platform/upstream/libxkbcommon.git
3 years agoatom: rewrite as a hash table
Ran Benita [Sat, 2 Nov 2019 14:19:32 +0000 (16:19 +0200)]
atom: rewrite as a hash table

While the previous 1987-style[0] scheme was fun (and I reasonably
optimized it for a fair comparison), this task is more suited to a hash
table. Even a simple implementation beats the old one.

[0] Seems to have first appeared in X11R1, released September 1987.
    See server/dix/atom.c here: https://www.x.org/releases/X11R1/X.V11R1.tar.gz

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agobench: add atom benchmark
Ran Benita [Thu, 8 Apr 2021 07:51:07 +0000 (10:51 +0300)]
bench: add atom benchmark

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agobench/rulescomp: decrease benchmark iterations
Ran Benita [Tue, 27 Apr 2021 09:31:37 +0000 (12:31 +0300)]
bench/rulescomp: decrease benchmark iterations

No need for more, and easier to convert to milliseconds per iteration.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agobuild: fix missing includes
Ran Benita [Tue, 27 Apr 2021 07:53:36 +0000 (10:53 +0300)]
build: fix missing includes

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agoMove include files to include/ subdirectory
Ran Benita [Tue, 27 Apr 2021 07:10:26 +0000 (10:10 +0300)]
Move include files to include/ subdirectory

This way we don't specify `include_directorories('.')` which brings in
more than needed.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agobuild: move the subproject variables to a common section at the end
Ran Benita [Tue, 27 Apr 2021 07:02:00 +0000 (10:02 +0300)]
build: move the subproject variables to a common section at the end

As suggested in:
https://github.com/xkbcommon/libxkbcommon/pull/240#discussion_r620784021

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agoMeson: Allow building as subproject
Adrian Perez de Castro [Mon, 26 Apr 2021 14:27:01 +0000 (17:27 +0300)]
Meson: Allow building as subproject

Specify where to find the headers for libxkbcommon_dep,
libxkbcommon_x11_dep, and libxkbregistry_dep,  which allows other
projects to correctly locate the headers when libxkbcommon is being
built as a Meson subproject.

The dep_libxkbregistry variable is renamed to libxkbregistry_dep,
to follow the usual convention for variables which hold declared
dependencies to be used from subproject builds.

3 years agoWindows: Pass list of symbols to export to MSVC
Adrian Perez de Castro [Mon, 26 Apr 2021 14:38:48 +0000 (17:38 +0300)]
Windows: Pass list of symbols to export to MSVC

Arrange for passing .def files with the lists of symbols to export from
DLLs when building on Windows with MSVC. Without this no symbols were
being exported at all.

The .def files are generated from the .map files at build time using
scripts/map-to-def, which avoids needing to maintain two different sets
of files.

3 years agotools: change xkbcli list to output YAML
Peter Hutterer [Thu, 22 Apr 2021 02:52:55 +0000 (12:52 +1000)]
tools: change xkbcli list to output YAML

We have a lot of keyboard layouts and the current output format is virtually
useless at searching for a specific one to debug any issues with either the
layout list or the output from libxkbregistry.

Let's use YAML instead because that can easily be post-processed to extract the
specific layouts wanted, e.g. to get the list of all layouts:
  xkbcli-list | yq -r ".layouts[].layout"
to get the list of all variants of the "us" layout:
  xkbcli-list | yq -r '.layouts[] | select(.layout == "us") | .variant
and the number of option groups:
  xkbcli-list | yq -r '.option_groups[] | length'

Note that the top-level nodes have been de-capitalized, so where it was "Models"
before it is now "models" and the "Options" node is now "option_groups".

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agotools: change the list separator handling
Peter Hutterer [Thu, 22 Apr 2021 02:42:00 +0000 (12:42 +1000)]
tools: change the list separator handling

Slightly easier to read than the "bool first" approach.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agokeysym: fix underflow in binary searches
Isaac Freund [Wed, 21 Apr 2021 22:37:03 +0000 (00:37 +0200)]
keysym: fix underflow in binary searches

This is hit when passing an empty string and XKB_KEYSYM_CASE_INSENSITIVE
to xkb_keysym_from_name currently if `(lo + hi) / 2` is 0 and `cmp < 0`,
causing mid to underflow and the the array access into name_to_keysym on
the next iteration of the loop to be out of bounds .

We *would* use ssize_t here as it is the appropriate type, but windows
unfortunately does not define it.

3 years agotest: allow skipping errors in certain conditions
Peter Hutterer [Tue, 20 Apr 2021 23:02:40 +0000 (09:02 +1000)]
test: allow skipping errors in certain conditions

Add a skipError test function that can analyse rc, stdout, stderr to skip a test
even when we expect an error. We use that to skip if we couldn't find a keyboard
in the interactive-evdev test.

