/dev/shm/tdelibs/tdeio/tdeio/backtrace_symbols.c:57:22: fatal error: demangle.h: No such file or directory
We ran into this before with tdebase and we added DWITH_KDESKTOP_LOCK_BACKTRACE as a configure option in commit 936d3cec 2013-01-26.
demangle.h is not supported in all distros.
I'll try to massage 936d3cec into tdelibs.
Darrell
On Friday 16 of August 2013 00:39:24 Darrell Anderson wrote:
/dev/shm/tdelibs/tdeio/tdeio/backtrace_symbols.c:57:22: fatal error: demangle.h: No such file or directory
We ran into this before with tdebase and we added DWITH_KDESKTOP_LOCK_BACKTRACE as a configure option in commit 936d3cec 2013-01-26.
demangle.h is not supported in all distros.
I'll try to massage 936d3cec into tdelibs.
Darrell
As soon as I saw Tim's commit, it was clear to me that we will need to add such a choice. You have prepared a patch or should I do it?
Slavek --
On Friday 16 of August 2013 00:39:24 Darrell Anderson wrote:
/dev/shm/tdelibs/tdeio/tdeio/backtrace_symbols.c:57:22: fatal error: demangle.h: No such file or directory
We ran into this before with tdebase and we added DWITH_KDESKTOP_LOCK_BACKTRACE as a configure option in commit 936d3cec 2013-01-26.
demangle.h is not supported in all distros.
I'll try to massage 936d3cec into tdelibs.
Darrell
As soon as I saw Tim's commit, it was clear to me that we will need to add such a choice. You have prepared a patch or should I do it?
Slavek
Sorry about that guys, I only had a short time to work on the crash bug and therefore didn't add in the compile time checks/flags. Thanks for cleaning this up!
Tim
On Friday 16 of August 2013 01:56:03 Timothy Pearson wrote:
On Friday 16 of August 2013 00:39:24 Darrell Anderson wrote:
/dev/shm/tdelibs/tdeio/tdeio/backtrace_symbols.c:57:22: fatal error: demangle.h: No such file or directory
We ran into this before with tdebase and we added DWITH_KDESKTOP_LOCK_BACKTRACE as a configure option in commit 936d3cec 2013-01-26.
demangle.h is not supported in all distros.
I'll try to massage 936d3cec into tdelibs.
Darrell
As soon as I saw Tim's commit, it was clear to me that we will need to add such a choice. You have prepared a patch or should I do it?
Slavek
Sorry about that guys, I only had a short time to work on the crash bug and therefore didn't add in the compile time checks/flags. Thanks for cleaning this up!
Tim
Two questions: 1) backtrace handler we want by default ${WITH_ALL_OPTIONS} or OFF? 2) and for Debian/Ubuntu packages by default ON or OFF?
Slavek
2013/8/16 Slávek Banko slavek.banko@axis.cz
On Friday 16 of August 2013 01:56:03 Timothy Pearson wrote:
On Friday 16 of August 2013 00:39:24 Darrell Anderson wrote:
/dev/shm/tdelibs/tdeio/tdeio/backtrace_symbols.c:57:22: fatal error: demangle.h: No such file or directory
We ran into this before with tdebase and we added DWITH_KDESKTOP_LOCK_BACKTRACE as a configure option in commit 936d3cec 2013-01-26.
demangle.h is not supported in all distros.
I'll try to massage 936d3cec into tdelibs.
Darrell
As soon as I saw Tim's commit, it was clear to me that we will need to add such a choice. You have prepared a patch or should I do it?
Slavek
Sorry about that guys, I only had a short time to work on the crash bug and therefore didn't add in the compile time checks/flags. Thanks for cleaning this up!
Tim
Two questions:
- backtrace handler we want by default ${WITH_ALL_OPTIONS} or OFF?
- and for Debian/Ubuntu packages by default ON or OFF?
Slavek
Wait a minute, don't apply it... IMO, it place there the system configuration check fits better than prompting another option. Isn't it?
On Friday 16 of August 2013 02:16:40 Fat-Zer wrote:
2013/8/16 Slávek Banko slavek.banko@axis.cz
On Friday 16 of August 2013 01:56:03 Timothy Pearson wrote:
On Friday 16 of August 2013 00:39:24 Darrell Anderson wrote:
/dev/shm/tdelibs/tdeio/tdeio/backtrace_symbols.c:57:22: fatal error: demangle.h: No such file or directory
We ran into this before with tdebase and we added DWITH_KDESKTOP_LOCK_BACKTRACE as a configure option in commit 936d3cec 2013-01-26.
demangle.h is not supported in all distros.
