graphics/rawtherapee: r342622 crashes on HEAD

classic Classic list List threaded Threaded
48 messages Options
123
Reply | Threaded
Open this post in threaded view
|

graphics/rawtherapee: r342622 crashes on HEAD

Rainer Hurling
Many thanks for the update of graphics/rawtherapee, r342622. This
program is really important for photographers.

It builds and installs just fine on recent HEAD amd64, but unfortunately
it crashes immediately, when started.

I tried to build rawtherapee and some of its dependencies with
WITH_DEBUG=yes and then to have a look with gdb, but with only little
luck. Obviously there is a problem with DWARF(?) and many libs without
debug symbols?


# gdb rawtherapee
[..SNIP..]
This GDB was configured as "amd64-marcel-freebsd"...Dwarf Error: wrong
version in compilation unit header (is 4, should be 2) [in module
/usr/local/bin/rawtherapee]
[..SNIP..]

(gdb) r
Dwarf Error: wrong version in compilation unit header (is 4, should be
2) [in module /usr/local/lib/gcc48/libgcc_s.so.1]
[..SNIP..]
[New LWP 101478]
[New Thread 4ec06400 (LWP 101478/rawtherapee)]
Error while reading shared library symbols:
Dwarf Error: wrong version in compilation unit header (is 4, should be
2) [in module /usr/local/lib/gcc48/libgcc_s.so.1]
[..SNIP..]
Terminating due to uncaught exception 0x4fe78700 of type Glib::ConvertError
Program received signal SIGABRT, Aborted.
[Switching to Thread 4ec06400 (LWP 101478/rawtherapee)]
0x0000000048b847ba in thr_kill () from /lib/libc.so.7

(gdb) bt full
#0  0x0000000048b847ba in thr_kill () from /lib/libc.so.7
No symbol table info available.
#1  0x0000000048c3b029 in abort () from /lib/libc.so.7
No symbol table info available.
#2  0x00000000484d37da in __cxa_rethrow () from /lib/libcxxrt.so.1
No symbol table info available.
#3  0x00000000423aa8f8 in Glib::ConvertError::throw_func () from
/usr/local/lib/libglibmm-2.4.so.1
No symbol table info available.
#4  0x00000000423baa0f in Glib::Error::throw_exception () from
/usr/local/lib/libglibmm-2.4.so.1
No symbol table info available.
#5  0x00000000423c60b4 in Glib::operator<< () from
/usr/local/lib/libglibmm-2.4.so.1
No symbol table info available.
#6  0x000000000064fce7 in ?? ()
No symbol table info available.
#7  0x0000000000fa7c00 in ?? ()
No symbol table info available.
#8  0x00007fffffffbae0 in ?? ()
No symbol table info available.
#9  0x00007fffffffbc00 in ?? ()
No symbol table info available.
#10 0x000000000064f765 in ?? ()
No symbol table info available.
#11 0x00007fffffffbb00 in ?? ()
No symbol table info available.
#12 0x0000000000b8f988 in ?? ()
No symbol table info available.
#13 0x0000000000fa7c00 in ?? ()
No symbol table info available.
#14 0x00007fffffffbd70 in ?? ()
No symbol table info available.
#15 0x0000000000f853b8 in ?? ()
No symbol table info available.
#16 0x0000000000f85490 in ?? ()
No symbol table info available.
#17 0x00000000484c0ee0 in std::__1::locale::id::__next_id () from
/usr/local/lib/libc++.so.1
No symbol table info available.
#18 0x0000000000000000 in ?? ()
No symbol table info available.


I know, that this output is only of little help. But I could need some
advise what to do next to get more info.

Any help is really appreciated. Thanks in advance,
Rainer Hurling
_______________________________________________
[hidden email] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "[hidden email]"
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Matthias Andree-2
Am 05.02.2014 20:46, schrieb Rainer Hurling:
> Many thanks for the update of graphics/rawtherapee, r342622. This
> program is really important for photographers.
>
> It builds and installs just fine on recent HEAD amd64, but unfortunately
> it crashes immediately, when started.

Rainer,

I don't see the crash on FreeBSD 10.0-RELEASE amd64 and 9.2-RELEASE
amd64 - those are versions I tried and where I could successfully open a
Sony ARW file and click a few UI items.  Note sure what 11 changed that
it would break.

> I tried to build rawtherapee and some of its dependencies with
> WITH_DEBUG=yes and then to have a look with gdb, but with only little
> luck. Obviously there is a problem with DWARF(?) and many libs without
> debug symbols?

It's rather that the base /usr/bin/gdb cannot deal with the newer debug
symbol formats...

> # gdb rawtherapee
> [..SNIP..]
> This GDB was configured as "amd64-marcel-freebsd"...Dwarf Error: wrong
> version in compilation unit header (is 4, should be 2) [in module
> /usr/local/bin/rawtherapee]
> [..SNIP..]

...do you have more luck with gdb built from the ports collection
(devel/gdb), which is version 7.6.2, as opposed to the base system gdb
6.1.1?

Also, if you recompile rawtherapee without the highly aggressive
compiler flags, does that help?

I saw warnings about undefined behaviour in aggressive loop
optimization, not sure if those are the culprit.  If they are, we might
need to tune down optimization a bit.

Thanks.

Cheers,
Matthias
_______________________________________________
[hidden email] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "[hidden email]"
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Dimitry Andric-4
In reply to this post by Rainer Hurling
On 05 Feb 2014, at 20:46, Rainer Hurling <[hidden email]> wrote:

> Many thanks for the update of graphics/rawtherapee, r342622. This
> program is really important for photographers.
>
> It builds and installs just fine on recent HEAD amd64, but unfortunately
> it crashes immediately, when started.
>
> I tried to build rawtherapee and some of its dependencies with
> WITH_DEBUG=yes and then to have a look with gdb, but with only little
> luck. Obviously there is a problem with DWARF(?) and many libs without
> debug symbols?
It looks like you have built the port with gcc 4.8, which defaults to
DWARF4 format.  Our gdb in base is too old to understand this format, so
just install devel/gdb76 instead.


> # gdb rawtherapee
...
> #3  0x00000000423aa8f8 in Glib::ConvertError::throw_func () from
> /usr/local/lib/libglibmm-2.4.so.1
> No symbol table info available.
> #4  0x00000000423baa0f in Glib::Error::throw_exception () from
> /usr/local/lib/libglibmm-2.4.so.1
> No symbol table info available.
> #5  0x00000000423c60b4 in Glib::operator<< () from
> /usr/local/lib/libglibmm-2.4.so.1

Looks like that operator encountered something it cannot handle, so it
throws an exception, and nobody seems to catch it.  But more information
from a backtrace in gdb 7.6 could possibly help.


...
> #17 0x00000000484c0ee0 in std::__1::locale::id::__next_id () from
> /usr/local/lib/libc++.so.1

Hmm, is this a ports version of libc++?  I was not aware Baptiste had
already committed this? :)

-Dimitry


signature.asc (210 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Matthias Andree-2
Am 05.02.2014 21:08, schrieb Dimitry Andric:

>> #17 0x00000000484c0ee0 in std::__1::locale::id::__next_id () from
>> /usr/local/lib/libc++.so.1
>
> Hmm, is this a ports version of libc++?  I was not aware Baptiste had
> already committed this? :)

Yes, it is (as a build requisite, but apparently remained installed on
the destination machine), because we need to match the libraries that
the requisites use (Glibmm for one).

I have given up on compiling RawTherapee with clang++ for now, and use
GCC 4.8 on all systems.  RawTherapee is somewhat demanding, especially
at higher optimization level, and kills the 10.0-RELEASE base clang and
Port GCC 4.6 and 4.7, all with internal compiler errors.  Since GCC 4.8
worked for me, I did not bother to send Gerald the details.

We may want to retry with clang if we've got the next clang version.
Feel free to use Rawtherapee as compiler system test ;)


signature.asc (205 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Rainer Hurling
In reply to this post by Matthias Andree-2
Hi Matthias,

thanks for answering.

Am 05.02.2014 21:03, schrieb Matthias Andree:

> Am 05.02.2014 20:46, schrieb Rainer Hurling:
>> Many thanks for the update of graphics/rawtherapee, r342622. This
>> program is really important for photographers.
>>
>> It builds and installs just fine on recent HEAD amd64, but unfortunately
>> it crashes immediately, when started.
>
> Rainer,
>
> I don't see the crash on FreeBSD 10.0-RELEASE amd64 and 9.2-RELEASE
> amd64 - those are versions I tried and where I could successfully open a
> Sony ARW file and click a few UI items.  Note sure what 11 changed that
> it would break.
>
>> I tried to build rawtherapee and some of its dependencies with
>> WITH_DEBUG=yes and then to have a look with gdb, but with only little
>> luck. Obviously there is a problem with DWARF(?) and many libs without
>> debug symbols?
>
> It's rather that the base /usr/bin/gdb cannot deal with the newer debug
> symbol formats...
>
>> # gdb rawtherapee
>> [..SNIP..]
>> This GDB was configured as "amd64-marcel-freebsd"...Dwarf Error: wrong
>> version in compilation unit header (is 4, should be 2) [in module
>> /usr/local/bin/rawtherapee]
>> [..SNIP..]
>
> ...do you have more luck with gdb built from the ports collection
> (devel/gdb), which is version 7.6.2, as opposed to the base system gdb
> 6.1.1?