Fixes #235

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agotest: add extra info to the layout-tester --help output
Peter Hutterer [Tue, 20 Apr 2021 02:21:28 +0000 (12:21 +1000)]
test: add extra info to the layout-tester --help output

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agotest: print the layout-tester progress bar to stdout by default
Peter Hutterer [Thu, 15 Apr 2021 23:24:20 +0000 (09:24 +1000)]
test: print the layout-tester progress bar to stdout by default

tqdm prints to stderr by default but we're using that for failed keymap
compiles (which are the ones that really matter). Plus, whether we're using tqdm
is dependent on isatty(sys.stdout) anyway.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agotest: add an LVO argument to the XKB layout tester
Peter Hutterer [Thu, 15 Apr 2021 21:42:27 +0000 (07:42 +1000)]
test: add an LVO argument to the XKB layout tester

Slightly easier to debug if we can have it only parse one single layout.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agotest: print the compiled keymaps to a given directory
Peter Hutterer [Thu, 15 Apr 2021 01:07:33 +0000 (11:07 +1000)]
test: print the compiled keymaps to a given directory

With --keymap-output-dir, the given directory will contain a list of files named
after the layout + variant ('us', 'us(euro)', ...) that contain the keymaps for
each variant + option combination compiled.

This is still a lot, but better to sift through hundreds of keymaps than tens of
thousands.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agotest: rework the output for the xkeyboard-config layout tester
Peter Hutterer [Thu, 15 Apr 2021 00:39:05 +0000 (10:39 +1000)]
test: rework the output for the xkeyboard-config layout tester

The previous output is largely unusable. The result in the CI test runs is a 6GB
file with every compiled keymap in it and while we can grep for ERROR, it's not
particularly useful.

Let's change this and print out YAML instead - that can be machine-processed.
This patch adds a new parent class that prints itself in YAML format,
the tool invocations are child classes of that class. The result looks like this:

Example output:
- rmlvo: ["evdev", "pc105", "us", "haw", "grp:rwin_switch"]
  cmd: "xkbcli-compile-keymap --verbose --rules evdev --model pc105 --layout us --variant haw --options grp:rwin_switch"
  status: 0
- rmlvo: ["evdev", "pc105", "us", "foo", ""]
  cmd: "xkbcli-compile-keymap --verbose --rules evdev --model pc105 --layout us --variant foo"
  status: 1
  error: "failed to compile keymap"

Special status codes are: 99 for "unrecognized keysym" and 90 for "Cannot open
display" in the setxkbmap case.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agotest: add proper --verbose handling to the xkeyboard-config tester
Peter Hutterer [Wed, 14 Apr 2021 22:57:51 +0000 (08:57 +1000)]
test: add proper --verbose handling to the xkeyboard-config tester

Instead of defaulting to verbose on/off depending on isatty, make it an
argument instead.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agointeractive-wayland: fallback to ftruncate() if needed
Evgeniy Khramtsov [Wed, 14 Apr 2021 09:56:07 +0000 (12:56 +0300)]
interactive-wayland: fallback to ftruncate() if needed

Fallback to ftruncate() if the underlying filesystem does not
support posix_fallocate().

Idea by: Jan Beich <jbeich@FreeBSD.org>, Niclas Zeising <zeising@FreeBSD.org>
Inspired by: Wayland cursor/os-compatibility.c
[ran: small adjustments]
Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agodoc: fix user-configuration sample file
Jouke Witteveen [Sat, 10 Apr 2021 21:28:06 +0000 (23:28 +0200)]
doc: fix user-configuration sample file

Support copy-pasting from the docs to get something functional.

Signed-off-by: Jouke Witteveen <j.witteveen@gmail.com>
3 years agoscripts: update license note in perfect_hash.py
Ran Benita [Thu, 8 Apr 2021 07:13:27 +0000 (10:13 +0300)]
scripts: update license note in perfect_hash.py

Ref: https://github.com/ilanschnell/perfect-hash/issues/5

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agoBump version to 1.2.1 xkbcommon-1.2.1
Ran Benita [Wed, 7 Apr 2021 16:23:55 +0000 (19:23 +0300)]
Bump version to 1.2.1

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agox11: fix xkb_x11_keymap_new_from_device failing when a level name is empty
Ran Benita [Wed, 7 Apr 2021 07:47:15 +0000 (10:47 +0300)]
x11: fix xkb_x11_keymap_new_from_device failing when a level name is empty

The numpad:mac option doesn't specify a name for the first level:

    // On Mac keypads, level 1 and 2 are swapped.
    partial xkb_types "mac" {
        type "KEYPAD" {
            modifiers = None;
            map[None] = Level2;
            level_name[Level2] = "Number";
        };
        include "extra(keypad)"
    };

This means the atom for level name is XCB_ATOM_NONE. We tried to get its
name, which fails. This regressed in 40c00b472144d1684d2fb97cafef39.

