Sun Feb 6 21:59:09 2005 Manish Singh <yosh@gimp.org>
* glib/abicheck.sh: feed the contents of config.h and glibconfig.h
into the preprocessor, since glib.symbols could need #defines
from them.
Sun Feb 6 12:09:55 2005 Manish Singh <yosh@gimp.org>
* glib/gatomic.c: Don't use matching constraints for asm mem
parameters. This makes the code match with current glibc, and
quiets some warnings with gcc 3.4.
2005-02-05 Tor Lillqvist <tml@novell.com>
* configure.in (G_LIBS_EXTRA)
* glib/giowin32.c: I think we can include <winsock2.h> and link
with -lws2_32. Only Windows 95 originally didn't have Winsock 2,
and it's available as an update for it. Not that we use any actual
Winsock 2 -only features, but still, simpler to link directly with
ws2_32.dll instead of going through wsock32.dll.
2005-02-05 Tor Lillqvist <tml@novell.com>
* glib/gfileutils.c (g_file_get_contents): In the Win32 ABI
stability version, call g_file_get_contents_utf8(), not
itself. (#166386)
2005-02-04 Matthias Clasen <mclasen@redhat.com>
* configure.in: Define a HAVE_GNUC_VISIBILITY automake
conditional.
* glib/Makefile.am (TESTS): Only check the abi if
the compiler supports visibility, otherwise we
know there will be (harmless) extra symbols.
(#166181, Ed Avis)
2005-02-02 Tor Lillqvist <tml@novell.com>
* glib/gwin32.h: Remove unnecessary MSVC-only typedef for
pid_t. GLib does not pretend to be a POSIX emulation library.
2005-02-02 Tor Lillqvist <tml@novell.com>
* glib/gfileutils.c: Don't use potentially incorrect errno in
several places. (#165951)
(g_mkstemp): Set errno appropriately.
2005-02-02 Tor Lillqvist <tml@novell.com>
* glib/gstdio.c: In the Win32 part of all functions, handle
invalid UTF-8 more robustly. Don't pass the resulting NULL
wide-char or codepage strings on to C library functions, but bail
out early. (#166084)
2005-02-01 Matthias Clasen <mclasen@redhat.com>
* tests/keyfile-test.c (test_key_remove): Add test case for
key removal.
* glib/gkeyfile.c (g_key_file_remove_key): Actually remove
the key from the list of pairs. (#165980, David Hoover)
2005-02-01 Matthias Clasen <mclasen@redhat.com>
* tests/keyfile-test.c (test_group_remove): Add test case for
group removal.
* glib/gkeyfile.c (g_key_file_remove_group): Don't segfault
if the group doesn't exist. (#165887, Mathias Hasselmann)
2005-01-28 Christian Rose <menthos@menthos.com>
* configure.in: Added "tl" to ALL_LINGUAS.
* po/tl.po: Added Tagalog translation by
Eric Pareja <xenos@upm.edu.ph>.
2005-01-24 Matthias Clasen <mclasen@redhat.com>
* tests/markups/fail-40.gmarkup: Add a test with a long entity
name.
* glib/gmarkup.c (unescape_text_state_inside_entity_name): Don't
copy the entity name into a short buffer of fixed length. Instead,
compare it in place with strncmp(), and do a full strdup() in the
error path. (#165100, Simon Budig)
2005-01-22 Tor Lillqvist <tml@novell.com>
* glib/gdate.c (g_date_set_time): Don't g_assert that localtime()
returns non-NULL. It does return NULL at least on Win32 if you
pass it a negative time_t, which although wrong, shouldn't cause
an application to abort. Instead return 2000-01-01 as a default
date. Print a warning unless G_DISABLE_CHECKS. (#164622)
2005-01-20 Matthias Clasen <mclasen@redhat.com>
* glib/gkeyfile.c (g_key_file_parse_value_as_integer): Don't
interpret leading zeros as octal. The specification requires
%f parsing, and %f doesn't allow octal.
* tests/keyfile-test.c: Add some more tests.
2005-01-20 Matthias Clasen <mclasen@redhat.com>
* tests/keyfile-test.c: Add some more tests.
* glib/gkeyfile.c (g_key_file_get_keys): Return keys in
the order found in the file, rather than the opposite.
(g_key_file_parse_value_as_string): Fix error reporting.
2005-01-19 Matthias Clasen <mclasen@redhat.com>
* glib/gkeyfile.c (g_key_file_parse_data): Accept \r\n as
line end. (#163030, Bastian Nocera)
(g_key_file_load_from_data): Make -1 work as a size.
* tests/keyfile-test.c: Some unit tests for the keyfile
parser.
* tests/Makefile.am (test_programs): Add keyfile-test.
2005-01-19 Tor Lillqvist <tml@novell.com>
* glib/gbacktrace.c (g_on_error_stack_trace): On Win32 call
G_BREAKPOINT() if being debugged.