Okay, here it comes. RawTherapee from before, with newer gdb:

#gdb762 rawtherapee
GNU gdb (GDB) 7.6.2 [GDB v7.6.2 for FreeBSD]
Copyright (C) 2013 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later
<http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-portbld-freebsd11.0".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /usr/local/bin/rawtherapee...done.
(gdb) r
Starting program: /usr/local/bin/rawtherapee
[New LWP 100312]
[New Thread 4ec06400 (LWP 100312)]
Terminating due to uncaught exception 0x4fe78700 of type Glib::ConvertError

Program received signal SIGABRT, Aborted.
[Switching to Thread 4ec06400 (LWP 100312)]
0x0000000048b847ba in thr_kill () from /lib/libc.so.7
(gdb) bt full
#0  0x0000000048b847ba in thr_kill () from /lib/libc.so.7
No symbol table info available.
#1  0x0000000048c3b029 in abort () from /lib/libc.so.7
No symbol table info available.
#2  0x00000000484d37da in ?? () from /lib/libcxxrt.so.1
No symbol table info available.
#3  0x00000000423aa8f8 in Glib::ConvertError::throw_func(_GError*) ()
from /usr/local/lib/libglibmm-2.4.so.1
No symbol table info available.
#4  0x00000000423baa0f in Glib::Error::throw_exception(_GError*) () from
/usr/local/lib/libglibmm-2.4.so.1
No symbol table info available.
#5  0x00000000423c60b4 in
Glib::operator<<(std::__1::basic_ostream<wchar_t,
std::__1::char_traits<wchar_t> >&, Glib::ustring const&) () from
/usr/local/lib/libglibmm-2.4.so.1
No symbol table info available.
#6  0x000000000064fce7 in
Glib::ustring::FormatStream::stream<Glib::ustring> (this=0x7fffffffbae0,
value=...) at /usr/local/include/glibmm-2.4/glibmm/ustring.h:1057
No locals.
#7  0x000000000064f765 in Glib::ustring::format<Glib::ustring, char [9]>
(a1=..., a2=...) at /usr/local/include/glibmm-2.4/glibmm/ustring.h:1145
        buf = {stream_ =
    {<std::__1::basic_ostream<wchar_t, std::__1::char_traits<wchar_t> >>
= {<std::__1::basic_ios<wchar_t, std::__1::char_traits<wchar_t> >> =
{<std::__1::ios_base> = {<No data fields>}, __tie_ = 0x0,
                __fill_ = -1}, _vptr.basic_ostream = 0xf853b8 <vtable
for std::__1::basic_ostringstream<wchar_t,
std::__1::char_traits<wchar_t>, std::__1::allocator<wchar_t> >+24>},
            __sb_ = {<std::__1::basic_streambuf<wchar_t,
std::__1::char_traits<wchar_t> >> = {
                _vptr.basic_streambuf = 0xf85490 <vtable for
std::__1::basic_stringbuf<wchar_t, std::__1::char_traits<wchar_t>,
std::__1::allocator<wchar_t> >+16>, __loc_ = {static none = 0,
                  static collate = 1, static ctype = 2, static monetary
= 8, static numeric = 16, static time = 32, static messages = 4, static
all = 63, __locale_ = 0x484c0ee0}, __binp_ = 0x0,
                __ninp_ = 0x0, __einp_ = 0x0, __bout_ = 0x7fffffffbb2c
L"", __nout_ = 0x7fffffffbb2c L"", __eout_ = 0x7fffffffbb3c L""},
              __str_ = {<std::__1::__basic_string_common<true>> = {<No
data fields>},
                __r_ =
{<std::__1::__libcpp_compressed_pair_imp<std::__1::basic_string<wchar_t,
std::__1::char_traits<wchar_t>, std::__1::allocator<wchar_t> >::__rep,
std::__1::allocator<wchar_t>, 2u>> = {<std::__1::allocator<wchar_t>> =
{<No data fields>}, __first_ = {{__l = {__cap_ = 8, __size_ = 0, __data_
= 0x0}, __s = {{__size_ = 8 '\b', __lx = 8 L'\b\000\000\000'},
                          __data_ = L'\000' <repeats 16 times>}, __r =
{__words = {8, 0, 0}}}}}, <No data fields>}, static npos =
18446744073709551615}, __hm_ = 0x7fffffffbb2c L"", __mode_ = 16}}}
#8  0x000000000064c798 in RTImage::setPaths (opt=...) at
/usr/ports/graphics/rawtherapee/work/rawtherapee-4.0.12/rtgui/rtimage.cc:101
        configFilename = {static npos = 18446744073709551615, string_ =
{<std::__1::__basic_string_common<true>> = {<No data fields>},
            __r_ =
{<std::__1::__libcpp_compressed_pair_imp<std::__1::basic_string<char,
std::__1::char_traits<char>, std::__1::allocator<char> >::__rep,
std::__1::allocator<char>, 2u>> = {<std::__1::allocator<char>> = {<No
data fields>}, __first_ = {{__l = {__cap_ = 0, __size_ = 0, __data_ =
0x0}, __s = {{__size_ = 0 '\000', __lx = 0 '\000'}, __data_ = '\000'
<repeats 22 times>}, __r = {__words = {0,
                        0, 0}}}}}, <No data fields>}, static npos =
18446744073709551615}}
        keyFile = {<Glib::KeyFile> = {gobject_ = 0x4ec15d80,
owns_gobject_ = true}, <No data fields>}
        hasKeyFile = true
#9  0x0000000000696aaf in main (argc=1, argv=0x7fffffffd6d0) at
/usr/ports/graphics/rawtherapee/work/rawtherapee-4.0.12/rtgui/main.cc:239
        m = <incomplete type>
        icon_path = {static npos = 18446744073709551615, string_ =
{<std::__1::__basic_string_common<true>> = {<No data fields>},
            __r_ =
{<std::__1::__libcpp_compressed_pair_imp<std::__1::basic_string<char,
std::__1::char_traits<char>, std::__1::allocator<char> >::__rep,
std::__1::allocator<char>, 2u>> = {<std::__1::allocator<char>> = {<No
data fields>}, __first_ = {{__l = {__cap_ = 49, __size_ = 35, __data_ =
0x4fe3d220 "/usr/local/share/rawtherapee/images"}, __s = {{__size_ = 49
'1', __lx = 49 '1'},
                      __data_ =
"\000\000\000\000\000\000\000#\000\000\000\000\000\000\000
\322\343O\000\000\000"}, __r = {__words = {49, 35, 1340330528}}}}}, <No
data fields>},
            static npos = 18446744073709551615}}
        rtWindow = 0x1
        defaultIconTheme = {pCppObject_ = 0x4fde90a0}
        settings = {pCppObject_ = 0x0}


>
> Also, if you recompile rawtherapee without the highly aggressive
> compiler flags, does that help?

This version was compiled without the option OPTIMIZED_CFLAGS. Is this
enough?

Rainer

>
> I saw warnings about undefined behaviour in aggressive loop
> optimization, not sure if those are the culprit.  If they are, we might
> need to tune down optimization a bit.
>
> Thanks.
>
> Cheers,
> Matthias

_______________________________________________
[hidden email] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "[hidden email]"
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Baptiste Daroussin-2
In reply to this post by Matthias Andree-2
On Wed, Feb 05, 2014 at 09:19:51PM +0100, Matthias Andree wrote:

> Am 05.02.2014 21:08, schrieb Dimitry Andric:
>
> >> #17 0x00000000484c0ee0 in std::__1::locale::id::__next_id () from
> >> /usr/local/lib/libc++.so.1
> >
> > Hmm, is this a ports version of libc++?  I was not aware Baptiste had
> > already committed this? :)
>
> Yes, it is (as a build requisite, but apparently remained installed on
> the destination machine), because we need to match the libraries that
> the requisites use (Glibmm for one).
>
> I have given up on compiling RawTherapee with clang++ for now, and use
> GCC 4.8 on all systems.  RawTherapee is somewhat demanding, especially
> at higher optimization level, and kills the 10.0-RELEASE base clang and
> Port GCC 4.6 and 4.7, all with internal compiler errors.  Since GCC 4.8
> worked for me, I did not bother to send Gerald the details.
>
> We may want to retry with clang if we've got the next clang version.
> Feel free to use Rawtherapee as compiler system test ;)
>
try with something like this in libmap.conf
libc++.so.1 /usr/local/lib/libc++.so.1
If that fixes the problem, then a rpath with /usr/local/lib should be set while
building the port

regards,
Bapt

attachment0 (203 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Matthias Andree-2
In reply to this post by Rainer Hurling
Hi Rainer,

this is more useful as a backtrace in itself, but I don't know how to
make heads or tails of it; the interesting parts appear to be in frames
#5 (meaning that you might need to reinstall glibmm WITH_DEBUG=yes)  and
#8/#9 (where the whole call chain starts).

I can say that rawtherapee starts properly for me on 9.2-RELEASE and
10.0-RELEASE amd64 (the former has packages built from source, the
latter uses binary packages installed with pkg upgrade), even with
aggressive optimization.

Unfortunately, I do not have the time to debug highly complex ports on
any STABLE/UNSTABLE/HEAD tree. My ports work is limited to what I can do
on -RELEASE.

So I propose to
1. first trying the libc++.so mapping that Baptiste Daroussin has
proposed, and if that does not help,

2. hack the Makefile to use only -O for optimization and then see in the
frames #9/#8 if the string passed down is properly initialized, and in
frame #5 what data arrives and why glibmm fails to convert it.

Also make sure that all requisites of rawtherapee are up to date, there
have been many changes to a few of the requisites lately.

> Okay, here it comes. RawTherapee from before, with newer gdb:
> Terminating due to uncaught exception 0x4fe78700 of type Glib::ConvertError
>
> Program received signal SIGABRT, Aborted.
...