Instead, translate it to XKB_ATOM_NONE, same as the previous behavior.

Fixes: https://github.com/xkbcommon/libxkbcommon/issues/229
Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agoBump version to 1.2.0 xkbcommon-1.2.0
Ran Benita [Fri, 2 Apr 2021 19:12:00 +0000 (22:12 +0300)]
Bump version to 1.2.0

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agocompose: fix max compose nodes check
Ran Benita [Fri, 2 Apr 2021 12:00:46 +0000 (15:00 +0300)]
compose: fix max compose nodes check

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agokeysym: speed up the perfect hash function
Ran Benita [Thu, 1 Apr 2021 19:46:56 +0000 (22:46 +0300)]
keysym: speed up the perfect hash function

Make it use a bit operation instead of an expensive modulo.

perf diff:

    Baseline  Delta Abs  Shared Object      Symbol
    ........  .........  .................  ...................................

      28.15%     -6.57%  bench-compose      [.] xkb_keysym_from_name

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agokeysym: avoid strtoul in xkb_keysym_from_name
Ran Benita [Thu, 1 Apr 2021 19:07:28 +0000 (22:07 +0300)]
keysym: avoid strtoul in xkb_keysym_from_name

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agokeysym: use a perfect hash function for case sensitive xkb_keysym_from_name
Ran Benita [Sun, 28 Mar 2021 17:22:54 +0000 (20:22 +0300)]
keysym: use a perfect hash function for case sensitive xkb_keysym_from_name

In 7d84809fdccbb5898d0838849ec7c321410182d5 I added a fast path for the
case-sensitive case, but it is still slowing down Compose parsing.

Instead of the binary search, use a perfect hash function, computed with
a simple python module I found (vendored).

It is faster -- perf diff is:

   Baseline  Delta Abs  Shared Object      Symbol
   ........  .........  .................  ...................................

     22.35%    -14.04%  libc-2.33.so       [.] __strcmp_avx2
     16.75%    +10.28%  bench-compose      [.] xkb_keysym_from_name
     20.72%     +2.40%  bench-compose      [.] parse.constprop.0
      2.29%     -1.97%  bench-compose      [.] strcmp@plt
      2.56%     +1.81%  bench-compose      [.] resolve_name
      2.37%     +0.92%  libc-2.33.so       [.] __GI_____strtoull_l_internal
     26.19%     -0.63%  bench-compose      [.] lex
      1.45%     +0.56%  libc-2.33.so       [.] __memchr_avx2
      1.13%     -0.31%  libc-2.33.so       [.] __strcpy_avx2

Also reduces the binary size:

Before:

      text    data     bss     dec     hex filename
    341111    5064       8  346183   54847 build/libxkbcommon.so.0.0.0

After:

      text    data     bss     dec     hex filename
    330215    5064       8  335287   51db7 build/libxkbcommon.so.0.0.0

Note however that it's still larger than before 7d84809fdccbb5898d08388:

      text    data     bss     dec     hex filename
    320617    5168       8  325793   4f8a1 build/libxkbcommon.so.0.0.0

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agocompose: use a ternary tree instead of a regular trie
Ran Benita [Mon, 29 Mar 2021 13:05:14 +0000 (16:05 +0300)]
compose: use a ternary tree instead of a regular trie

Previously we used a simple trie with a linked list for each chain.
Unfortunately most compose files have very long chains which means the
constructions performs an almost quadratic number of comparisons.

Switch to using a ternary search tree instead. This is very similar to a
trie, only the linked list is essentially replaced with a binary tree.

On the en_US/Compose file, the perf diff is the following (the modified
function is `parse`):

  Event 'cycles:u'

  Baseline  Delta Abs  Shared Object     Symbol
  ........  .........  ................  .................................

    39.91%    -17.62%  bench-compose     [.] parse.constprop.0
    20.54%     +6.47%  bench-compose     [.] lex
    17.28%     +5.55%  libc-2.33.so      [.] __strcmp_avx2
    12.78%     +4.01%  bench-compose     [.] xkb_keysym_from_name
     2.30%     +0.83%  libc-2.33.so      [.] __GI_____strtoull_l_internal
     3.36%     +0.78%  bench-compose     [.] strcmp@plt

Thanks to some careful packing, the memory usage is pretty much the
same.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agoast: remove comment re. anonymous struct
Ran Benita [Tue, 30 Mar 2021 17:34:11 +0000 (20:34 +0300)]
ast: remove comment re. anonymous struct

