http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
Jonathan Wakely changed:
What|Removed |Added
CC||arch.jslin at gmail dot com
--- Comment
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
Jonathan Wakely changed:
What|Removed |Added
Summary|shared_ptr consuming too|resource leaks due to
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #13 from Paolo Carlini 2010-11-15
18:32:23 UTC ---
Thanks a lot Jon for tracking this down! Do you think we can fix it in time for
4.6.0?
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
Jonathan Wakely changed:
What|Removed |Added
Status|NEW |ASSIGNED
AssignedTo|unassigned
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #15 from Jonathan Wakely 2010-11-15
21:54:40 UTC ---
Butenhof's book says you don't need to destroy a mutex/condvar that was
statically initialized, so given the FreeBSD bug I will only define the
destructor when the __GTHREAD_XXX_INI
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #16 from Jonathan Wakely 2010-11-16
00:48:52 UTC ---
Created attachment 22413
--> http://gcc.gnu.org/bugzilla/attachment.cgi?id=22413
add destructors in
could you try applying this patch to ext/concurrence.h and let me know if it
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #17 from Zouzou 2010-11-16 08:37:02
UTC ---
(In reply to comment #16)
> Created attachment 22413 [details]
> add destructors in
> could you try applying this patch to ext/concurrence.h and let me know if it
> works on Windows? I'm t
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
Jonathan Wakely changed:
What|Removed |Added
Target Milestone|--- |4.6.0
--- Comment #18 from Jonathan Wak
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #19 from Paolo Carlini 2010-11-16
10:25:40 UTC ---
Jon, sometimes finding a reviewer for those gthr changes takes a bit of time...
and we are already in Stage 3... Thus, I would recommend doing our best to
figure out first a decent li
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #20 from Jonathan Wakely 2010-11-16
11:05:44 UTC ---
OK, that will be ugly though. The cast from __gthread_recursive_mutex_t* to
__gthread_mutex_t* is not correct, because the "sema" member (the actual Win32
handle) is at a different
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #21 from Paolo Carlini 2010-11-16
11:52:20 UTC ---
Argh, I see. I think we should keep the option open, anyway, together with a
huge FIXME in the code, of course. I also think we should try to explain the
problem to the people actuall
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #22 from Jonathan Wakely 2010-11-16
12:32:36 UTC ---
... and when using gthr-mipssde.h / gthr-posix95.h / gthr-solaris.h:
static inline int
__recursive_mutex_destroy(__gthread_recursive_mutex_t* __rmutex)
{
return ___gthread_mutex_
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
Jonathan Wakely changed:
What|Removed |Added
Attachment #22413|0 |1
is obsolete|
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #24 from Zouzou 2010-11-16 22:29:35
UTC ---
(In reply to comment #23)
> Created attachment 22424 [details]
> add destructors in
> here's another patch, this one uses SFINAE to select an appropriate overload
> based on the type of __g
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #25 from Jonathan Wakely 2010-11-16
22:36:16 UTC ---
(In reply to comment #24)
> apparently the 1st overload doesn't match so it falls back to the 3rd one.
bah! ok, thanks, back to the drawing board...
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #26 from Jonathan Wakely 2010-11-16
23:03:08 UTC ---
Oops, that patch was broken - could you change the recursive mutex destructor
so it calls _S_destroy(&_M_mutex, 0) instead of _S_destroy(&_M_mutex)
As I can only test on POSIX syst
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #27 from Jonathan Wakely 2010-11-16
23:22:24 UTC ---
I don't think that's right either ... I need to find a better way to test
this
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
Jonathan Wakely changed:
What|Removed |Added
Attachment #22424|0 |1
is obsolete|
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #29 from Zouzou 2010-11-17 07:38:39
UTC ---
(In reply to comment #28)
> Created attachment 22427 [details]
> add destructors in
> almost the same as the last patch but the destructor calls:
> _S_destroy<__gthread_recursive_mutex_t>
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #30 from Jonathan Wakely 2010-11-17
09:37:37 UTC ---
(In reply to comment #29)
> because __gthread_mutex_destroy returns void.
Argh, why is gthr-win32.h different here?!
Nevermind ...
> after changing the return types of the _S_des
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
Jonathan Wakely changed:
What|Removed |Added
Attachment #22427|0 |1
is obsolete|
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #32 from Zouzou 2010-11-17 22:38:53
UTC ---
(In reply to comment #31)
> Created attachment 22436 [details]
> add destructors in
> a simpler version - could you test this on mingw? thanks
great, compiles and fixes the resource leaks
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #33 from Paolo Carlini 2010-11-17
23:55:22 UTC ---
If we need __is_same / __are_same, maybe we can just include
bits/cpp_type_traits.h: isn't that big (and probably something else includes it
anyway in most of the cases)
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #34 from Jonathan Wakely 2010-11-18
00:29:20 UTC ---
(In reply to comment #33)
> If we need __is_same / __are_same, maybe we can just include
> bits/cpp_type_traits.h
aha, thanks for the pointer to are_same, I was going to move
__gnu
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #35 from Jonathan Wakely 2010-11-18
02:45:09 UTC ---
patch posted to http://gcc.gnu.org/ml/gcc-patches/2010-11/msg01875.html
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
--- Comment #36 from Jonathan Wakely 2010-11-18
18:56:34 UTC ---
Author: redi
Date: Thu Nov 18 18:56:29 2010
New Revision: 166917
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=166917
Log:
2010-11-18 Jonathan Wakely
PR libstdc++/4
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46455
Jonathan Wakely changed:
What|Removed |Added
Status|ASSIGNED|RESOLVED
Resolution|
27 matches
Mail list logo
