profile/ivi/libxkbcommon.git
12 years agoChange xkb_key_get_syms to just return a bare int
Daniel Stone [Wed, 9 May 2012 19:51:37 +0000 (20:51 +0100)]
Change xkb_key_get_syms to just return a bare int

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoMove KcCGST API to internal-only
Daniel Stone [Wed, 9 May 2012 19:49:04 +0000 (20:49 +0100)]
Move KcCGST API to internal-only

And don't export it.  We don't need it for X11 support, let alone
anything else.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoRename serialise to serialize
Daniel Stone [Wed, 9 May 2012 19:20:12 +0000 (20:20 +0100)]
Rename serialise to serialize

Yes, British English is correct, but unfortunately we've lost that
battle.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoAdd common LED names to xkbcommon-names.h
Daniel Stone [Wed, 9 May 2012 19:18:30 +0000 (20:18 +0100)]
Add common LED names to xkbcommon-names.h

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoInclude xkbcommon-names.h from xkbcommon.h
Daniel Stone [Wed, 9 May 2012 19:17:13 +0000 (20:17 +0100)]
Include xkbcommon-names.h from xkbcommon.h

So clients only have one file to include.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoRemove keycode_range_is_legal
Daniel Stone [Wed, 9 May 2012 19:12:18 +0000 (20:12 +0100)]
Remove keycode_range_is_legal

It was a pretty pointless check.  Also sanitise the _x11 variant to
actually do what it says on the box.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoMerge remote-tracking branch 'krh/keysyms'
Daniel Stone [Wed, 9 May 2012 19:12:12 +0000 (20:12 +0100)]
Merge remote-tracking branch 'krh/keysyms'

Conflicts:
src/keysym.c
src/misc.c
src/text.h
src/xkbcomp/expr.c
src/xkbcomp/parser.y
src/xkbcomp/parseutils.c
src/xkbcomp/symbols.c

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoUse our own keysyms
Kristian Høgsberg [Wed, 9 May 2012 13:05:00 +0000 (09:05 -0400)]
Use our own keysyms

12 years agoAdd XKB version of X11 keysyms
Kristian Høgsberg [Wed, 9 May 2012 12:33:04 +0000 (08:33 -0400)]
Add XKB version of X11 keysyms

With this we're now completely standalone.

add vendor keysyms

12 years agoRemove support for xkb_layout and xkb_semantics file types
Ran Benita [Wed, 9 May 2012 14:54:37 +0000 (17:54 +0300)]
Remove support for xkb_layout and xkb_semantics file types

These are two aggregate file types which are not used anywhere. We
maintain useful-enough backward compatibility in the parser, by treating
them as xkb_keymap. The keymap type allows for all types of components,
so they will still compile fine if they ever come up.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoShorten context to ctx
Ran Benita [Wed, 9 May 2012 12:03:11 +0000 (15:03 +0300)]
Shorten context to ctx

(This breaks the API.)

"context" is really annoying to type all the time (and we're going to
type it a lot more :). "ctx" is clear, concise and common in many other
libraries. Use it!

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Fix for xkb -> keymap change.]

12 years agoContextualize the atom table
Ran Benita [Wed, 9 May 2012 11:02:26 +0000 (14:02 +0300)]
Contextualize the atom table

Each context gets its own table, i.e. interning a string in one context
does not affect any other context.

The existing xkb_atom_* functions are turned into wrappers around a new
standalone atom_table object.

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Updated for xkb -> keymap.]

12 years agoMake the context available for XkbcAtomText
Ran Benita [Wed, 9 May 2012 10:50:05 +0000 (13:50 +0300)]
Make the context available for XkbcAtomText

And rename the function to xkb_atom_text.

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Updated for xkb -> keymap.]

12 years agoMake the context available for XkbcAtomGetString
Ran Benita [Wed, 9 May 2012 09:01:03 +0000 (12:01 +0300)]
Make the context available for XkbcAtomGetString

In preparation of contextualizing atom handling.
Since we touch every function call, we also rename the function to
xkb_atom_strdup to match xkb_atom_intern, and be more descriptive.

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Updated for xkb -> keymap.]

