| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When twelve years ago, in commit c28367a15, the code points 0x100FE67
and further were added for "dead a" and such, the intention was that
they were raw key codes (0xfe67 and further), not Unicode codepoints.
See the original patch that requested adding very similar key codes
to xorgproto three years earlier:
https://bugs.freedesktop.org/attachment.cgi?id=18679
The compose sequences in libX11 have been updated to use the dead-vowel
symbols instead of the erroneous Unicode codepoints (most of which are
valid codepoints for certain characters, that thus became unusable in
the Amharic locale), so now it's time to update the keyboard layout.
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
|
|
|
|
| |
All other such layouts are named in this manner: Language (Country).
|
|
|
|
|
| |
For most people a cuneiform script like Ugaritic is more likely to
represent the "canonical" ancient script than something like Gothic.
|
|
|
|
|
|
|
|
| |
Six years ago, commit 10cf5acd90 added the word "the" in the phrase
"Key to choose the 3rd level", but not in its sister phrase, the one
for the 5th-level chooser.
To compare, run: `grep -R "Key to choose"`
|
| |
|
| |
|
|
|
|
|
| |
Fix spelling, add helpful little word, s/pressing/releasing/,
drop duplicate word, and improve punctuation.
|
|
|
|
|
|
|
|
|
|
| |
- Create Python bindings to xkbcommon.
- Create a regression test framework using pytest.
- Add regression tests for issues 90, 346, 382 and 383.
- Document how to write tests.
- CI: Create a separate job for the libxkbcommon build that share
its artifacts.
- CI: Add the tests to the keymap_tests job.
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
Four days ago commit 43a420f003 dropped the redefinition of `AB09` from
the Canadian OLPC layout, thinking that the Unicode HYPHEN (U+2010) is
the same as keysym 'hyphen'. But the latter is U+00AD, the SOFT HYPHEN.
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
|
|
|
|
|
| |
Also, adjust some whitespace, delete two stray characters from the
"graphical" representation, correct one, and add a missing one.
|
|
|
|
| |
Also, replace two Unnnn codes with the corresponding symbolic names.
|
|
|
|
| |
(This should have been part of commit db45bd75 from four days ago.)
|
|
|
|
| |
Additionally, report successful tests count.
|
|
|
|
|
|
|
| |
Required to run pipelines after some infrastructure changes, see
https://gitlab.freedesktop.org/freedesktop/freedesktop/-/issues/540
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This Gothic layout is put in a new symbols file, called 'ancient',
so that there will be a place for other layouts for antique scripts.
The name "ancient" is chosen because that is how Gothic is marked
in https://en.wikipedia.org/wiki/List_of_ISO_639-2_codes.
This fulfills issue #55.
Proposed-by: Tim Nelson
Original-patch-by: Tim Nelson
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Just like Lock does not swap levels 3 and 4 when just LevelThree and
maybe Shift are in effect for EIGHT_LEVEL_ALPHABETIC_WITH_LEVEL5_LOCK,
Lock should *also* not swap those levels when also the LevelFive-Lock
(NumLock) *and* LevelFive itself are in effect.
This fixes issue #381.
Problem existed since the EIGHT_LEVEL_ALPHABETIC_WITH_LEVEL5_LOCK type
was introduced, in commit 39124ccf14, thirteen years ago.
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
|
|
|
|
| |
Users have had enough time to get used to the new names.
|
| |
|
|
|
|
| |
Also reshuffle three keys into order, and drop an unneeded redefinition.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Also delete three pointless NoSymbol specifications, and tweak comments.
|
|
|
|
| |
Also, reshuffle a few keys into their normal order.
|
|
|
|
| |
Also add comments after the Unnnn codes, and drop redundant includes.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Nineteen years ago, commit 8e1e2d8195 baptized these four variants with
the words "extended backslash". Nine months later, commit 999e981f4c
synced these names to the relevant symbol files.
But just two days later, commit 3ad7cee7c5 changed one of the names to
read "with <\|> key", in response to https://bugs.freedesktop.org/2319.
(A week later, commit 1b8556fe6f synced this name to the XML file.)
This gave one of the four very similar variants a deviant name.