> No symbol table info available.
> #3  0x00000000423aa8f8 in Glib::ConvertError::throw_func(_GError*) ()
> from /usr/local/lib/libglibmm-2.4.so.1
> No symbol table info available.
> #4  0x00000000423baa0f in Glib::Error::throw_exception(_GError*) () from
> /usr/local/lib/libglibmm-2.4.so.1
> No symbol table info available.
> #5  0x00000000423c60b4 in
> Glib::operator<<(std::__1::basic_ostream<wchar_t,
> std::__1::char_traits<wchar_t> >&, Glib::ustring const&) () from
> /usr/local/lib/libglibmm-2.4.so.1
> No symbol table info available.
> #6  0x000000000064fce7 in
> Glib::ustring::FormatStream::stream<Glib::ustring> (this=0x7fffffffbae0,
> value=...) at /usr/local/include/glibmm-2.4/glibmm/ustring.h:1057
> No locals.
> #7  0x000000000064f765 in Glib::ustring::format<Glib::ustring, char [9]>
> (a1=..., a2=...) at /usr/local/include/glibmm-2.4/glibmm/ustring.h:1145
>         buf = {stream_ =
>     {<std::__1::basic_ostream<wchar_t, std::__1::char_traits<wchar_t> >>
> = {<std::__1::basic_ios<wchar_t, std::__1::char_traits<wchar_t> >> =
> {<std::__1::ios_base> = {<No data fields>}, __tie_ = 0x0,
>                 __fill_ = -1}, _vptr.basic_ostream = 0xf853b8 <vtable
> for std::__1::basic_ostringstream<wchar_t,
> std::__1::char_traits<wchar_t>, std::__1::allocator<wchar_t> >+24>},
>             __sb_ = {<std::__1::basic_streambuf<wchar_t,
> std::__1::char_traits<wchar_t> >> = {
>                 _vptr.basic_streambuf = 0xf85490 <vtable for
> std::__1::basic_stringbuf<wchar_t, std::__1::char_traits<wchar_t>,
> std::__1::allocator<wchar_t> >+16>, __loc_ = {static none = 0,
>                   static collate = 1, static ctype = 2, static monetary
> = 8, static numeric = 16, static time = 32, static messages = 4, static
> all = 63, __locale_ = 0x484c0ee0}, __binp_ = 0x0,
>                 __ninp_ = 0x0, __einp_ = 0x0, __bout_ = 0x7fffffffbb2c
> L"", __nout_ = 0x7fffffffbb2c L"", __eout_ = 0x7fffffffbb3c L""},
>               __str_ = {<std::__1::__basic_string_common<true>> = {<No
> data fields>},
>                 __r_ =
> {<std::__1::__libcpp_compressed_pair_imp<std::__1::basic_string<wchar_t,
> std::__1::char_traits<wchar_t>, std::__1::allocator<wchar_t> >::__rep,
> std::__1::allocator<wchar_t>, 2u>> = {<std::__1::allocator<wchar_t>> =
> {<No data fields>}, __first_ = {{__l = {__cap_ = 8, __size_ = 0, __data_
> = 0x0}, __s = {{__size_ = 8 '\b', __lx = 8 L'\b\000\000\000'},
>                           __data_ = L'\000' <repeats 16 times>}, __r =
> {__words = {8, 0, 0}}}}}, <No data fields>}, static npos =
> 18446744073709551615}, __hm_ = 0x7fffffffbb2c L"", __mode_ = 16}}}
> #8  0x000000000064c798 in RTImage::setPaths (opt=...) at
> /usr/ports/graphics/rawtherapee/work/rawtherapee-4.0.12/rtgui/rtimage.cc:101
>         configFilename = {static npos = 18446744073709551615, string_ =
> {<std::__1::__basic_string_common<true>> = {<No data fields>},
>             __r_ =
> {<std::__1::__libcpp_compressed_pair_imp<std::__1::basic_string<char,
> std::__1::char_traits<char>, std::__1::allocator<char> >::__rep,
> std::__1::allocator<char>, 2u>> = {<std::__1::allocator<char>> = {<No
> data fields>}, __first_ = {{__l = {__cap_ = 0, __size_ = 0, __data_ =
> 0x0}, __s = {{__size_ = 0 '\000', __lx = 0 '\000'}, __data_ = '\000'
> <repeats 22 times>}, __r = {__words = {0,
>                         0, 0}}}}}, <No data fields>}, static npos =
> 18446744073709551615}}
>         keyFile = {<Glib::KeyFile> = {gobject_ = 0x4ec15d80,
> owns_gobject_ = true}, <No data fields>}
>         hasKeyFile = true
> #9  0x0000000000696aaf in main (argc=1, argv=0x7fffffffd6d0) at
> /usr/ports/graphics/rawtherapee/work/rawtherapee-4.0.12/rtgui/main.cc:239
>         m = <incomplete type>
>         icon_path = {static npos = 18446744073709551615, string_ =
> {<std::__1::__basic_string_common<true>> = {<No data fields>},
>             __r_ =
> {<std::__1::__libcpp_compressed_pair_imp<std::__1::basic_string<char,
> std::__1::char_traits<char>, std::__1::allocator<char> >::__rep,
> std::__1::allocator<char>, 2u>> = {<std::__1::allocator<char>> = {<No
> data fields>}, __first_ = {{__l = {__cap_ = 49, __size_ = 35, __data_ =
> 0x4fe3d220 "/usr/local/share/rawtherapee/images"}, __s = {{__size_ = 49
> '1', __lx = 49 '1'},
>                       __data_ =
> "\000\000\000\000\000\000\000#\000\000\000\000\000\000\000
> \322\343O\000\000\000"}, __r = {__words = {49, 35, 1340330528}}}}}, <No
> data fields>},
>             static npos = 18446744073709551615}}
>         rtWindow = 0x1
>         defaultIconTheme = {pCppObject_ = 0x4fde90a0}
>         settings = {pCppObject_ = 0x0}
>
>
>>
>> Also, if you recompile rawtherapee without the highly aggressive
>> compiler flags, does that help?
>
> This version was compiled without the option OPTIMIZED_CFLAGS. Is this
> enough?

Perhaps not good enough for debugging...

_______________________________________________
[hidden email] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "[hidden email]"
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Rainer Hurling
In reply to this post by Baptiste Daroussin-2
Am 05.02.2014 22:20 (UTC+1) schrieb Baptiste Daroussin:

> On Wed, Feb 05, 2014 at 09:19:51PM +0100, Matthias Andree wrote:
>> Am 05.02.2014 21:08, schrieb Dimitry Andric:
>>
>>>> #17 0x00000000484c0ee0 in std::__1::locale::id::__next_id () from
>>>> /usr/local/lib/libc++.so.1
>>>
>>> Hmm, is this a ports version of libc++?  I was not aware Baptiste had
>>> already committed this? :)
>>
>> Yes, it is (as a build requisite, but apparently remained installed on
>> the destination machine), because we need to match the libraries that
>> the requisites use (Glibmm for one).
>>
>> I have given up on compiling RawTherapee with clang++ for now, and use
>> GCC 4.8 on all systems.  RawTherapee is somewhat demanding, especially
>> at higher optimization level, and kills the 10.0-RELEASE base clang and
>> Port GCC 4.6 and 4.7, all with internal compiler errors.  Since GCC 4.8
>> worked for me, I did not bother to send Gerald the details.
>>
>> We may want to retry with clang if we've got the next clang version.
>> Feel free to use Rawtherapee as compiler system test ;)
>>
>
> try with something like this in libmap.conf
> libc++.so.1 /usr/local/lib/libc++.so.1
> If that fixes the problem, then a rpath with /usr/local/lib should be set while
> building the port

Hmm, I am not very familiar with libmapping. After adding it to
/etc/libmap.conf I get

#rawtherapee
Shared object "/usr/local/lib/libc++.so.1" not found, required by
"rawtherapee"

Thanks for the tip,
Rainer

>
> regards,
> Bapt
>
_______________________________________________
[hidden email] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "[hidden email]"
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Rainer Hurling


Am 06.02.2014 07:03 (UTC+1) schrieb Rainer Hurling:

> Am 05.02.2014 22:20 (UTC+1) schrieb Baptiste Daroussin:
>> On Wed, Feb 05, 2014 at 09:19:51PM +0100, Matthias Andree wrote:
>>> Am 05.02.2014 21:08, schrieb Dimitry Andric:
>>>
>>>>> #17 0x00000000484c0ee0 in std::__1::locale::id::__next_id () from
>>>>> /usr/local/lib/libc++.so.1
>>>>
>>>> Hmm, is this a ports version of libc++?  I was not aware Baptiste had
>>>> already committed this? :)
>>>
>>> Yes, it is (as a build requisite, but apparently remained installed on
>>> the destination machine), because we need to match the libraries that
>>> the requisites use (Glibmm for one).
>>>
>>> I have given up on compiling RawTherapee with clang++ for now, and use
>>> GCC 4.8 on all systems.  RawTherapee is somewhat demanding, especially
>>> at higher optimization level, and kills the 10.0-RELEASE base clang and
>>> Port GCC 4.6 and 4.7, all with internal compiler errors.  Since GCC 4.8
>>> worked for me, I did not bother to send Gerald the details.
>>>
>>> We may want to retry with clang if we've got the next clang version.
>>> Feel free to use Rawtherapee as compiler system test ;)
>>>
>>
>> try with something like this in libmap.conf
>> libc++.so.1 /usr/local/lib/libc++.so.1
>> If that fixes the problem, then a rpath with /usr/local/lib should be set while
>> building the port
>
> Hmm, I am not very familiar with libmapping. After adding it to
> /etc/libmap.conf I get
>
> #rawtherapee
> Shared object "/usr/local/lib/libc++.so.1" not found, required by
> "rawtherapee"

I just recognized, that in my CURRENT boxes in base their are two
versions of libc++:

#ll /usr/lib/libc++.so*
-r--r--r--  1 root  wheel  -    134  3 Aug 22:33:00 2013 /usr/lib/libc++.so
-r--r--r--  1 root  wheel  - 768248  4 Feb 18:08:00 2014
/usr/lib/libc++.so.1

Shouldn't libc++.so be a link to libc++.so.1 or at least also come from
the newest built?

>
> Thanks for the tip,
> Rainer
>
>>
>> regards,
>> Bapt
_______________________________________________
[hidden email] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "[hidden email]"
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Dimitry Andric-4
On 06 Feb 2014, at 07:48, Rainer Hurling <[hidden email]> wrote:
...

