2000-09-01 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
* gstring.c (g_string_free): Use g_return_val_if_fail instead of
g_return_if_fail, as the function now is supposed to return
something.
2000-09-01 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
* glib.h: Include gerror.h before it is used for some g_thread_*
functions.
* gthread.c, gthreadpool.c, glib.h: Enable error reporting for
thread creation, namly for g_thread_create, g_thread_pool_new,
g_thread_pool_push and g_thread_pool_set_max_threads.
* tests/thread-test.c, tests/threadpool-test.c: Adapted
accordingly.
* gthread-posix.c (g_thread_create_posix_impl): Use GError to
report errors.
2000-09-01 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
* gerror.c, gerror.h (g_propagte_error): Added function
g_propagte_error to hand over local errors to the calling
function.
2000-08-31 Tor Lillqvist <tml@iki.fi>
* glib.h
* glib.def
* giowin32.c (g_io_channel_win32_make_pollfd): New function, to
make a GPollFD from a GIOChannel. Creates the events and starts
the reader thread if necessary.
2000-08-27 Tor Lillqvist <tml@iki.fi>
Hmm, the parameter to g_io_channel_win32_poll() should be
a GPollFD array, not GIOChannel. That way the caller can know which
channels cause the callback.
2000-08-27 Tor Lillqvist <tml@iki.fi>
* giowin32.c (g_io_channel_win32_poll): New function, otherwise
like g_io_channel_win32_wait_for_condition(), but accept several
GIOChannels.
(g_io_channel_win32_wait_for_condition): Call
g_io_channel_win32_poll().
* glib.h: Declare g_io_channel_win32_poll().
* gwin32.c (g_win32_error_message): Don't believe return value
from FormatMessage.
2000-08-25 Elliot Lee <sopwith@redhat.com>
* glib.h, gunicode.h, gmodule/gmodule.h:
Mark the following functions G_GNUC_CONST (to allow
betteroptimization) because their results are a function of only their parameters:
g_int_hash, g_int_equal, g_direct_hash, g_direct_equal, g_quark_to_string,
g_date_is_leap_year, g_date_days_in_month, g_date_monday_weeks_in_year,
g_date_sunday_weeks_in_year, g_spaced_primes_closest, g_unichar_is*,
g_unichar_to*, g_unichar_*digit_value, g_unichar_type
Mon Aug 21 14:46:23 2000 Owen Taylor <otaylor@redhat.com>
* tests/gio-test.c: Fix a couple of trivial bugs that
were causing warnings.
Mon Aug 21 14:39:36 2000 Owen Taylor <otaylor@redhat.com>
* glib.h: Use C99 varargs macros where possible
(check __STDC_VERSION__), otherwise, on gcc, use an alternate
form of gcc varargs which is more likely
to be supported going forward. (Based on some code
from Raja Harinath)
* glib.h:
* garray.c: (g_array_free), (g_ptr_array_free),
(g_byte_array_free): Return the data left behind.
* gstring.c: (g_string_free): Return the data left behind.
Changed the free calls that leave data behind so they
return a pointer to the left-behind data, NULL if told not
to leave anything behind. This makes these calls easier
to use correctly, without any incompatible API change for
callers that don't know about the return value. Of course,
it would be even clearer if the free calls weren't dual-purpose
in the first place.
2000-08-12 Tor Lillqvist <tml@iki.fi>
* giowin32.c: Some indentation and spacing fixes. Add some more
logging.
(g_io_win32_add_watch): New function, with common code from
g_io_win32_fd_add_watch and g_io_win32_sock_add_watch. Don't start
more than one reader thread for a GIOChannel. We should obviously
have just one reader thread reading a file descriptor or socket.
2000-08-07 Tor Lillqvist <tml@iki.fi>
* tests/gio-test.c (shutdown_source): New function, that calls
g_source_remove(). Check return value of g_source_remove(), and
decrement running subprocess counter only if g_source_remove()
actually did remove the source.
(recv_message): Call shutdown_source() on EOF condition. Return
FALSE on G_IO_HUP and G_IO_ERR condition. Fix printf format typo.