I'll try to massage 936d3cec into tdelibs.
Darrell
As soon as I saw Tim's commit, it was clear to me that we will need to add such a choice. You have prepared a patch or should I do it?
Slavek
Sorry about that guys, I only had a short time to work on the crash bug and therefore didn't add in the compile time checks/flags. Thanks for cleaning this up!
Tim
Two questions:
- backtrace handler we want by default ${WITH_ALL_OPTIONS} or OFF?
- and for Debian/Ubuntu packages by default ON or OFF?
Slavek
Wait a minute, don't apply it... IMO, it place there the system configuration check fits better than prompting another option. Isn't it?
Not exactly - even if the system is suitable for backtrace handler may be deliberately specified building without backtrace handler.
Slavek --
2013/8/16 Slávek Banko slavek.banko@axis.cz
On Friday 16 of August 2013 02:16:40 Fat-Zer wrote:
2013/8/16 Slávek Banko slavek.banko@axis.cz
On Friday 16 of August 2013 01:56:03 Timothy Pearson wrote:
On Friday 16 of August 2013 00:39:24 Darrell Anderson wrote:
/dev/shm/tdelibs/tdeio/tdeio/backtrace_symbols.c:57:22: fatal error: demangle.h: No such file or directory
We ran into this before with tdebase and we added DWITH_KDESKTOP_LOCK_BACKTRACE as a configure option in commit 936d3cec 2013-01-26.
demangle.h is not supported in all distros.
I'll try to massage 936d3cec into tdelibs.
Darrell
As soon as I saw Tim's commit, it was clear to me that we will need to add such a choice. You have prepared a patch or should I do it?
Slavek
Sorry about that guys, I only had a short time to work on the crash
bug
and therefore didn't add in the compile time checks/flags. Thanks
for
cleaning this up!
Tim
Two questions:
- backtrace handler we want by default ${WITH_ALL_OPTIONS} or OFF?
- and for Debian/Ubuntu packages by default ON or OFF?
Slavek
Wait a minute, don't apply it... IMO, it place there the system configuration check fits better than prompting another option. Isn't it?
Not exactly - even if the system is suitable for backtrace handler may be deliberately specified building without backtrace handler.
Slavek
To unsubscribe, e-mail: trinity-devel-unsubscribe@lists.pearsoncomputing.net For additional commands, e-mail: trinity-devel-help@lists.pearsoncomputing.net Read list messages on the web archive: http://trinity-devel.pearsoncomputing.net/ Please remember not to top-post: http://trinity.pearsoncomputing.net/mailing_lists/#top-posting
For that purpose there is a NDEBUG deffinition used all over the sources.
On Friday 16 of August 2013 02:22:50 Fat-Zer wrote:
2013/8/16 Slávek Banko slavek.banko@axis.cz
On Friday 16 of August 2013 02:16:40 Fat-Zer wrote:
2013/8/16 Slávek Banko slavek.banko@axis.cz
On Friday 16 of August 2013 01:56:03 Timothy Pearson wrote:
On Friday 16 of August 2013 00:39:24 Darrell Anderson wrote: > /dev/shm/tdelibs/tdeio/tdeio/backtrace_symbols.c:57:22: fatal > error: demangle.h: No such file or directory > > We ran into this before with tdebase and we added > DWITH_KDESKTOP_LOCK_BACKTRACE as a configure option in commit > 936d3cec 2013-01-26. > > demangle.h is not supported in all distros. > > I'll try to massage 936d3cec into tdelibs. > > Darrell
As soon as I saw Tim's commit, it was clear to me that we will need to add such a choice. You have prepared a patch or should I do it?
Slavek
Sorry about that guys, I only had a short time to work on the crash
bug
and therefore didn't add in the compile time checks/flags. Thanks
for
cleaning this up!
Tim
Two questions:
- backtrace handler we want by default ${WITH_ALL_OPTIONS} or OFF?
- and for Debian/Ubuntu packages by default ON or OFF?
