admin/unidata/unidata-gen.el (unidata-split-name): Don't give any NAME
to <control> characters: the Unicode Standard says they have no
name.
doc/lispref/nonascii.texi (Character Properties): NAME or OLD-NAME
properties can be nil (there's no empty string).
* admin/grammars/c.by (expr-binop): Add MOD.
(variablearg): Add 'opt-assign'.
(variablearg, varnamelist): Add default values so that it can be
later expanded into the tag.
(opt-stuff-after-symbol): Rename to 'brackets-after-symbol' and
remove empty match.
(multi-stage-dereference): Adapt to above rename.
(unaryexpression): Use 'symbol' instead of 'namespace-symbol',
since the latter also leads to an empty match at the end which
would make this too greedy.
(variablearg-opt-name): Support parsing of function pointers
inside an argument list.
* semantic/analyze.el
(semantic-analyze-find-tag-sequence-default): Always add scope to
the local miniscope for each type. Otherwise, structure tags are
not analyzed correctly. Also, always search the extended
miniscope even when not dealing with types.
* semantic/ctxt.el (semantic-get-local-variables-default): Also
try to parse local variables for buffers which are currently
marked as unparseable. Otherwise, it is often impossible to
complete local variables.
* semantic/scope.el (semantic-analyze-scoped-types-default): If we
cannot find a type in the typecache, also look into the the types
we already found. This is necessary since in C++, a 'using
namespace' can be dependend on a previous one.
(semantic-completable-tags-from-type): When creating the list of
completable types, pull in types which are referenced through
'using' statements, and also preserve their filenames.
* semanitc/bovine/c.el (semantic/analyze/refs): Require.
(semantic-analyze-tag-references): New override. Mainly copied
from the default implementation, but if nothing could be found (or
just the tag itself), drop all namespaces from the scope and
search again. This is necessary for implementations which are
defined outside of the namespace and only pull those in through
'using' statements.
(semantic-ctxt-scoped-types): Go through all tags around point and
search them for using statements. In the case for using
statements outside of function scope, append them in the correct
order instead of using 'cons'. This is important since using
statements may depend on previous ones.
(semantic-expand-c-tag-namelist): Do not try to parse struct
definitions as default values. The grammar parser seems to return
the point positions slightly differently (as a cons instead of a
list). Also, set parent for typedefs to 'nil'. It does not
really make sense to set a parent class for typedefs, and it can
also lead to endless loops when calculating scope.
(semantic-c-reconstitute-token): Change handling of function
pointers; instead of seeing them as variables, handle them as
functions with a 'function-pointer' attribute. Also, correctly
deal with function pointers as function arguments.
(semantic-c-reconstitute-function-arglist): New function to parse
function pointers inside an argument list.
(semantic-format-tag-name): Use 'function-pointer' attribute
instead of the old 'functionpointer-flag'.
(semantic-cpp-lexer): Use new `semantic-lex-spp-paren-or-list'.
* semantic/bovine/gcc.el (semantic-gcc-setup): Add 'features.h' to
the list of files whose preprocessor symbols are included. This
pulls in things like __USE_POSIX and similar.
* semantic/format.el (semantic-format-tag-prototype-default):
Display default values if available.
* semantic/analyze/refs.el (semantic-analyze-refs-impl)
(semantic-analyze-refs-proto): Add 'default-value' as ignorable in
call to `semantic-tag-similar-p'.
* semantic/db-mode.el (semanticdb-semantic-init-hook-fcn): Always
set buffer for `semanticdb-current-table'.
* semantic/db.el (semanticdb-table::semanticdb-refresh-table): The
previous change turned up a bug in this method. Since the current
table now correctly has a buffer set, the first clause in the
`cond' would be taken, but there was a `save-excursion' missing.
* semantic/lex-spp.el (semantic-c-end-of-macro): Declare.