12 years agoMake the context available to xkb_intern_atom
Ran Benita [Wed, 9 May 2012 08:47:20 +0000 (11:47 +0300)]
Make the context available to xkb_intern_atom

In preparation of contextualizing the atom table.
Since we touch every function call, also rename the function to
xkb_atom_intern, to match better with the rest (which will also be
renamed).

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Fixed for 'xkb' -> 'keymap'.]

12 years agoContextualize XkbFile IDs
Ran Benita [Wed, 9 May 2012 08:29:04 +0000 (11:29 +0300)]
Contextualize XkbFile IDs

Currently the IDs are assigned from a static variable inside
CreateXKBFile. This can lead to some unpleasantness with threads, so
maintain the counter in the context instead.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoMake the context available to the parser
Ran Benita [Wed, 9 May 2012 08:12:30 +0000 (11:12 +0300)]
Make the context available to the parser

We will need the context to remove some global state.
Also make the Parse* function just return bool while wer'e at it.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agomakekeys: fix cross-compilation
Pekka Paalanen [Tue, 8 May 2012 08:46:58 +0000 (11:46 +0300)]
makekeys: fix cross-compilation

makekeys must be built with the build-native compiler, not with $(CC)
which is the cross-compiler. The only sane way to achieve this seems to
be to use a separate Makefile.am for it.

This patch fixes the problem apparently caused by:

commit b5efe41f190cbb76eb1ca8ddf0c96990ddb83704
Author: Ran Benita <ran234@gmail.com>
Date:   Sat Mar 24 04:48:31 2012 +0200

    Make build non-recursive

There is no such thing as makekeys_makekeys_CC in automake.

Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
12 years agoMake build directory includes more explicit
Daniel Stone [Wed, 9 May 2012 14:23:44 +0000 (15:23 +0100)]
Make build directory includes more explicit

We depend on parser.h which is generated by flex during the build, so
use an explicit $(top_builddir) rather than a relative path.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoChange all 'xkb' xkb_keymap names to 'keymap'
Daniel Stone [Wed, 9 May 2012 14:15:30 +0000 (15:15 +0100)]
Change all 'xkb' xkb_keymap names to 'keymap'

To make it a bit more clear what it actually is.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoRename keysym <-> string API
Daniel Stone [Wed, 9 May 2012 12:22:34 +0000 (13:22 +0100)]
Rename keysym <-> string API

Change them to refer to the string representation of the keysym's name
as a name rather than a string, since we want to add API to get the
Unicode printable representation as well.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoAdd multiple modifier state matching API
Daniel Stone [Wed, 9 May 2012 00:06:10 +0000 (01:06 +0100)]
Add multiple modifier state matching API

Two new calls allow users to test the exact modifier state, including
verifying that no other modifiers but the ones you wanted are down.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoAdd pre-defined names database
Daniel Stone [Tue, 8 May 2012 16:59:35 +0000 (17:59 +0100)]
Add pre-defined names database

xkbcommon-names.h right now just contains a set of hardcoded modifier
strings that are most commonly used for the usual modifiers.  Provide
definitions of these so people don't have to worry about typoing a
string or mixing up Mod1 and Mod4.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoAdd context flag to inhibit default include paths
Daniel Stone [Tue, 8 May 2012 16:52:45 +0000 (17:52 +0100)]
Add context flag to inhibit default include paths

Which will make the context start with no include paths at all.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoAdd flags to context creation
Daniel Stone [Tue, 8 May 2012 16:51:16 +0000 (17:51 +0100)]
Add flags to context creation

None defined as yet, but why not.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoAdd flags to keymap compilation entrypoints
Daniel Stone [Tue, 8 May 2012 16:48:29 +0000 (17:48 +0100)]
Add flags to keymap compilation entrypoints

No use as yet, but might as well ...

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoReintroduce $(builddir)/src/xkbcomp include
Daniel Stone [Tue, 8 May 2012 16:47:54 +0000 (17:47 +0100)]
Reintroduce $(builddir)/src/xkbcomp include

This is required for parser.h, which is autogenerated and thus only in
the build directory.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoRename XKBcommonint.h to xkb-priv.h and use it
Ran Benita [Tue, 8 May 2012 11:52:23 +0000 (14:52 +0300)]
Rename XKBcommonint.h to xkb-priv.h and use it

Make the files in the src/* directory use their own header or a
consilidated private header. This makes the file dependencies clearer.

