1 2008-10-23 Matthias Clasen <mclasen@redhat.com>
3 Bug 557210 – g_compute_checksum_for_* asserts with less than 2 bytes
5 * glib/gchecksum.c (g_compute_checksum_for_data): Accept
6 lengths < 2. Patch by Tommi Komulainen
8 * tests/checksum-test.c: Add a test for this
10 2008-10-23 Matthias Clasen <mclasen@redhat.com>
12 Bug 556921 – gpoll.h breaks hal compilation
14 * glib/gpoll.h: Include gtypes.h. Pointed out by Anis Elleuch
16 2008-10-20 Christian Persch <chpe@gnome.org>
18 Bug 557087 – mem leak in g_content_types_get_registered
20 * gio/gcontenttype.c: Plug a mem leak.
22 2008-10-19 Michael Natterer <mitch@imendio.com>
24 Bug 556186 – gpoll.h breaks gmain.h inclusion
26 * glib/gpoll.h: different fix: disallow direct inclusion
27 unconditionally again but make an exception if included from
28 gmain.h to maintain compatibility.
30 2008-10-16 Matthias Clasen <mclasen@redhat.com>
32 * configure.in: Bump version to 2.19.1
34 * === Released 2.19.0 ===
38 2008-10-14 Christian Persch <chpe@gnome.org>
40 Bug 556186 – gpoll.h breaks gmain.h inclusion
42 * glib/gpoll.h: Only disallow direct inclusion when
43 G_DISABLE_SINGLE_INCLUDES is defined.
45 2008-10-13 Sven Herzberg <sven@imendio.com>
47 Bug 556101 – static mutex yields warnings with g++
50 * configure.in: added an intermediate cast to gpointer
52 2008-10-13 Christian Persch <chpe@gnome.org>
54 Bug 555311 – format not a string literal and no format arguments
57 * glib/gshell.c: Use literal errors where appropriate.
59 2008-10-10 Behdad Esfahbod <behdad@gnome.org>
61 Bug 551355 – [PATCH] Make glib build with libtool 2.2
63 * autogen.sh: Accept libtool 2.2. We are moving towards having
66 2008-10-10 Behdad Esfahbod <behdad@gnome.org>
68 * configure.in: With libtool 2.x, regenerate libtool early.
69 Based on bug 552107, patch by Patryk Zawadzki
71 2008-10-10 Matthias Clasen <mclasen@redhat.com>
73 Bug 554557 – Patch to fix gcc warnings about missing format
78 * glib/grel.c: Fix gcc warnings. Patch by Kjartan Maraas
80 2008-10-10 Matthias Clasen <mclasen@redhat.com>
82 Bug 552861 – glib-2.0.m4 calls system(3) without storing its result
84 * m4macros/glib-2.0.m4: Cosmetic change to make -Werror happy.
85 Patch by Andreas Köhler
87 2008-10-08 Christian Persch <chpe@gnome.org>
89 Bug 555313 – GFileAttribute boxed type get_type function should
90 use usual get_type pattern
92 * gio/gfileattribute.c (g_file_attribute_info_list_get_type): Use
93 g_once_init_enter/leave.
95 2008-10-08 Christian Persch <chpe@gnome.org>
97 Bug 555314 – mem leak in gmarkup
99 * glib/gmarkup.c: (g_markup_parse_context_parse): Plug a mem leak.
101 2008-10-08 Tor Lillqvist <tml@novell.com>
103 Bug 554790 - g_convert() misbehaves with winiconv versions
105 * glib/win_iconv.c (kernel_mbtowc): If converting from ASCII,
106 explicitly check for and reject 8bit chars. MultiByteToWideChar()
107 doesn't, at least not on XP.