C11 is not sufficient for this, needs `--ms-extensions` which we don't
want to enable.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agocompose: use anonymous union
Ran Benita [Tue, 30 Mar 2021 17:12:08 +0000 (20:12 +0300)]
compose: use anonymous union

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agobuild: require C11
Ran Benita [Tue, 30 Mar 2021 17:08:42 +0000 (20:08 +0300)]
build: require C11

I'd really like to use anonymous unions/structs. Supposedly even MSVC
supports it now. Let's try and see.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agokeysym: properly handle overflow in 0x keysym names
Ran Benita [Tue, 30 Mar 2021 16:11:59 +0000 (19:11 +0300)]
keysym: properly handle overflow in 0x keysym names

Relatedly, strtoul allows a lot of unwanted stuff (spaces, +/- sign,
thousand seperators), we really ought not use it. But that's for another
time.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agocompose: add a limit of 65535 sequences
Ran Benita [Tue, 30 Mar 2021 14:52:36 +0000 (17:52 +0300)]
compose: add a limit of 65535 sequences

Fits in uint16_t, which enables some future optimizations. But also a
good idea to have some limit. Not aware of any compose files which come
close.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agoutils: assert on streq for NULL pointers
Peter Hutterer [Mon, 29 Mar 2021 22:09:37 +0000 (08:09 +1000)]
utils: assert on streq for NULL pointers

We have streq_null for that purpose

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agotest: move an assert up to before the strlen() use
Peter Hutterer [Mon, 29 Mar 2021 22:08:58 +0000 (08:08 +1000)]
test: move an assert up to before the strlen() use

../../../test/keysym.c:80:24: warning: Null pointer passed to 1st parameter
expecting 'nonnull' [core.NonNullParamChecker]
  (unsigned) strlen(expected));

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agoxkbcomp: remove useless assignment
Peter Hutterer [Mon, 29 Mar 2021 21:56:09 +0000 (07:56 +1000)]
xkbcomp: remove useless assignment

../../../src/xkbcomp/compat.c:693:16: warning: Although the value stored to
'merge' is used in the enclosing expression, the value is never actually read
from 'merge' [deadcode.DeadStores]
    si.merge = merge = (def->merge == MERGE_DEFAULT ? merge : def->merge);

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agoxkbcomp: use memcpy over strncpy to avoid analyzer warnings
Peter Hutterer [Mon, 29 Mar 2021 06:23:28 +0000 (16:23 +1000)]
xkbcomp: use memcpy over strncpy to avoid analyzer warnings

The target buffer is 7 bytes long, null-termination is optional (as the comment
already suggests). Coverity is unhappy about this though so let's use memset and
memcpy instead.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agoman: add missing pointer to xkbcli-compile-keymap(1)
Ran Benita [Mon, 29 Mar 2021 17:59:12 +0000 (20:59 +0300)]
man: add missing pointer to xkbcli-compile-keymap(1)

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agokeysym: fast path for case sensitive xkb_keysym_from_name
Ran Benita [Sun, 28 Mar 2021 12:51:01 +0000 (15:51 +0300)]
keysym: fast path for case sensitive xkb_keysym_from_name

xkb_keysym_from_name() is called a lot in Compose file parsing. The
lower case handling slows things down a lot (particularly given we can't
use the optimized strcasecmp() due to locale issues). So add separate
handling for the non-case-sensitive case which is used by Compose.

To do this we need to add another version of the ks_tables table. This
adds ~20kb to the shared library binary. We can probably do something
better here but I think it's fine.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agokeysym: inline find_sym function
Ran Benita [Sun, 28 Mar 2021 12:21:47 +0000 (15:21 +0300)]
keysym: inline find_sym function

It's easier when everything is in one place.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agokeysym: use a more descriptive argument name
Ran Benita [Sun, 28 Mar 2021 12:23:40 +0000 (15:23 +0300)]
keysym: use a more descriptive argument name

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agokeysym: open-code bsearch
Ran Benita [Sun, 28 Mar 2021 12:03:31 +0000 (15:03 +0300)]
keysym: open-code bsearch

We want to optimize things here which requires messing with the binary
search some.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agointeractive-evdev: fix missing initialization
Ran Benita [Sun, 28 Mar 2021 13:10:52 +0000 (16:10 +0300)]
interactive-evdev: fix missing initialization

Accidentally got lost in 6b65be4.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agointeractive-evdev: switch from epoll(2) to poll(2)
Ran Benita [Sun, 28 Mar 2021 09:55:08 +0000 (12:55 +0300)]
interactive-evdev: switch from epoll(2) to poll(2)

Turns out FreeBSD supports evdev, so this toll can work on it; however
it does not support epoll, so switch to poll, which is portable.