Also drop the pointless "xkb" file name prefix, add split a few
declarations to their own files (atom.h and text.h).

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoRename xkbcomp/misc.h to xkbcomp-priv.h and use it
Ran Benita [Tue, 8 May 2012 10:57:07 +0000 (13:57 +0300)]
Rename xkbcomp/misc.h to xkbcomp-priv.h and use it

The include dependencies were quite convoluted, where you change the
order and get a ton of errors. Instead, change one file to act as the
internal interface for the xkbcomp files, and make every file use it.

Also drop the pointless "xkb" prefix to file names.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoRefactor Compile<component> functions
Ran Benita [Mon, 7 May 2012 22:08:07 +0000 (01:08 +0300)]
Refactor Compile<component> functions

The error handling was not ideal, so unify it. Also makes the functions
a bit easier to read.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoDon't use typeof
Ran Benita [Mon, 7 May 2012 11:44:30 +0000 (14:44 +0300)]
Don't use typeof

clang complains with the xorg-macros warning flags:
src/context.c:58:36: error: extension used [-Werror,-pedantic,-Wlanguage-extension-token]
                                   typeof(new_paths));

This was not entirely correct, too. So bring back the casts to the
results of the allocation macros; might as well make them a bit more
type safe.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoRemove left over filecomp.sh file
Ran Benita [Mon, 7 May 2012 11:08:34 +0000 (14:08 +0300)]
Remove left over filecomp.sh file

Should have removed it in a previous commit.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoUpdate .gitignore for automake 1.12
Ran Benita [Mon, 7 May 2012 11:23:08 +0000 (14:23 +0300)]
Update .gitignore for automake 1.12

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoUncomment sections of test/rulescomp.c
Ran Benita [Mon, 7 May 2012 12:07:41 +0000 (15:07 +0300)]
Uncomment sections of test/rulescomp.c

Fixes an 'unused' warning. There seems to be nothing wrong with these
sections though, all the tests pass.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoRevert "Unconstify xkb_rules_names"
Ran Benita [Mon, 7 May 2012 11:54:12 +0000 (14:54 +0300)]
Revert "Unconstify xkb_rules_names"

This reverts commit d007cd0a1f3f4b9c927175771ff79aae6fe4ab8b.

This is in fact more restrictive, because it breaks the (common) case
where the strings are const themselved, e.g. "evdev", "us", etc. As is
you must either duplicate the strings or suppress the warnings.

If the user needs to retain the non-const strings, he should instead
just keep them in some other struct and use xkb_rules_names just as
a temporary parameter for xkb_map_new_from_names. Mildly annoying but
acceptable.

12 years agoUse AC_ARG_VAR for X11 header and XKB paths
Daniel Stone [Mon, 7 May 2012 16:30:47 +0000 (17:30 +0100)]
Use AC_ARG_VAR for X11 header and XKB paths

And don't call pkg-config unless we need to.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoRename YYLTYPE to struct YYLTYPE
Daniel Stone [Mon, 7 May 2012 13:23:46 +0000 (14:23 +0100)]
Rename YYLTYPE to struct YYLTYPE

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoMerge src/alloc.c and src/malloc.c
Ran Benita [Tue, 10 Apr 2012 23:02:45 +0000 (02:02 +0300)]
Merge src/alloc.c and src/malloc.c

The two files do exactly the same sort of things, without any discernible
reason for splitting them.

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Updated for xkb_desc -> xkb_keymap changes.]

12 years agoRemove unused stuff from xkbrules.h
Ran Benita [Tue, 10 Apr 2012 23:28:57 +0000 (02:28 +0300)]
Remove unused stuff from xkbrules.h

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoRemove unused stuff from XKBcommonint.h
Ran Benita [Tue, 10 Apr 2012 20:08:49 +0000 (23:08 +0300)]
Remove unused stuff from XKBcommonint.h

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoRemove unused 'which' and 'merge' arguments
Ran Benita [Tue, 10 Apr 2012 22:55:50 +0000 (01:55 +0300)]
Remove unused 'which' and 'merge' arguments

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Updated for xkb_desc -> xkb_keymap changes.]