Sun Aug 6 20:06:02 2000 Tim Janik <timj@gtk.org>
* gmessages.c (g_log_domain_check_free): keep *last updated while
running through the domain list, so we don't screw up the removal,
patch provided by Gady Kozma <gadykozma@hotmail.com>.
Sun Aug 6 20:03:41 2000 Tim Janik <timj@gtk.org>
* gmessages.c (g_log_remove_handler): keep *last updated while running
through the handler list, so we don't screw up the removal.
Sun Jul 30 16:54:13 2000 Owen Taylor <otaylor@redhat.com>
* gunicode.h: Fix stray character
* gutf8.c (g_unichar_to_utf8): Allow outbuf to be NULL, in
which case we just compute the length.
2000-07-31 Tor Lillqvist <tml@iki.fi>
* giowin32.c (buffer_read): The code didn't compile (must
have been sleepy when committing). "return" instead of "break"
(g_io_win32_fd_add_watch): Cannot check if the file descriptor is
readable by calling ReadFile to read zero bytes. ReadFile blocks
on NT even if trying to read nothing at all. So, don't check if
file descriptor is readable; assume this function isn't called
otherwise.
Sun Jul 30 10:44:16 2000 Tim Janik <timj@gtk.org>
* gmain.c (g_get_current_time): fix tor's recent changes which
got rid of a required variable in the non-windows path.
2000-07-30 Tor Lillqvist <tml@iki.fi>
* giowin32.c: Compile in the debugging code all the time, but only
output debug messages if told so. Add (unadvertised) function to
turn on/off debug messages for a channel.
(buffer_read): Don't loop. It is expected behaviour to return a
short read occasionally, for instance when reading from
pipes. It's the calling code that should loop if it *knows* how
much the writer has written.
* tests/gio-test.c: Correct the program's name in the output.
(recv_message): Loop calling g_io_channel_read() (in a new
function read_all()) until we have all the bytes we want (that we
know the writer has written/will write).
2000-07-30 Tor Lillqvist <tml@iki.fi>
Finally, a new and improved IO Channel and condition watch
implementation for Win32. Based on code provided by Craig Setera.
When watching file descriptors, for which there is no select()
like functionality on Win32 that would work on all Win32 platforms
for all types of file descriptors (including anonymous pipes), we
start a new thread that blocks while trying to read from the file
descriptor. When the read returns, a Win32 Event is signalled that
the polling routine eventually notices. Meanwhile, the data being
read is stored in a circular buffer, from where the IO channel's
read() method picks it up.
If the buffer fills up the reading thread has to wait for space
becoming available. For this another Win32 Event is used. The IO
Channel's read() method signals this when it has read some data
out of the buffer.
The separate reader thread(s), and the circular buffer(s) with
associated events mean lots of possibilities for fun parallellism
errors. But it seems to work OK, i.e. GIMP runs.
* gmain.c: Small changes to the Win32 polling function.
(g_main_win32_get_poll_func): New function. Perhaps it would be a
good idea to provide this on all platforms.
* giowin32.c: The bulk of the new implementation.
(g_io_channel_win32_wait_for_condition): New function. To be used
where on Unix one does a select() on the channel's fd, like
libgimp's gimp_extension_process(). Could be provided on all
platforms.
* glib.h: Update documentation for IO Channels on Win32. Remove
the declarations for the as of now obsolete old functions related
to IO Channels for pipes with "wakeup" messages.
* glib.def: Some new functions.
* tests/gio-test.c: New file, to test GIOChannel and main loop.
* tests/Makefile.am
* tests/makefile.mingw.in: Add it.
2000-07-27 Tor Lillqvist <tml@iki.fi>
* gobject.c: No, don't mark glib_debug_objects for export here,
it's handled specially in gtktypeutils.c.
* gobject.def: Export g_type_fundamental_last (the function).
Wed Jul 26 12:59:31 2000 Tim Janik <timj@gtk.org>
* *.[hc]: applied patch from Andreas Persenius <ndap@swipnet.se> that
updates the license headers to the GNU Lesser General Public License,
as well as updating the copyright year to 2000.