(semantic-lex-spp-token-macro-to-macro-stream): Deal with macros
which open/close a scope. For this, leave an overlay if we
encounter a single open paren and return a semantic-list in the
lexer. When this list gets expanded, retrieve the old position
from the overlay. See the comments in the function for further
details.
(semantic-lex-spp-find-closing-macro): New function to find the
next macro which closes scope (i.e., has a closing paren).
(semantic-lex-spp-replace-or-symbol-or-keyword): Go to end of
closing macro if necessary.
(semantic-lex-spp-paren-or-list): New lexer to specially deal with
parens in macro definitions.
* semantic/decorate/mode.el (semantic-decoration-mode): Do not
decorate available tags immediately but in an idle timer, since
EDE will usually not be activated yet, which will make it
impossible to find project includes.
* semantic/decorate/include.el
(semantic-decoration-on-includes-highlight-default): Remove
'unloaded' from throttle when decorating includes, otherwise all
would be loaded. Rename 'table' to 'currenttable' to make things
clearer.
* ede/linux.el (cl): Require during compile.
* ede/linux.el (project-linux-build-directory-default)
(project-linux-architecture-default): Add customizable variables.
(ede-linux-project): Add additional slots to track Linux-specific
information (out-of-tree build directory and selected
architecture).
(ede-linux--get-build-directory, ede-linux--get-archs)
(ede-linux--detect-architecture, ede-linux--get-architecture)
(ede-linux--include-path): Added function to detect Linux-specific
information.
(ede-linux-load): Set new Linux-specific information when creating
a project.
(ede-expand-filename-impl): Use new and more accurate include
information.
* semantic/scope.el (semantic-calculate-scope): Return a clone of
the scopecache, so that everyone is working with its own (shallow)
copy. Otherwise, if one caller is resetting the scope, it would
be reset for all others working with the scope cache as well.
* build-aux/dir_top: Move here from admin/.
* build-aux/make-info-dir: New script.
* Makefile.in (bootstrap-clean): Delete info/.
(info-dir, ${srcdir}/info/dir): New rules.
(info): Also make info-dir.
(check-info): Rename from check-info-dir.
Instead of info/dir entries, check @dircategory in info/*.info.
* make-dist: Use `info' rule rather than `info-real'.
No more info/COPYING (not even the right license for info/ files).
* info/: Remove from repository.
* admin/update_autogen (info_dir):
Use dir_top from build-aux/ rather than admin/.
* .bzrignore: Ignore info/ altogether.
* doc/emacs/Makefile.in: Comment.
This scorches the earth and waits for spring;
see Ted Zlatanov and Stefan Monnier in
<http://lists.gnu.org/archive/html/emacs-devel/2013-12/msg00323.html>.
* configure.ac (with_openssl_default, HAVE_LIB_CRYPTO): Remove.
Do not say whether Emacs is configured to use a crypto library,
since it's no longer an option.
(gl_CRYPTO_CHECK): Define a dummy.
* lib/gl_openssl.h, m4/gl-openssl.m4: Remove.
* admin/merge-gnulib: Remove lib/gl_openssh.h and m4/gl-openssl.m4.
* src/Makefile.in (LIB_CRYPTO): Remove.
(LIBES): Don't use it.
src/bidi.c (bidi_get_type, bidi_get_category): Handle the isolate
directional control characters. Update type and category
determination according to the UBA from Unicode v6.3.
(bidi_category_t): New category EXPLICIT_FORMATTING.
src/dispextern.h (bidi_type_t): Update to include new bidirectional
properties introduced with Unicode v6.3.
admin/unidata/unidata-gen.el (unidata-prop-alist): Update bidi-class
to include the new isolate-related classes introduced with Unicode
v6.3.
(unidata-encode-val): Accept an additional optional argument, a
warning message to emit when UnicodeData.txt defines bidi-class
values that are not in unidata-prop-alist. Add a comment
explaining what should maintainers do if/when such a warning ever
appears.