Reported-by: Evgeniy Khramtsov <evgeniy@khramtsov.org>
Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agox11: fix comparison of integer expressions of different signedness
Ran Benita [Sun, 28 Mar 2021 10:01:57 +0000 (13:01 +0300)]
x11: fix comparison of integer expressions of different signedness

src/x11/keymap.c:980:26: warning: comparison of integer expressions of different signedness: â€˜size_t’ {aka â€˜long unsigned int’} and â€˜int’ [-Wsign-compare]
  980 |     for (size_t i = 0; i < length; i++) {
      |                          ^

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agoxkbcomp: plug a potential memory leak
Peter Hutterer [Thu, 18 Mar 2021 02:22:34 +0000 (12:22 +1000)]
xkbcomp: plug a potential memory leak

libxkbcommon-1.0.3/src/xkbcomp/ast-build.c:526: leaked_storage: Variable "file"
going out of scope leaks the storage it points to.

Where we exit the loop early, we don't release the various allocated memory.

Make this patch more obvious my moving the declaration for those into the loop
as well, this way we know that they aren't used outside the loop anywhere.

Found by coverity

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agoregistry: plug a potential memleak for invalid rules files
Peter Hutterer [Thu, 18 Mar 2021 01:22:38 +0000 (11:22 +1000)]
registry: plug a potential memleak for invalid rules files

If the name is missing in a configItem, we'd fail and leak the memory for
description, brief and vendor.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agotest: fix missing va_end in case of test failures
Peter Hutterer [Thu, 18 Mar 2021 01:13:43 +0000 (11:13 +1000)]
test: fix missing va_end in case of test failures

Found by coverity

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agobench: silence coverity complaint about a double free
Peter Hutterer [Thu, 18 Mar 2021 01:08:16 +0000 (11:08 +1000)]
bench: silence coverity complaint about a double free

False positive because we rely on xkb_components_from_rules() to initalize this
struct, but let's localize the variable anyway to shut coverity up.

libxkbcommon-1.0.3/bench/rules.c:59:9: warning[-Wanalyzer-double-free]:
double-free of kccgst.symbols

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agoInline x11_atom_interner_adopt_atoms() into callers
Uli Schlachter [Tue, 9 Mar 2021 08:44:25 +0000 (09:44 +0100)]
Inline x11_atom_interner_adopt_atoms() into callers

Signed-off-by: Uli Schlachter <psychon@znc.in>
3 years agoSave another GetAtomName round trip
Uli Schlachter [Sun, 7 Mar 2021 08:24:40 +0000 (09:24 +0100)]
Save another GetAtomName round trip

Both get_atom_name() and the new atom interner required a round trip. Move
get_atom_name() into the atom interner to save one more round trip. This brings
xkb_x11_keymap_new_from_device() down to two round trips, which is the minimum
possible number.

(Also, I think the new code in keymap.c is more readable than the mess I
previously created)

With this last commit in the series, this definitely:

Fixes: https://github.com/xkbcommon/libxkbcommon/pull/217
Signed-off-by: Uli Schlachter <psychon@znc.in>
3 years agoAlso batch the XKB GetNames request
Uli Schlachter [Sun, 7 Mar 2021 07:48:34 +0000 (08:48 +0100)]
Also batch the XKB GetNames request

This gets rid of another round trip.

Signed-off-by: Uli Schlachter <psychon@znc.in>
3 years agoAlso batch the XKB-GetMapMap request
Uli Schlachter [Sun, 7 Mar 2021 07:46:19 +0000 (08:46 +0100)]
Also batch the XKB-GetMapMap request

This gets rid of one more round trip.

Signed-off-by: Uli Schlachter <psychon@znc.in>
3 years agoRemove three more round trips
Uli Schlachter [Sun, 7 Mar 2021 07:42:54 +0000 (08:42 +0100)]
Remove three more round trips

There are a number of XKB requests needed to request all the information from
the X11 server. So far, the code was sending one request and waiting for the
reply. This commit starts batching the request so that we get multiple replies
with one round trip.

This removes three round trips.

Only the simple requests are converted. get_map() and get_names() use some
bitmasks that are needed for both the request and the reply. These will be dealt
with separately.

Signed-off-by: Uli Schlachter <psychon@znc.in>
3 years agoSave three more round trips in xkb_x11_keymap_new_from_device()
Uli Schlachter [Sun, 7 Mar 2021 07:01:49 +0000 (08:01 +0100)]
Save three more round trips in xkb_x11_keymap_new_from_device()

Instead of asking for an atom name and waiting for the reply four times, this
now sends four GetAtomName requests and waits for all the replies at once. Thus,
this saves three round trips.