Slavek
Wait a minute, don't apply it... IMO, it place there the system configuration check fits better than prompting another option. Isn't it?
Not exactly - even if the system is suitable for backtrace handler may be deliberately specified building without backtrace handler.
Slavek
For that purpose there is a NDEBUG deffinition used all over the sources.
Hmm, it starts to get complicated :)
And if I want to backtrace handler, but I do not want NDEBUG because it enables additional debugging code, that I not want in this situation?
Slavek --
2013/8/16 Slávek Banko slavek.banko@axis.cz
On Friday 16 of August 2013 02:22:50 Fat-Zer wrote:
2013/8/16 Slávek Banko slavek.banko@axis.cz
On Friday 16 of August 2013 02:16:40 Fat-Zer wrote:
2013/8/16 Slávek Banko slavek.banko@axis.cz
On Friday 16 of August 2013 01:56:03 Timothy Pearson wrote:
> On Friday 16 of August 2013 00:39:24 Darrell Anderson wrote: >> /dev/shm/tdelibs/tdeio/tdeio/backtrace_symbols.c:57:22: fatal >> error: demangle.h: No such file or directory >> >> We ran into this before with tdebase and we added >> DWITH_KDESKTOP_LOCK_BACKTRACE as a configure option in commit >> 936d3cec 2013-01-26. >> >> demangle.h is not supported in all distros. >> >> I'll try to massage 936d3cec into tdelibs. >> >> Darrell > > As soon as I saw Tim's commit, it was clear to me that we will > need to add such a choice. You have prepared a patch or should
I
> do it? > > Slavek > --
Sorry about that guys, I only had a short time to work on the
crash
bug
and therefore didn't add in the compile time checks/flags.
Thanks
for
cleaning this up!
Tim
Two questions:
- backtrace handler we want by default ${WITH_ALL_OPTIONS} or OFF?
- and for Debian/Ubuntu packages by default ON or OFF?
Slavek
Wait a minute, don't apply it... IMO, it place there the system configuration check fits better than prompting another option. Isn't it?
Not exactly - even if the system is suitable for backtrace handler may
be
deliberately specified building without backtrace handler.
Slavek
For that purpose there is a NDEBUG deffinition used all over the sources.
Hmm, it starts to get complicated :)
And if I want to backtrace handler, but I do not want NDEBUG because it enables additional debugging code, that I not want in this situation?
Slavek
To unsubscribe, e-mail: trinity-devel-unsubscribe@lists.pearsoncomputing.net For additional commands, e-mail: trinity-devel-help@lists.pearsoncomputing.net Read list messages on the web archive: http://trinity-devel.pearsoncomputing.net/ Please remember not to top-post: http://trinity.pearsoncomputing.net/mailing_lists/#top-posting
For now there is only one call to backtrace_symbols and it's already protected by NDEBUG ifdef.
Tim, the code of backtrace_symbols.c is quite complex is it possible to make it work without demangle.h?
On Friday 16 of August 2013 02:38:43 Fat-Zer wrote:
Tim, the code of backtrace_symbols.c is quite complex is it possible to make it work without demangle.h?
I saw similar code, where is used bfd_demangle instead of cplus_demangle. I'm not determine if it can be useful.
http://sourcecodebrowser.com/mutrace/0.2.0/backtrace-symbols_8c_source.html
Slavek --
2013/8/17 Slávek Banko slavek.banko@axis.cz
On Friday 16 of August 2013 02:38:43 Fat-Zer wrote:
Tim, the code of backtrace_symbols.c is quite complex is it possible to make it work without demangle.h?
I saw similar code, where is used bfd_demangle instead of cplus_demangle. I'm not determine if it can be useful.
http://sourcecodebrowser.com/mutrace/0.2.0/backtrace-symbols_8c_source.html
Yes, thanks, I've already noticed bfd_demangle, but the third argument is flags which defined in demangle.h (of coarse it's only flags, but defining them manually seems quite ugly for me). I found another alternative: abi::__cxx_demangle() witch presents at least in gcc.