Wed Jul 26 05:47:48 2000 Tim Janik <timj@gtk.org>
* configure.in:
* testglib.c:
* gstrfuncs.c:
* glib.h: added g_strlcat() and g_strlcpy() wrappers, supplied by
David Wheeler <dwheeler@ida.org>:
* glib.h, gstrfuncs.c: added g_strlcpy and g_strlcat to support
safe manipulation of fixed-length string buffers.
These functions were originally developed by Todd Miller to simplify
development of security-related programs, and
are available on many (but not all) Unix-like systems,
including OpenBSD, FreeBSD, and Solaris. See
ftp://ftp.openbsd.org/pub/OpenBSD/src/lib/libc/string/strlcpy.3
and http://www.openbsd.org/security.html.
If there's a strlcpy/strlcat on the system, it's called, otherwise
an implementation is provided.
* testglib.c: Added tests for g_strlcpy, g_strlcat.
Wed Jul 26 05:03:24 2000 Tim Janik <timj@gtk.org>
* acglib.m4 (GLIB_SIZEOF): include <stdlib.h> and <stddef.h> if
STDC_HEADERS is defined.
* glib.h:
* glibconfig.h: define gsize and gssize in terms of GLIB_SIZEOF_SIZE_T
* glib.h (g_return_if_reached): applied darin's fix for copy'n
paste error in the macro implementation.
Sun Jul 23 17:23:35 2000 Tim Janik <timj@gtk.org>
* gtype.[hc]: make g_type_fundamental_last() a function, avoiding all
that extern variable clutter and avoiding further problems with thread
safety.
2000-07-22 Tor Lillqvist <tml@iki.fi>
* build-dll: Fix resource handling, the resource file got left out
from the DLL after all... Remove the WIN32APIHEADERS, not needed
with current windres.
* glib.def: Add new functions.
2000-07-20 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
* gutils.c, glib.h: Mark the functions g_basename and g_dirname
deprecated. They will issue an warning once, when compiled with
G_ENABLE_DEBUG, but continue to work as before. Instead the
functions g_path_get_basename and g_path_get_dirname should be
used, which BOTH return newly allocated memory, that has to freed
by g_free. The new g_path_get_basename now strips trailing slashes
from the path. This fixes#5097. For discussion see
http://mail.gnome.org/pipermail/gtk-devel-list/2000-April/003139.html
* gwin32.c, testglib.c, tests/dirname-test.c: Use the new
functions instead of the old ones.
* gmodule/libgplugin_a.c, gmodule/testgmodule.c: Use
g_path_get_basename instead of the deprecated g_basename.
2000-07-20 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
* ghash.c, gscanner.c, glib.h: Mark the functions
g_hash_table_freeze, g_hash_table_thaw and thus
g_scanner_freeze_symbol_table and g_scanner_thaw_symbol_table
deprecated. They will issue an warning once, when compiled with
G_ENABLE_DEBUG. This fixes Bug #3883. For discussion see
http://mail.gnome.org/pipermail/gtk-devel-list/2000-April/003139.html
2000-07-19 Sebastian Wilhelmi <wilhelmi@ira.uka.de>
* configure.in, glib.h: glibconfig.h and glib.h now include files
outside of the extern "C" block. Makes some C++ compiler
happy. Reported by Denis Vakatov <vakatov@peony.nlm.nih.gov>.
* gmodule/gmodule.h: include glib.h before doing extern "C".
Makes some C++ compiler happy. Reported by Denis Vakatov
<vakatov@peony.nlm.nih.gov>.
2000-07-19 Tor Lillqvist <tml@iki.fi>
* gmodule-win32.c: Don't #include <tlhelp32.h> when compiling with gcc, as it
isn't provided. Declaration of a needed struct (from
www.microsoft.com) inserted instead.
Sat Jul 15 22:44:22 2000 Owen Taylor <otaylor@redhat.com>
* configure.in: Add build/Makefile and
build/win32/Makefile to AC_OUTPUT() so things build
again.
Sat Jul 15 09:11:46 2000 Tim Janik <timj@gtk.org>
* gstrfuncs.c (g_strncasecmp): fixed an off by 0 error (yeah,
the function went off when the while (n--) loop failed due to
n==0 ;), reported by Jean-Louis HAMEL <jlhamel@club-internet.fr>.