Signed-off-by: Uli Schlachter <psychon@znc.in>
3 years agoxkb_x11_keymap_new_from_device: Less X11 round-trips
Uli Schlachter [Sun, 7 Mar 2021 06:42:28 +0000 (07:42 +0100)]
xkb_x11_keymap_new_from_device: Less X11 round-trips

On my system, calling xkb_x11_keymap_new_from_device() did 78 round trips to the
X11 server, which seems excessive. This commit brings this number down to about
9 to 10 round trips.

The existing functions adopt_atom() and adopt_atoms() guarantee that the atom
was adopted by the time they return. Thus, each call to these functions must do
a round-trip. However, none of the callers need this guarantee.

This commit makes "atom adopting" asynchronous: Only some time later is the atom
actually adopted. Until then, it is in some pending "limbo" state.

This actually fixes a TODO in the comments.

Fixes: https://github.com/xkbcommon/libxkbcommon/issues/216
Signed-off-by: Uli Schlachter <psychon@znc.in>
3 years agoBump version to 1.1.0 xkbcommon-1.1.0
Ran Benita [Sat, 27 Feb 2021 20:48:12 +0000 (22:48 +0200)]
Bump version to 1.1.0

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agodoc: add keymap-format-text-v1.md to the HTML documentation
Ran Benita [Sat, 27 Feb 2021 20:38:21 +0000 (22:38 +0200)]
doc: add keymap-format-text-v1.md to the HTML documentation

It's incomplete but might be helpful for someone.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agodoc: add note on variants in rule names
Ran Benita [Sat, 27 Feb 2021 20:12:08 +0000 (22:12 +0200)]
doc: add note on variants in rule names

Fixes: https://github.com/xkbcommon/libxkbcommon/issues/208
Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agoFix a few keysymtab entries to match their comment in xkbcommon-keysyms.h
Pierre Le Marre [Fri, 26 Feb 2021 09:54:31 +0000 (10:54 +0100)]
Fix a few keysymtab entries to match their comment in xkbcommon-keysyms.h

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agoscripts: remove meson-junit-report.py
Ran Benita [Sat, 27 Feb 2021 19:38:02 +0000 (21:38 +0200)]
scripts: remove meson-junit-report.py

Not used since ed5a0b4fede69b8e6dc4db53d97ea4ae0a73956d.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agomeson.build: replace the remaining join_paths() with the nicer / syntax
Peter Hutterer [Mon, 22 Feb 2021 10:07:45 +0000 (20:07 +1000)]
meson.build: replace the remaining join_paths() with the nicer / syntax

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agotest: add a keysym tester
Peter Hutterer [Mon, 22 Feb 2021 02:54:15 +0000 (12:54 +1000)]
test: add a keysym tester

A simple script that creates a new layout with the given keysym replacing TLDE.
Then we compile a keymap and search for the keysym being assigned to TLDE and
bail if that fails.

The list of keysyms is manually maintained but we only need to add one or two to
spot-check whenever the xorgproto is updated.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agoFix documentation for XKB_CONFIG_EXTRA_PATH
Peter Hutterer [Mon, 22 Feb 2021 02:33:35 +0000 (12:33 +1000)]
Fix documentation for XKB_CONFIG_EXTRA_PATH

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agoUpdate keysym definitions to latest xorgproto
Peter Hutterer [Sun, 21 Feb 2021 22:20:40 +0000 (08:20 +1000)]
Update keysym definitions to latest xorgproto

As of xorgproto commit e5d8af9711516385f8346c9e077692b29c914478

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agoscripts: update makeheader script for the _EVDEVK keysym defines
Peter Hutterer [Tue, 16 Feb 2021 23:06:57 +0000 (09:06 +1000)]
scripts: update makeheader script for the _EVDEVK keysym defines

As of xorgproto commit 5dbb5b76597f [1], the 0x10081XXX keycode range is defined
for direct evdev kernel keycode mapping. For example, KEY_MACRO1 (0x290) is
mapped to 0x10081290. The format of the #define lines for these keys is
stable to allow for parsing:

  #define XF86XK_FooBar _EVDEVK(0x123) /* optional comment */

Update our script so we detect these new lines. Our keysym generation is a
two-step process: makeheader and then makekeys. Replacing the key with its full
value in the makeheader script means we don't have to update makekeys to handle
the _EVDEVK macro and our header file is fully resolved.

[1] https://gitlab.freedesktop.org/xorg/proto/xorgproto/-/merge_requests/23

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agoUpdate imported comments in xkbcommon-keysyms.h
Peter Hutterer [Sun, 21 Feb 2021 22:17:58 +0000 (08:17 +1000)]
Update imported comments in xkbcommon-keysyms.h

Matches xorgproto commits 39eb5974 and 09602b213.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agotest: fix interactive evdev test invocation
Peter Hutterer [Thu, 21 Jan 2021 22:18:13 +0000 (08:18 +1000)]
test: fix interactive evdev test invocation

rmlvos is the parent list which then fails during a list join because, well,
it's a list of lists.