109 2008-10-06 Matthias Clasen <mclasen@redhat.com>
111 * glib/gtypes.h: Properly include gmacros.h
113 2008-10-06 Behdad Esfahbod <behdad@gnome.org>
115 Bug 555309 – giochannel breaks on error
116 Patch from Christian Persch
118 * glib/giounix.c (g_io_unix_read), (g_io_unix_write),
119 (g_io_unix_seek), (g_io_unix_close), (g_io_unix_set_flags):
120 Don't shadow err. Oops!
122 2008-10-06 Christophe Fergeau <teuf@gnome.org>
124 Bug 555224 – Improve g_format_size_for_display doc
126 * glib/gfileutils.c: change g_format_size_for_display API doc to
127 explicitly say that the returned string has to be freed. Change
128 spelling of "newly allocated" to "newly-allocated" in g_file_read_link
129 API doc to be more consistent with what is done in that file.
131 2008-10-01 David Zeuthen <davidz@redhat.com>
133 * README.in: Add "Notes about glib 2.20" section detailing the
134 ramifications of the patch from bug #528670.
136 2008-09-30 Behdad Esfahbod <behdad@gnome.org>
138 Bug 554092 – glib doesn't return G_FILE_ERROR_NOENT et al on OS X
140 * glib/giounix.c (g_io_unix_read), (g_io_unix_write),
141 (g_io_unix_seek), (g_io_unix_close), (g_io_unix_set_flags),
142 (g_io_unix_get_flags), (g_io_channel_new_file):
143 Like mclasen says: "well, thats the way errno works...,
144 save it or loose it". Save errno.
146 2008-09-30 Tor Lillqvist <tml@novell.com>
149 * glib/tests/Makefile.am: Bypass gtester related stuff on Windows.
151 2008-09-30 Tor Lillqvist <tml@novell.com>
154 * glib/gnulib/vasnprintf.c: Don't define _GNU_SOURCE on Windows,
155 as _GNU_SOURCE has unintended side effects when compiling against
156 newest mingw headers.
158 2008-09-26 Dan Winship <danw@gnome.org>
160 Bug 553447 – g_assert_no_error()
162 * glib/gtestutils.h (g_assert_no_error, g_assert_error): Macros to
163 assert that a GError is not set, or else is set to a particular
166 * glib/gtestutils.c (g_assertion_message_error): utility for
169 * glib/tests/keyfile.c:
170 * tests/asyncqueue-test.c:
171 * tests/bookmarkfile-test.c:
172 * tests/convert-test.c:
173 * tests/file-test.c: Use g_assert_error/g_assert_no_error
175 2008-09-26 Dan Winship <danw@gnome.org>
177 * glib/gthreadpool.c (wakeup_thread_marker): make this a "const
178 gpointer" rather than a gconstpointer to avoid warnings later
180 * glib/pcre/pcre_ucp_searchfuncs.c:
181 * glib/pcre/pcre_valid_utf8.c: #include "config.h"
183 * glib/tests/printf.c (test_d): fool gcc into not warning about
184 some printf format strings that we know are dubious
186 2008-09-26 Matthias Clasen <mclasen@redhat.com>
188 Bug 553857 – gbacktrace.h requires signal.h
190 * glib/gbacktrace.h: Include signal.h for raise().
191 Pointed out by Sebastien Bacher
193 2008-09-26 Matthias Clasen <mclasen@redhat.com>
195 Bug 553724 – python interpretter path not patched in correctly
197 * glib/Makefile.am: Fix the sed magic to replace python.
199 2008-09-26 Matthias Clasen <mclasen@redhat.com>
201 * glib/gmain.c: Add some more docs.
203 * glib/giochannel.c: Move more docs inline, and improve them
206 2008-09-25 Tor Lillqvist <tml@novell.com>
208 Bug 553820 - gpoll.c: undeclared identifier
211 * glib/gpoll.c: Make the g_poll() function non-static also on
212 Windows. Prefix an underscore to the g_main_poll_debug variable
213 and make it non-static in gmain.c so that it can be used in
214 gpoll.c. Add back missing variable declaration.
216 2008-09-25 Tor Lillqvist <tml@novell.com>
218 * glib/gspawn-win32.c (do_spawn_with_pipes) (do_spawn_directly):
219 Just ignore the child_setup function, never call it. The is no
220 situation in which it could be useful on Windows. Do print a
221 warning, like before.
