macro.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:17:51 +10:00
Peter Hutterer
f35b591b9c libei: ei_get_user_data() is a public API 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:17:51 +10:00
Peter Hutterer
b309ebcc66 libei: define a cleanup helper for the internal message 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:17:51 +10:00
Peter Hutterer
f149a93743 libei: unify the wire message logging in one place 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:17:51 +10:00
Peter Hutterer
81cb24124d libei: drop the nested event types 
There's relatively little benefit here since we won't have a lot of different
events and any caller will do the switch based on the event type anyway.
So let's just export a single event type and have everything contained in
that.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:17:51 +10:00
Peter Hutterer
b547fdc529 Improve naming for the DeviceAdded message 
DeviceAdded is more self-explanatory than just "Added"

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:17:51 +10:00
Peter Hutterer
325622ed7c libeis: unify the wire message logging in one place 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:17:51 +10:00
Peter Hutterer
6c823b1308 libeis: keep a ref of the client during dispatch 
On error, we call eis_client_disconnect() which unrefs the client. Since we're
using it afterwards for debugging logs, we need to keep at least one ref
going.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:17:51 +10:00
Peter Hutterer
fb7bbbd4ce libeis: plug a memory leak if we get more than one message per dispatch 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:17:51 +10:00
Peter Hutterer
2e944450a0 tools: change the default socket to $XDG_RUNTIME_DIR/eis-0 
This matches the future portal implementation

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:17:51 +10:00
Peter Hutterer
ce91763683 util: add a errno wrapper 
Converts into a negative errno or the respective value

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-13 16:15:09 +10:00
Peter Hutterer
8cdbcda7a7 test: add EIS integration tests 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-12 14:42:09 +10:00
Peter Hutterer
dd426757fb test: fix the suite name for the EI integration tests 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-12 14:36:10 +10:00
Peter Hutterer
efa01ffc38 test: move the with_server/with_client helper macros to eierpecken.h 
We want to re-use those.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-12 14:35:41 +10:00
Peter Hutterer
d38f9c82ce meson.build: rename the test binary eitest 
Faster to type, no other reason.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-12 14:35:19 +10:00
Peter Hutterer
c0ced9f234 meson.build: eierpecken helper lib must link against ei and eis 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-12 14:34:46 +10:00
Peter Hutterer
d7bc93cb4f libei: send the name in the original connect message 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-12 14:17:36 +10:00
Peter Hutterer
bfee550834 More log messages for debugging 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-12 14:16:09 +10:00
Peter Hutterer
017260798b README: update the blurb about batched processing 
Short story, we can't do it but it's simple enough to work around in the
caller so let them do it.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-12 14:16:09 +10:00
Peter Hutterer
78b0b99883 README: add a blurb about libreis 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-12 14:16:09 +10:00
Peter Hutterer
7d75b3eb8c test: use drop caps, not reject caps as terminology 
And add a test that tests for dropped caps

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-11 13:01:36 +10:00
Peter Hutterer
7ae661a029 test: add a test for a device dropped to zero caps and refused 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-11 13:01:36 +10:00
Peter Hutterer
a95d2bdd8b test: fix the accept/reject behaviour for peck eis contexts 
Rejecting means disabling the accept and vice versa

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-11 13:01:36 +10:00
Peter Hutterer
21d6b6a5ca test: fix the default behaviour of the peck EI context 
This is a flag, not a value.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-11 13:01:36 +10:00
Peter Hutterer
3db476930b Add a bunch of log_debug messages 
We can remove those when we have a working implementation, for now it's too
painful to debug when an exchange doesn't  work for some reason.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-11 13:01:36 +10:00
Peter Hutterer
57a9892f5d eis: send the removed deviceid down the wire 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-11 11:13:14 +10:00
Peter Hutterer
2a98906b32 util: add name lookup to tristate handling 
And make it more obvious that _val isn't something you should access directly.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-11 11:06:34 +10:00
Peter Hutterer
b717837b14 Downgrade the ECANCELED warning to a mere info 
ECANCELED is used for any time the other end disconnects, so it's not really
an error.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-11 11:06:07 +10:00
Peter Hutterer
ce51392241 test: fix the device add test 
Check that we're actually adding the device and getting the message back

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-11 10:51:30 +10:00
Peter Hutterer
55246512f0 Add capabilities to the Added event 
This makes it possible for the server to limit capabilities.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-10 19:32:34 +10:00
Peter Hutterer
fb707a3395 Add a prefix to the logger 
Makes it easier to distinguish between ei and eis when everything in the tests
is just printed to stderr anyway.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-10 13:50:13 +10:00
Peter Hutterer
6f75463261 Rename "accepted" to "added" for consistency with the public API 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-10 08:49:22 +10:00
Peter Hutterer
2631854dbb libreis: update the API docs a bit 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-10 08:42:25 +10:00
Peter Hutterer
12a04fa657 libei: update the API docs a bit 
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-10 08:40:53 +10:00
Peter Hutterer
fe22b9b5a5 proto: specify that the protocol isn't batch-compatible after all 
The protocol itself could work like originally described here but there's a
stumbling block: the decision on whether to accept a device is made by the
caller through EIS_EVENT_DEVICE_ADDED and the following eis_device_connect()
call. We cannot process any events from that device until that call is
complete and that effectively disallows batch submission of requests.

To allow batching we'd have to pause the protocol but that means missing out
on other devices (and their events) and disconnect events. The alternative to
that would be for libeis to peek at incoming requests and sort them by device
ID so we only pause one device's stream but now we're also mangling the device
event order and potentially triggering all sorts of side-effects.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
 2020-08-10 08:04:05 +10:00