Here is completely rewritten implementation of backtrace() functionality.
some reasons why WITH_LIBBFD should be a separate option and should be OFF by default: - It's a yet another dependence. - it's provides very specific and minor functionality. - The functionality may be interested only for developers and some testers. End users even won't have any ability to see the difference. - Strictly speaking there is only one thing on linux-glibc-gcc system (IMO the most of trinity systems) which cannot be handled without libbfd: the discovery of source code file and library if the binary was build with debug info.
the backtrase format is gdb-like :
[WITH_LIBBFD=ON] no matter present demangle.h on system or not #0 0x00007f0f16b6efe3 in kdBacktrace(int) in /tmp/trinity/tdelibs/tdecore/kdebug.cpp:791 #1 0x0000000000400f4b in MyNamespace::Foo::doFoo(int) from ./a.out:0x00000f4b #2 0x0000000000400e7d in MyNamespace::bar() from ./a.out:0x00000e7d #3 0x0000000000400e88 in main from ./a.out:0x00000e88
[WITH_LIBBFD=OFF] the demangling is handled by abi::__cxx_demangle() #0 0x00007f0f16b6efe3 in kdBacktrace(int) from /tmp/tdelibs-build/tdecore/libtdecore.so:0x00139b55 #1 0x0000000000400f4b in MyNamespace::Foo::doFoo(int) from ./a.out:0x00000f4b #2 0x0000000000400e7d in MyNamespace::bar() from ./a.out:0x00000e7d #3 0x0000000000400e88 in main from ./a.out:0x00000e88
2013/8/17 Slávek Banko slavek.banko@axis.cz
On Friday 16 of August 2013 02:38:43 Fat-Zer wrote:
Tim, the code of backtrace_symbols.c is quite complex is it possible
to
make it work without demangle.h?
I saw similar code, where is used bfd_demangle instead of cplus_demangle. I'm not determine if it can be useful.
http://sourcecodebrowser.com/mutrace/0.2.0/backtrace-symbols_8c_source.html
Yes, thanks, I've already noticed bfd_demangle, but the third argument is flags which defined in demangle.h (of coarse it's only flags, but defining them manually seems quite ugly for me). I found another alternative: abi::__cxx_demangle() witch presents at least in gcc.
Here is completely rewritten implementation of backtrace() functionality.
some reasons why WITH_LIBBFD should be a separate option and should be OFF by default:
- It's a yet another dependence.
- it's provides very specific and minor functionality.
- The functionality may be interested only for developers and some
testers. End users even won't have any ability to see the difference.
- Strictly speaking there is only one thing on linux-glibc-gcc system
(IMO the most of trinity systems) which cannot be handled without libbfd: the discovery of source code file and library if the binary was build with debug info.
the backtrase format is gdb-like :
[WITH_LIBBFD=ON] no matter present demangle.h on system or not #0 0x00007f0f16b6efe3 in kdBacktrace(int) in /tmp/trinity/tdelibs/tdecore/kdebug.cpp:791 #1 0x0000000000400f4b in MyNamespace::Foo::doFoo(int) from ./a.out:0x00000f4b #2 0x0000000000400e7d in MyNamespace::bar() from ./a.out:0x00000e7d #3 0x0000000000400e88 in main from ./a.out:0x00000e88
[WITH_LIBBFD=OFF] the demangling is handled by abi::__cxx_demangle() #0 0x00007f0f16b6efe3 in kdBacktrace(int) from /tmp/tdelibs-build/tdecore/libtdecore.so:0x00139b55 #1 0x0000000000400f4b in MyNamespace::Foo::doFoo(int) from ./a.out:0x00000f4b #2 0x0000000000400e7d in MyNamespace::bar() from ./a.out:0x00000e7d #3 0x0000000000400e88 in main from ./a.out:0x00000e88
Looks good to me!
Tim
2013/8/19 Timothy Pearson kb9vqf@pearsoncomputing.net
2013/8/17 Slávek Banko slavek.banko@axis.cz
On Friday 16 of August 2013 02:38:43 Fat-Zer wrote:
Tim, the code of backtrace_symbols.c is quite complex is it possible
to
make it work without demangle.h?
I saw similar code, where is used bfd_demangle instead of cplus_demangle. I'm not determine if it can be useful.
http://sourcecodebrowser.com/mutrace/0.2.0/backtrace-symbols_8c_source.html
Yes, thanks, I've already noticed bfd_demangle, but the third argument is flags which defined in demangle.h (of coarse it's only flags, but
defining
them manually seems quite ugly for me). I found another alternative: abi::__cxx_demangle() witch presents at least in gcc.