> I just recognized, that in my CURRENT boxes in base their are two
> versions of libc++:
>
> #ll /usr/lib/libc++.so*
> -r--r--r--  1 root  wheel  -    134  3 Aug 22:33:00 2013 /usr/lib/libc++.so
> -r--r--r--  1 root  wheel  - 768248  4 Feb 18:08:00 2014
> /usr/lib/libc++.so.1
>
> Shouldn't libc++.so be a link to libc++.so.1 or at least also come from
> the newest built?
No, libc++.so is a linker script, similar to libc.so:

$ cat /usr/lib/libc++.so
/* $FreeBSD: head/lib/libc++/libc++.ldscript 253917 2013-08-03 16:23:43Z dim $ */
GROUP ( /usr/lib/libc++.so.1 /usr/lib/libcxxrt.so )

-Dimitry


signature.asc (210 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Baptiste Daroussin-2
In reply to this post by Rainer Hurling
On Thu, Feb 06, 2014 at 07:03:22AM +0100, Rainer Hurling wrote:

> Am 05.02.2014 22:20 (UTC+1) schrieb Baptiste Daroussin:
> > On Wed, Feb 05, 2014 at 09:19:51PM +0100, Matthias Andree wrote:
> >> Am 05.02.2014 21:08, schrieb Dimitry Andric:
> >>
> >>>> #17 0x00000000484c0ee0 in std::__1::locale::id::__next_id () from
> >>>> /usr/local/lib/libc++.so.1
> >>>
> >>> Hmm, is this a ports version of libc++?  I was not aware Baptiste had
> >>> already committed this? :)
> >>
> >> Yes, it is (as a build requisite, but apparently remained installed on
> >> the destination machine), because we need to match the libraries that
> >> the requisites use (Glibmm for one).
> >>
> >> I have given up on compiling RawTherapee with clang++ for now, and use
> >> GCC 4.8 on all systems.  RawTherapee is somewhat demanding, especially
> >> at higher optimization level, and kills the 10.0-RELEASE base clang and
> >> Port GCC 4.6 and 4.7, all with internal compiler errors.  Since GCC 4.8
> >> worked for me, I did not bother to send Gerald the details.
> >>
> >> We may want to retry with clang if we've got the next clang version.
> >> Feel free to use Rawtherapee as compiler system test ;)
> >>
> >
> > try with something like this in libmap.conf
> > libc++.so.1 /usr/local/lib/libc++.so.1
> > If that fixes the problem, then a rpath with /usr/local/lib should be set while
> > building the port
>
> Hmm, I am not very familiar with libmapping. After adding it to
> /etc/libmap.conf I get
>
> #rawtherapee
> Shared object "/usr/local/lib/libc++.so.1" not found, required by
> "rawtherapee"
>
> Thanks for the tip,
> Rainer
>
> >
> > regards,
> > Bapt
> >
try reinstalling devel/libc++ and keeping the libmap.conf entry, that should do
the trick

as it was a build only dep it may have been removed.
just remove the line from libmap.conf before reinstalling devel/libc++ and readd
it once it is installed.

Bapt

attachment0 (203 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Rainer Hurling


Am 06.02.2014 08:52 (UTC+1) schrieb Baptiste Daroussin:

> On Thu, Feb 06, 2014 at 07:03:22AM +0100, Rainer Hurling wrote:
>> Am 05.02.2014 22:20 (UTC+1) schrieb Baptiste Daroussin:
>>> On Wed, Feb 05, 2014 at 09:19:51PM +0100, Matthias Andree wrote:
>>>> Am 05.02.2014 21:08, schrieb Dimitry Andric:
>>>>
>>>>>> #17 0x00000000484c0ee0 in std::__1::locale::id::__next_id () from
>>>>>> /usr/local/lib/libc++.so.1
>>>>>
>>>>> Hmm, is this a ports version of libc++?  I was not aware Baptiste had
>>>>> already committed this? :)
>>>>
>>>> Yes, it is (as a build requisite, but apparently remained installed on
>>>> the destination machine), because we need to match the libraries that
>>>> the requisites use (Glibmm for one).
>>>>
>>>> I have given up on compiling RawTherapee with clang++ for now, and use
>>>> GCC 4.8 on all systems.  RawTherapee is somewhat demanding, especially
>>>> at higher optimization level, and kills the 10.0-RELEASE base clang and
>>>> Port GCC 4.6 and 4.7, all with internal compiler errors.  Since GCC 4.8
>>>> worked for me, I did not bother to send Gerald the details.
>>>>
>>>> We may want to retry with clang if we've got the next clang version.
>>>> Feel free to use Rawtherapee as compiler system test ;)
>>>>
>>>
>>> try with something like this in libmap.conf
>>> libc++.so.1 /usr/local/lib/libc++.so.1
>>> If that fixes the problem, then a rpath with /usr/local/lib should be set while
>>> building the port
>>
>> Hmm, I am not very familiar with libmapping. After adding it to
>> /etc/libmap.conf I get
>>
>> #rawtherapee
>> Shared object "/usr/local/lib/libc++.so.1" not found, required by
>> "rawtherapee"
>>
>> Thanks for the tip,
>> Rainer
>>
>>>
>>> regards,
>>> Bapt
>>>
>
> try reinstalling devel/libc++ and keeping the libmap.conf entry, that should do
> the trick
>
> as it was a build only dep it may have been removed.
> just remove the line from libmap.conf before reinstalling devel/libc++ and readd
> it once it is installed.

I commented out libmap.conf entry, reinstalled devel/libc++ and readded
libmap.conf entry.

After that, I get the same error, when starting rawtherapee.

In a second step I tried to rebuild graphics/rawtherapee with the entry
in /etc/libmap.conf active. That also fails with:

[..snip..]
/bin/mkdir -p /usr/ports/graphics/rawtherapee/work/.build
Shared object "/usr/local/lib/libc++.so.1" not found, required by "cmake"
*** Error code 1
Stop.
make[1]: stopped in /usr/ports/graphics/rawtherapee
*** Error code 1

Rainer

>
> Bapt
>
_______________________________________________
[hidden email] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "[hidden email]"
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Rainer Hurling
In reply to this post by Dimitry Andric-4
Am 06.02.2014 08:23 (UTC+1) schrieb Dimitry Andric:

> On 06 Feb 2014, at 07:48, Rainer Hurling <[hidden email]> wrote:
> ...
>> I just recognized, that in my CURRENT boxes in base their are two
>> versions of libc++:
>>
>> #ll /usr/lib/libc++.so*
>> -r--r--r--  1 root  wheel  -    134  3 Aug 22:33:00 2013 /usr/lib/libc++.so
>> -r--r--r--  1 root  wheel  - 768248  4 Feb 18:08:00 2014
>> /usr/lib/libc++.so.1
>>
>> Shouldn't libc++.so be a link to libc++.so.1 or at least also come from
>> the newest built?
>
> No, libc++.so is a linker script, similar to libc.so:
>
> $ cat /usr/lib/libc++.so
> /* $FreeBSD: head/lib/libc++/libc++.ldscript 253917 2013-08-03 16:23:43Z dim $ */
> GROUP ( /usr/lib/libc++.so.1 /usr/lib/libcxxrt.so )

Oh, yes. I did not look enough at it, sorry. So no problem here in this
place.

Thanks,
Rainer

>
> -Dimitry
>
_______________________________________________
[hidden email] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "[hidden email]"
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Erich Dollansky-7
In reply to this post by Rainer Hurling
Hi,

I just faced the same problem.

On Thu, 06 Feb 2014 10:27:12 +0100
Rainer Hurling <[hidden email]> wrote:

>
>
> Am 06.02.2014 08:52 (UTC+1) schrieb Baptiste Daroussin:
> > On Thu, Feb 06, 2014 at 07:03:22AM +0100, Rainer Hurling wrote:
> >> Am 05.02.2014 22:20 (UTC+1) schrieb Baptiste Daroussin:
> >>> On Wed, Feb 05, 2014 at 09:19:51PM +0100, Matthias Andree wrote:
> >>>> Am 05.02.2014 21:08, schrieb Dimitry Andric:
> >>>>
> >>>>>> #17 0x00000000484c0ee0 in std::__1::locale::id::__next_id ()
> >>>>>> from /usr/local/lib/libc++.so.1
> >>>>>
> >>>>> Hmm, is this a ports version of libc++?  I was not aware
> >>>>> Baptiste had already committed this? :)
> >>>>
> >>>> Yes, it is (as a build requisite, but apparently remained
> >>>> installed on the destination machine), because we need to match
> >>>> the libraries that the requisites use (Glibmm for one).
> >>>>
> >>>> I have given up on compiling RawTherapee with clang++ for now,
> >>>> and use GCC 4.8 on all systems.  RawTherapee is somewhat
> >>>> demanding, especially at higher optimization level, and kills
> >>>> the 10.0-RELEASE base clang and Port GCC 4.6 and 4.7, all with
> >>>> internal compiler errors.  Since GCC 4.8 worked for me, I did
> >>>> not bother to send Gerald the details.
> >>>>
> >>>> We may want to retry with clang if we've got the next clang
> >>>> version. Feel free to use Rawtherapee as compiler system test ;)
> >>>>
> >>>
> >>> try with something like this in libmap.conf
> >>> libc++.so.1 /usr/local/lib/libc++.so.1
> >>> If that fixes the problem, then a rpath with /usr/local/lib
> >>> should be set while building the port
> >>
> >> Hmm, I am not very familiar with libmapping. After adding it to
> >> /etc/libmap.conf I get
> >>
> >> #rawtherapee
> >> Shared object "/usr/local/lib/libc++.so.1" not found, required by
> >> "rawtherapee"
> >>
> >> Thanks for the tip,
> >> Rainer
> >>
> >>>
> >>> regards,
> >>> Bapt
> >>>
> >
> > try reinstalling devel/libc++ and keeping the libmap.conf entry,
> > that should do the trick
> >
> > as it was a build only dep it may have been removed.
> > just remove the line from libmap.conf before reinstalling
> > devel/libc++ and readd it once it is installed.
>
> I commented out libmap.conf entry, reinstalled devel/libc++ and
> readded libmap.conf entry.
>
> After that, I get the same error, when starting rawtherapee.
>
> In a second step I tried to rebuild graphics/rawtherapee with the
> entry in /etc/libmap.conf active. That also fails with:
>
> [..snip..]
> /bin/mkdir -p /usr/ports/graphics/rawtherapee/work/.build
> Shared object "/usr/local/lib/libc++.so.1" not found, required by
> "cmake" *** Error code 1
> Stop.
> make[1]: stopped in /usr/ports/graphics/rawtherapee
> *** Error code 1
>
It looks to me that the entry in libmap.conf is not even needed as
there is a link in /usr/local/lib anyway.

Rawtherapee is a very sensitive program from my point of view. It works
after one update and it crashes after the next. It might also be just
random.

Erich
_______________________________________________
[hidden email] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "[hidden email]"
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Matthias Andree-2
Am 07.02.2014 05:22, schrieb Erich Dollansky:
> Hi,
>
> I just faced the same problem.
...

> It looks to me that the entry in libmap.conf is not even needed as
> there is a link in /usr/local/lib anyway.
>
> Rawtherapee is a very sensitive program from my point of view. It works
> after one update and it crashes after the next. It might also be just
> random.