Fixes #206

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agotest: fill in srcdir/builddir when not set in the environment
Peter Hutterer [Thu, 21 Jan 2021 22:05:09 +0000 (08:05 +1000)]
test: fill in srcdir/builddir when not set in the environment

Makes this test easier to run from the commandline. Where either of top_srcdir
or top_builddir isn't set, fill them in from the CWD or fail otherwise.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agodoc/keymap: some slight editing
Ran Benita [Sun, 27 Dec 2020 07:47:08 +0000 (09:47 +0200)]
doc/keymap: some slight editing

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agodoc/keymap: add documentation for xkb_symbols (#205)
Simon Zeni [Sun, 27 Dec 2020 07:48:39 +0000 (02:48 -0500)]
doc/keymap: add documentation for xkb_symbols (#205)

3 years agoBump version to 1.0.3 sandbox/jeon/upstream upstream_bkup_20220207 xkbcommon-1.0.3
Ran Benita [Mon, 23 Nov 2020 18:30:13 +0000 (20:30 +0200)]
Bump version to 1.0.3

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agokeymap-dump: follow xkbcomp in printing affect=both in pointer actions
Ran Benita [Mon, 23 Nov 2020 17:51:04 +0000 (19:51 +0200)]
keymap-dump: follow xkbcomp in printing affect=both in pointer actions

It is equivalent to nothing but good to match up.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agox11: fix type level names missing
Ran Benita [Mon, 23 Nov 2020 16:49:49 +0000 (18:49 +0200)]
x11: fix type level names missing

When reading the keymap, the level names would get discarded.
Regressed in 26453b84732da870f5695ee347970b337cfea9c1.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agotest/data: update host.xkb to match keymap-dump style
Ran Benita [Mon, 23 Nov 2020 16:35:27 +0000 (18:35 +0200)]
test/data: update host.xkb to match keymap-dump style

This is needed for fixing the x11comp test.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agox11/keymap: fix case with no actions
Ran Benita [Mon, 23 Nov 2020 16:24:52 +0000 (18:24 +0200)]
x11/keymap: fix case with no actions

Possible regression in f41e609bbea8447fc82849a1a6ea0d116189f2f8 (not
confirmed yet).

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agoBump version to 1.0.2 xkbcommon-1.0.2
Ran Benita [Fri, 20 Nov 2020 19:01:56 +0000 (21:01 +0200)]
Bump version to 1.0.2

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agox11: cache X11 atoms
Ran Benita [Wed, 18 Nov 2020 22:28:37 +0000 (00:28 +0200)]
x11: cache X11 atoms

On every keymap notify event, the keymap should be refreshed, which
fetches the required X11 atoms. A big keymap might have a few hundred of
atoms.

A profile by a user has shown this *might* be slow when some intensive
amount of keymap activity is occurring. It might also be slow on a
remote X server.

While I'm not really sure this is the actual bottleneck, caching the
atoms is easy enough and only needs a couple kb of memory, so do that.

On the added bench-x11:

Before: retrieved 2500 keymaps from X in 11.233237s
After : retrieved 2500 keymaps from X in 1.592339s

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agox11: eliminate slow divisions
Ran Benita [Fri, 20 Nov 2020 10:45:42 +0000 (12:45 +0200)]
x11: eliminate slow divisions

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agobuild: combine test & bench internal deps
Ran Benita [Thu, 19 Nov 2020 14:15:42 +0000 (16:15 +0200)]
build: combine test & bench internal deps

No need to duplicate this really.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agoci: fix windows CI failure due to add-path deprecation
Ran Benita [Wed, 18 Nov 2020 23:08:47 +0000 (01:08 +0200)]
ci: fix windows CI failure due to add-path deprecation

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agotest: catch unrecognized keysyms in the xkeyboard-config test
Peter Hutterer [Mon, 19 Oct 2020 01:51:27 +0000 (11:51 +1000)]
test: catch unrecognized keysyms in the xkeyboard-config test

Prompted by
https://gitlab.freedesktop.org/xkeyboard-config/xkeyboard-config/-/merge_requests/127

We run the keymap tool with --verbose which will print the messages from the
compiler to the log file as well. And then we can search for the warning
regarding an unrecognized keysym and fail our test based on that.

3 years agoxkbcomp: where a keysym cannot be resolved, set it to NoSymbol
Peter Hutterer [Mon, 19 Oct 2020 00:49:37 +0000 (10:49 +1000)]
xkbcomp: where a keysym cannot be resolved, set it to NoSymbol

Where resolve_keysym fails we warn but use the otherwise uninitialized variable
as our keysym. That later ends up in the keymap as random garbage hex value.