Here is completely rewritten implementation of backtrace() functionality.
some reasons why WITH_LIBBFD should be a separate option and should be
OFF
by default:
- It's a yet another dependence.
- it's provides very specific and minor functionality.
- The functionality may be interested only for developers and some
testers. End users even won't have any ability to see the difference.
- Strictly speaking there is only one thing on linux-glibc-gcc system
(IMO the most of trinity systems) which cannot be handled without libbfd: the discovery of source code file and library if the binary was build with debug info.
the backtrase format is gdb-like :
[WITH_LIBBFD=ON] no matter present demangle.h on system or not #0 0x00007f0f16b6efe3 in kdBacktrace(int) in /tmp/trinity/tdelibs/tdecore/kdebug.cpp:791 #1 0x0000000000400f4b in MyNamespace::Foo::doFoo(int) from ./a.out:0x00000f4b #2 0x0000000000400e7d in MyNamespace::bar() from ./a.out:0x00000e7d #3 0x0000000000400e88 in main from ./a.out:0x00000e88
[WITH_LIBBFD=OFF] the demangling is handled by abi::__cxx_demangle() #0 0x00007f0f16b6efe3 in kdBacktrace(int) from /tmp/tdelibs-build/tdecore/libtdecore.so:0x00139b55 #1 0x0000000000400f4b in MyNamespace::Foo::doFoo(int) from ./a.out:0x00000f4b #2 0x0000000000400e7d in MyNamespace::bar() from ./a.out:0x00000e7d #3 0x0000000000400e88 in main from ./a.out:0x00000e88
Looks good to me!
Tim
To unsubscribe, e-mail: trinity-devel-unsubscribe@lists.pearsoncomputing.net For additional commands, e-mail: trinity-devel-help@lists.pearsoncomputing.net Read list messages on the web archive: http://trinity-devel.pearsoncomputing.net/ Please remember not to top-post: http://trinity.pearsoncomputing.net/mailing_lists/#top-posting
if there is no objections, please commit it sooner because for now, as I said before, tdelibs FTBFS with following error if compiled on system with binutils >= 2.23.1 (may be also with 2.23) as discribed in upstream bug http://sourceware.org/bugzilla/show_bug.cgi?id=14243.
=========================== In file included from /var/tmp/portage/trinity-base/tdelibs-9999/work/tdelibs/tdeio/tdeio/backtrace_symbols.c:53:0: /usr/include/bfd.h:37:2: error: #error config.h must be included before this header ===========================
I've tested the patch once more and found that it contains a small error. Here is a fixed one.
2013/8/24 Slávek Banko slavek.banko@axis.cz
On Saturday 24 of August 2013 00:41:10 Fat-Zer wrote:
I've tested the patch once more and found that it contains a small error. Here is a fixed one.
Should there be attachment? :)
It was a late night in me timezone... =) Here it is.
On Saturday 24 of August 2013 12:18:42 Fat-Zer wrote:
2013/8/24 Slávek Banko slavek.banko@axis.cz
On Saturday 24 of August 2013 00:41:10 Fat-Zer wrote:
I've tested the patch once more and found that it contains a small error. Here is a fixed one.
Should there be attachment? :)
It was a late night in me timezone... =) Here it is.
Tim suggested to enable backtrace generation, regardless of the option NDEBUG. For me it also seems like a good idea. What do you think?
It seems that the enhanced capabilities kdBacktrace could be used also for backtrace in kdesktop_lock (tdebase). Please, can you prepare a patch?
Slavek
2013/8/24 Slávek Banko slavek.banko@axis.cz
Tim suggested to enable backtrace generation, regardless of the option NDEBUG. For me it also seems like a good idea. What do you think?
If so, we supposed to introduce another function rather than making kdBacktrace produce output regardless of NDEBUG because in most cases it's used for non-critical errors in conjunction with kDebug(), so it will cause useless overhead if build with NDEBUG.
It seems that the enhanced capabilities kdBacktrace could be used also for backtrace in kdesktop_lock (tdebase). Please, can you prepare a patch?
Slavek
yes sure... I'm also going to fix those creepy hacks with "#define protected public". No objections?