Sorry, but that won't help.  Either someone (not me) can debug this, or
we'll just not use the port on 11.  I have marked the port BROKEN on
OSVERSION >= 1100000.

If someone wants to use rawtherapee on 11, he'll have to do some
digging/debugging; and the port build is still possible by adding
TRYBROKEN=yes to the make command line, and possibly WITH_DEBUG=yes on
several of the requisites and on rawtherapee itself, or I'll need to
collect at least three plausible reports from FreeBSD 11 systems that
rawtherapee works there.

_______________________________________________
[hidden email] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "[hidden email]"
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Waitman Gobble-3
In reply to this post by Rainer Hurling

On Wed, February 5, 2014 10:03 pm, Rainer Hurling wrote:

> Am 05.02.2014 22:20 (UTC+1) schrieb Baptiste Daroussin:
>
>> On Wed, Feb 05, 2014 at 09:19:51PM +0100, Matthias Andree wrote:
>>
>>> Am 05.02.2014 21:08, schrieb Dimitry Andric:
>>>
>>>
>>>>> #17 0x00000000484c0ee0 in std::__1::locale::id::__next_id () from
>>>>>  /usr/local/lib/libc++.so.1
>>>>>
>>>>
>>>> Hmm, is this a ports version of libc++?  I was not aware Baptiste
>>>> had already committed this? :)
>>>
>>> Yes, it is (as a build requisite, but apparently remained installed
>>> on the destination machine), because we need to match the libraries
>>> that the requisites use (Glibmm for one).
>>>
>>> I have given up on compiling RawTherapee with clang++ for now, and
>>> use GCC 4.8 on all systems.  RawTherapee is somewhat demanding,
>>> especially at higher optimization level, and kills the 10.0-RELEASE
>>> base clang and Port GCC 4.6 and 4.7, all with internal compiler
>>> errors.  Since GCC 4.8 worked for me, I did not bother to send Gerald
>>> the details.
>>>
>>> We may want to retry with clang if we've got the next clang version.
>>> Feel free to use Rawtherapee as compiler system test ;)
>>>
>>>
>>
>> try with something like this in libmap.conf libc++.so.1
>> /usr/local/lib/libc++.so.1
>> If that fixes the problem, then a rpath with /usr/local/lib should be
>> set while building the port
>
> Hmm, I am not very familiar with libmapping. After adding it to
> /etc/libmap.conf I get
>
>
> #rawtherapee
> Shared object "/usr/local/lib/libc++.so.1" not found, required by
> "rawtherapee"
>
>
> Thanks for the tip,
> Rainer
>
>
>>
>> regards, Bapt
>>
>>
> _______________________________________________
> [hidden email] mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-ports
> To unsubscribe, send any mail to "[hidden email]"
>
>


Here are a few points regarding rawtherapee on 11.0-CURRENT.

1) it builds fine using clang (removing all the gcc4.8 from makefile). but
still crashes the same.

2) the parse errors such as
/usr/local/share/rawtherapee/themes/25-Gray-Gray.gtkrc:96: error:
unexpected identifier `colorize_scrollbar', expected character `}'

can be eliminated by installing ports/x11-themes/clearlooks-phenix-theme,
this is the clearlooks theme ported to gtk3

3) I believe i'be narrowed the crash down to Glib::ustring::format. I've
experimented with various examples and they all crash. Currently exploring
this problem to see what's up.. anyone have some clue?

here is a sample:

#include <iostream>
#include <iomanip>
#include <glibmm.h>

int main() {

        double value = 22.0 / 7.0;
        std::cout << value << std::endl;
        Glib::ustring test =
Glib::ustring::format(std::fixed,std::setprecision(2),value);
        std::cout << test << std::endl;

}


$ clang++ -o testglib `pkg-config --cflags --libs glibmm-2.4` testglib.cpp
$ ./testglib
3.14286
Abort trap (core dumped)


$ /usr/local/bin/gdb testglib
GNU gdb (GDB) 7.6.2 [GDB v7.6.2 for FreeBSD]
Copyright (C) 2013 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-portbld-freebsd11.0".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /usr/home/waitman/testglib...done.
(gdb) run
Starting program: /usr/home/waitman/testglib
[New LWP 100327]
3.14286
[New Thread 803006400 (LWP 100327)]

Program received signal SIGABRT, Aborted.
[Switching to Thread 803006400 (LWP 100327)]
0x0000000801e313ba in kill () from /lib/libc.so.7
(gdb) bt
#0  0x0000000801e313ba in kill () from /lib/libc.so.7
#1  0x0000000801e30069 in abort () from /lib/libc.so.7
#2  0x00000008016ce7da in ?? () from /lib/libcxxrt.so.1
#3  0x000000080085de68 in Glib::ConvertError::throw_func(_GError*) ()
   from /usr/local/lib/libglibmm-2.4.so.1
#4  0x000000080086ed2f in Glib::Error::throw_exception(_GError*) ()
   from /usr/local/lib/libglibmm-2.4.so.1
#5  0x000000080087bcc2 in Glib::ustring::FormatStream::to_string() const ()
   from /usr/local/lib/libglibmm-2.4.so.1
#6  0x0000000000401de0 in Glib::ustring::format<std::__1::ios_base&
(std::__1::ios_base&), std::__1::__iom_t5, double>(std::__1::ios_base& (
const&)(std::__1::ios_base&), std::__1::__iom_t5 const&, double const&) (
    a1= 0x401e50: {std::__1::ios_base &(std::__1::ios_base &)} 0x401e50
<std::__1::fixed(std::__1::ios_base&)>, a2=..., a3= 0x7fffffffd860:
3.1428571428571428)
    at /usr/local/include/glibmm-2.4/glibmm/ustring.h:1158
#7  0x0000000000401696 in main () at testglib.cpp:9
(gdb)



$ uname -a
FreeBSD akira.waitman.net 11.0-CURRENT FreeBSD 11.0-CURRENT #1 r261643:
Sat Feb  8 22:11:05 PST 2014
root akira waitman net:/usr/obj/usr/src/sys/AKIRA  amd64

$ ldd testglib
testglib:
        libglibmm-2.4.so.1 => /usr/local/lib/libglibmm-2.4.so.1 (0x80081e000)
        libgobject-2.0.so.0 => /usr/local/lib/libgobject-2.0.so.0
(0x800a9d000)
        libglib-2.0.so.0 => /usr/local/lib/libglib-2.0.so.0 (0x800cea000)
        libintl.so.9 => /usr/local/lib/libintl.so.9 (0x800ff2000)
        libsigc-2.0.so.0 => /usr/local/lib/libsigc-2.0.so.0 (0x8011fb000)
        libc++.so.1 => /usr/lib/libc++.so.1 (0x801400000)
        libcxxrt.so.1 => /lib/libcxxrt.so.1 (0x8016be000)
        libm.so.5 => /lib/libm.so.5 (0x8018d8000)
        libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x801b00000)
        libc.so.7 => /lib/libc.so.7 (0x801d0d000)
        libffi.so.6 => /usr/local/lib/libffi.so.6 (0x80208a000)
        libgmodule-2.0.so.0 => /usr/local/lib/libgmodule-2.0.so.0
(0x802291000)
        libpcre.so.3 => /usr/local/lib/libpcre.so.3 (0x802494000)
        libthr.so.3 => /lib/libthr.so.3 (0x8026fb000)


Thank you,



--
Waitman Gobble
San Jose California USA
+1.510-830-7975
_______________________________________________
[hidden email] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "[hidden email]"
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Waitman Gobble-2
In reply to this post by Rainer Hurling

On Wed, February 5, 2014 10:03 pm, Rainer Hurling wrote:

> Am 05.02.2014 22:20 (UTC+1) schrieb Baptiste Daroussin:
>
>> On Wed, Feb 05, 2014 at 09:19:51PM +0100, Matthias Andree wrote:
>>
>>> Am 05.02.2014 21:08, schrieb Dimitry Andric:
>>>
>>>
>>>>> #17 0x00000000484c0ee0 in std::__1::locale::id::__next_id () from
>>>>>  /usr/local/lib/libc++.so.1
>>>>>
>>>>
>>>> Hmm, is this a ports version of libc++?  I was not aware Baptiste
>>>> had already committed this? :)
>>>
>>> Yes, it is (as a build requisite, but apparently remained installed
>>> on the destination machine), because we need to match the libraries
>>> that the requisites use (Glibmm for one).
>>>
>>> I have given up on compiling RawTherapee with clang++ for now, and
>>> use GCC 4.8 on all systems.  RawTherapee is somewhat demanding,
>>> especially at higher optimization level, and kills the 10.0-RELEASE
>>> base clang and Port GCC 4.6 and 4.7, all with internal compiler
>>> errors.  Since GCC 4.8 worked for me, I did not bother to send Gerald
>>> the details.
>>>
>>> We may want to retry with clang if we've got the next clang version.
>>> Feel free to use Rawtherapee as compiler system test ;)
>>>
>>>
>>
>> try with something like this in libmap.conf libc++.so.1
>> /usr/local/lib/libc++.so.1
>> If that fixes the problem, then a rpath with /usr/local/lib should be
>> set while building the port
>
> Hmm, I am not very familiar with libmapping. After adding it to
> /etc/libmap.conf I get
>
>
> #rawtherapee
> Shared object "/usr/local/lib/libc++.so.1" not found, required by
> "rawtherapee"
>
>
> Thanks for the tip,
> Rainer
>
>
>>
>> regards, Bapt
>>
>>
> _______________________________________________
> [hidden email] mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-ports
> To unsubscribe, send any mail to "[hidden email]"
>
>


Here are a few points regarding rawtherapee on 11.0-CURRENT.

1) it builds fine using clang (removing all the gcc4.8 from makefile). but
still crashes the same.