12 years agoRemove unused 'compiled' field in XkbFile
Ran Benita [Wed, 11 Apr 2012 16:58:03 +0000 (19:58 +0300)]
Remove unused 'compiled' field in XkbFile

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoRemove unused debugging function
Ran Benita [Wed, 11 Apr 2012 11:13:24 +0000 (14:13 +0300)]
Remove unused debugging function

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoRemove unused function XkbcNameMatchesPattern
Ran Benita [Tue, 10 Apr 2012 23:10:35 +0000 (02:10 +0300)]
Remove unused function XkbcNameMatchesPattern

Because the function is recursive, the compiler didn't say anything, eh.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoUse LT_INIT instead of deprecated AC_PROG_LIBTOOL
Ran Benita [Tue, 10 Apr 2012 19:04:05 +0000 (22:04 +0300)]
Use LT_INIT instead of deprecated AC_PROG_LIBTOOL

And add some warnings to the autoreconf step, for the developers to
see.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoUse stdbool.h
Ran Benita [Fri, 6 Apr 2012 00:38:55 +0000 (03:38 +0300)]
Use stdbool.h

'Cause defining your own True and False is so 1990's.

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Fixed for xkb_desc -> xkb_keymap changes.]

12 years agoFix formatting in xkbcomp headers
Ran Benita [Thu, 5 Apr 2012 08:24:39 +0000 (11:24 +0300)]
Fix formatting in xkbcomp headers

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Fixed for xkb_desc -> xkb_keymap change.]

12 years agoDump include paths when we can't find rules
Daniel Stone [Thu, 26 Apr 2012 14:02:49 +0000 (16:02 +0200)]
Dump include paths when we can't find rules

Since the most common failure mode here is a failure to properly set the
XKB data path, dump the include path so people at least have a clue
where to look.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoInclude strings.h in XKBcommonint.h for strcasecmp etc
Alan Coopersmith [Thu, 26 Apr 2012 11:49:00 +0000 (13:49 +0200)]
Include strings.h in XKBcommonint.h for strcasecmp etc

POSIX specifies that these functions require <strings.h>, but we were
only including <string.h>.  It did work, but still.

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
12 years agoOverhaul test suite
Ran Benita [Fri, 30 Mar 2012 23:44:39 +0000 (02:44 +0300)]
Overhaul test suite

Rewrite all of the current tests in the following ways:
- Instead of the current mix of C and shell, just use single-process
  pure C file per test. All of the .sh files are removed, but everything
  that was tested is ported.
- Instead of handling the test logs ourselves, use Automake's
  "parallel-test" mechanism. This will create a single log file for each
  test with it's stdout+stderr, and a top level "test-suite.log" file
  for all the failed tests.
- The "parallel-tests" directive also makes the test run in parallel,
  so "make check" runs faster.
- Also use the "color-tests" directive to have the "make check" output
  colorized. Who doesn't like to see PASS in green?
- All of the test data files are moved into the test/data subdirectory.
  That way we can just put the directory in EXTRA_DIST and forget about
  it.
- The test/Makefile.am file is consolidated into the main Makefile.am,
  for a completely non-recursive build.

Right now the tests are completely independent and just use simple
assert()'s. More sophistication can be added as needed.

It should also be noted that it's still possible to use shell, python,
etc. if a test wants more flexibility than C can provide, just do as
before.

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Updated for xkb_keymap changes.]

12 years agoReformat actionHandler dispatch table
Ran Benita [Fri, 6 Apr 2012 01:33:43 +0000 (04:33 +0300)]
Reformat actionHandler dispatch table

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoAdd xkb_state_get_map()
Ran Benita [Sat, 7 Apr 2012 23:08:37 +0000 (02:08 +0300)]
Add xkb_state_get_map()

This is very useful because it avoids redundent pointers in structs
and/or parameter passing in the application.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoImplicitly include config.h in all files
Ran Benita [Fri, 6 Apr 2012 00:12:50 +0000 (03:12 +0300)]
Implicitly include config.h in all files

The definitions in config.h should be available in all files an
implementation detail; it can be included through the build system
instead of having each file pull it every time.