Here is yet another patchset...
tdelibs patches: - 0001-kdebug-add-default-message-printing-destinations.patch That one introduces Default runtime-configurable debug/error/warning message destinations. - 0004-fix-an-FTBFS-then-build-with-clang.patch This one fixes yet another FTBFS which appears when being built with clang-3.3 ( builds ok with gcc ) ============================ /home/alexander/projects/trinity/tdelibs/networkstatus/testservice.cpp:95:17: error: call to constructor of 'TDEApplication' is ambiguous TDEApplication app(argc, argv, "testdcop"); ^ ~~~~~~~~~~~~~~~~~~~~~~ /home/alexander/projects/trinity/tdelibs/tdecore/tdeapplication.h:131:3: note: candidate constructor TDEApplication( bool allowStyles=true, bool GUIenabled=true, bool SMenabled=true); ^ /home/alexander/projects/trinity/tdelibs/tdecore/tdeapplication.h:270:3: note: candidate constructor TDEApplication(int& argc, char** argv, ^ =============================
tdebase patches: - 0003-kdesktop-lock-use-kdBacktrace-in-case-of-fail.patch That one will utilize the kdBacktrace. - 0004-remove-a-useless-check-from-kdesktop-ConfgureChecks.patch It's just a small fixup... the second check was useless. - 0005-kdesktop-lock-remove-tqwidget-include-hacks.patch This one removes hack I noticed before.
On Saturday 24 of August 2013 12:18:42 Fat-Zer wrote:
2013/8/24 Slávek Banko slavek.banko@axis.cz
On Saturday 24 of August 2013 00:41:10 Fat-Zer wrote:
I've tested the patch once more and found that it contains a small error. Here is a fixed one.
Should there be attachment? :)
It was a late night in me timezone... =) Here it is.
Oops, with the patch is a problem. When testing on Debian Squeeze is not available CheckCXXSymbolExists in cmake 2.8.2. How to test differently?
Slavek --
2013/8/24 Slávek Banko slavek.banko@axis.cz
On Saturday 24 of August 2013 12:18:42 Fat-Zer wrote:
2013/8/24 Slávek Banko slavek.banko@axis.cz
On Saturday 24 of August 2013 00:41:10 Fat-Zer wrote:
I've tested the patch once more and found that it contains a small error. Here is a fixed one.
Should there be attachment? :)
It was a late night in me timezone... =) Here it is.
Oops, with the patch is a problem. When testing on Debian Squeeze is not available CheckCXXSymbolExists in cmake 2.8.2. How to test differently?
Slavek
To unsubscribe, e-mail: trinity-devel-unsubscribe@lists.pearsoncomputing.net For additional commands, e-mail: trinity-devel-help@lists.pearsoncomputing.net Read list messages on the web archive: http://trinity-devel.pearsoncomputing.net/ Please remember not to top-post: http://trinity.pearsoncomputing.net/mailing_lists/#top-posting
Does it provides CheckCxxSourceCompiles? If yes you can test it with next snipset: ============================================ check_cxx_source_compiles( "#include <cxxabi.h> int main() { abi::__cxa_demangle(0, 0, 0, 0); return 0; }" HAVE_ABI_CXA_DEMANGLE ) ============================================
On Saturday 24 of August 2013 17:07:34 Fat-Zer wrote:
2013/8/24 Slávek Banko slavek.banko@axis.cz
On Saturday 24 of August 2013 12:18:42 Fat-Zer wrote:
2013/8/24 Slávek Banko slavek.banko@axis.cz
On Saturday 24 of August 2013 00:41:10 Fat-Zer wrote:
I've tested the patch once more and found that it contains a small error. Here is a fixed one.
Should there be attachment? :)
It was a late night in me timezone... =) Here it is.
Oops, with the patch is a problem. When testing on Debian Squeeze is not available CheckCXXSymbolExists in cmake 2.8.2. How to test differently?
Slavek
Does it provides CheckCxxSourceCompiles? If yes you can test it with next snipset: ============================================ check_cxx_source_compiles( "#include <cxxabi.h> int main() { abi::__cxa_demangle(0, 0, 0, 0); return 0; }" HAVE_ABI_CXA_DEMANGLE ) ============================================
Great, it looks good:
-- Performing Test HAVE_ABI_CXA_DEMANGLE -- Performing Test HAVE_ABI_CXA_DEMANGLE - Success
Slavek
-
Darrell Anderson -
Fat-Zer -
Slávek Banko -
Timothy Pearson