2) the parse errors such as
/usr/local/share/rawtherapee/themes/25-Gray-Gray.gtkrc:96: error:
unexpected identifier `colorize_scrollbar', expected character `}'

can be eliminated by installing ports/x11-themes/clearlooks-phenix-theme,
this is the clearlooks theme ported to gtk3

3) I believe i'be narrowed the crash down to Glib::ustring::format. I've
experimented with various examples and they all crash. Currently exploring
this problem to see what's up.. anyone have some clue?

here is a sample:

#include <iostream>
#include <iomanip>
#include <glibmm.h>

int main() {

        double value = 22.0 / 7.0;
        std::cout << value << std::endl;
        Glib::ustring test =
Glib::ustring::format(std::fixed,std::setprecision(2),value);
        std::cout << test << std::endl;

}


$ clang++ -o testglib `pkg-config --cflags --libs glibmm-2.4` testglib.cpp
$ ./testglib
3.14286
Abort trap (core dumped)


$ /usr/local/bin/gdb testglib
GNU gdb (GDB) 7.6.2 [GDB v7.6.2 for FreeBSD]
Copyright (C) 2013 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-portbld-freebsd11.0".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /usr/home/waitman/testglib...done.
(gdb) run
Starting program: /usr/home/waitman/testglib
[New LWP 100327]
3.14286
[New Thread 803006400 (LWP 100327)]

Program received signal SIGABRT, Aborted.
[Switching to Thread 803006400 (LWP 100327)]
0x0000000801e313ba in kill () from /lib/libc.so.7
(gdb) bt
#0  0x0000000801e313ba in kill () from /lib/libc.so.7
#1  0x0000000801e30069 in abort () from /lib/libc.so.7
#2  0x00000008016ce7da in ?? () from /lib/libcxxrt.so.1
#3  0x000000080085de68 in Glib::ConvertError::throw_func(_GError*) ()
   from /usr/local/lib/libglibmm-2.4.so.1
#4  0x000000080086ed2f in Glib::Error::throw_exception(_GError*) ()
   from /usr/local/lib/libglibmm-2.4.so.1
#5  0x000000080087bcc2 in Glib::ustring::FormatStream::to_string() const ()
   from /usr/local/lib/libglibmm-2.4.so.1
#6  0x0000000000401de0 in Glib::ustring::format<std::__1::ios_base&
(std::__1::ios_base&), std::__1::__iom_t5, double>(std::__1::ios_base& (
const&)(std::__1::ios_base&), std::__1::__iom_t5 const&, double const&) (
    a1= 0x401e50: {std::__1::ios_base &(std::__1::ios_base &)} 0x401e50
<std::__1::fixed(std::__1::ios_base&)>, a2=..., a3= 0x7fffffffd860:
3.1428571428571428)
    at /usr/local/include/glibmm-2.4/glibmm/ustring.h:1158
#7  0x0000000000401696 in main () at testglib.cpp:9
(gdb)



$ uname -a
FreeBSD akira.waitman.net 11.0-CURRENT FreeBSD 11.0-CURRENT #1 r261643:
Sat Feb  8 22:11:05 PST 2014
root akira waitman net:/usr/obj/usr/src/sys/AKIRA  amd64

$ ldd testglib
testglib:
        libglibmm-2.4.so.1 => /usr/local/lib/libglibmm-2.4.so.1 (0x80081e000)
        libgobject-2.0.so.0 => /usr/local/lib/libgobject-2.0.so.0
(0x800a9d000)
        libglib-2.0.so.0 => /usr/local/lib/libglib-2.0.so.0 (0x800cea000)
        libintl.so.9 => /usr/local/lib/libintl.so.9 (0x800ff2000)
        libsigc-2.0.so.0 => /usr/local/lib/libsigc-2.0.so.0 (0x8011fb000)
        libc++.so.1 => /usr/lib/libc++.so.1 (0x801400000)
        libcxxrt.so.1 => /lib/libcxxrt.so.1 (0x8016be000)
        libm.so.5 => /lib/libm.so.5 (0x8018d8000)
        libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x801b00000)
        libc.so.7 => /lib/libc.so.7 (0x801d0d000)
        libffi.so.6 => /usr/local/lib/libffi.so.6 (0x80208a000)
        libgmodule-2.0.so.0 => /usr/local/lib/libgmodule-2.0.so.0
(0x802291000)
        libpcre.so.3 => /usr/local/lib/libpcre.so.3 (0x802494000)
        libthr.so.3 => /lib/libthr.so.3 (0x8026fb000)


Thank you,



--
Waitman Gobble
San Jose California USA
+1.510-830-7975
_______________________________________________
[hidden email] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "[hidden email]"
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Waitman Gobble-3

On Sun, February 9, 2014 4:54 am, Waitman Gobble wrote:
>

> On Wed, February 5, 2014 10:03 pm, Rainer Hurling wrote:
>
>> Am 05.02.2014 22:20 (UTC+1) schrieb Baptiste Daroussin:
>>
>>
>>> On Wed, Feb 05, 2014 at 09:19:51PM +0100, Matthias Andree wrote:
>>>
>>>
>>>> Am 05.02.2014 21:08, schrieb Dimitry Andric:
>>>>
>>>>
>>>>
>>>>>> #17 0x00000000484c0ee0 in std::__1::locale::id::__next_id ()
>>>>>> from /usr/local/lib/libc++.so.1
>>>>>>
>>>>>>
>>>>>
>>>>> Hmm, is this a ports version of libc++?  I was not aware Baptiste
>>>>>  had already committed this? :)
>>>>
>>>> Yes, it is (as a build requisite, but apparently remained installed
>>>>  on the destination machine), because we need to match the
>>>> libraries that the requisites use (Glibmm for one).
>>>>
>>>> I have given up on compiling RawTherapee with clang++ for now, and
>>>> use GCC 4.8 on all systems.  RawTherapee is somewhat demanding,
>>>> especially at higher optimization level, and kills the 10.0-RELEASE
>>>>  base clang and Port GCC 4.6 and 4.7, all with internal compiler
>>>> errors.  Since GCC 4.8 worked for me, I did not bother to send
>>>> Gerald
>>>> the details.
>>>>
>>>> We may want to retry with clang if we've got the next clang
>>>> version. Feel free to use Rawtherapee as compiler system test ;)
>>>>
>>>>
>>>>
>>>
>>> try with something like this in libmap.conf libc++.so.1
>>> /usr/local/lib/libc++.so.1
>>> If that fixes the problem, then a rpath with /usr/local/lib should be
>>> set while building the port
>>
>> Hmm, I am not very familiar with libmapping. After adding it to
>> /etc/libmap.conf I get
>>
>>
>>
>> #rawtherapee
>> Shared object "/usr/local/lib/libc++.so.1" not found, required by
>> "rawtherapee"
>>
>>
>>
>> Thanks for the tip,
>> Rainer
>>
>>
>>
>>>
>>> regards, Bapt
>>>
>>>
>> _______________________________________________
>> [hidden email] mailing list
>> http://lists.freebsd.org/mailman/listinfo/freebsd-ports
>> To unsubscribe, send any mail to "[hidden email]"
>>
>>
>>
>
>
> Here are a few points regarding rawtherapee on 11.0-CURRENT.
>
>
> 1) it builds fine using clang (removing all the gcc4.8 from makefile).
> but still crashes the same.
>
> 2) the parse errors such as
> /usr/local/share/rawtherapee/themes/25-Gray-Gray.gtkrc:96: error:
> unexpected identifier `colorize_scrollbar', expected character `}'
>
> can be eliminated by installing ports/x11-themes/clearlooks-phenix-theme,
>  this is the clearlooks theme ported to gtk3
>
> 3) I believe i'be narrowed the crash down to Glib::ustring::format. I've
> experimented with various examples and they all crash. Currently exploring
>  this problem to see what's up.. anyone have some clue?
>
> here is a sample:
>
> #include <iostream>
> #include <iomanip>
> #include <glibmm.h>
>
>
> int main() {
>
> double value = 22.0 / 7.0; std::cout << value << std::endl;
> Glib::ustring test =
> Glib::ustring::format(std::fixed,std::setprecision(2),value);
> std::cout << test << std::endl;
>
>
> }
>
>
>
> $ clang++ -o testglib `pkg-config --cflags --libs glibmm-2.4`
> testglib.cpp $ ./testglib
> 3.14286
> Abort trap (core dumped)
>
>
>
> $ /usr/local/bin/gdb testglib
> GNU gdb (GDB) 7.6.2 [GDB v7.6.2 for FreeBSD]
> Copyright (C) 2013 Free Software Foundation, Inc.
> License GPLv3+: GNU GPL version 3 or later
> <http://gnu.org/licenses/gpl.html>
> This is free software: you are free to change and redistribute it.
> There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
>  and "show warranty" for details. This GDB was configured as
> "x86_64-portbld-freebsd11.0".
> For bug reporting instructions, please see:
> <http://www.gnu.org/software/gdb/bugs/>...
> Reading symbols from /usr/home/waitman/testglib...done.
> (gdb) run
> Starting program: /usr/home/waitman/testglib
> [New LWP 100327]
> 3.14286
> [New Thread 803006400 (LWP 100327)]
>
>
> Program received signal SIGABRT, Aborted.
> [Switching to Thread 803006400 (LWP 100327)]
> 0x0000000801e313ba in kill () from /lib/libc.so.7
> (gdb) bt
> #0  0x0000000801e313ba in kill () from /lib/libc.so.7
> #1  0x0000000801e30069 in abort () from /lib/libc.so.7
> #2  0x00000008016ce7da in ?? () from /lib/libcxxrt.so.1
> #3  0x000000080085de68 in Glib::ConvertError::throw_func(_GError*) ()
> from /usr/local/lib/libglibmm-2.4.so.1 #4  0x000000080086ed2f in
> Glib::Error::throw_exception(_GError*) ()
> from /usr/local/lib/libglibmm-2.4.so.1 #5  0x000000080087bcc2 in
> Glib::ustring::FormatStream::to_string() const ()
> from /usr/local/lib/libglibmm-2.4.so.1 #6  0x0000000000401de0 in
> Glib::ustring::format<std::__1::ios_base&
> (std::__1::ios_base&), std::__1::__iom_t5, double>(std::__1::ios_base& (
> const&)(std::__1::ios_base&), std::__1::__iom_t5 const&, double const&) (
> a1= 0x401e50: {std::__1::ios_base &(std::__1::ios_base &)} 0x401e50
> <std::__1::fixed(std::__1::ios_base&)>, a2=..., a3= 0x7fffffffd860:
> 3.1428571428571428)
> at /usr/local/include/glibmm-2.4/glibmm/ustring.h:1158 #7
> 0x0000000000401696 in main () at testglib.cpp:9
> (gdb)
>
>
>
>
> $ uname -a
> FreeBSD akira.waitman.net 11.0-CURRENT FreeBSD 11.0-CURRENT #1 r261643:
> Sat Feb  8 22:11:05 PST 2014
> root akira waitman net:/usr/obj/usr/src/sys/AKIRA  amd64
>
> $ ldd testglib
> testglib:
> libglibmm-2.4.so.1 => /usr/local/lib/libglibmm-2.4.so.1 (0x80081e000)
> libgobject-2.0.so.0 => /usr/local/lib/libgobject-2.0.so.0 (0x800a9d000)
> libglib-2.0.so.0 => /usr/local/lib/libglib-2.0.so.0 (0x800cea000)
> libintl.so.9 => /usr/local/lib/libintl.so.9 (0x800ff2000) libsigc-2.0.so.0
> => /usr/local/lib/libsigc-2.0.so.0 (0x8011fb000)
> libc++.so.1 => /usr/lib/libc++.so.1 (0x801400000) libcxxrt.so.1 =>
> /lib/libcxxrt.so.1 (0x8016be000)
> libm.so.5 => /lib/libm.so.5 (0x8018d8000) libgcc_s.so.1 =>
> /lib/libgcc_s.so.1 (0x801b00000)
> libc.so.7 => /lib/libc.so.7 (0x801d0d000) libffi.so.6 =>
> /usr/local/lib/libffi.so.6 (0x80208a000)
> libgmodule-2.0.so.0 => /usr/local/lib/libgmodule-2.0.so.0 (0x802291000)
> libpcre.so.3 => /usr/local/lib/libpcre.so.3 (0x802494000) libthr.so.3 =>
> /lib/libthr.so.3 (0x8026fb000)
>
>
>
> Thank you,
>
>
>
>
> --
> Waitman Gobble
> San Jose California USA
> +1.510-830-7975
> _______________________________________________
> [hidden email] mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-ports
> To unsubscribe, send any mail to "[hidden email]"
>
>