This is especially helpful with AC_USE_SYSTEM_EXTENSIONS, as _GNU_SOURCE
and friends can have an effect by merely being defined, which can lead
to some confusion if its effective for only half the files.

And we don't really support a build _without_ config.h; so, one less
thing to worry about.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoRemove xproto and kbproto from pkg-config file
Ran Benita [Thu, 5 Apr 2012 07:52:47 +0000 (10:52 +0300)]
Remove xproto and kbproto from pkg-config file

These are no longer needed for using the library, only building it. Most
users would still want xproto though, for the keysym definitions.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoRemove Xfuncproto.h and XKB.h from xkbcommon/xkbcommon.h
Ran Benita [Thu, 5 Apr 2012 07:47:43 +0000 (10:47 +0300)]
Remove Xfuncproto.h and XKB.h from xkbcommon/xkbcommon.h

The kbproto header is already not needed here anymore.

Move the _X_EXPORT's to the corresponding function definitions, and use
straight extern "C" clauses instead of _XFUNCPROTOBEGIN/END.

It also makes more sense to have the EXPORT's in the source files, as it
provides some documentation to the reader, whereas in the header it's
obvious.

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Updated for xkb_keymap changes.]

12 years agoConstify the syms_out argument to xkb_key_get_syms()
Ran Benita [Sun, 8 Apr 2012 12:40:12 +0000 (15:40 +0300)]
Constify the syms_out argument to xkb_key_get_syms()

The caller should not mess around with these as they come directly from
our internal structs.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoImplement missing xkb_state_ref and add return value
Ran Benita [Thu, 5 Apr 2012 07:13:24 +0000 (10:13 +0300)]
Implement missing xkb_state_ref and add return value

xkb_state_ref was missing.

Also modify the _ref functions to return the object instead of being
void. This is a useful idiom:

    struct my_object my_object_new(struct xkb_state *state)
    {
        [...]
        my_object->state = xkb_state_ref(state);
        [...]
    }

Essentially "taking" a reference, such that you don't forget to
increment it and it's one line less (see example in our own code).

A case could also be made for _unref to return the object or NULL, but
this is quite uncommon.

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Updated for xkb_keymap changes.]

12 years agoDo not enable yacc trace facility unconditionally
Ran Benita [Fri, 6 Apr 2012 02:09:17 +0000 (05:09 +0300)]
Do not enable yacc trace facility unconditionally

Define DEBUG and set yydebug if you want that.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoConstify a global table and function arguments
Ran Benita [Fri, 30 Mar 2012 22:18:55 +0000 (01:18 +0300)]
Constify a global table and function arguments

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoRemove unused struct definition
Ran Benita [Fri, 6 Apr 2012 02:02:35 +0000 (05:02 +0300)]
Remove unused struct definition

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoReplace remaining use of Atom instead of xkb_atom_t
Ran Benita [Fri, 6 Apr 2012 01:20:35 +0000 (04:20 +0300)]
Replace remaining use of Atom instead of xkb_atom_t

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoFix possible overflow in scanner
Ran Benita [Sun, 8 Apr 2012 17:58:39 +0000 (20:58 +0300)]
Fix possible overflow in scanner

Also reduce the size of scanBuf given that it's allocated on the stack,
and 1024 is enough.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoFix KeyInfo's syms array size calculation
Ran Benita [Sun, 8 Apr 2012 15:46:58 +0000 (18:46 +0300)]
Fix KeyInfo's syms array size calculation

When merging group info from two KeyInfo's, the new size of the keysym
array was off. Fix it to match how it is used a few lines below.

There are also some peripheral fixes, and some comments (took me a
few minutes to get what's going on).

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoSet xorg-macros CFLAGS correctly
Ran Benita [Sun, 8 Apr 2012 12:38:04 +0000 (15:38 +0300)]
Set xorg-macros CFLAGS correctly

A previous commit messed up all warning flags. Oops.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoFix compiler warnings
Ran Benita [Sun, 8 Apr 2012 12:37:36 +0000 (15:37 +0300)]
Fix compiler warnings

(They were not reported, see next commit).
The reset function declaration didn't match its name in the definition;
the _defaults variant matches better with the rest.