* glib/gmessages.c (g_logv): For fatal messages on Win32, use
G_BREAKPOINT() also without G_ENABLE_DEBUG, but only if being
debugged. Suggestion by Ulf Lamping. Otherwise use FatalAppExit().
2005-01-18 Matthias Clasen <mclasen@redhat.com>
* glib/gqsort.c: Don't include alloca.h. It is not needed,
and does not exist on win32. (#164468, Kazuki IWAMOTO)
2005-01-11 Owen Taylor <otaylor@redhat.com>
* configure.in: Add gmodule-export-2.0.pc to AC_CONFIG_FILES.
* glib/Makefile.am gobject/Makefile.am gthread/Makefile.am:
Use grep | head -n 1 instead of grep -m 1 which doesn't work
with slightly older versions of GNU grep in
install-libtool-import-lib rule.
2005-01-09 Hans Breuer <hans@breuer.org>
* glib/gfileutils.c : make it compile with mvc6 default sdk,
(#define INVALID_FILE_ATTRIBUTES, FILE_ATTRIBUTE_DEVICE)
2005-01-07 Matthias Clasen <mclasen@redhat.com>
* configure.in:
* Makefile.am: Generate and distribute gmodule-export-2.0.pc,
which is currently just a copy of gmodule-2.0.pc, but makes
it explicit that it adds --export-dynamic.
* gmodule-export-2.0.pc.in: Copy of gmodule-2.0.pc.in.
2005-01-06 Tor Lillqvist <tml@iki.fi>
* glib/Makefile.am
* gobject/Makefile.am: glib.def and gobject.def are generated, not
in $(srcdir). (#163143, J. Ali Harlow)
2005-01-04 Tor Lillqvist <tml@iki.fi>
* glib/guniprop.c (get_locale_type): Use g_win32_getlocale()
instead of setlocale() on Windows. setlocale() returns strings
like "Turkish_Turkey".
2005-01-04 Matthias Clasen <mclasen@redhat.com>
Fix the inlining magic. (#157536, Jens Hatlak, and
#149907, Morten Welinder)
* configure.in: Define G_CAN_INLINE in glibconfig.h
* glib/gutils.h: Streamline the inlining magic a bit,
don't use extern when implementing the non-inlined
version.
2005-01-01 J. Ali Harlow <ali@juiblex.co.uk>
* configure.in: Don't use AC_TRY_RUN to test for long long format
when using the MSVCRT.DLL runtime since we know the answer anyway
and it causes some mild inconvience when cross compiling.
* README.win32: Add a reference to the cross compiling section
of the reference manual.
* docs/reference/glib/cross.sgml: MinGW uses MSVCRT.DLL which
requires %I64i instead of %lli. (#161306)
2005-01-01 Tor Lillqvist <tml@iki.fi>
* glib/glib.symbols
* glib/gutils.h
* glib/gutils.c: Make also g_getenv(), g_setenv(), g_unsetenv()
and g_find_program_in_path() take and return UTF-8 strings on
Win32. Implement DLL ABI backward compatility for them, too. Move
all the DLL ABI stability wrappers to the end of the file. Use
wide character API when available in inner_find_program_in_path().
* glib/gfileutils.c: With the UTF-8ness of g_getenv() above, just
use g_getenv() to get PATHEXT. (Yeah, it's probably overdoing it
to consider somebody actually having anything else than ASCII
in PATHEXT, but...)
2004-12-31 Tor Lillqvist <tml@iki.fi>
* glib/gfileutils.c (g_file_test): Rewrite the Win32 version to
use GetFileAttributes() instead of stat(). stat() is unreliable
for corner cases like '\\server\share' or '.\'. Part of fixing
#161797. When testing for executability, in addition to the fixed
set of executable file name extensions also check the PATHEXT
environment variable.
2004-12-30 Tor Lillqvist <tml@iki.fi>
* glib/gutils.c (g_get_current_dir): Use GetCurrentDirectory()
directly for simpler buffer length management. I don't trust
getcwd() getting it right all the time.
2004-12-30 Tor Lillqvist <tml@iki.fi>
* glib/gutils.c (g_path_skip_root): Require UNC paths to start
with exactly two slashes.
(g_get_current_dir): Use wide character API when available.
(g_path_get_dirname): Handle UNC paths better. Part of fix for
#161797.
2004-12-27 Matthias Clasen <mclasen@redhat.com>
* configure.in: Add an OS_LINUX conditional.
* glib/Makefile.am (TESTS): Only check the ABI on
linux, since the shell script is not portable, and libtool
deficiencies may distort the ABI on other platforms. (#161741)
2004-12-21 Matthias Clasen <mclasen@redhat.com>
* glib/goption.c (print_entry): Don't show
the special G_OPTION_REMAINING entry. (#161934,
Matthew F. Barnes)
2004-12-20 Matthias Clasen <mclasen@redhat.com>
* tests/option-test.c: Add some tests for '--'
stripping.