update, if i wrap the segfaulting 101 line in a try block, in rtimage.cc
(which has the format code that bombs out), i get a Glib::Error::what() of

"Conversion from character set 'UTF-8' to 'WCHAR_T'"

so I think this is a hint about the culprit..

ideas?

--
Waitman Gobble
San Jose California USA
+1.510-830-7975
_______________________________________________
[hidden email] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "[hidden email]"
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Waitman Gobble-2
In reply to this post by Rainer Hurling


On Sun, February 9, 2014 4:54 am, Waitman Gobble wrote:
>

> On Wed, February 5, 2014 10:03 pm, Rainer Hurling wrote:
>
>> Am 05.02.2014 22:20 (UTC+1) schrieb Baptiste Daroussin:
>>
>>
>>> On Wed, Feb 05, 2014 at 09:19:51PM +0100, Matthias Andree wrote:
>>>
>>>
>>>> Am 05.02.2014 21:08, schrieb Dimitry Andric:
>>>>
>>>>
>>>>
>>>>>> #17 0x00000000484c0ee0 in std::__1::locale::id::__next_id ()
>>>>>> from /usr/local/lib/libc++.so.1
>>>>>>
>>>>>>
>>>>>
>>>>> Hmm, is this a ports version of libc++?  I was not aware Baptiste
>>>>>  had already committed this? :)
>>>>
>>>> Yes, it is (as a build requisite, but apparently remained installed
>>>>  on the destination machine), because we need to match the
>>>> libraries that the requisites use (Glibmm for one).
>>>>
>>>> I have given up on compiling RawTherapee with clang++ for now, and
>>>> use GCC 4.8 on all systems.  RawTherapee is somewhat demanding,
>>>> especially at higher optimization level, and kills the 10.0-RELEASE
>>>>  base clang and Port GCC 4.6 and 4.7, all with internal compiler
>>>> errors.  Since GCC 4.8 worked for me, I did not bother to send
>>>> Gerald
>>>> the details.
>>>>
>>>> We may want to retry with clang if we've got the next clang
>>>> version. Feel free to use Rawtherapee as compiler system test ;)
>>>>
>>>>
>>>>
>>>
>>> try with something like this in libmap.conf libc++.so.1
>>> /usr/local/lib/libc++.so.1
>>> If that fixes the problem, then a rpath with /usr/local/lib should be
>>> set while building the port
>>
>> Hmm, I am not very familiar with libmapping. After adding it to
>> /etc/libmap.conf I get
>>
>>
>>
>> #rawtherapee
>> Shared object "/usr/local/lib/libc++.so.1" not found, required by
>> "rawtherapee"
>>
>>
>>
>> Thanks for the tip,
>> Rainer
>>
>>
>>
>>>
>>> regards, Bapt
>>>
>>>
>> _______________________________________________
>> [hidden email] mailing list
>> http://lists.freebsd.org/mailman/listinfo/freebsd-ports
>> To unsubscribe, send any mail to "[hidden email]"
>>
>>
>>
>
>
> Here are a few points regarding rawtherapee on 11.0-CURRENT.
>
>
> 1) it builds fine using clang (removing all the gcc4.8 from makefile).
> but still crashes the same.
>
> 2) the parse errors such as
> /usr/local/share/rawtherapee/themes/25-Gray-Gray.gtkrc:96: error:
> unexpected identifier `colorize_scrollbar', expected character `}'
>
> can be eliminated by installing ports/x11-themes/clearlooks-phenix-theme,
>  this is the clearlooks theme ported to gtk3
>
> 3) I believe i'be narrowed the crash down to Glib::ustring::format. I've
> experimented with various examples and they all crash. Currently exploring
>  this problem to see what's up.. anyone have some clue?
>
> here is a sample:
>
> #include <iostream>
> #include <iomanip>
> #include <glibmm.h>
>
>
> int main() {
>
> double value = 22.0 / 7.0; std::cout << value << std::endl;
> Glib::ustring test =
> Glib::ustring::format(std::fixed,std::setprecision(2),value);
> std::cout << test << std::endl;
>
>
> }
>
>
>
> $ clang++ -o testglib `pkg-config --cflags --libs glibmm-2.4`
> testglib.cpp $ ./testglib
> 3.14286
> Abort trap (core dumped)
>
>
>
> $ /usr/local/bin/gdb testglib
> GNU gdb (GDB) 7.6.2 [GDB v7.6.2 for FreeBSD]
> Copyright (C) 2013 Free Software Foundation, Inc.
> License GPLv3+: GNU GPL version 3 or later
> <http://gnu.org/licenses/gpl.html>
> This is free software: you are free to change and redistribute it.
> There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
>  and "show warranty" for details. This GDB was configured as
> "x86_64-portbld-freebsd11.0".
> For bug reporting instructions, please see:
> <http://www.gnu.org/software/gdb/bugs/>...
> Reading symbols from /usr/home/waitman/testglib...done.
> (gdb) run
> Starting program: /usr/home/waitman/testglib
> [New LWP 100327]
> 3.14286
> [New Thread 803006400 (LWP 100327)]
>
>
> Program received signal SIGABRT, Aborted.
> [Switching to Thread 803006400 (LWP 100327)]
> 0x0000000801e313ba in kill () from /lib/libc.so.7
> (gdb) bt
> #0  0x0000000801e313ba in kill () from /lib/libc.so.7
> #1  0x0000000801e30069 in abort () from /lib/libc.so.7
> #2  0x00000008016ce7da in ?? () from /lib/libcxxrt.so.1
> #3  0x000000080085de68 in Glib::ConvertError::throw_func(_GError*) ()
> from /usr/local/lib/libglibmm-2.4.so.1 #4  0x000000080086ed2f in
> Glib::Error::throw_exception(_GError*) ()
> from /usr/local/lib/libglibmm-2.4.so.1 #5  0x000000080087bcc2 in
> Glib::ustring::FormatStream::to_string() const ()
> from /usr/local/lib/libglibmm-2.4.so.1 #6  0x0000000000401de0 in
> Glib::ustring::format<std::__1::ios_base&
> (std::__1::ios_base&), std::__1::__iom_t5, double>(std::__1::ios_base& (
> const&)(std::__1::ios_base&), std::__1::__iom_t5 const&, double const&) (
> a1= 0x401e50: {std::__1::ios_base &(std::__1::ios_base &)} 0x401e50
> <std::__1::fixed(std::__1::ios_base&)>, a2=..., a3= 0x7fffffffd860:
> 3.1428571428571428)
> at /usr/local/include/glibmm-2.4/glibmm/ustring.h:1158 #7
> 0x0000000000401696 in main () at testglib.cpp:9
> (gdb)
>
>
>
>
> $ uname -a
> FreeBSD akira.waitman.net 11.0-CURRENT FreeBSD 11.0-CURRENT #1 r261643:
> Sat Feb  8 22:11:05 PST 2014
> root akira waitman net:/usr/obj/usr/src/sys/AKIRA  amd64
>
> $ ldd testglib
> testglib:
> libglibmm-2.4.so.1 => /usr/local/lib/libglibmm-2.4.so.1 (0x80081e000)
> libgobject-2.0.so.0 => /usr/local/lib/libgobject-2.0.so.0 (0x800a9d000)
> libglib-2.0.so.0 => /usr/local/lib/libglib-2.0.so.0 (0x800cea000)
> libintl.so.9 => /usr/local/lib/libintl.so.9 (0x800ff2000) libsigc-2.0.so.0
> => /usr/local/lib/libsigc-2.0.so.0 (0x8011fb000)
> libc++.so.1 => /usr/lib/libc++.so.1 (0x801400000) libcxxrt.so.1 =>
> /lib/libcxxrt.so.1 (0x8016be000)
> libm.so.5 => /lib/libm.so.5 (0x8018d8000) libgcc_s.so.1 =>
> /lib/libgcc_s.so.1 (0x801b00000)
> libc.so.7 => /lib/libc.so.7 (0x801d0d000) libffi.so.6 =>
> /usr/local/lib/libffi.so.6 (0x80208a000)
> libgmodule-2.0.so.0 => /usr/local/lib/libgmodule-2.0.so.0 (0x802291000)
> libpcre.so.3 => /usr/local/lib/libpcre.so.3 (0x802494000) libthr.so.3 =>
> /lib/libthr.so.3 (0x8026fb000)
>
>
>
> Thank you,
>
>
>
>
> --
> Waitman Gobble
> San Jose California USA
> +1.510-830-7975
> _______________________________________________
> [hidden email] mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-ports
> To unsubscribe, send any mail to "[hidden email]"
>
>

update, if i wrap the segfaulting 101 line in a try block, in rtimage.cc
(which has the format code that bombs out), i get a Glib::Error::what() of

"Conversion from character set 'UTF-8' to 'WCHAR_T'"

so I think this is a hint about the culprit..

ideas?