(unidata-gen-table): Call unidata-encode-val with 3rd arg non-nil
when generating uni-bidi.el.
Note: I used Emacs 24.3 to generate the uni-*.el files, owing to
unexplained (possibly harmless) differences that the current trunk's
use of lexical-binding for subr.el causes. Ref:
http://debbugs.gnu.org/15825#29
* admin/unidata/BidiMirroring.txt, admin/unidata/UnicodeData.txt:
Update to 6.3.0.
* lisp/international/uni-bidi.el, lisp/international/uni-category.el:
* lisp/international/uni-name.el, lisp/international/uni-numeric.el:
Regenerate.
Ref: http://lists.gnu.org/archive/html/emacs-devel/2013-10/msg00806.html
* autogen: Remove directory. Move update_autogen to admin/.
* autogen.sh: Remove reference to copy_autogen.
* GNUmakefile (configure):
* Makefile.in (bootstrap): Do not try to run copy_autogen.
* config.bat: Use msdos/autogen rather than autogen.
* admin/update_autogen: Move here from ../autogen.
(usage): Update. Remove -l, add -A.
(autogendir): New variable.
(ldefs_flag): Default to set.
(genfiles): Reduce to only ms-dos relevant files.
(main): Make checking autogen sources optional.
Make copying of autogen files optional.
* msdos/autogen/config.in:
* msdos/autogen/Makefile.in: Move here from ../autogen.
* nt/INSTALL: Remove reference to copy_autogen.
* nt/config.nt: Comment.
* admin/unidata/unidata-gen.el (unidata-gen-files): Use pop.
Also take the output directory as an argument.
* admin/unidata/Makefile.in: Simplify now that unidata-gen-files takes
the output directory as an argument (no need to cd, etc).
(abs_srcdir, abs_builddir): Remove.
(abs_top_builddir): Replace by top_builddir.
(${DSTDIR}/charprop.el): No need to cd. Pass dest as argument.
(${DSTDIR}/charprop.el, charprop.el):
No need to pass unidata.txt as argument.
(unidata-gen-table-name, unidata-gen-table-decomposition)
(unidata-gen-files): Use unidata--ensure-compiled.
Add FSF copyright years based on when this file first appeared in
Emacs trunk.
Do this by using the Gnulib modules for this.
This should generate faster code on non-GCC, non-MSC platforms,
and make the code a bit more portable, at least in theory.
* admin/merge-gnulib (GNULIB_MODULES): Add count-one-bits
and count-trailing-zeros.
* lib/count-one-bits.c, lib/count-one-bits.h:
* lib/count-trailing-zeros.c, lib/count-trailing-zeros.h:
* m4/count-one-bits.m4, m4/count-trailing-zeros.m4:
New files, copied from gnulib.
* lib/gnulib.mk, m4/gnulib-comp.m4: Regenerate.
* nt/gnulib.mk: Merge changes from lib/gnulib.mk.
* src/data.c: Include <count-one-bits.h>, <count-trailing-zeros.h>.
(USE_MSC_POPCOUNT, POPCOUNT_STATIC_INLINE)
(NEED_GENERIC_POPCOUNT, popcount_size_t_generic)
(popcount_size_t_msc, popcount_size_t_gcc):
Remove; now done by Gnulib.
(popcount_size_t): Now a macro that defers to Gnulib.
(count_trailing_zero_bits): Return int, for consistency with
Gnulib and because Emacs prefers signed to unsigned int.
Don't assume that size_t is either unsigned int or unsigned long
or unsigned long long.
(size_t_to_host_endian): Do not assume that size_t is either
exactly 32 or exactly 64 bits wide.
* src/lisp.h (BITS_PER_SIZE_T): Define consistently with BITS_PER_LONG
etc., so that it's now an enum constant, not a macro.
No need to assume that it's either 32 or 64.
Fixes: debbugs:15550