* glib/goption.c (g_option_context_parse): Don't
strip '--' if it would be needed by a second option
parser. (#161701)
2004-12-19 Matthias Clasen <mclasen@redhat.com>
* glib/goption.c (g_option_context_parse): Call
post-parse hooks also if argv is NULL. (#161668,
Marcin Krzyzanowski)
2004-12-12 Tor Lillqvist <tml@iki.fi>
* glib/glib.symbols
* glib/gconvert.[ch]: Implement Windows DLL ABI stability also for
g_filename_{to,from}_uri().
2004-12-11 Tor Lillqvist <tml@iki.fi>
* glib/gstdio.c (g_rename, g_unlink, g_remove): Add doc comments
about Windows restrictions on renaming and removing.
(g_remove): Mimic POSIX remove() on Windows: Try also to rmdir if
removal as a file fails. Document this.
* glib/gstdio.h: Clarify comment about file name encoding on
Windows.
2004-12-09 Matthias Clasen <mclasen@redhat.com>
* glib/goption.c (print_help): Don't print help options
if the options of a specific group have been
requested. (#160645, Glynn Foster)
2004-12-07 Matthias Clasen <mclasen@redhat.com>
* glib/gutils.c (g_get_language_names): Update the returned
value after locale changes. (#160271, Christian Persch)
(_g_utils_thread_init): Initialize the language name cache
before going threaded.
* glib/gthread.c (g_thread_init_glib): Call _g_utils_thread_init().
* glib/gthreadinit.h: Add _g_utils_thread_init().
2004-12-06 Tor Lillqvist <tml@iki.fi>
* glib/giowin32.c (g_io_win32_finalize): Fix a write to freed
memory: Unlock the channel's mutex before unrefing the channel.
2004-12-05 Tor Lillqvist <tml@iki.fi>
* glib/glib.symbols
* glib/gwin32.[hc]: Implement DLL ABI stability also for
g_win32_get_package_installation_directory()) and
g_win32_get_package_installation_subdirectory().
2004-12-05 Tor Lillqvist <tml@iki.fi>
* glib/Makefile.am (install-libtool-import-lib): Remove the DLL
binary compatibilty entries from the import library. They aren't
used by newly compiled code, so no need to have them in the import
library. (The "PRIVATE" keyword in the .def file is supposed to do
this, but not implemented in the GNU linker yet.)
2004-12-02 Matthias Clasen <mclasen@redhat.com>
* glib/glib.symbols: Protect the the various push/pop allocator
functions by #ifndef DISABLE_MEM_POOLS. (#160215, Frederic Crozat)
* glib/makegalias.pl: Support #ifndef'ed sections.
2004-11-30 Tor Lillqvist <tml@iki.fi>
* glib/Makefile.am (glib.def)
* gobject/Makefile.am (gobject.def): gcc -E complains "linker
unused because linking not done" if told to preprocess an
*.symbols file. Use '-' and redirection to pass it as standard
input instead.
2004-11-30 Tor Lillqvist <tml@iki.fi>
* glib/gspawn-win32-helper.c (WinMain): Also check errno to detect
true errors from spawn*(). (#157258, reported by Bruce Hochstetler)
2004-11-29 Matthias Clasen <mclasen@redhat.com>
* glib/gmarkup.c: Remove leftover noinline attributes.
(is_name_start_char, is_name_char): Avoid possible reads
beyond the end of g_ascii_table.
Sun Nov 28 13:13:56 2004 Manish Singh <yosh@gimp.org>
* glib/abicheck.sh: filter G_GNUC before PRIVATE so $ is still true.
* tests/utf8-validate.c: cast pointer math to gint for error print
message.
2004-11-28 Tor Lillqvist <tml@iki.fi>
* glib/gutils.[hc]
* glib/glib.symbols: [Win32] Make also g_get_user_name() and
g_get_real() name return UTF-8. As for the similar changes to fix
#101792, for DLL ABI stability we use preprocessor defines to get
the new UTF-8 versions, and keep the old names for versions
returning strings in the system codepage.
Fix g_get_tmp_dir() and g_get_home_dir() to actually return UTF-8
as was intended in the fix for bug #101792. (#159664, noticed by
Robert Ögren)
2004-11-27 Matthias Clasen <mclasen@redhat.com>
* glib/gspawn.c (g_spawn_async_with_pipes): Update the @flags
documentation with references to GChildWatch and
g_spawn_close_pid(). (#136255, noted by Owen Taylor)
2004-11-25 Matthias Clasen <mclasen@redhat.com>
* tests/strtod-test.c (main): Add a testcase for the previous fix.
* glib/gstrfuncs.c (g_ascii_strtod): Make it work again for floats
starting with a decimal point, like .75 (#156421, Hans Breuer)