The string "with <\|> key" is difficult for translators. For a long
time I have thought that "<" and ">" were the relevant symbols, and
that the "|" was an OR symbol, escaped by the "\". In other words:
I thought that the string referred to the LSGT key. :/
What the four variants actually do is: put the backslash and vertical
bar characters on the BKSL key (replacing the dead diaeresis and the
apostrophe), for those keyboards that lack the LSGT key, to make typing
"\" and "|" more straightforward -- these symbols are present also on
the third level of the Q and W keys. So what those variants do is:
add an *extra* backslash and vertical bar character to the layout,
for easier access, at the cost of having the apostrophe accessible
only via AltGr and the dead diaeresis only via Shift+AltGr.
So... name all four variants more fittingly and in the same manner.
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In order to make the option `grp:shifts_toggle` effective also with
the few layouts that define the Shift keys too, the option should be
applied to each group separately, and the option should specify the
key type explicitly.
This fixes issue #380.
Problem existed since the T3 layout was added in 2013,
and worsened when the E1 layout was added in 2020.
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Many people do not need the CapsLock functionality, so leaving the
first level of the `<CAPS>` key undefined in these two options allows
users to combine `grp:shift_caps_toggle` or `grp:alt_caps_toggle` with
most of the one-level `caps:` options, for more versatility.
(Because of the order of the rules, any `caps:` option will be applied
first, and then the `grp:` option will add a second level.)
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
|
| |
|
|
|
|
|
|
|
|
|
|
| |
This looks better and should mostly have no effect as the options are
largely independent. Only the moved `lv3:` options will now overrule
some of the `caps:`, `ctrl:` and `compose:` options, instead of being
overruled by them. In my opinion this is more consistent: all level3
options overrule most other options.
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
|
|
|
|
|
|
|
|
|
|
| |
Arguably, when options conflict, the group-switching options should
overrule others, because when one has multiple groups one definitely
wants to be able to switch.
This fixes issue #384.
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Mapping modifiers to keys instead of to symbols causes problems when
those keys carry symbols that are associated with *other* modifiers,
as this leads to two or more modifiers getting linked together and
thus activated in unison -- which one normally doesn't want.
This fixes issue #382.
(Issue #81 will have to be addressed in some other way.)
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
|
|
|
|
|
|
|
|
|
|
|
|
| |
When using a `grp:` option, one doesn't want other options to interfere
with this option, so the rules for `grp:` options should come last in
the list.
This fixes issue #383.
Reported-by: Sergey Stolyarov
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
|
|
|
|
|
|
|
|
|
|
|
| |
In Persian, an isolated form of the letter Ye is added after nouns
that are connected to adjectives, and if the noun ends with the
letter He, it is often merged with the following Ye, into the letter
He with Ye above (U+06C0).
This character exists on the second level of AC05 in the Persian
layout of Microsoft Windows, but not in the standard layout.
This commit extends the standard layout to include U+06C0 on the
third level of AC05.
|
|
|
|
|
|
|
|
|
|
|
|
| |
Mapping a modifier to a symbol had the tiny advantage of not needing
to explicitly specify a modifier mapping when putting the symbol onto
some other key through some option. But it had the disadvantage of
risking to lose the modifier mapping for the original key when the
other key was just an addition and not a replacement.
This generalizes the fix from the previous commit.
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When a modifier is mapped to a symbol, then it will effectively be
mapped to the first-found key that contains the relevant symbol,
and that key might not be the key that the modifier is usually
mapped to, leaving this latter key without a mapping.
So, map the Super modifier to `<LWIN>` and `<RWIN>` instead of to
`Super_L` and `Super_R` -- the Windows keys will then not lack their
Super modifier when some other key is made into a Super key too.
This fixes issue #81.
Reported-by: Josh Triplett
Signed-off-by: Benno Schulenberg <bensberg@telfort.nl>
|
| |
|
| |
|
| |
|
|
|
|
| |
Also, add some level specifications for clarity, although unneeded.
|
|
|
|
|
|
|
| |
This makes it easier to see where which symbols are assigned to a key
when grepping for its code.
Also improve a couple of comments.
|
| |
|