Commit by Raphael Manfredi on devel :: rd7facbf / (4 files in 2 dirs): (link)

Factorized ADNS message definition and added magic number in message header. - https://github.com/gtk-gnutella/gtk-gnutella/commit/d7facbfa41f7eb9e10687bd515a132d85ec70f8c

Commit by Raphael Manfredi on devel :: r10583f7 / src/lib/adns.c : (link)

ADNS: normalized messages, using G_STRFUNC and %m. - https://github.com/gtk-gnutella/gtk-gnutella/commit/10583f7f5b148a72171ce36e2dc73a4b115fa6ae

Commit by Raphael Manfredi on devel :: rb02332d / (10 files): (link)

Rename mutex_get() and mutex_release() to bear more conventional names. - https://github.com/gtk-gnutella/gtk-gnutella/commit/b02332d20812282436899d5455acb150e2ac65ba

Commit by Raphael Manfredi on devel :: r8080325 / (src/lib/mutex.c src/lib/spinlock.c): (link)

Properly account spinlock / mutex destruction at shutdown time. - https://github.com/gtk-gnutella/gtk-gnutella/commit/80803256c76b7b9878dd94b08a6c35b459f397be

Commit by Raphael Manfredi on devel :: r790ac69 / src/lib/xmalloc.c : (link)

Perform xmalloc_split_setup() as soon as possible after freelist is initialized. - https://github.com/gtk-gnutella/gtk-gnutella/commit/790ac69b087e2121524dba30856fbd0565733bb6

Commit by Raphael Manfredi on devel :: r8632ac1 / src/lib/mingw32.c : (link)

Avoid too many exceptions in a row on Windows. - https://github.com/gtk-gnutella/gtk-gnutella/commit/8632ac10a47f44c0fab052432c5590d8461279e4

Commit by Raphael Manfredi on devel :: r0e67814 / (3 files): (link)

Adjusted initialization procedures for FreeBSD pthread idiosyncracies. - https://github.com/gtk-gnutella/gtk-gnutella/commit/0e67814698f4d29bcc33abebaf6e38440bfd399a

Commit by Raphael Manfredi on devel :: r7072d3d / src/lib/mutex.c : (link)

mutex_ungrab(): trace releasing thread ID on assertion failure. - https://github.com/gtk-gnutella/gtk-gnutella/commit/7072d3da30548a8c26bf50967d3c805ccc3a0244

Commit by Raphael Manfredi on devel :: rb6ee6e4 / src/lib/mutex.c : (link)

mutex_is_owned_by_fast(): forgot type (was defaulting to "int", which was OK). - https://github.com/gtk-gnutella/gtk-gnutella/commit/b6ee6e463ca3fe3c3d9509a62d2c83fc7c39fd22

Commit by Raphael Manfredi on devel :: ra2c1e16 / src/lib/bfd_util.c : (link)

bfd_util_compute_offset(): don't substract void * pointers. - https://github.com/gtk-gnutella/gtk-gnutella/commit/a2c1e16ac153fb0fd477c464d570c15ae8a988fc

Commit by Raphael Manfredi on devel :: rca4013c / (src/lib/thread.c src/lib/thread.h): (link)

Made thread quasi ID truly unique on 64-bit machines. - https://github.com/gtk-gnutella/gtk-gnutella/commit/ca4013c84d56e57c02c3cb316df141d3b776f227

Commit by Raphael Manfredi on devel :: ra245110 / (src/lib/elist.h src/lib/erbtree.h): (link)

Use G_GNUC_EXTENSION to shut up gcc warnings about braced expressions. - https://github.com/gtk-gnutella/gtk-gnutella/commit/a245110352508946878d0489797ab382a8d1737b

Commit by Raphael Manfredi on devel :: raa66e82 / src/gtk-gnutella.h : (link)

Bumped date. (+41 more commits...) - https://github.com/gtk-gnutella/gtk-gnutella/commit/aa66e82119ff9a9438b35c01fcd4456bb6837548

Commit by Raphael Manfredi on devel :: ra5a5f89 / src/lib/random.c : (link)

Fixed outdated comment. (+7 more commits...) - https://github.com/gtk-gnutella/gtk-gnutella/commit/a5a5f89ee9048e111625beb8fa442cfc0082ed40

Commit by Raphael Manfredi on devel :: r4a6985f / src/dht/routing.c : (link)

Disable move_node() now that we use a hikset to store DHT nodes in buckets.

This fixes bugs #3522465 and #3522466. (+20 more commits...) - https://github.com/gtk-gnutella/gtk-gnutella/commit/4a6985f08bc419d4b3ba50459ad55757bc18e34e

Commit by Raphael Manfredi on devel :: r20c8be9 / src/lib/thread.c : (link)

thread_is_single(): fixed inverted logic in previous commit. - https://github.com/gtk-gnutella/gtk-gnutella/commit/20c8be96ac02d5de67ff5f5bfd57343f34ea2238

Commit by Raphael Manfredi on devel :: r1b34f93 / src/lib/thread.c : (link)

thread_single(): properly handle calls in the middle of a thread creation. - https://github.com/gtk-gnutella/gtk-gnutella/commit/1b34f939dbd0ac915ea20ed6304b48aa7f753f86

Commit by Raphael Manfredi on devel :: r19cc7ab / src/lib/thread.c : (link)

thread_find(): was computing wrong quick thread ID now that arg can be NULL. - https://github.com/gtk-gnutella/gtk-gnutella/commit/19cc7abaf7b4fac45c2f507b01271df244c7a127

Commit by Raphael Manfredi on devel :: r4026dcd / src/lib/thread.c : (link)

thread_find(): now use TID lookup most of the time, which should be safer.

We reserve the location of the thread based on its stack pointer to stack
pointer validation routines. (+5 more commits...) - https://github.com/gtk-gnutella/gtk-gnutella/commit/4026dcdab8df891a28809cd0f05e4b7c430796ec

Commit by Raphael Manfredi on devel :: r38e63ff / src/lib/thread.c : (link)

Revisited the new thread recording logic to avoid race conditions.

The old code was not behaving correctly when we were recording a new thread:
any mutex grabbed before the insertion of the new thread in the global hash
table was accounted for in the wrong thread, leading to assertion failures
later when the mutex was released by the thread.

Also changed thread_find() to make sure it now scan the thread array by
small thread ID looking for a matching thread ID when we are in the process
of creating a new thread. This is slower because it involves a call to
thread_self() which may again do the same thing at the lower pthread level,
but this only happens when a new thread is in the critical section where it
registers itself. - https://github.com/gtk-gnutella/gtk-gnutella/commit/38e63fffc4dd8d7a520674779ea21b1ee29c7e5b