Simplest test case, set this in the 'us' keymap:
    key <TLDE>               {      [        xyz ] };

And without this patch we get random garbage:
./build/xkbcli-compile-keymap --layout us | grep TLDE:
    key <TLDE>               {      [      0x018a5cf0 ] };

With this patch, we now get NoSymbol:
./build/xkbcli-compile-keymap --layout us | grep TLDE:
    key <TLDE>               {      [        NoSymbol ] };

3 years agodoc: add note about "short" layout names (#194)
Ran Benita [Sun, 11 Oct 2020 06:55:14 +0000 (09:55 +0300)]
doc: add note about "short" layout names (#194)

Fixes: https://github.com/xkbcommon/libxkbcommon/issues/193
Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agoBump version to 1.0.1 xkbcommon-1.0.1
Ran Benita [Fri, 11 Sep 2020 12:07:39 +0000 (15:07 +0300)]
Bump version to 1.0.1

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agoparser: fix another format string for int64_t (#191)
hhb [Fri, 11 Sep 2020 12:06:23 +0000 (05:06 -0700)]
parser: fix another format string for int64_t (#191)

3 years agotools: align `xkbcli how-to-type` output ourselves
Ran Benita [Wed, 9 Sep 2020 07:09:13 +0000 (10:09 +0300)]
tools: align `xkbcli how-to-type` output ourselves

Can possibly add a machine-parsable format if desired, but for now just
have it work nicely.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agotest/tool-option-parsing: skip testing of disabled tools
Ran Benita [Mon, 7 Sep 2020 09:27:26 +0000 (12:27 +0300)]
test/tool-option-parsing: skip testing of disabled tools

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agotest/tool-option-parsing: switch from pytest to unittest
Ran Benita [Mon, 7 Sep 2020 07:36:18 +0000 (10:36 +0300)]
test/tool-option-parsing: switch from pytest to unittest

For me, installing pytest for libxkbcommon is a bit problematic, so I
end up skipping it which is not great.

Switch to unittest which is built in to Python. It's not as nice as
pytest but good enough in this case.

Note: I was too lazy to switch the plain asserts to unittest
assertions...

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agotest/tool-option-parsing: keep isolated by using our own test data
Ran Benita [Mon, 7 Sep 2020 08:10:19 +0000 (11:10 +0300)]
test/tool-option-parsing: keep isolated by using our own test data

Make it possible to run the test on all machines.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agobuild: move tests to after tools
Ran Benita [Mon, 7 Sep 2020 09:19:17 +0000 (12:19 +0300)]
build: move tests to after tools

So tests can refer to stuff set by the tools section.

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agoutils: include unistd.h where we have it
Peter Hutterer [Mon, 7 Sep 2020 09:49:36 +0000 (19:49 +1000)]
utils: include unistd.h where we have it

MacOS doesn't have eaccess/euidaccess but it does have unistd.h, so let's
include it to silence the R_OK redefinition compiler warnings.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agocontext: fix a compiler warning
Peter Hutterer [Mon, 7 Sep 2020 09:47:42 +0000 (19:47 +1000)]
context: fix a compiler warning

../src/context.c:57:9: warning: variable 'err' is used uninitialized whenever
'if' condition is true [-Wsometimes-uninitialized]

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agoregistry: mark the rxkb_log function as attribute printf
Peter Hutterer [Mon, 7 Sep 2020 09:38:47 +0000 (19:38 +1000)]
registry: mark the rxkb_log function as attribute printf

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agoci: enable xkbregistry on macos job
Ran Benita [Mon, 7 Sep 2020 08:54:55 +0000 (11:54 +0300)]
ci: enable xkbregistry on macos job

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agotest: include unstd.h in the registry test to cut down the MacOS warnings
Peter Hutterer [Mon, 7 Sep 2020 09:24:12 +0000 (19:24 +1000)]
test: include unstd.h in the registry test to cut down the MacOS warnings

mkdtmp, rmdir and unlink are in unstd.h on MacOS. Since including that it
doesn't hurt us on Linux, let's do it without ifdefs.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
3 years agotest/data: add rule registry files
Ran Benita [Mon, 7 Sep 2020 08:38:00 +0000 (11:38 +0300)]
test/data: add rule registry files

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agotest/data: sync from xkeyboard-config 2.30
Ran Benita [Mon, 7 Sep 2020 08:35:22 +0000 (11:35 +0300)]
test/data: sync from xkeyboard-config 2.30

Signed-off-by: Ran Benita <ran@unusedvar.com>
3 years agotest/data: change quartz.xkb from CRLF to LF
Ran Benita [Mon, 7 Sep 2020 08:15:43 +0000 (11:15 +0300)]
test/data: change quartz.xkb from CRLF to LF

Signed-off-by: Ran Benita <ran@unusedvar.com>