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Updated to current master.]

12 years agoUnconstify xkb_rules_names
Daniel Stone [Tue, 3 Apr 2012 16:08:57 +0000 (17:08 +0100)]
Unconstify xkb_rules_names

Since we never return an xkb_rules_names and it's all user-provided
strings, seems a bit harsh to have it const.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoRename xkb_desc to xkb_keymap
Daniel Stone [Tue, 3 Apr 2012 14:14:16 +0000 (15:14 +0100)]
Rename xkb_desc to xkb_keymap

struct xkb_desc was just a hangover from the old XkbDescRec, which isn't
a very descriptive name.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoUse xorg CFLAGS and fix AC_USE_SYSTEM_EXTENSIONS
Daniel Stone [Tue, 3 Apr 2012 13:29:57 +0000 (14:29 +0100)]
Use xorg CFLAGS and fix AC_USE_SYSTEM_EXTENSIONS

autotools was warning that AC_USE_SYSTEM_EXTENSIONS was being called too
late, so move it earlier.  Also shove BASE_CFLAGS into CFLAGS so we get
all the added warning flags from xorg.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoEliminate remaining gcc warnings
Daniel Stone [Tue, 3 Apr 2012 13:26:04 +0000 (14:26 +0100)]
Eliminate remaining gcc warnings

Various one-liners (mostly removing unused variables) to make the code
safe for the full set of warnings used by the xorg macros.

On Debian-based systems, flex generates incorrect code resulting in two
warnings about yy_getcolumn and yy_setcolumn having no previous
declaration despite being non-static.  Fedora carries a patch to fix
this, and a bug has been filed on Debian's flex to add the patch:
http://bugs.debian.org/667027

Aside from this, it's now safe for --enable-strict-compilation.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoDon't parse nonsense Unicode/hex keycodes
Daniel Stone [Tue, 3 Apr 2012 12:57:44 +0000 (13:57 +0100)]
Don't parse nonsense Unicode/hex keycodes

If a keysym was specified as "U1039andsomeextrastuffontheend", return
NoSymbol rather than 0x10001039; similarly, return NoSymbol for
"0xdeadbeefhitherehowsyourdaybeen" rather than 0xdeadbeef.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoRemove unused device_spec
Daniel Stone [Tue, 3 Apr 2012 12:47:08 +0000 (13:47 +0100)]
Remove unused device_spec

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoinclude resets group compatibility modifiers #43091
Daniel Stone [Tue, 3 Apr 2012 12:44:21 +0000 (13:44 +0100)]
include resets group compatibility modifiers #43091

This change makes sure that include does not overwrite previous
compatibility modifier settings when the included files does not
explicitly specify them.

Signed-off-by: Andreas Wettstein <wettstein509@solnet.ch>
Reviewed-by: Daniel Stone <daniel@fooishbar.org>
[Cross-picked from xkbcomp commit 14470719.]

12 years agoAdd cscope output to .gitignore
Daniel Stone [Tue, 3 Apr 2012 12:01:10 +0000 (13:01 +0100)]
Add cscope output to .gitignore

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoProperly document xkb_key_get_syms
Daniel Stone [Tue, 3 Apr 2012 11:48:05 +0000 (12:48 +0100)]
Properly document xkb_key_get_syms

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoConstify a global table
Ran Benita [Fri, 30 Mar 2012 22:18:55 +0000 (01:18 +0300)]
Constify a global table

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoMake temporary fix for stale xkb_atoms
Ran Benita [Fri, 30 Mar 2012 22:26:15 +0000 (01:26 +0300)]
Make temporary fix for stale xkb_atoms

When xkb_free_keymap is called the atoms are all free'd, but action.c
keeps a global copy of interned "true" and "false", which remains stale.

The correct fix is to remove the need for the ActionsInit function
entirely.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoFix bugs to allow multiple keymaps in one process
Ran Benita [Fri, 30 Mar 2012 21:02:12 +0000 (00:02 +0300)]
Fix bugs to allow multiple keymaps in one process

These were several initializations that were forgotten in the previous
memory leak fixes.