223 * glib/gspawn.c (g_spawn_async_with_pipes): Corresponding change
226 2008-09-24 Sven Herzberg <sven@imendio.com>
228 Be a little more explcit in the docs. Includes Owen's requested
231 * glib/gmain.c: improved documentation for g_source_attach() and
234 2008-09-23 Michael Natterer <mitch@imendio.com>
236 * glib/glib.h: #include <glib/gpoll.h>
238 * glib/gpoll.h: #error out if gpoll.h is included directly.
240 * glib/gpoll.c: remove trailing whitespace.
242 2008-09-23 Dan Winship <danw@gnome.org>
244 * glib/gpoll.[ch] (g_poll): Move this out of gmain.c and make it part
245 of the public API. (Part of Bug 505361 - gunixinputstream.c assumes
248 2008-09-23 Tor Lillqvist <tml@novell.com>
250 * glib/gmain.c (poll_rest) [Win32]: Fix embarrassing bug: I was
251 passing an incorrect third parameter to memmove(), had forgotten
252 to multiply by the size of the table entry. Just use a for loop
253 instead, clearer. Odd I didn't notice when testing this code.
255 2008-09-19 Hans Petter Jansson <hpj@novell.com>
257 Rewrite most of GHashTable to use open addressing with quadratic
258 probing instead of chaining. This has the potential to reduce memory
259 fragmentation significantly, while being slightly faster due to
260 better locality and no need to call alloc/free functions for nodes.
261 Benchmarks suggest it also uses less memory overall.
263 * glib/ghash.c (prime_mod): Table of suitable primes for
264 initial-probe distribution.
265 (g_hash_table_set_shift): New function.
266 (g_hash_table_find_closest_shift): New function.
267 (g_hash_table_set_shift_from_size): New function.
268 (g_hash_table_lookup_node_for_insertion): New function.
269 (g_hash_table_lookup_node): Rewritten to return node index instead of
270 pointer, use quadratic probe on flat table, and not return insertion
271 data. The latter saves some computation for read-only lookups.
272 (g_hash_table_remove_node): Rewrite to take a pointer directly to the
273 node structure to remove, and clear that. Remove unlinking code.
274 (g_hash_table_remove_all_nodes): Rewrite to not clear nodes
275 individually, but en masse using memset () after potentially calling
277 (iter_remove_or_steal): Use new data structure and algorithm. Vastly
278 simplified - now just a call to g_hash_table_remove_node ().
279 (g_hash_table_resize): New resize code, re-indexing with new prime
280 and cleaning up tombstones.
281 (g_hash_table_maybe_resize): Table may hold 8 buckets minimum, no less
282 than 1/4 load excluding tombstones, and no more than 15/16 load
283 including tombstones. These numbers are the results of a lot of
284 benchmarking with multiple complex applications, and should not be
286 (g_hash_table_iter_next)
287 (g_hash_table_lookup)
288 (g_hash_table_lookup_extended)
289 (g_hash_table_insert_internal)
290 (g_hash_table_remove_internal)
291 (g_hash_table_foreach_remove_or_steal)
292 (g_hash_table_foreach)
294 (g_hash_table_get_keys)
295 (g_hash_table_get_values): Use new data structure and algorithm,
296 fairly trivial changes.
298 2008-09-19 Tor Lillqvist <tml@novell.com>
300 * glib-zip.in: Look for man pages in share/man.
302 * glib/gutils.c (_glib_get_dll_directory)
303 * glib/gspawn-win32.c (do_spawn_with_pipes): Be a bit less
304 restrictive, look for the helper programs in the same folder where
305 the GLib DLL is, not necessarily in a "bin" subfolder of the top
306 GLib installation folder.
308 2008-09-18 Matthias Clasen <mclasen@redhat.com>
310 * configure.in: Bump version to 2.19.0
312 * ChangeLog.pre-2-18: rotate ChangeLog
314 * === branch for 2.18 ===