--
Waitman Gobble
San Jose California USA
+1.510-830-7975


--
Waitman Gobble
San Jose California USA
(510) 830-7975
_______________________________________________
[hidden email] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-ports
To unsubscribe, send any mail to "[hidden email]"
Reply | Threaded
Open this post in threaded view
|

Re: graphics/rawtherapee: r342622 crashes on HEAD

Baptiste Daroussin-2
On Sun, Feb 09, 2014 at 06:30:02AM -0800, Waitman Gobble wrote:

>
>
> On Sun, February 9, 2014 4:54 am, Waitman Gobble wrote:
> >
>
> > On Wed, February 5, 2014 10:03 pm, Rainer Hurling wrote:
> >
> >> Am 05.02.2014 22:20 (UTC+1) schrieb Baptiste Daroussin:
> >>
> >>
> >>> On Wed, Feb 05, 2014 at 09:19:51PM +0100, Matthias Andree wrote:
> >>>
> >>>
> >>>> Am 05.02.2014 21:08, schrieb Dimitry Andric:
> >>>>
> >>>>
> >>>>
> >>>>>> #17 0x00000000484c0ee0 in std::__1::locale::id::__next_id ()
> >>>>>> from /usr/local/lib/libc++.so.1
> >>>>>>
> >>>>>>
> >>>>>
> >>>>> Hmm, is this a ports version of libc++?  I was not aware Baptiste
> >>>>>  had already committed this? :)
> >>>>
> >>>> Yes, it is (as a build requisite, but apparently remained installed
> >>>>  on the destination machine), because we need to match the
> >>>> libraries that the requisites use (Glibmm for one).
> >>>>
> >>>> I have given up on compiling RawTherapee with clang++ for now, and
> >>>> use GCC 4.8 on all systems.  RawTherapee is somewhat demanding,
> >>>> especially at higher optimization level, and kills the 10.0-RELEASE
> >>>>  base clang and Port GCC 4.6 and 4.7, all with internal compiler
> >>>> errors.  Since GCC 4.8 worked for me, I did not bother to send
> >>>> Gerald
> >>>> the details.
> >>>>
> >>>> We may want to retry with clang if we've got the next clang
> >>>> version. Feel free to use Rawtherapee as compiler system test ;)
> >>>>
> >>>>
> >>>>
> >>>
> >>> try with something like this in libmap.conf libc++.so.1
> >>> /usr/local/lib/libc++.so.1
> >>> If that fixes the problem, then a rpath with /usr/local/lib should be
> >>> set while building the port
> >>
> >> Hmm, I am not very familiar with libmapping. After adding it to
> >> /etc/libmap.conf I get
> >>
> >>
> >>
> >> #rawtherapee
> >> Shared object "/usr/local/lib/libc++.so.1" not found, required by
> >> "rawtherapee"
> >>
> >>
> >>
> >> Thanks for the tip,
> >> Rainer
> >>
> >>
> >>
> >>>
> >>> regards, Bapt
> >>>
> >>>
> >> _______________________________________________
> >> [hidden email] mailing list
> >> http://lists.freebsd.org/mailman/listinfo/freebsd-ports
> >> To unsubscribe, send any mail to "[hidden email]"
> >>
> >>
> >>
> >
> >
> > Here are a few points regarding rawtherapee on 11.0-CURRENT.
> >
> >
> > 1) it builds fine using clang (removing all the gcc4.8 from makefile).
> > but still crashes the same.
> >
> > 2) the parse errors such as
> > /usr/local/share/rawtherapee/themes/25-Gray-Gray.gtkrc:96: error:
> > unexpected identifier `colorize_scrollbar', expected character `}'
> >
> > can be eliminated by installing ports/x11-themes/clearlooks-phenix-theme,
> >  this is the clearlooks theme ported to gtk3
> >
> > 3) I believe i'be narrowed the crash down to Glib::ustring::format. I've
> > experimented with various examples and they all crash. Currently exploring
> >  this problem to see what's up.. anyone have some clue?
> >
> > here is a sample:
> >
> > #include <iostream>
> > #include <iomanip>
> > #include <glibmm.h>
> >
> >
> > int main() {
> >
> > double value = 22.0 / 7.0; std::cout << value << std::endl;
> > Glib::ustring test =
> > Glib::ustring::format(std::fixed,std::setprecision(2),value);
> > std::cout << test << std::endl;
> >
> >
> > }
> >
> >
> >
> > $ clang++ -o testglib `pkg-config --cflags --libs glibmm-2.4`
> > testglib.cpp $ ./testglib
> > 3.14286
> > Abort trap (core dumped)
> >
> >
> >
> > $ /usr/local/bin/gdb testglib
> > GNU gdb (GDB) 7.6.2 [GDB v7.6.2 for FreeBSD]
> > Copyright (C) 2013 Free Software Foundation, Inc.
> > License GPLv3+: GNU GPL version 3 or later
> > <http://gnu.org/licenses/gpl.html>
> > This is free software: you are free to change and redistribute it.
> > There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
> >  and "show warranty" for details. This GDB was configured as
> > "x86_64-portbld-freebsd11.0".
> > For bug reporting instructions, please see:
> > <http://www.gnu.org/software/gdb/bugs/>...
> > Reading symbols from /usr/home/waitman/testglib...done.
> > (gdb) run
> > Starting program: /usr/home/waitman/testglib
> > [New LWP 100327]
> > 3.14286
> > [New Thread 803006400 (LWP 100327)]
> >
> >
> > Program received signal SIGABRT, Aborted.
> > [Switching to Thread 803006400 (LWP 100327)]
> > 0x0000000801e313ba in kill () from /lib/libc.so.7
> > (gdb) bt
> > #0  0x0000000801e313ba in kill () from /lib/libc.so.7
> > #1  0x0000000801e30069 in abort () from /lib/libc.so.7
> > #2  0x00000008016ce7da in ?? () from /lib/libcxxrt.so.1
> > #3  0x000000080085de68 in Glib::ConvertError::throw_func(_GError*) ()
> > from /usr/local/lib/libglibmm-2.4.so.1 #4  0x000000080086ed2f in
> > Glib::Error::throw_exception(_GError*) ()
> > from /usr/local/lib/libglibmm-2.4.so.1 #5  0x000000080087bcc2 in
> > Glib::ustring::FormatStream::to_string() const ()
> > from /usr/local/lib/libglibmm-2.4.so.1 #6  0x0000000000401de0 in
> > Glib::ustring::format<std::__1::ios_base&
> > (std::__1::ios_base&), std::__1::__iom_t5, double>(std::__1::ios_base& (
> > const&)(std::__1::ios_base&), std::__1::__iom_t5 const&, double const&) (
> > a1= 0x401e50: {std::__1::ios_base &(std::__1::ios_base &)} 0x401e50
> > <std::__1::fixed(std::__1::ios_base&)>, a2=..., a3= 0x7fffffffd860:
> > 3.1428571428571428)
> > at /usr/local/include/glibmm-2.4/glibmm/ustring.h:1158 #7
> > 0x0000000000401696 in main () at testglib.cpp:9
> > (gdb)
> >
> >
> >
> >
> > $ uname -a
> > FreeBSD akira.waitman.net 11.0-CURRENT FreeBSD 11.0-CURRENT #1 r261643:
> > Sat Feb  8 22:11:05 PST 2014
> > root akira waitman net:/usr/obj/usr/src/sys/AKIRA  amd64
> >
> > $ ldd testglib
> > testglib:
> > libglibmm-2.4.so.1 => /usr/local/lib/libglibmm-2.4.so.1 (0x80081e000)
> > libgobject-2.0.so.0 => /usr/local/lib/libgobject-2.0.so.0 (0x800a9d000)
> > libglib-2.0.so.0 => /usr/local/lib/libglib-2.0.so.0 (0x800cea000)
> > libintl.so.9 => /usr/local/lib/libintl.so.9 (0x800ff2000) libsigc-2.0.so.0
> > => /usr/local/lib/libsigc-2.0.so.0 (0x8011fb000)
> > libc++.so.1 => /usr/lib/libc++.so.1 (0x801400000) libcxxrt.so.1 =>
> > /lib/libcxxrt.so.1 (0x8016be000)
> > libm.so.5 => /lib/libm.so.5 (0x8018d8000) libgcc_s.so.1 =>
> > /lib/libgcc_s.so.1 (0x801b00000)
> > libc.so.7 => /lib/libc.so.7 (0x801d0d000) libffi.so.6 =>
> > /usr/local/lib/libffi.so.6 (0x80208a000)
> > libgmodule-2.0.so.0 => /usr/local/lib/libgmodule-2.0.so.0 (0x802291000)
> > libpcre.so.3 => /usr/local/lib/libpcre.so.3 (0x802494000) libthr.so.3 =>
> > /lib/libthr.so.3 (0x8026fb000)
> >
> >
> >
> > Thank you,
> >
> >
> >
> >
> > --
> > Waitman Gobble
> > San Jose California USA
> > +1.510-830-7975
> > _______________________________________________
> > [hidden email] mailing list
> > http://lists.freebsd.org/mailman/listinfo/freebsd-ports
> > To unsubscribe, send any mail to "[hidden email]"
> >
> >
>
> update, if i wrap the segfaulting 101 line in a try block, in rtimage.cc
> (which has the format code that bombs out), i get a Glib::Error::what() of
>
> "Conversion from character set 'UTF-8' to 'WCHAR_T'"
>
> so I think this is a hint about the culprit..
>
> ideas?
>
> --
> Waitman Gobble
> San Jose California USA
> +1.510-830-7975
>
>
> --
> Waitman Gobble
> San Jose California USA
> (510) 830-7975
> _______________________________________________
> [hidden email] mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-ports
> To unsubscribe, send any mail to "[hidden email]"
Sounds like base iconv ;)

regards,
Bapt

attachment0 (203 bytes) Download Attachment
123