Now several xkb_desc's can coexist (relatively) peacefully.

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Only the atom.c is relevant with the new context API.]

12 years agoRemove useless check from xkb_intern_atom
Ran Benita [Fri, 30 Mar 2012 23:26:24 +0000 (02:26 +0300)]
Remove useless check from xkb_intern_atom

The "makeit" variable is always true. Remove it and de-indent.
(Also change the type of the "len" variable to size_t to avoid some
useless casting).

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoAvoid leak in CompileKeymap error path
Ran Benita [Fri, 30 Mar 2012 23:30:33 +0000 (02:30 +0300)]
Avoid leak in CompileKeymap error path

The NULL check is unneeded, and prevented the atoms from being free'd.

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Updated for xkb_map_unref.]

12 years agomakekeys: use correct format strings
Ran Benita [Sun, 1 Apr 2012 13:48:34 +0000 (16:48 +0300)]
makekeys: use correct format strings

The new glibc (2.15) appear to cause trouble, particularly the sscanf
call, where makekeys will output empty hash tables. Using the appropriate
macros from inttypes.h makes it work again.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoAdd a NULL check before before strcmp'ing
Ran Benita [Sun, 1 Apr 2012 13:51:48 +0000 (16:51 +0300)]
Add a NULL check before before strcmp'ing

The names array can have NULL entries for some virtual modifier indexes.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agomakekeys: free strings before exiting
Ran Benita [Sun, 1 Apr 2012 13:57:31 +0000 (16:57 +0300)]
makekeys: free strings before exiting

Just to make valgrind happy.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoUse straight strtok_r instead of wrapper
Ran Benita [Sun, 1 Apr 2012 14:56:17 +0000 (17:56 +0300)]
Use straight strtok_r instead of wrapper

strtok_r is POSIX-2001 and should be fine.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoAvoid use of partly initialized VModInfo in error path
Ran Benita [Sun, 1 Apr 2012 14:07:56 +0000 (17:07 +0300)]
Avoid use of partly initialized VModInfo in error path

Such as:

Compiling path: ./test/data/bad.xkb  mapName:
==1300== Conditional jump or move depends on uninitialised value(s)
==1300==    at 0x4E46166: HandleVModDef (vmod.c:90)
==1300==    by 0x4E3FEC9: HandleKeyTypesFile (keytypes.c:1035)
==1300==    by 0x4E3FBE1: HandleIncludeKeyTypes.constprop.11 (keytypes.c:387)
==1300==    by 0x4E401DD: HandleKeyTypesFile (keytypes.c:1022)
==1300==    by 0x4E3FBE1: HandleIncludeKeyTypes.constprop.11 (keytypes.c:387)
==1300==    by 0x4E401DD: HandleKeyTypesFile (keytypes.c:1022)
==1300==    by 0x4E4026F: CompileKeyTypes (keytypes.c:1150)
==1300==    by 0x4E3DF9B: CompileKeymap (keymap.c:169)
==1300==    by 0x4E465E9: compile_keymap (xkbcomp.c:205)
==1300==    by 0x4E46BE4: xkb_compile_keymap_from_file (xkbcomp.c:290)
==1300==    by 0x400B37: test_file (filecomp.c:47)
==1300==    by 0x4008E3: main (filecomp.c:90)
==1300==  Uninitialised value was created by a stack allocation
==1300==    at 0x4E3FB3F: HandleIncludeKeyTypes.constprop.11 (keytypes.c:366)

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoFail gracefully on failure to find component include
Daniel Stone [Thu, 29 Mar 2012 16:39:11 +0000 (17:39 +0100)]
Fail gracefully on failure to find component include

If we can't find the component of the include file we're looking for,
make sure we don't return success when we meant failure, segfault, or
spectacularly leak everything.

Tested with incorrect component includes for keycodes, compat, symbols,
and types.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
Reported-by: David Herrmann <dh.herrmann@googlemail.com>
12 years agoTests: Release context on failure to build keymap
Daniel Stone [Thu, 29 Mar 2012 16:38:44 +0000 (17:38 +0100)]
Tests: Release context on failure to build keymap

No practical effect since they exit(1) regardless, but it keeps valgrind
happy.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoFull support for multiple keysyms per level
Daniel Stone [Thu, 29 Mar 2012 15:31:09 +0000 (16:31 +0100)]
Full support for multiple keysyms per level

Which also involved moving the global symbol map to be per-key instead;
this should probably be split out into a separate commit.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoParsing support for multiple keysyms per entry
Daniel Stone [Thu, 29 Mar 2012 00:57:06 +0000 (01:57 +0100)]
Parsing support for multiple keysyms per entry

Instead of allowing only one keysym per level per group, do as the
external API indicates and allow multiples.  The existing syntax is:
    key <AD01> { [ q, Q ] };
where the new syntax is:
    key <AD01> { [ q, Q, { H, E, L, L, O },
                 { Y, E, S, space, T, H, I, S, space, I, S, space, D, O, G } };

to make the key in the extreme top left of the keyboard do pretty
surprising things in levels 3 and 4.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoDon't leak strings in merge declarations
Daniel Stone [Thu, 29 Mar 2012 00:49:55 +0000 (01:49 +0100)]
Don't leak strings in merge declarations

Introduced by the reentrant scanner commit.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoDon't leak include_paths in context
Daniel Stone [Thu, 29 Mar 2012 00:41:40 +0000 (01:41 +0100)]
Don't leak include_paths in context

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoRemove unused rtrnValue declaration
Daniel Stone [Wed, 28 Mar 2012 18:50:45 +0000 (19:50 +0100)]
Remove unused rtrnValue declaration

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoMake parser and scanner reentrant
Ran Benita [Tue, 27 Mar 2012 21:41:22 +0000 (22:41 +0100)]
Make parser and scanner reentrant

All global state is removed from the parser and scanner.
This makes use of the standard facilities in Bison and Flex for
reentrant/pure scanner/lexer and location tracking.

Signed-off-by: Ran Benita <ran234@gmail.com>
[daniels: Updated to current sources.]

12 years agoUse xkb_contexts in keymap compilation
Daniel Stone [Tue, 27 Mar 2012 16:22:35 +0000 (17:22 +0100)]
Use xkb_contexts in keymap compilation

Primarily for the include path, but also for the logging in future.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoAdd new context API
Daniel Stone [Tue, 27 Mar 2012 15:59:01 +0000 (16:59 +0100)]
Add new context API

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoUpdate xkb_filter_group_lock_func for xkb_key_direction
Daniel Stone [Tue, 27 Mar 2012 15:15:06 +0000 (16:15 +0100)]
Update xkb_filter_group_lock_func for xkb_key_direction

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoFix distcheck with non-recursive Makefiles
Daniel Stone [Tue, 27 Mar 2012 14:51:28 +0000 (15:51 +0100)]
Fix distcheck with non-recursive Makefiles

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoFix build for X11 modifier masks
Daniel Stone [Tue, 27 Mar 2012 14:50:59 +0000 (15:50 +0100)]
Fix build for X11 modifier masks

Exposed by include path changes, oops.

Signed-off-by: Daniel Stone <daniel@fooishbar.org>
12 years agoHandle group lock actions
Ran Benita [Fri, 23 Mar 2012 15:50:37 +0000 (17:50 +0200)]
Handle group lock actions

The spec is simple here, as this action has no effect on key releases.

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoMake build non-recursive
Ran Benita [Sat, 24 Mar 2012 02:48:31 +0000 (04:48 +0200)]
Make build non-recursive

Unify all the different Makefile.am into a single short top level one
(the test/Makefile.am file is left intact though).

This makes the build system simpler to look and should encourage
unifying more currently-disparate code.

Some further motivation can be found in this page:
http://www.flameeyes.eu/autotools-mythbuster/automake/nonrecursive.html

Signed-off-by: Ran Benita <ran234@gmail.com>
12 years agoDefine our own NoSymbol value and use it
Ran Benita [Sat, 24 Mar 2012 11:27:48 +0000 (13:27 +0200)]
Define our own NoSymbol value and use it

Since we have our own xkb_keysym_t type, it makes sense to have our own
NoSymbol value instead of the one from X11/X.h.

Signed-off-by: Ran Benita <ran234@gmail.com>